Beating the sicilian 3 by Jose Manuel Vargas

Beating the Sicilian 3 John Νυnn and Joe Gallagher

Β. Τ.

Batsford Ltd, London

Pirst published 1995 © John Νυηn and Joe Gallagher 1995 ISBN Ο 7134 7844 6 Βήtίsh

Library Cataloguing-in-Publication Data. catalogue record for this book is available from the Βήtίsh Library.

ΑlI ήghts reserνed. Νο part of this book may be reproduced, by any means, without Ρήοr permission of the publisher.

Typeset by John Νυηη and Ρήnted ίη Great Βήtaίn by Redwood Books, Trowbridge, Wilts for the publishers, Β. Τ. Batsford Ltd, 4 Pitzhardinge Street, London WIH ΟΑΗ

BATSPORD CHESS

ΒΟΟΚ

Editoriαl Pαnel: Mark DvoretSky, John Νυηη, Ιοη Speelman Generαl Adviser:

Raymond Keene ΟΒΕ Graham Burgess

Mαnαging Editor:

Contents

Symbols Introduction 1 Ν ajdorf Variation 2 Scheveningen Variation 3 Classica1 Variation 4 Pelikan Variation 5 The Dragon 6 Kan Variation 7 Maroczy Bind 8 Taimanov Variation 9 Sicilian Four Κnights 10 Lδwentha1 and Ka1ashnikov Variations 11 Ρίη Variation 12 Nimzowitsch Variatίon: 2 ... llJf6 13 O'Kelly Variation: 2 ... a6 14 Unusua1 Lines Index of Variations

4 5 7 28 60 95 110 123 144 164 180 188 196 200 206 210 222

Symbols +

++ # ;t (=i=)

± (+) +- (-+) = 00

! ? !! ?? !? ?! Ch Cht ιι

jr wom rpd Wch Ζ

ΙΖ

Ct OL Cοπ

(D)

Check Double check Checkmate Slίght advantage Ιο White (Black) Clear advantage Ιο White (Black) Winning advantage Ιο White (Black) Level position Unclear position Goodmove Badmove Outstanding move Blunder Interesting move Dubious move Championship Team championship Team tournament Junior Event Women's event Rapidplay World Championship Zonal Interzona1 Candidates Olympiad Posta1 game Diagram follows

Introduction This third edition of Beαting the Siciliαn (BTS) does ηοι need a lengthy introduction, since repertoire books are now a familiar concept. Νο one will be surpήsed Ιο learn that the aim ()f this book is Ιο provide a complete repertoire for White against the Sicilian, sufficiently detailed for most players Ιο meet any line of the Sicilian with confidence. As ίη BTS2, the lines we recommend are well-established main variations. Τοο many opening books advocate one or another sideline, claiming ίι Ιο be ηο worse than the usual variations, and offering analy~is which appearS convincing - υηtίl you actually play the line over-thehoard. Μαίη lines of the type analysed ίη this book cannot be refuted, although the whims of fashion may lead Ιο them becoming more or less ρορυlαι as the years go by. The lines we recommend should remain valid t·or yearS Ιο come (υηιΗ Beαting the Siciliαn 41) and readers may be confident that the effort ρυΙ ίηΙο studying them will earn a long-term reward. We would like Ιο use the rest of this introductίon Ιο descήbe the changes ίη the proposed repertoire. The lines given against the Najdorf, Scheveningen, Classical and Pelikan variations are broadly the

same as ίη BTS2, although all have been thoroughly updated Ιο reflect both developments ίη the pre-existing theory and new ideas for Black which have arisen ίη the intervening years. Retaining the original recommendations was ηοΙ done ιο minίmίse the authors' efforts, but because most players do ηοΙ have the time Ιο change their repertoires completely overnight. Even grandmasters normally change their openings incrementally, because of the considerable study required Ιο grasp both the ideas behind openings and the concrete variations embodying those ideas. The repertoire against the Dragon has substantially changed. The main line of the Yugoslav Attack has become just Ιοο vast for most players Ιο study ίη depth, and anyone intending ιο venture ίηΙο such tricky waters needs Ιο have an intimate knowledge of all the hidden reefs. Therefore, we have switched from 9 .ic4 Ιο 9 0-0-0 d5 1Ο 'ii'e 1, a promίs ing but relatίvely undeveloped line. The Kan and Maroczy Bind chapters include many new finesses, but the broad outline remains little changed. Against the Taimanov, we have abandoned BTS2's 5 tίJb5, both because recent practical results have

6 Introduction been disappointing for White, and because the stodgy positions which result don't fit ίη very well with the style οί the rest οί the repertoire. The cunent recommendation οί 5 lbc3 'fIc7 6 f4 has had a knock-on effect, ίη that lines also need Ιο be provided against 5 ... d6 and 5 ... a6. This new material is also covered ίη Chapter 8, and those who have earlier editions of BTS are recommended Ιο study this chapter carefully, ίη order ηοΙ Ιο be left stranded high and dry if Black does ηοΙ play 5 ...'fIc7. The Sicilian Four Κnights experienced a bit of a revival a few years ago, which resulted ίη some new developments. Although ίι has now faded again ίηιο semί-obscurity, readers should be aware that thίs chapter has altered subsιantially.

Perhaps the greatest change since BTS2 has been the rise of the Kalashnikov variation, the οηlΥ really new Black system Ιο have been developed ίη the Sicilian for the past 15 years. Ιι may be that the popularity of this line has now passed its peak, but ίι is still relatively common and deserves careful attention. Chapter Ι Ο has the details. The remaining chapters constitute the Odds and Ends οί the Sicilian. Maybe they are ηοΙ really so bad, but the tide οί fashion has turned fιrmly against them and new ideas are few and far between. Perhaps the one exception is the line 1 e4 c5 2 lbf3 g6, covered ίη Chapter 14, whίch is often used for transpositional purposes; although still a littIe unusual, ίι appears Ιο be increasing ίη popularity. JohnNunn London J oe Gallagher NeuchΔtel

ΝajdοrfVaήatίοn

Of a11 the lίnes ίη the Sicilίan which Black can adopt, the Najdorfhas developed the largest body of theory. Whole books have been written οη mere sub-variations of the Najdorf, for example the Polugaevsky νΜί ation and the infamous Poisoned Pawn. The Najdorf starts with the moves 1 e4 c5 2 ~f3 d6 3 d4 cxd4 4 ~xd4 ~f6 5 ~c3 a6. Black's first aim is to play ... e5 without allowing the reply .i.b5+, while the secondary ροίηΙ is Ιο prepare queenside expansion by ... b5. Some ofWhite's systems against the Najdorf are specifica11y aimed at preventing ... e5, while others allow Black to play this move ίη the hope of exploiting the backward d-pawn later. Devising a counter which is viable ίη tournament play, while at the same time necessitating relatively slίght book knowledge, has proved especially hard. These days the five major systems against the Najdorf, 6 .tg5, 6 .i.e2, 6 .i.c4, 6 .te3 and 6 f4 seem Ιο occur with about equal frequency. As ίη BTS2 we have opted for 6 f4, a system which offers good attacking chances while retaining an element of solίdity. Another advantage of 6 f4 is that as ίι has ηοΙ been popular for as long as some of the other moves there is ηοΙ quite as much Ιο learn. Ιη Games 1 to 4 we investigate the

replies most commonly encountered ίη practice. The main division is between those lines involving an early ... e5 and those ίη which Black delays this move or omits ίι entirely. We will postpone consideration of the former lίnes until Games 3 and 4, and concentrate first οη the alternatives ιο ... e5. Game 1

J. Polgar - Pliester Aruba 1992 Ι e4 2 ~f3 3 d4 4 lDxd4

d6 cxd4

5 lbc3

lί)f6

6 f4(D)

6 f4 is a flexible move; White gives little away regarding his piece deployment, and waits for Black's

NαjdorfVariαtion

reply before deciding where Ιο ρυι his bishops. 6 ..• lbbd7 This move is designed Ιο reserve the ορΙίοn of playing ... e5 under more favourable circumstances if White should develop his pieces Ιο unsuitable squares. 6•••'fIJc7 is covered ίn Game 2, but there are other playable moves: 1) 6•••lbc6 7 lbxc6 bxc6 8 e5 lbd7 (8 ... dxe5 9 'ifxd8+ ΦΧd8 10 fxe5lbd5 11 ~d2 is good for White) 9 .ic4!? dxe5 100-0 e6 11 f5.ic5+ 12 ΦhΙ and White has good attacking chances. 2) 6•.• g6 7 .id3 and after a subsequent ... 'ifc7 or ... lbbd7 there will probably be a transposition ίηΙο lines considered ίη Game 2. 3) 6•••e6 (after this White may transpose ίnΙο various lines of the Scheveningen, but since these lines do ηοΙ form part of the repertoire recommended ίη this book, we suggest an independent a1ternative which promises good chances for White) 7 'fIJf3 (White's advantage over similar lines is that his bishops are ηοΙ committed, so he can force through g4g5 very quickly) and now: 3a) 7.•• lbbd7 8 g4 h6 (8 ... e5 9 lbf5!? and now 9 ... exf4?! 10 .ixf4 lbe5 11 .ixe5 dxe5 12 g5 lbd7 13 .ic4 gives White a tremendous attack, whίlst 9... g610 g5 gxf5 11 exf5 leaves Black under heavy pressure) 9 h4 (9 .ie2 1t'b6 10 lbb3 1t'c7 11 'ifg2 :b8 12 .ie3 b5 13 g5 hxg5 14 fxg5lbh5 15 g6lbe5 was distinctly unclear ίη Short-Kasparov, Belgrade

1989) 9..•e5 10 lbb3 exf4 11 .ixf4

lbes 12 ~xe5 dxe5 and now: 3al) 13 g5 hxg5 14 hxg5 1:txhl 15 'ifxhl lbg4 16 lbd5 is given as unclear by Kasparov. 3a2) 13 .ih3 looks more testing Ιο me. Now 13 ... g5 loses Ιο 14 :η .ie7 15lbd5 and 13 ....ie6 14 g5 is at least slightly better for White. 13 ...h5 should be met by 14 :dl! (14 g5 .ixh3 15 gxf6 .ig4 16 fxg7 .ixg7 17 'ifg3 ο-ο is ΟΚ for Black) followed by g5. This leaves the cήti cal line as 13••• .ib4 14 g5 (D) and now:

3a21) 14•...ixc3+ (14 ... JLxh3 is met by 15 gxf6!) 15 'ifxc3 (15 bxc3 lbg8100ks ΟΚ for Black) 15 ...lbxe4 (15 ...hxg5 16:d 1 ! is very strong) 16 'ifxe5+ 'ife7 and the big question is whether White can get away with the cheeky 17 'ifxg7. Ι believe he can, although strong nerves are required. Here is a sample variation: 17 ... lbg3+ (17 ...:f8 18 .ixc8 favours White after 18 ...:xc8 19 0-0-0 or 18 ...lbg3+ 19 Φf2lbe4+ 20 ~f3! :Ιχc8 21 :hel) 18 Φd2 'ife2+ 19 Φc3! 'ife3+ (19 ... lbe4+ 20 Φd4!)

NαjdorfVαriαtion

20 ~M "b6+ (20 ... a5+ 21 ~b5 .02+ 22 c4) 21 ~a3 "d6+ 22l2)c5! .KcS+ 23 b4 "e3+ 24 ~b2 and the chocks have run ουΙ and although Hluck is a piece υρ he is ίη trouble as White threatens both 'ifxh8 and :el. 3α22) 14•••hxg5 15 :dl! (after Ι ~ hxg5 .i.xc3+! both 16 'ifxc3 /t)g4! and 16 bxc3 :xh3! 17 :xh3 .tKh3 18 gxf6 .i.e6 19 fxg7 "g5! nrc ΟΚ for Black) 15•• :i'c7 16 hxg5 Ilnd now: 3a221) 16•••.i.g4 17 gxf6! is winIIing for White. 3a222) 16•••l2)h7 17 .i.xc8 11xc8 Ι Μ 11d2! and Black is caught ίη a falul ρίη οη the h-file. 3a223) 16•••l2)g4 and now the re111Υ 17 .i.xg4 is ηοι so good οη account of 17 ... .i.xc3+ 18 bxc3 11xhl + 19 "xhl "xc3+ 20 11d2 .i.xg4! with a winning attack for Black. However, White does have the very ιιurpήsίηg 17 ~e2! (17 ~d2 'ifd8+! ίιι unclear) which seems Ιο win muteria1. Α possible continuation is 17 ... .i.xc3+ 18 bxc3 11h4 19 ..g3 "c4+ 20 Φd2 11xh3 (20 .....xe4 21 .tg2!) 21 11xh3 'ii'xe4 22 11h8+ ~e7 23 11el (safest) 23 .....f5 24 ~cl! "xg5+ 25 ~b2 when the dominant I"uctor ίη assessing this position is the Nufety of the respective kίngs. 3b) 7•• :i'c7 8 g4 b5 (8 ... l2)c6 9 li::Jb3 b5 10 g5 l2)d7 11 .i.e3 b4 12 li::Ja4.i.b7 13 "f2 g6 14 l2)b6l2)xb6 15 .i.xb6 "e7 160-0-0 was pleasant ΙΌr White ίη Zso.Polgar-Plίester, Aruba 1992, whilst 9l2)xc6 bxc6 1Ο g5l2)d7 11 b3!? a1so comes ίηΙο conNideration) 9 g5 and now:

3bl) 9••• lMd7 10 a3 .i.b7 ll.i.e3 (11.i.g2g612'ii'f2~613~e2h6

14 f5 was a1so promising ίη Smyslov-Kamsky,NewYorkOpen 1989) 11 ... l2)c6 12 .i.h3 b4 13 axb4l2)xb4 14 0-0 ~5 15 :adl g6 16 :d2 .i.e7 17 :df2 with advantage Ιο White, Timman-Hjartarson, Belfort World Cup 1988 . 3b2) 9••• b4 10 l2)cb5 axb5 11 gxf6 gxf6 12 f5 (12 .i.xb5+ .i.d7 13 .i.d3 is also quite good) 12 ... exf5?! (12 ... e5 13l2)xb5 "c6 14 c4 is given as;!; by Rogers) 13 .i.xb5+ .i.d7 14 "xf5 l2)c6 (14 ... :a5 15 'ii'xf6 :g8 16 ~6! 'it'b6 17 :f1 is winning) 15 .i.xc6! .i.xc6, Rogers-Ehlvest, Biel ΙΖ 1993, and now Rogers gives 16 :gl! .te7 17l2)xc6 'ifxc6 18 .i.d2 or 18 .i.e3 as clearly better for White. 3c) 7.•:i'b6 is the main lίne, after which White has: 3cl) 8 a3, with the blunt ροίηι 8••:i'xd4?? 9 .i.e3, was very fashionable for a few years, but the most recent evidence suggests that Black has managed ιο overcome his initial difficulties. For example, ΑΙ masi-Kasparov, Lyon 1994 continued 8•.. ~! (8 ... l2)bd7 9 .i.e2 g6 10 .i.e3! is good for White as 10..."xb2 loses Ιο 11 ~d2!) 9l2)xc6 (9l2)b3 is quite often played, and may ίη fact be better than 9l2)xc6, but ifyou're going Ιο play this ίι seems more logical Ιο play 8 l2)b3; Ι can't believe, as some people do, that the inclusion of a3 and ... l2)c6 is ίη White's favour) 9 ... bxc6 10 b3 .i.b7 11 .i.b2 d5! 120-0-0 'fi'a5! 13 e5 l2)d7 14 ~4 'it'c7 15 :el? (15 c4 d4

NαjdorfVαriαtion

16 J.d3 c5 17 J.e4 is given as unclear by Kasparov) 15 ... g6 16 g4 c5 17 J.g2 :b8 18 Wdl c4 and Black was clearly οη top. 3c2) 8 tbb3 'ikc7 9 g4 b510 J.d3 (1 Ο g5 b4 is less good for White than the similar line above as his knight is now more passively placed οη b3) 10..•J.b7 (after 1O... h6 11 h4, both 11 ... h5 12 g5 tbg4 13 g6! J.b7 14 gxf7+ 1ί'Χf7 15 J.d2 and 11 ... b4 12 tbe2 h5 13 g5 lbg4, Yudasin-Dorfman, USSR 1981, 14 J.d2! Wb6 15 g6! fxg6 16 e5 .tb7 17 .txg6+ ~d7 18 J.e4 .txe4 19 1ί'χe4 1ί'f2+ 20 Wd 1 are good for White) 11 g5lill'd7 12 'iί'h3 (perhaps the most accurate move order as Black has been doing ΟΚ after 12 J.e3 tbc5) 12••. g6 (D) (White was threatening ... g6) and now there are a couple of ways for White ιο handle the position:

3c21) 13 :n!? οccuπed ίη the game Ulybin-Κhurtsidze, Oakham 1992. White, who is a leading specialist ίη this variation, decided Ιο develop his kingside initiative before turning his attention Ιο the mundane matter of king safety. Play continued

13 ... b4 14 tbe2 lbc6 15 f5 gxf5 16 exf5 e5 17 f6 0-0-0 18 1ί'h5 tbb6 19 J.e3 d5 20 J.f5+ Φb8 21 J.xb6 'ikxb6 22 0-0-0 a5 23 ΦbΙ and then 23•.•.ta6 24 tbbcl 1ί'b7 25 g6 fxg6 26 .txg6 J.c5 27 J.d3 with an indίs putable advantage for White, a1though his means of achieving this advantage were ηοΙ totaHy convincing. For example, 23•.•a4 24 tbbcl a3 looks very unclear Ιο me. Ιι will be interesting Ιο see whether Ulybin repeats 13 :η ίη his next outing with this line. 3c22) 13 .te3 b4 (13 .. .lbc5 is ηοΙ so good now οη account of 14 J.d4 and after 13 ...tbc6 White has the ορ tion of 14 f5) 14 tbe2 (14 tbd5 exd5 15 J.d4 dxe4 16 J.e2 tbe5 is good forBlack) 14 ...tbc615 0-0-0.tg716 :hf1! (more accurate than 16 ΦbΙ, which gave Black adequate counterplay ίη Nijboer-Van Wely, Dutch Ch 1993 after 16... 0-0! 17 f5 exf5 18 exf5 :fe8 19 :hf1 tbce5 as 20 J:[f4 is met by 20...:ac8) 16...0-0-0 (now 16 ... 0-0 17 f5 exf5 18 exf5 :fe8 19 :f4 is more dangerous for Black) 17 f5 tbce5 18 ΦbΙ with a slight advantage for White, U1Υbίn-Μageπamοv, Uzhgorod 1988. 7 J.e2 (D) This position frequently arίses via the move order 6 J.e2 tbbd7 (instead of the more common 6 ...e5) 7 f4. 7 ..• e5 This move is most common, the following Iines being somewhat unpleasant for Black: 1) 7.••g6 (an attempt ιο reach ρο sitions akin Ιο the Dragon, but here

NαjdoήVαriαtion

the extra pawn was soon converted victory. 1b) Ι think the best way for White Ιο proceed ίη this line is 11 ~Ι3!, preventing ... b6 and providing the οριίοη of 'ii'e2. Ιη that case White could cIaim a sizeable advantage. 2) 7.••lL)cS 8 ~f3 'ii'b6 9 lL)b3 (once again this move, putting the question Ιο the c5-knight, gives White the advantage) 9 ... lL)xb3 10 axb3 g6 11 e5 dxe5 12 fxe5lL)d7 13 lL)d5 'ii'd8 14 ~g5, Gipslis-Quinteros, Οlοι 1973 and White stands very well since 14 ... lL)xe5? loses ιο 15lL)f6+ exf6 16 'ii'xd8+ ΦΧd8 17 i.xf6+. 3) 7.....b6 8lL)b3 g6 9 "d3 (preparing ~e3, when Black is driven back ίη confusion) 9 .....c7 10 g4 lL)c5 11lL)xc5 'ii'xc5 12 i.e3 'ii'a5 13 b4! with advantage Ιο White, Τοπe Quinteros, Leningrad ΙΖ 1973. 4) 7••• bS (Ι {JN} suggested this move ίη 1982 but a few months later found a good repIy) 8lL)d5! ~b7 (9 lL)c6 was the threat and 8 ... lL)xd5 9 exd5 gives White a superb outpost at c6) 9lL)xf6+ lL)xf6 (or else White has a positional advantage) 10 e5 dxe5 11 fxe5 followed by e6, when Black has problems deveIoping his kingside pieces. 8 fxe5 Inexperienced players sometimes try 8 tί)b3 but after the reply 8 ... b5 White should be thinking about equalising! Ν atural moves like 8 lL)b3 quite often turn out badIy ίη the Najdorf, which is one reason why ίι is so popular. ίηΙΟ

White can exploit an interesting tactical resource) 8 g4 tί)cS (8 ... h6 9 f5 lL)c5 10 i-f3 e5 11 lL)b3 gxf5 12 cxf5 e4 13 ~g2lL)xb3 14 axb3 :g8 15 h3 ~xf5 16 'ii'd4 ~e6 was unclear ίη Beliavsky-Ljubojevic, Bugojno 1984, but this line is certainly risky for Black) 9 lL)b3! lL)xb3 (9 ... lL)fxe4? loses Ιο 10 lL)xe4lL)xe4 11 'ii'd4lL)f6 12 g5, and after 9 ... b6 lO g5lL)fd7, Arnasοn-Τήngοv, ΡΙον div 1986, White should play 11 i-f3 ~b7 12 'ii'e2 supporting e4 and prepaήng ~d2 and 0-0-0 with a promίs ing position) 9... lL)xb3 10 axb3 i-g7 and now: la) 11 gS lL)d7 12 ~e3 is given as a clear plus for White ίη BTS2. Perhaps this is true, but the continuation of the game Langumina-Gallagher, Forli 1992, shows that Black is ηοΙ withoutcounterchances: 12 ... b6! 13 'ii'd2 ~b7 14 i-f3?! (140-0-0 b5 15 ΦbΙ) 14 ... b5 15 h4 b4 16lL)a4 'ii'a5 17 h5 :g8 18 hxg6 hxg6 and now my opponent played 19 i-g4?, falling straight ίηΙο my trap: 19 ...~xb2! 20i-χd7+(20:bΙ'ii'χa4!)20 ... Φχd7 21 lL)c5+ 'iί'xc5! 22 ~xc5 ~c3 23 ~xb4 ~xd2+ 24 ΦΧd2 i.xe4 and

NαjdorfVαriαtion

BTS2 recom.mended 8 lΔΙ5, but the imιnedίate exchange οη e5 is creatίng more trouble for Black. The problem with 8 lDf5 is that after 8... lDc5 9 lΔg3 'i'b6 10 :bl J.d7 11 fxe5 dxe5 12 J.e3 'ii'c6 13 ο-ο, Black can actually take the pawn οη e4. After 13 ... lΔcxe4 14 lDcxe4lDxe4 15 J.f3 (15 lΔxe4 was played ίη Short-Gelfand, Tilburg 1990 but White never had enough for the pawn) 15 ...J.c5! 16 lΔxe4 J.xe3+ 17 ςPh 1 ο-ο 18 1Dc3 "e6! (BTS2 only took 18 ... 'i'c8? ίηΙο account) 19lΔd5 e4! White may win some materia1 but Black is going Ιο obtain a lot of compensation. cuπentΙΥ

8 ... lΔxe5 This has been played most often ίη practice, but perhaps the alternatίve recapture, 8...dxe5, is better. After 9lDf5 (D) Black has played:

1) 9.....c7 10 ο-ο lΔc5 11 lΔg3 J.e6 12 J.g51Dcd7, Wahls-Gelfand, Munich 1991, and now Gelfand gives 13 ςPh Ι! as good for White. For example, 13 ... h6 (13 ...:d8 14 lDds!) 14 J.xf6lΔxf6 15 lDds lDxd5 16 exd5 0-0-017 c4 ±.

2) 9...'iVb6, aiming Ιο interfere with White's smooth development, is the most recent try. Κasparov-Gel fand, Horgen 1994 now continued 10 lΔd5 lΔxd5 11 "xd5 'i'c5! 12 'ii'b3 lΔf6 13 J.c4 "b4+ 14 "xb4 J.xb4+ 15 c3 J.f8 16 J.d3! with a roughly level game. Ιη his notes, though, Kasparov criticises his 10th move, giving instead 10 J.f3lΔc5 11 lΔe3 J.e6 12 ο-ο :d8 13 'ii'el as ;1;. This remains Ιο be tested a1though it's hard Ιο see what Black can do against the 10gica1 follow-up, 14 ςPhl and 15 lΔed5. 9 J.g5 J.e7 10 "d2 ο-ο The bishop οη g5 exerts a lοι of pressure οη Black's position, which ηο doubt explains the urge Ιο drive ίι away with 1O...h6, as was played ίη the game J.Polgar-I.Gurevich, Hastings 1992/3. After 11 J.f4 J.e6 12 lΔf5 J.xf5 13 exf5 :c8 14 0-0-0 'ii'aS 15 ςPbl 'i'c5 16 :hel ο-ο 17 J.f1 White's positίon was preferable. 11 ο-ο h6?! Once Black has castled he should leave this pawn well a1one. 11•••J.e6 is more solid, although ίη J.PolgarGelfand, Munich 1991 White stiIl obtained an edge after 12lΔf5 J.xf5 13 :xf5 :c8 (13 ... 'i'd7? 14 J.xf6! J.xf6 15 :xf6 gxf6 16lΔd5 'ii'e6 17 'ii'h6lΔd7 18 :f1 is a typica1 sacrifίce ίη this variation, and one which Black should usua1ly take care ηοΙ Ιο a1low) 14 ςPhl 'ii'b6 15 :bllDed7 16 J.d3 1:tfe8 17 J.e3 'i'd8 18 1:tbf1. 12 J.e3 J.e6

NajdorjVariation 13

13 ΦhΙ Whίte plans lbf5 but must first take this precautionary measure as after 13lbf5 .txf5 14 ':'xf5lbfg4 15 .td4 .tg5 with ... .te3 or ... lbe3 Ιο follow, Black has clearly overcome his opening difficulties. 13 ':'c8 14 lbf5 .txf5 15 ':'ΧΙ5 ':'e8 Ιι is easy Ιο be wise after the event and cήtίcίse this move for weakening f7, but the alternatives are also ηοΙ so palatable. 15.•• lbfg4 16 .tgl .tg5 17 'iί'd4 now gets nowhere as Black can't infiltrate οη e3 and after 15...'ii'd7 16 ':'afl White enjoys a considerable kingside space advantage and will be able Ιο start thinking about punishing Black for his 11 th move. lbh7? 16 .:.an Black wants Ιο exchange the darksquared bishops but this is a hopelessly optimistic plan. 17 'iί'd5! Thanks Ιο Black's last move the white queen is able Ιο take υρ a menacing position, creating threats οη both the kingside and queenside. 17 ... ':'c6 17... b5 is well met by 18 'iί'b7!. 18 'iί'b3 b5 19 lbd5 'ii'a8 (D) Black neutralises the threat of lbb4, but allows something much more devastating. 20 ':'xf7!! lbxf7 21':'xf7 ΦΧf7 21 ....td8 is ηο help at all: 22 ':'a7! 'iί'xa7 23 lbe7+ and 24lbg6#.

22 lbb6+ Φg6 Or 22...Φf8 23 lbd7 and 22...ΦΙ6 23lbd7+! Φg6 24 .th5+!, which is similar Ιο the game. 23 .th5+! Φό5 24 'ii'f7+ ι-ο 24 ... Φh4 25 g3+ is instant mate. Game2 Νοnn - Grίinfeld

England-lsrael Telex Match 1981

1 e4 c5 2 lbf3 d6 cxd4 3 d4 4 lbxd4 lbf6 a6 5 lbc3 6 Ι4 'ii'c7 If Black wishes Ιο delay ... e5 (or even dispense with ίι altogether) then this is probably the most reliable way Ιο go about ίι Black avoids the tactical problems resulting from a quick e5 by White and can continue his development by ... g6• ....tg7.... lbbd7 and maybe ... b5 and ... 0-0 as well. 7 lbf3 (D) This is more accurate than 7 .td3, when 7 ... e5 8 lbf3 b5 transposes ίηΙο

NαjdorjVαriαtion

a relatively comfortable line for Black.

7 ... lbbd7 After this we reach a standard ρο sition which can arise by a wide range of move-orders. The main question is whether Black can exρΙοίι White's early lbf3 by playing 7 ... e6. The analysis runs 7.•.e6 8 .i.d3 and now: 1) 8••• b5 (this is thought Ιο be very risky, but ίι was recently tried by Kasparov) 9 e5! (ίη Renet-Kasparov, French League 1993 White played the less cήtίcal move 9 'ii'e2) 9 ... dxe5 (9 ... b4 10 lbe4 lbxe4 11 .i.xe4 d5 12 .i.d3 is slight1y better for White, while Sax-Guena, Dubai OL 1986continued9... lbfd7100-0 lbc6 11 ~hl .i.e7 12 'ii'el ο-ο 13 'ii'g3 f5 14 exf6lbxf6 15 .i.d2 with some advantage Ιο White) 10 fxe5 lbfd7 (1O... lbg4 11 'ii'e2 .i.b7 12 .i.e4 also gives White an edge) 11 ο-ο lbc6 12 .i.f4 .i.b7 13 ~hllbc5 14lbe4lbxe4 15 .i.xe4lbb4, ReehSchulz, West German Ch 1987, and now 16lbg5!? gives White dangerous attacking chances.

2) 8•••lbc6 9 ο-ο and now:

2a) 9•••.i.e7 10 1We1 and Black has a range of possibilities. 2al) 10•••lbd7 (passive) 11 'ii'g3 ο-ο 12 ~hl :e8?! 13 e5! lbb4 14 f5! with a strong attack for White, Hazai-Karolyi, Hungary Ch 1986. 2a2) 10•••0-0 11 e5 lbd7 (the line 11 ... dxe5 12 fxe5lbd7 13 .i.f4lbc5 14 'ii'g3lbxd3 15 cxd3 ςf,j>h8 16lbe4 is better for White, according Ιο Sax and Hazai) 12lbg5 .i.xg5 13 exd6 'ii'xd6 14 fxg5 lbc5 15 .i.e4 e5 16 .i.xc6 bxc6 17 .i.e3 gave White an edge ίη Kindermann-Schlosser, AItensteig 1992. 2a3) 10•••lbb4 (perhaps the best) 11 e5 lbxd3 12 cxd3 lbd5 13 lbxd5 exd5 14 'ii'g3 ο-ο 15 ΦhΙ dxe5 16 fxe5.i.f5 17lbd4 .i.g6 18 .i.d2 with a level position ίη Sax-De Firmian, New York Open 1987, although nobody has since cared Ιο repeat this with Black. 2b) 9••. b5 10 'ii'el .i.b7 11 ~hl .i.e7 12 e5! dxe5 13 fxe5 lbd7 14 .i.f4lbc5 15lbe4! (just as ίη line , l' above, except for the additional moves "el and ....i.e7) 15 ...lbxe4 (15 ... lbxd3 16 cxd3 ο-ο 17 "g3 ~h8 18 lbf6! 'ii'd8 19lbg5 is very good for White) 16 .i.xe4 h6?! (16 ... lbb4 17 .i.xb7 'ii'xb7 18 "g3 g6 may be better, although 19 .i.h6 cuts Black off from the kingside), Sax-A.Sokolov, Reykjavik 1988, and now 17 a4 b4 18 'ii'f2 gives White good attacking prospects. 3) 8•••lbbd7 9 ο-ο and now: 3a) 9••• b5 may be met by the simpIe 10 "el, or even by 10 e5!? dxe5

NαjdoifVαriαtion

11 fxe5 ~xe5 12 ~xe5 1Wxe5 13 1Wf3 and now 13 ...:b8?! 14 .tf4 1Wc5+ 15 ~hl .tb7 16 .te4! ~xe4 17 .txb8 f5 18 :ael was favourable for White ίη Ni~evski-Markiewicz, Dembica 1987, so Black should have continued 13 ...:a7 14 .tf41Wh5 15 1Wg3 with an unclear position; it is worth noting that the natura1 continuation 15 ....tc5+ 16 ~hl ο-ο loses material to 17 .tb8!, with a double attack οη the rook at a7 and the knight οη f6. 3b) 9 ....te7 10 1Wel ο-ο 11 e5! ~e8 121Wg3 ~c5 13 .te3 .td7?! 14 .txh7+! ~xh715 ~g5+~g8 (after 15 ... .txg5, 16 fxg5 :h8 17 :xf7 is unpleasant) 161Wh4 .txg5 17 fxg5 .tc6, Wedberg-Ionescu, Berlin 1988, and after 18 exd6 ~xd6 (but not 18.....xd619 :adl "e7 20"b4) 19 .txc5 ~f5 Black does not have enough for the pawn. 8 .td3 g6 8 •••e5 9 a4 transposes to Game 3, while 8.••e6 is lίne '3' ίη the above analysis. 9 ο-ο White's strategy ίη this line is rather crude. He intends a straightforward attacking build-up οη the kingside by 1Wel-h4, f5, .th6, and ~g5. Of course Black is a1so playing moves while all this is going οη but if he continues naively with his development without taking specific countermeasures he can easily fa11 victim to White's attack. 9 .tg7 10 1i'el (D) 10 ο-ο

This move is probably a1ready an inaccuracy. The a1tematives are: 1) 10•••e5 and after 11 a4 we haνe traπsposed to Game 3. 2) 10•••~c5 11 e5 dxe5 12 fxe5 ~fd7 13 .tf4 ~e6 14 .tg3 ~b6 (after 14 ... 0-0 15 ~hl Black has to find aπ answer to ~d5) 15 a4 .td7 16 a5 ~c8 17 ~4 and White's ίη itiative proved decisive ίη SaxMinic, Rovinj-Zagreb 1975. 3) 10•••b5 (probably the best move, aiming to complete Black's development before he gives White a target Ιο attack by castling) 11 e5 (attacking moves like 1i'h4 serve ηο function while Black's king is still ίη the centre) 11 ... dxe5 12 fxe5 ~g4 13 e6 fxe6 14 1Wh4 with an unclear ρο sition. For the pawn White has some inίtiatίve and Black has problems finding a refuge for his king. White also has the ορΙίοη of opening lines οη the queenside by a timely a4, and ίη practice Black will not have an easy defensive task ahead of him. The above paragraph remains unchanged from BTS2 as this pawn sacήfice has stίll ηοΙ been seriously tested, mainly because 1O... b5 is

16 NajdorfVariation believed ιο be Ιοο risky. However, Daniel Κing, ίη Winning with the Najdorj (Batsford 1993), examines this line ίη some detail and concludes that White has insufficient compensation for the material. Ι think, though, that he is overestimating Black's chances. For example, he gives 14 ... lbde5 (after 14 "'h4) 15 lbxe5 lbxe5 and then 16 .t.h6 .t.xh6 17 "'xh6 .t.b7 followed by ... 0-0-0 as very good for Black; but after 18 :ael 0-0-0 19 We3lbxd3 20 'iί'xe6+, Svendsen-Gallagher, Lenk 1995, the position is about equal. Moreover, at the board Ι was more concerned about 16 .t.e4 .t.b7 (otherwise Black 10ses his ήght ΙΟ castle long) 17 .t.xb7 "'xb7 18 .t.h6 .t.xh6 19 "'xh6 0-0-0 20 :ael when the assessment from the above paragraph sums υρ this position perfectly. 11 ~4 b5 12 Ι5 Αι one time White invariably played ~hl before proceeding with his attack. This type of consolidating move is often a symptom of chess laziness, ίη that White does ηοΙ want Ιο be bothered with calculating the consequences of Black's queen check ίη every vaήation and so simρΙΥ rules ίι ουΙ, even though ίι may cost him a vita1 tempo. 12 ••• .t.b7? Black continues Ιο play normal Sicilian moves without realising how cήtica1 his position has become. The ροίηι is that after White's fxg6 Black does ηοΙ want ιο play ... hxg6 when lbg5 gives White a permanent

mating threat οη h7. However, the recapture ... fxg6 invites White's knight Ιο come ίη at e6 and Black's ... Jιb7 removes a vital defence from this square. 12...lbcs was essential, when 13 .t.h6 b4 may enable Black Ιο defend. 13 fxg6 Ιη a game Velίkov- Valenti, Pemik 1979, White played 13 .t.e3 (laziness again) when Black missed his second chance Ιο play ...lbc5 and 10st after 13 ... b4? 14lbd5! .t.xd5 15 exd5 lbxd5 16 fxg6 hxg6 17 lbg5 lb5f6 18 %Σf3 with a crushing attack. 13 ... fxg6 After 13...hxg6 14 lbg5 Black cannot move either knight since ... lbh5 is met by g4. White can just build υρ by and :afl Ιο eliminate the defensive knights at f6. 14 lbg5 lbcs (D) Τοο late! 14... 'iί'b6+ (14 ... lbh5 15 .t.e3 is good for White) 15 ΦhΙ lbh5 was best, but even then 16 .t.d2 threatening both lbe6 and lbd5 gives White a promising attack.

:f3

15 hf6! 16 'iVxh7+

hf6 Φf8

NajdorfVαriαtion

17 ~e3 White's material inνestment is νery slight for such a strong attack. The main threat is 18lZ)d5 ~xd5 19 exd5 attacking g6 and preparing b4 followed by the occupation of e6 by White's knight. 17 ... lbxd3 17••• eS 18lZ)d5 ~xd5 19 exd5 e4 20 ~e2 :e8 21 b4 followed by lZ)e6+ is also winning. 18 cxd5 'iνd7 Black cannot meet the threat of lZ)d5 by 18..•e6 since 19 lZ)xe6+! :xe6 20 :f1+ <j;Je8 (20 ...:f6 21 :xf6+ wins the queen) 21 'ίWg8+ is decisiνe.

19 lZ)ds ~xd5 20 exd5 21 lbe6+ %':xe6 22 dxe6 'ii'xe6 23 ~h6 ι-ο 23 ... ~xh6 24 'i!kh8+ ~fΊ 25 :n + wins eνerything.

-'rs

Game3 Nunn-Csema Lugano 1984 1 e4 c5 d6 2 lZ)f3 cxd4 3 d4 lZ)f6 4 lbxd4 a6 5 lZ)c3 e5 6 Ι4 Black's most popular moνe. After all, ... a6 was designed Ιο prepare ... e5 and the determined Najdorf player will generally play ... e5 υη less ίι is absolutely impossible. 7 m 'ii'c7

Αι one time ίι was held that Black should preνent the actiνe deνelopment of White's bishop at c4 and so this moνe was almost uniνer sal, Βυι more recently 7 •••lZ)bd7 has become the most popular moνe. We consider this ίη Game 4. 8 a4 The altematiνe is 8 ~d3. The continuations after 8 a4 and 8 ~d3 are rather similar, but there are some important differences. Firstly 8 a4 expends a tempo, but this is ηοΙ especially seήοus since White can often omit <j;Jh 1 (after 8 ~d3 b5 White usually has Ιο play 'ίthl since Black's check οη b6 giνes him extra defensiνe possibilities). More significantly, a4 reserνes the c4-square for White's use (by 'iί'e2 and ~c4 or lZ)d2 and ~c4) and ίη some lines the added pressure White can exert οη f7 by these manoeuνres improνes his chances considerably. lZ)bd7 8 9 ~d3(D)

9 ••• g6 Black's main decision is whether the f8-bishop should go Ιο e7 or g7.

NαjdoifVαriαtion

There are two other lines, one ίη which Black commits hiιnself Ιο ...~e7 at once and one ίη which he postρones the decision: 1) 9 ... ~e7 100·00·011 tbh4!? (White can also play 11 ΦhΙ and after 11 ... tbc5 we have transposed Ιο Game 4) when Black has played: la) 11...~d8?! 12 tbf5 g6 13 tbh6+ Φg7 14 f5 b6 15 g4 with a dangerous attack for White, SaxBukic, Vrbas 1980. lb) 11 ... g6 (the usual response Ιο tbh4) 12 f5 d5!? 13 exd5 with a further branch: lbl) 13... tbxd5? 14 tbxd5 'ifc5+ 15 ~e3 'ifxd5 16 fxg6 hxg6, Cramlίng-Gallagher, Oakham 1984, and now instead of 17 tbxg6? fxg6 18 'ifg4 %:ιχf1 + 19 %:ιχf1 tbf8! when Black was able Ιο repulse the attack, White could have won by 17 'ii'g4!, e.g. 17 ... tbf6 18 tbxg6 ~xg4 19 tbxe7+ Φh8 20 tbxd5 tbxd5 21 ~h6, 17 ... ~c5 18 tbxg6 ~xe3+ 19 ΦhΙ 'ife6 20 ~f5, 17 ... tbb6 18 'ifg3 and White has threats at b6 and g6, or finally 17 ... ~xh4 18 'ifxh4 followed by ~c4 and ~h6, and ίη every case White has a winning position. lb2) 13...e4 14 ~e2 ~d6 15 g3 b6, Sax-Andersson, London 1980, and now 16 tbg2 is best, followed by ~h6 and tbe3 supporting the pawns at d5 and f5, when White should have the advantage. lc) 11...d5!? 12 tbf5 (12 exd5 e4! 13 ~e2 'ifc5+ 14 ΦhΙ tbb6 is fine for Black) 12...~c5+ 13 ΦhΙ dxe4 14 ~xe4 ~b4 15 fxe5 ~xc3 (15 ... tbxe5 16 tbd5 tbxd5 17 'ifxd5

appears Ιο be a lίttle better for White, whilst the hackers can investίgate 16 tbxg7!?) 16 exf6 tbxf6 17 ~d3 ~e5 18 ~g5 with some pressure for White, Mortensen-Zso.Polgar, Hastίngs

1992/3.

2) 9...b6 (the problem with this delaying move is that Black may have trouble getting castled) 10 0-0 ~b7 11 'ifel g6 (11 ... ~e7 12 ΦhΙ ο-ο 13 %4 g6 14 fxe5 dxe5 15 ~h6 :e816tbf5! ~f817 ~χf8%:ιχf818 tbe3 is slightly better for White, Sznapik-Ostermeyer, Oslo 1983) 12 fxe5 dxe5 13 'ifh4 ~e7 (13 ...~g7 14 ~h6 ο-ο transposes ίηΙο the maίn line) 14 ~g5 h6 15 ΦhΙ (15 tbd2? ~c5+ and 16... tbh5) 15 ... Φf8 (or 15 ... 0-0-0 16 ~e3 and White has an automatic attack agaίnst Black's weakened queenside) 16 tbd2! Φg7 17 ~e3 tbc5 18 ~c4 lIaf8 19 'iff2 ~d8 20 ~xc5 'ifxc5 21 'ifxc5 bxc5 22 a5 1/2-1/2 Rantanen-Nunn, Helsinki 1981, although White is distinctly better. This is an example of the advantage of having c4 avaίlable for the bishop. ~g7 10 0·0 11 'ii'el 0·0 12 fxe5 dxe5 13 1t'h4 b6 14 .th6 ~b7 15 tbg5 (D) This position is the natural result of White's blunt play. Although its evaluation has f1uctuated over the years, the scales have recently tipped ίη White's favour. It is very rarely seen today because Black players steer well clear of ίι

NαjdorfVαriαtion

15 000 :ΙcS The two lines ISoootDhS? 16 .ιχg7 ~xg7 17 :xf7+ and 15ooo:&e8 16 g4! are best avoided, while 15000:Ιe8 16 .ιχg7 ~xg7 17 tDxf7! ~xf7 18 'iWxh7+ ~e6 19 :xf6+! ~xf6 20 :f1+ ~e6 21 .ιc4+ ~d6 22 :dl+ ~c6 23 .ιd5+ ~c5 24 .ιΧb7 1-0 Rantanen-Monis, Gausdal 1978, was a devasιatίng win for White. The οηlΥ other reasonable move is IS.o:ii'd6, but 16 :adl causes serious problems. After 16...tDh5 White plays 17 .ιχg7 Φχg7 18 .ιe2 'iνc5+ 19 ~hl, when 19 ... tDhf610ses Ιο 20 :xd7 and 19 ... tDdf6 loses Ιο 20 .ιχh5 tDxh5 22 :xf7+. Other 16th moves are almost as bad. 16 ~hl Black intended Ιο meet moves such as 16 g4 and 16 :adl with ... 'iWc5+ and ... 'iWf8 whereafter, with the rook οη c8 preventing .ιc4, he should be able Ιο defend his kingside. Unfortunately White has a tactical idea which cuts across Black's plan Ιο bήηg his queen Ιο f8. 'iνd6 16 17 .ιχg7 ~xg7 18 tDxt7!

This hardly counts as a sacrifice, since White immediately gains three pawns for the piece, while Black's king is left floating around ίη the middle of the board. <i;xrι 18 19 'iνύ7+ ~e6 20 'iνxg6 (D) There was even a second good line ίη 20 :xf6+ tDxf6 21 'iWxb7 since the attempt Ιο liquidate by 21 ... 'iWc7 allows 22 .ιχa6.

20 ο.. 'ii'e7 21 :adl As is so often the case, it is better ιο spend time cutting off the king's escape route than Ιο give pointless checks which only serve Ιο dήve the king ίηΙΟ safety. 'iνh7 21 .. ο 22 'iνg3 ~e7 Black cannot play 22.. o:g8 because of the check at c4. 23 tDdS+ .ιχdS 24 exd5 'iνh6? This enor allows White Ιο force the king οηΙο the back rank, cutting off both black rooks from the kingside. However even the best line

Nαjdoιf Vαriαtion

24•.•e4 is very good for White after 25 ':'del ':'g8 26 'ii'c7! 'ii'g6 27 g3 1:tac8 28 d6+ ~e6 29 'ii'b7, and with the fall of the e4-pawn White 's rooks can at last get Ιο gήΡS with the black king. 25 d6+ ~d8 26 ~Ι5! 1%a7 Ιι is hardly surprising that Black has ηο reasonable move. The immediate threat is 27 .txd7 ~xd7 28 'ii'xe5. 27 ~xd7 ':'xd7 28 ':'ΧΙ6 1-0 Game4 Nunn-Κing Bundesligα

19860

1 e4 c5 2 lDf3 d6 cxd4 3 d4 4 lDxd4 lDr6 5 1Dc3 a6 6 Ι4 e5 7 lDf3 lDbd7 Ιη my (JN) view this is Black's best reply Ιο 6 f4. Although White's bishop can now be developed more actively at c4, Black saves a vitaI tempo by missing ουΙ .. :flc7 and this gives him good chances Ιο equaIise. Ιη fact cunent theory suggests that White's best plan is Ιο ignore the ορ ιίοη Ιο play .tc4, and Ιο proceed with his normaI development by .td3. Admittedly Black benefits from missing ουΙ .. :flc7, but ίι is ηοΙ clear that the alternatives Ιο .. :flc7 fully equaIise. 8 a4

White cannot do without this as 8 .tc4 allows 8 ... b5 9 .td5 1:tb8 10 lDg5 (after 10 fxe5 dxe5 11 .tg5

.tb4 Black was slightly better ίη Hort-Andersson, Wijk aan Zee 1979) 10... lDxd5 11 'ii'xd5 'ii'e7 12 ο-ο h6 with equaIity, Korchnoi-Hort, Zurich 1984, while 8 .td3 aIlows 8 ... b5, just the line White is trying Ιο avoid. 8 .•. .te7 If Black relents by 8.....c7, we reach Game 3. 8...d5 enjoyed a bήef spell of popularity a few years ago but has now virtuaIly disappeared from practice. The best line for White seems to be 9 exd5 e410 lDe5.tb4 (10 ....tc5 I1lDc4 followed by .te3 is good for White) 11 .tc4 when Black has: 1) 11•••lDb6 12 .tb3 when both 12... lDbxd5 13 ο-ο .txc3 14 bxc3 ο-ο 15 'ii'd4 and 12••• lDrxd5 13 a5 lDxc3 14 'ii'xd8+ ~xd8 15 lDxf7+ Φe7 18 bxc3 .txc3+ 17 ~f2 are good for White. 2) 11 •.•0-0 12 ο-ο 'ii'c7 13 ~hl! lDxe5 14 fxe5 .txc3 15 bxc3 'ii'xe5 16 'ii'd4! 'ii'xd4 17 cxd4 with a better ending for White, Reeh-Gallagher, Κecskemet 1990. 3) 11 •••lDxe5 12 fxe5 'ii'c7 13 ο-ο! 'ii'xc4 (13 ... .tg4 14 .te2 .txe2 15 'ii'xe2 'ii'xe5 16 .tf4 is good for White) 14 exf6 gxf6 15 1:txf6 .txc3 16 bxc3 'ii'xc3 17 .tg5 and White was clearl Υ better ίη Adams-Van der Wiel, Wijk aan Zee 1991. 9 .td3 ο-ο 10 ο-ο (D)

NαjdorfVαriation

10 ... Grabbing a pawn with 10...exf4 1 1 ~xΙ4 "b6+ 12 ~hl "xb2 is a risky business, but that hasn't prevented it becoming a more and more frequent visitor to the tournament hall over the last few years: 1) BTS2 recommended 13 "el (still the main line) after which Blackhas: la) 13... ω 14 :bl lbxd3 15 cxd3 'ii'c2 16 d4 :e8 17 :f2! 'ii'd3 18 :cl ~f8 19 :e2 with advantage to White, Ciocaltea-Danner, Timisoara 1982. lb) 13.....b4 14 :bl 'ii'c5 15 lbd5lbxd5 16 exd5 ~f6 (16 ... ~d8 17 c4lbf6 18 ~e3 "c7 19 ~d4 gave White a dangerous attack ίη Ledermann-Lau, Ramat-Hasharon 1982) 17 c4 "fic7 18 'ifg3! lbe5 19 ~g5 ~xg5 20 ~xh7+! ~xh7 21lbxg5+ ~h6 (21 ... ~g8 22 'ifh4 :e8 23 :bel also gives White a winning attack) 22 1i'h4+ ~g6 23 :b3 f5 24 :g3 lbg4 25 :xg4 fxg4 26 :xf8 "ile7 27 :f7 "ile8 28 'ii'h7+ ~xg5 29 ::txg7+ ~f6 30 ~6+ ~f5 31 'ii'g5+ ~e4 32 :e7+ 1-0 Vogt-Womacka, E.Germany 1989.

lc) 13...'ii'b6 (thought to be best) andnow: lcl) 14 ~g5!? is the latest try: lcl1) 14.....d8?! 15 'ifh4 :e8 16 e5! dxe5 17 ~xf6lbxf6 18 lbxe5 ~e6 19 :ael 'ifa5 20 lbe4 and now 20••• h6? 21 lbxf6+ ~xf6 22 :xf6! gxf6 23 :e3! 'ifxe5 24 :xe5 fxe5 25 'ifxh6 left White close Ιο victory ίη the game Κveinys-Van Wely, Moscow OL 1994. However, it's doubtful whether White would have been able Ιο claim any advantage after 20...lbd5!. lcl2) Van Wely suggests the altemative 14...lbc5!, intending ...~e6 and providing the option of ... lbxd3. lc2) 14lbd5lbxd5 15 exd5 'ii'd8 16 c4 lbc5 17 ~c2 ~g4 18 'ii'g3 ~xf3 19 :xf3 ~h4 20 'ii'h3 g6 and now the best continuation, according Ιο Am. Rodriguez, is 21 ~h6 ~f6 22 :af1 ~g7 23 g4! (Jepson-Am.Rodriguez, Mondariz Balneario 1994) 23 ... 'ii'e7! 24 :e3! ~e5 25 ~xf8 :xf8 when Black has compensation for the exchange. 2) 13lbd5!? lbxd5 14 exd5 'ii'b6 15 "ile2! (15 'ii'el "fid8 transposes Ιο 'lc2', whilst now 15 ... "fid8 is met by 16 :ael as 16 ...:e8 allows 17 ~xd6) 15 ... ~f6 16 :abl "fic5 17 c4 'ii'c7 18 ~g5 'ii'd8 19 "fie4! g6 20 'ii'h4 h5 21 ~xf6 'ii'xf6 22lbg5 1i'e7 23 :bellbe5 24 c5! with very dangerous pressure for the pawn, Κίη dermann-Fta~nik, Pardubice 1994.

~hl

This is not the only reasonable move, but judging by recent results it is the most dangerous for Black.

22 NajdoifVariation

11 ... dS This is the tactical justifιcation of Black's play, but there are ηυί eter altematίves: 1) 11 ...~d3 12 cxd3 "a5 13 "el exf4 14 lΔd5 'ii'd8 15 lΔxf4 i.d7 16 i.d2 was good for White ίη Mateo-Byrne, New York 1986. 2) 11 ... 'ii'c7 12 'ii'el (D) (threatening 13 fxe5 dxe5 14 "g3, when Black has ηο natural way Ιο defend the e5-pawn) and now:

2a) 12••. i.d7 13 fxe5 dxe5 14 'ii'g3 J:ae8 15 "xe5 i.d6 16 "d4, with inadequate compensation ίη view ofWhite's control ofd5. 2b) 12••• J:e8 13 fxe5 dxe5 14 "g3 i.d8 15 lΔh4 ~h8 16 lΔf5 i.xf5 17 J:xf5lΔxd3?! (17 ...'ii'c6 18 i.e3 lΔxd3 19 cxd3 "d7 20 J..g5 lΔg8 ;t is a more accurate continuation) 18 cxd3 'ii'd7, Almasi-Novikov, Cattolίca 1993, and now instead of 19 i.g5lΔg8!, transposing ιο the note just above, White could have gained a more serious advantage with 19 'ii'f3! followed by J..g5. 2c) 12•••i.e613 Ι5 i.d714 g4!? (14 i.g5 i.c6 15 J..xf6 i.xf6 16 g4

is also slightly better for White) 14•••i.c6 (14 ... lΔxg4 10ses after 15 lΔd51fd8 16 J:gllΔf617 J..h6lΔe8 18 J:xg7+ lΔxg7 19 1fg3 J..f6 20 J:gl with a winning attack) 15 g5 lΔh5 (retreating ιο d7 gives White a completely free hand οη the kingside, e.g. 15 ... lΔfd7 16 f6 i.d8 17 1fh4lΔe6 18 lΔd5 and wins) 16 Ι6 i.d8 (not 16 ... gxf6? 17 'ii'h4 winning) 17 'ii'h4 g6 18 i.e3. White's kίngside attack has come to a temporary halt, so the time has come Ιο bήηg the remaίning pieces ίηΙο play. There is a positional threat of 19 J..xc5 dxc5 20 J..c4, followed by the occupation of d5. Thus the c5-knight mustmove. 2cl) However, 18••• lΔe6 19lΔd5 i.xd5 20 exd5lΔf4 21 i.xf4 exf4 22 J..f5! followed by i.g4 is good for White. 2c2) 18•••lΔxd3 19 cxd3 ~h8 20 lΔe2 and now: 2c21) 20•..i.d7, planning ...1fc2; is refuted by 21 J:fcl when 21 ...1fa5 surprisingly loses Ιο 22 J:c4! and Black has ηο defence Ιο 23 b4. 2c22) 2O•.•'ii'd7! is the best defence, aiming for counterplay by attacking a4, when White should continue 21 lΔg3 i.xa4 22 lΔxh5 gxh5 23 'ii'xh5 J..b5 24 J:a3 and Black still has Ιο find a defence Ιο the threat of J:gl-g4-h4. 2c23) 20.••d5? (a natural move aίming ιο weaken e4 and activate the c6-bishop agaίnst the white kίng, but ίι also weakens the important e5pawn and this tums ουΙ Ιο balance White's weak spot at e4) 21 lΔg3

NαjdorfVαriαtion

dxe4 (Black cannot play 21 .....d6 due Ιο 22 d4!, when both 22 ... dxe4 23 lbxe5 and 22 ... exd4 23 e5 followed by ~xd41eave the c6-hl diagonal blocked by a black pawn) 22 dxe4 "d6 23 ':adl! "b4 (23 .....e6 24 ~c5 ':g8 25 ~d61eads Ιο the loss of e5) 24 lbxe5 lbxg3+ 25 hxg3 "xe4+ (after 25 ... ~xe4+ 26 ΦgΙ ~b6 27 ~xb6 "xb6+ 28 ':f2 Black cannot meet the threats of "xe4 and "h6) 26 "xe4 ~xe4+ 27 Φh2 (the immediate threat is 28 ~c5; both 27 ... ':c8 28 :d7 Φg8 29lbg4! heading for h6 and 27 ... Φg8 28 ~c5 ':e8 29lbxf7! ΦΧf7 30 ':d7+ Φe6 31 f7 ':f8 32 ':d4 ':xf7 33 ':d6+ win for White) 27 ... ~c7 28 ~c5 ~xe5 29 ~xf8 ':xf8 30 :fel ~c2 31 ':d2 ~xa4 (Black cannot get two pawns for the exchange since 31 ... ~xg3+ 32 Φχg3 ~xa4 33 ':e7 threatens both ':xb7 and ':xf7) 32 ':xe5 h633 gxh6 Φh7 34 g4 ΦΧh6 35 g5+ Φh7 36 ':e4 ~c6 37 ':h4+ Φg8 38 Φg3 1-0 Nunn-Portisch, Brussels 1986. 2d) 12•••exf4 13 ~xf4 ':e8 (after 13 ... ~e6 14lbd4 "b6 15 ~e3lbg4 16 ~gllbe5 17lbf5 ~xf5 18lbd5 "d8 19 exf5 ~f6 20 ~e2 lbed7 21 lbxf6+ lbxf6 22 ~f3 White's two bishops gave him the advantage ίη Short-Gallagher, Βήtίsh Ch 1987) 14 lbd4 ~d7 15 .tg5 "d8, HazaiNovikov, Camaguey 1987, and now White should have taken the chance Ιο activate his bishop by 16 ~c4!, pointing ίι at the sensitive square f7. The key tacticalline 16 ... lbxa4 17 ':xa4 ~xa4 (17 ...b5 18lbcxb5 axb5 19 ':xa8 "xa8 20 .txb5 is good for

White) 18lbxa4 b5 19lbc6 "c7 20 lbxe7+ "xe7 21 ~xf6 gxf6 22lbb6 bxc4 23 lbd5 "e5 24 lbxf6+ Φg7 25"h4 h6 (25 ...':h8 26 "g4+) 26 ':f5! turns ουΙ well for White since 26 ... 'it'xb21oses Ιο 27 :g5+!. 3) 11 ..•exf4 12 ~xf4 (D) with two main possibilities:

3a) 12••• ~d7 13 'iνe2 ':c8 14 a5 ':e815 ~e3 ~f8 (perhaps 15 ...lbg4 is better, although there was ηο need for White Ιο agree a draw after 16 ~d4 ~f6 17 'iνd2, as he did ίη Κiη dermann-Brunner, Eurodata 1992) 16 .td4! (intending lbg5) with a total of five lines: 3al) 16••• ~e717b4lbe618~b6 with a clear plus for White. 3a2) 16•••h6 17 lbd2 lbg4 18 ~c4±.

3a3) 16•••lbe6 17 .tb6 "e7 18 lbh4 g6 19 ':ael ±. 3a4) 16•••~g4 17 "e3 ~h5 18 .txf6 "xf6 19 lbd5 "d8 20 lbd4 ~ι6 21 lbf5 ':e5 22 b4! lbxe4 23 ~xe4 ':c4 24 ':ael ':cxe4 25 "xe4 ':xe4 26 ':xe4 was very promising for White ίη Ulybin-Pigusov, Pavlodar 1987.

NαjdorjVariαtion

3a5) 16...lQfxe4 (Ulybin's recommendation ίη Informαtor. but it appears to have a tactical flaw) 17 ~xe4 ~xe4 18 .txe4 .tb5 19 c4! (Ulybin οηlΥ considered 19 'ii'e3. which leads to a draw) 19 ....:xc4 (or 19 ... .txc4 20 'ii'c2 .txf1 21.txh7+ 'iti>h8 22 'ii'f5 with a winning attack) 20 'ii'd3 ':'c5 21 .txh7+ 'iti>h8 22 'ii'b3 .txf1 23 .txc5 dxc5 24 ':xf1 ~xh7 25 'ii'xf7 with a large advantage for White. 3b) 12••..tg4!. with the idea of traηsfeπίηg the bishop Ιο g6. from where it will exert pressure οη e4 whίlst still cοveήηg the important f5-square. makes it diffίcult for White to achieve any advantage. Perhaps the best try is 13 .te3!? (the ideal square for this bishop is d4): 3bl) 13•...th5? is a serious eποr οη account of 14 .txc5! dxc5 15 e5 ~8 16 'ii'el ±. 3b2) Better is 13•••.:c8 14 'ii'd2 and now 14••..th5 15 ~d4 ~g4 16 ~d5 (thus far Svidler-Nebodora, Linares 1994) 16... ~xe3 17 'ii'xe3 permits White a slight advantage. but 14•.. ~d3! 15 'ii'xd3 (15 cxd3 d5 16 e5 .txf3 17 exf6 .txf6! 18 ':xf3 d4 19 ~e4 dxe3 20 ~xf6+ 'ii'xf6! slightly better for Black) 15 ....te6 is given as =ίη Informαtor. although Ι believe White may have a faint edge after 16 ~d5. 12 ~e5 (D) 12 ..• ~fxe4 Black has two important alternatives: 1) 12..• ~cxe4 13 .txe4 dxe4 14 'ii'e2andnow:

la) 14..•.tf5 15 g4.tc8 16 ':'dl 'ii'e8 17 g5 lZ:Id7 18 lZ:\c4 e3 (the lines 18 ... lZ:Ic5 19 b4 lZ:Ie6 20 lZ:Id5 and 18 ... b619.te3 .tb7 20~d5 arealso good for White) and now there are two favourable vaήations for White. either 19 .txe3 b5 20 axb5 .tb7+ 21 ~gl axb5. Kengis-Loginov. Pavlodar 1987. and now 22 ':xa8 .txa8 23 lZ:Ixb5 'ii'c8 24 lZ:Icd6 'ii'c6 25 ~f2 leaves Black with inadequate play for the two pawns. or the simple 19 'ii'xe3. lb) 14•..'ii'd4 15 ':'dl 'ii'b4 16 a5' .td8 17 ':a4 'ii'e7 is Loginov's suggestion ίη Informαtor. but now 18 b3! appears good for White. 2) 12••.dxe413 .te2 (Black gains time. but the pawn οη e4 obstructs Black's pieces) 13...'ii'c7 (13 ... 'ii'xdl 14 ':xdl .te6 15 .te3 ':fd8 16 g4 g6 17 g5 lZ:Id5 18 lZ:Ixd5 .txd5 19 b3 ~6 20 lZ:\c4 .txc4 21 .txc4 .tc5 22 .txc5 lZ:Ixc5 23 'iti>g2 ':ac8 24 'iti>f2 was marginally better for White ίη Kindermann-de Firmian. Biel1986. but 19 b4 lZ:Ie6 20 c4 .tc6 21 lZ:Ixc6 bxc6 22 ':'xd8 ':xd8 23 c5 looks more dangerous) 14 .te3 (White may also play 14 'ii'el fίrst. so as Ιο

NαjdorjVαriation

meet 14 ... lDe6 by 15 -tdl attacking e4) 14...b6 151νeΙ -tb7 161νg3 and now: 2a) Beliavsky-Chandler, Vienna 1986 continued 16•••lbe6?! (intending Ιο exchange bishops by ... -tc5, but this plan fails tactically) 17 ':adl (ηοΙ 17 f5 -td6!, but now f5 is a seή ous threat) 17 ... -tc5? (an eποr, but Black's position was uncomfortable ίη any case) 18 f5! -td6 (the ροίηι is that 18 ... -txe3 19 fxe6 fxe610ses Ιο 20 ':xf6! ':xf6 21 ':d7) 19 ':xd6 1Vxd6 20 fxe6 and White won. 2b) Black tried 16....:ad8 ίη Psakhis-Balashov, Irkutsk 1986, but after 17 ':adl ι!bcd7 18 -td4lbxe5 19 fxe5lbd7 20 b3! (20 -txa6! was even stronger) followed by -tc4 White had a decisive attack. Black should have played for exchanges by 17....:xdI18 ':xdl ':d8, but White is still slightly better. 13 -txe4 dxe4 14 lbd5! (D) White played 14 -te3 ίη Beliavsky-Portisch, Tilburg 1986 (acouple of weeks after the main game), but then 14 ... f6 15 -txc5 -txc5 16 lbxe4 'iWxdl 17 ':axdl fxe5 18lbxc5 -tg4 19 ':del exf4 led to equality. The idea of 14 lbd5 is to eliminate the e7-bishop; Black's remaining bishop will be obstructed by the e4pawn, whίle White's can become active along the b2-g7 diagonal. 14 ... -te6 Or: 1) 14...-td6 15 lbc4! and Black has immediate difficulties since the natural developing move 15 ... -te6

loses a piece Ιο 16lbxd6. Otherwise White can proceed with lbdb6, οι b4 followed by -tb2. 2) 14...f5 15 b4 lbd7 was suggested by Busch and Olthof ίη New in Chess, but 16 -te3! is promising for White. Then both 16 ... lbf6 and 16... lbxe5 17 fxe5 -te6 lose mateήal Ιο lbxe7+ and -tc5. 3) 14...f6! 15 lbxe7+ 1νχe7 16 lbc4 -te6 17 lDe3 f5 leads Ιο a structure similar to the game, except that Black has managed Ιο play ... f5. White can continue with b3, -tb2 and 'iWel, when his bishop is more effective than Black's. However, the opposite-coloured bishops will exert a drawish tendency, particularly if Black can exchange knights by means of ... lbd7-f6-d5. 15 lbxe7+ 'fIxe7 16 f5 This is the difference between 14 ... -te6and 14 ... f6!. Playing f4-f5 benefits White ίη three ways. First of all, ίι increases the scope of his bishop; secondly, it prevents Black supporting the e4-pawn with another pawn, and thirdly ίι gives White kingside attacking chances.

NαjdorjVariαtion

16 Ι6 17 4Jg4 17 fxe6 fxe5 18 ~e3 4Jxe6 19 'iί'd5 may give Whiιe a minute advantage, but the move played is far more combative. 17 ~f7 The reply 17ooo~c4 is ineffective after 18 Af4, and the bishop will soon be driven away by b3. 18 1i'el?! Ι decided that ίι was time Ιο start developing my queenside pieces, but Ι should have spent just one more tempo improving my position by 18 a5 !. Ιι looks strange Ιο put a pawn οη a dark square when White' s plan is Ιο block out Black's bishop using the pawns οη c2, b3, a4, e4 and f5, but ίι is very useful Ιο have the ορΙίοη of attacking the knight οη c5. Νοι only may White push it away by b4 at a later stage, but by preventing ... b6 White can also set up an awkward ρίη by ~a3. Moreover, the possible elimination of this knight gives White the οριίοη of playing for the win ofthe e4-pawn by 'iVel-h4, ~a3 and Aael. 18 aS! Black seizes οη the mistake and permanently secures the c5-square for his knight. 19 b3 Afd8 20 ~b2 Ad6 21 1i'g3 rJth8 22 Aael b6 23 1i'h4 Af8 (D) Ι decided Ιο make as much progress as Ι could without taking. any risk, by trying Ιο aπange a position 000

000

with ~c3, 4Jf2 and Ae3. This controls the d-file entry squares, and the bishop is defended so Black has ηο tricks. 24 Ae3?! White tries a little tήck; perhaps Black won't notice the threat of 25 Ah3 ~g8 264Je5. 24 1i'd8 Unfortunately he does. Moreover, now ίι is hard Ιο prevent a black rook invasion. 2S rJtgl Avoiding immediate back-rank problems and future long-diagonal troubles after a possible g4. Note that Black cannot initiate counterplay by 2Soo.Ad2 because of 264Jxf6 ~g8 27 4Jh5!. 2S ~g8 Now, however, 26 ... Ad2 is a tbreat. 26 ~c3 White must prevent ...Ad2, even if he thereby allows the altemative penetration ... Adl. Adl 26 Ad6?! 27 Aeel If cοπectlΥ followed up, this is a perfectly reasonable defence, but the 000

000

NαjdorjVαriαtion

alternative 27...':'xel 28 ':'xel 'ifd6 29 lΔe3 "c6 is safer. Swapping a pair ofrooks reduces White's attacking chances, and ίη this case Black would not have many problems drawing. If White had played less carelessly at move 24 (e.g. by 24 ~c3), then this possibility would not have existed. 28 till2 Sometimes an oversight is the best chance to win! Ιι seemed to me that this was the moment Ιο set υρ a favourable position with lΔf2 and J:te3, but this move should fail for tactical reasons. "cS! 28 ••• Since Black obviously cannot play ... 'ifxf5 because oflΔxe4, Ι decided Ιο continue with my plan. There was nothing better ίη any case, since 29 'ifg3 ':'c6 is awkward. 29 ':'e3 After having made this move, Ι suddenly noticed that Black could play 29.....xr5 30 lΔxe4 'ife6! with a certain draw, but perhaps the confident way Ι had made the move led my opponent Ιο believe that the pawn was invulnerable. 29 ':'d5? (D)

30 g4!

White ηοΙ only achieves the posihe has been aiming for, but ίη a very favourable form, since the rooks οη d5 and f8 are both vulnerable Ιο the manoeuvre lΔh3-f4-g6. ιίοη

30 ...

"c7

Perhaps Black could have offered more resistance, but the twin possibilities of lΔh3-f4 and g5 make his position very unpleasant. 31 lΔh3 lΔd3 32 cxd3 'iνxc3 33 lΔt'4 g5 34 lΔxd5 ~xd5 35 "el Thanks to the earlier ΦgΙ, Black has ηο real counterplay and he is soon forced Ιο give up. 35 'iνd4 36 dxe4 ~xe4 37 "d5

"c3

38 "c4

1-0

2 Scheveningen Variation This line is popular with many ofthe world's leading players, including Kasparov, and so one would hardly expect there ιο be a clear way for White ιο obtain an advantage. The characteristic feature of the Scheveningen is Black's pawn centre at d6 and e6 covering all the central squares οη Black's 4th rank. Thus Black avoids the slight weakness at d5 inherent ίη the Najdorf and Dragon systems. Behind the cover of his modest but solίd pawn centre Black intends ιο complete his development ίη peace. The most common move order for Black ιο adopt ifhe is aiming for a Scheveningen is 1 e4 c5 2 lί)f3 e6 3 d4 cxd4 4 lί)xd4 lί)f6 5 lLJc3 d6, but Black can invert his second and fifth moves ίη this line. The system we are recommending, the Keres Attack, is undoubtedly the line Scheveningen players fear most. Ιη fact they fear ίι so much that many of them are now trying ιο reach their beloved Scheveningen via a Najdorf move order (i.e. 2 ... d6 and 5 ... a6 and against moves such as 6 f4, 6 ~e2 or 6 ~e3 they reply 6 ... e6, whereas real Najdorf players prefer ... e5). The Keres Attack starts 1 e4 c5 2 lί)f3 e6 3 d4 cxd4 4 lί)xd4 lί)f6 5 lί)c3 d6 6 g4. This kingside pawn push aims firstly ιο drive the knight from f6, thereby making ίι

harder for Black ιο break open the centre by ... d5, and secondly to gain space οη the kingside and dissuade Black from castling there. Although White usually obtains good attacking chances with this system he must ηοΙ neglect his development since Black can often break the position open by ... d5 and even ifthis loses a pawn White can find his own king stuck ίη the centre. Ιη other words, a balance must be struck between furthering Whi te' s own kingside ambitions and restraining Black ίη other sectors of the board. There are two major options for Black after 6 g4. He may either prevent the further advance of the pawn by 6 ...h6, as covered ίη Games 5 and 6, or he may continue his own plans and allow the knight ιο be driven back Ιο d7. Ιη this case Black may choose 6 ... a6, 6 ... ~e7 or 6 ... lί)c6. The specific choice of sixth move may ηοΙ be very important because there are many transpositions. Ιη Game 7 we deal with lines specific Ιο 6 ... a6 (i.e. those involving an early ... b5), ίη Game 8 we deal with those specific Ιο 6 ... lLJc6 (i.e. those ίηνοlνing an early ... 'iνc7), and the other lines may be found ίη Game 9. Apart from these two major options, there is a third possibility for Black, namely Ιο counterattack by 6 ... e5.

Scheveningen With this move Black 10ses a tempo, but he hopes Ιο prove that White's g4 has οηlΥ served Ιο weaken his kingside. This is covered ίη Game 7. Game5 - Bischoff Lugano 1986

Νυηη

1 e4 2 lbf3 3 d4 4 lbxd4 5 lbc3 6 g4

cS e6 cxd4 lM6 d6 h6(D)

Variαtion

:gl and g5, dήving away the knight after all. 7 h4 a6!? Although this is a natural move, ίι οηlΥ became popular ίη the 1980s. The main variation is 7...lbc6, and this will be examined ίη Game 6, but there is one other important possibilίΙΥ, namely 7... ~e7 (7 ... e5 8 lbf5! ~e6 9 g5lbxe4 10 lbxg7+ ~xg7 11 lbxe4 d5 12 gxh6 ~xh6 13 ~xh6 :xh6 14 'iνd2 1:Ixh4 15 ~b5+ lbc6 16 0-0-0 was very good for White ίη Stanciu-Vegh, Ulan Bator 1986) 8 'iνf3 (D) and now:

Black avoids having his knight away from f6. For a 10ng time White almost always played 7 g5 hxg5 8 ~xg5, Anatoly Κarpoν being one supporter of White's ροίηι of view. Although this continuation gives White a lead ίη development it has defects, ηοΙ the least being that Black's rook οη h8 is activated and presses down οη White's weak hpawn. Now 7 h4 is considered Ιο give White better chances than 7 g5 and is cuπeηtΙΥ the most popular line. White wants Ιο continue with dήνeη

1) 8...lbc6?! 9 ~b5 ~d7 10 ~xc6 bxc6 (10 ... ~xc6 11 g5 is also good for White) 11 g5 hxg5 12 hxg5 :xh 1 13 'iνxhl lbg8 14 'ifh7 Φf8 15 f4, with advantage Ιο White, HowellTaher, Dublin 1991. 2) 8...lMd7 9 'iνg3 lbc6 10 ~e3 a6 11 0-0-0 'iνc7, Ljubojevit-Timman, Brussels SWIFf 1986, and now Ljubojevic recommends 12 ~e2 as slightly better for White. 3) 8... g6!? 9 g5 hxg5 lO~xg5 a6 11 0-0-0 e5 12lbde2 ~g4 13 "g3 lbbd7 and then 14 Ι3 ~e6 15 ~h3

30 Scheveningen

Vαriαtion

.i.xh3 16 ':xh3 ':c8 17 f4 was unclear ίη the garne De Wit-Oll, Groningen 1984/5. This interesting idea does ηοΙ seem Ιο have been repeated. Perhaps 14 Ι4 is better, hoping Ιο prove that the exposed position of the g4-bishop is a lίabίlity. 4) 8...h5 9 gxh5 and now: 4a) 9•••llJc6 10 .i.b5 (an attempt Ιο exploit Black's move-order; 10 ~xc6 bxc6 11 .i.g5 may be better, when Black has Ιο prove that he has something better than 11 ... ~xh5 transposing Ιο line '4b') 1O....i.d7 11 .i.xc6 bxc6 12 e5 ~d5?! (12 ... dxe5 13 ~xc6 .i.xc6 14 "xc6+ Φf8 15 h6 gxh6 16 .i.d2 ':b8 170-0-0 "b618 "xb6 ':xb6 is equal according Ιο Ljubojevic) 13 exd6 .i.xd6 14 .i.g5 "b6 15 0-0-0 .i.e5 16 ~xd5 cxd5 17 c3 with an edge for White, Ljubojevic-Timman, Bugojno 1986. 4b) 9...~xh5 10 .i.g5 (10 .i.e3!? is an interesting untested idea, offering the h-pawn ίη return for a quick attack with 0-0-0) 10... ~c6 11 ~xc6!? bxc6 120-0-0 .i.xg5+?! (accepting the sacrifice turns ουΙ Ιο be Ιοο risky; Black should develop by 12 ... ':b8) 13 hxg5 "xg5+ 14 ΦbΙ Φe7 (ηοι 14 ... d5? 15 exd5 cxd5 16 ~xd5 exd5 17 ':xd5 nor 14 .....c5? 15 e5! and White wins ίη both cases, whίle 14 .....e5 15 .i.e2 g6 16 "e3 intending f4 gives White a dangerous initiative) 15 .i.e2 g6 16 ':xd6! ΦΧd6 17 "χΠ! (somewhat surprisingly Black has ηο defence) 17 ... a5 18 ':dl + Φe5 19 .i.xh5 ':xh5 20 f4+ "xf4 21 "g7+ 1-0 Sobura-Pieniazek, Poland 1988.

8 .i.g2 (D) White abandons his plan Ιο play :gl and g5 because after 8 ':gl d5 9 exd5 ~xd5 10 ~xd5 "xd5 11 .i.g2 "c4 12 c3 .i.e7 13 g5 ~d7 14 "e2 "xe2+ 15 Φχe2 ~b6 White had ηο advantage ίη Karpov-Kindermann, Vienna 1986.

8 ...

tΩc6

Or:

1) 8...g6 9 g5 hxg5 10 .i.xg5 .i.e7 11 "d2 e5 12 ~e2 .i.e6 130-0-0 ~bd7 14 f4 "a5 (14 .....c7?! is ίη ferior after 15 fxe5! dxe5 16 ~d5 .i.xd5 17 exd5 ':c8 18 ':hf1! with some advantage for White, GhindaBonsch, Halle 1987) 15 ΦbΙ ':c8 (15 ... ~b616 b3;t Gufeld-Georgadze, USSR 1981) 16 ':hf1 b5 17 b3 exf4 18 ':xf4 ~h5? (Timman considers that 18 ... ':h5 followed by an exchange sacrifice οη g5 Ιο be Black's best chance) 19 .i.xe7! ~xf4 20 "xf4! Φχe7 21 "xd6+ Φe8 22 ~d5 .i.xd5 23 .i.h3!! (Black has been convincingly punished for his optimistic 18th move) 23 .....c7 24 .i.xd7+ "xd7 25 "e5+ "e6 26 "xh8 Φe7 27 "d4 with a winning

Scheveningen position for White, Ljubojevic-Timman, Amsterdam 1986. 2) 8...dS and then 9 exdS ltJxd5 1Ο ltJxd5 exd5 is given as unclear by ECO. However, ίη distinction Ιο 8 1:tgl d5, White's rook is still defending the h-pawn, so Whίte might consider 9 eS ltJfd7 10 f4, when 1Ο ... .ιe7 11 h5 and 10 ... h5 11 gxh5 100k good for White, so the cήtίcal reply is probably 10.....b6. 9 gS is also interesting. 9 gS bxgS 10 hxgS 1:txh1+ 11 .ιxh1 ltJd7 (D) If Black attacks the g5-pawn by 11•••ltJxd4 12 "xd4 ltJh7, White continues 13 e5! ltJxg5 (13 ...dxe5 14 "h4 traps the knight) 14 Wa4+! (14 exd6 is also promising) 14 ... .ιd7 15 Wg4.ιe7 (15 ... lL!h7 16 Wh510ses a piece while 15 ... f6 16 .ιχg5 fxg5 17 Wh5 is disastrous) 16 exd6 .ιf6 17 .ιΧb7 with a clear advanιage.

12 .ιg2 This move was the result of lengthy thought, but even so ίι wasn't the best. White has very few constructive moves apart from g6,

Vαriαtion

and the immediate 12 g6 lL!xd4 13 gxf7+ ΦΧf7 14 Wxd4 Wh4 15 .ιg2 lL!e5 is obscure. White might be able Ιο claim a slight plus after 16 .ιe3 Wg4 17 Φrι, but both kings are 00happily placed and Ι wanted ιο find something safer. When confronted with an unexpected move ίη the opening, players usually react by steering clear of very sharp lines which may have been well prepared by the opponent. 12 .ιe3 is bad due Ιο 12 ... ibde5 threatening ... lL!c4 (13 We2 ltJxd4 10ses a pawn). Thus the only direct altemative Ιο 12 g6 is 12 f4, but Ι was reluctant Ιο create a huge empty space around my king. However, later analysis showed that Black has ηο way Ιο exploit the temporary exposure of White's king, so ίη later games Whίte prefened the more accurate 12 Ι4!. Black has replied: 1) 12•••lL!xd4 13 Wxd4 "b6 14 Wxb6lL!xb615 a4!? .ιd7 16 a5lL!c8 17 .ιe3 .ιc6 18 0-0-0 Φd7 19 .ιπ lL!e7 20 .ιg4 with a clear plus for White, Ghinda-Vogt, Halle 1987. 2) 12...g6 13 .ιe3 Wb6 14 a3 Wc7 15 We2 lL!a5 160-0-0 1:tb8 17 e5!? (ηοι strictly necessary, but very dangerous) dxe5 18 lL!xe6 fxe6 19 "d3 with a strong attack for the piece, Nunn-Suba, London (Lloyds Bank) 1990. 3) 12••."b6 13 lL!de2 g6 14 b3 with a further branch: 3a) 14.....c7 15 .ιb2 b5 16 "d2 1:tb8 (16 ... .ιb7 17 lL!dl 0-0-0 18 ltJe3 .ιe7 19 0-0-0 lL!b6 20 ΦbΙ

32 Scheveningen

Vαriαtion

..ttb8 21 ~cl was also a little better for White ίη Grϋnfeld-Bischoff, Munich 1987) 17 .tg2 (17 0-0-0 ~b6 18 a3 .td7 19 .tg2 b4 20 axb4 ~xb4 was equal ίη Kir.Georgiev-Suba, Budapest Ζ 1993, but 18 :el b4 19 ~dl would have given White the edge, according ιο Stoica) 17 ... a5?! 18 ~dl! a4 19 ~3 b4 200-0-0 .ta6 21 ΦbΙ axb3 22 axb3 was good for White ίη Watson-Suba, Kuala Lumpur 1992. 3b) 14......c5 15 'it'd2 b5 16 .tb2 .tb7 170-0-00-0-018 ΦbΙ 'it'f2!? (or 18 ....te7 19 ~cl! f6 20 gxf6 ~xf6 21 ~3 with an edge) 19 a4 (19 :el .te7 20 lbdl 'it'c5 21lbe3 gave White a small but lasting advantage, Short-Κindermann, Dortmund 1986) 19 .....b6 20 axb5 axb5 21lbc Ilbc5 22 "h2! with a distinct plus for White, Anand-J.Polgar, Madήd 1993. Although 12 .tg2 is slightly infeήοr to 12 f4, ίι is still worth studying Nunn-Bischoff as the resulting positions are very similar. 12 ... g6! The idea behind 12 .tg2 was that White improves the position of his bishop (particularly ίη the g6 line given above, because ... 'it'h4 ηο longer gains a tempo), while Black has trouble finding useful moves. 12.•• ~e5 allows 13 f4 with gain of tempo (13 ...lbc4 14 b3 'it'b6 15 lbce2), 12..:l'c7 (or .te7) allows 13 g6 and 12...'l'b6 13 lbb3 loses time after a subsequent .te3. Black's reply is the best, cutting ουΙ g6 by White and again posing the question

as Ιο how White can improve his position.

Ι4

(D)

Now the defect of 12 .tg2 is revealed. Ιη the analysis of 12 f4 we saw that Black generally plays ...g6 ίη any case, while White's .tg2 is often ηοΙ necessary. Therefore Black may gain a tempo, although ίη this position the extra move is ηοΙ especially valuable. Now that e5 is denied Ιο the black knights, White threatens simply .te3, so Black's reply is more or less forced.

13 ... 'ib6 14 lbde2 Now White can only complete his development by "d2 (or 'it'd3), b3, .tb2 and 0-0-0, so before playing f4 Ι had Ιο make sure that Black couldn't use the four(!) free tempi Ιο harass White's centralised king. 14 ... 'it'cS Black settles for finishing his own development by ... b5 and ....tb7. This is clearly best, since 14..."'c7 allows 15 .te3 and 14...ω is met by 15 b3 ίη any case. 15 'it'd3

Scheveningen Ι5 'iVd2 would have been slightly better as the queen is exposed ιο possible knight attacks οη d3. ΜΥ idea was to leave open the possibility of .i.e3, but this is never feasible. Ι5 ... b5 Ι6 b3 Ι6 .i.e3 is effectively countered by 16...~M!. Ι6 ••• .i.b7 Ι7 .i.b2 Α ήsky move; Black decides ιο leave his king ίη the centre ίη order Ιο help his c-file counterplay. After Ι7•••0-0-0 18 O-O-OWhίtehasasmall space advantage, just as ίn the above analysis of 12 f4.

Vαriation

οη. Ιι goes without saying that al10wing the queens Ιο be exchanged would give White ηο advantage, because his main asset is the vulnerable position of Black's king, and this can οηlΥ be exploited ίη the middlegame. 23 .i.xc3 (D)

:c8

Ι8

0-0-0

Thanks Ιο Black's committaI decision last move, he has ιο follow υρ with active play. If White is al10wed Ιο consolidate, then ~bΙ,:Ιf1 and ~dl-e3-g4 will inevitably give White a strong attack against the centralised black king. Ι9 'iVd2 ~xc2 20 ~xc2 b4 2Ι

~bl

"iνf2!

This is the ροίηι of Black's combination. The immediate 2l ...bxc3 22 ~xc3 leaves Black ίη a poor posiιίοη, because his combination has faίled Ιο dent White's position, and the 10ng-term prospects lie with Whίte.

22 .i.hl bxc3 this stage Black surpήsίngΙΥ offered a draw, but although White must adopt the much less satisfactory recapture with the bishop, thus leaving e4 weak, Ι decided Ιο play Αι

23 •.• 'iVa7? misjudgement. The main merit of Black's combination is that his queen has become a nuisance by taking υρ residence ίη the heart of White's position, the more so as White cannot contemplate a queen exchange. Bischoff retreats ίι Ιο the passive square a8 ίη return for ίncon venient but ηοΙ really seήοus pressure agaίnst e4. He should have continued 23•••lDc5 24 .i.d4 "h2! (24 ... ~xe4 25 'fi'd3! 10ses a piece, while 24 .....h4 25 ~3 e5 26 i.e3 100ks good for White) when White has probIems with his e4-pawn. Then 25 'iVe3 e5 26 .i.b2 .i.g7 creates a very awkward threat of ... exf4, so Whίte would have Ιο pIay 2! ω 'fi'xd2 26 :Ιχd2 with equality. 24 .i.b2 'iVa8 Α

34 Scheveningen

Vαriαtion

"'e3

25 26 1Dc3

1Dc5 .ig7

27 lDd5! (D)

27 ••• .ixb2 The most natural move. The alternatives are: Ι) 27.• ..ixdS 28 exd5 .ixb2 29 dxe6!? (29 ΦΧb2 gives White an edge) 29 ... 'ifa7 (preventing ΦΧb2) 30 ':'xd6 Φf8 (30 ... .ia3 31 exf7+ Φf8 32 'ifc3 and 30 ... .ig7 31 .ic6+ Φf8 32 ':'d7 win for White) 31 ':'d7 'ifb6 32 e7+ Φe8 33 ':'d8+ ':'xd8 34 exd8'if+ ΦΧd8 35 'ifd2+ lZXt3 36 'ifxd3+ and the extra pawn gives White some winning chances ίη the ending. 2) 27•••exd5 28 .ixg7ll)xe4 (better than 28 ... dxe4 29 ':'xd6, as ίη this case the undefended c5-knight prevents 29 ... ':'d8, while 29 ... ll)d3 allows 30 ':'xd3) 29 .if3! is good for White. The threat is simply ':'hl-h8, and the bishop can move ιο g4 ιο cut off the escape of Black's king. 28 ll)b6 .ixe4+ 29 Φ"b2 "'a7 The only defence. Now 30 ll)c4 d5 31 .ixe4 is tempting, but Black

can reply 31 ... dxc4! 32 'ifd4 'ifc7, and ίι is doubtful if White has anything better than perpetual check. 30 ':'xd6 ll)d3+ Black's moves continue to be forced. 3O•••.ixhl 31 'ifd4! (attacking the rook οη c8 and threatening 'ifh8+) 31 ...':'b8 (31 ... ':'c7 32 'ifh8+ and mate at d8) 32 'ifxc5 (threatening ':'d7) 32 ... ':'d8 (32 ..."'e7 33ll)c8 'ifb7 34 ':'xe6+ wins) 33 'ife5 gives White a decisive attack, for example 33 ... ':'xd6I'ife7I'ifb8 34 'ifh8+ or 33 ... ':'b8 34 ':'xe6+ fxe6 35 'ifh8+. 31 Φa3 ':'c6! Ι had overlooked this ingenious defence when Ι played 26ll)c3. Altematives lose quickly: 1) 31 •••':'cS 32 ':'xd3 ':'a5+ (or 32 ... 'ifxb6 33 .ixe4) 33 Φb2 .ixd3 (33 ... .ixhl 34 'ifd4) 34 .ic6+ Φd8 (otherwise a knight check wins the queen) 35 b4 ':'f5 36 'ifxd3+ Φc7 37 'ifd7+ ΦΧb6 38 'ifd4+ and Black loses his queen. 2) 31 .••"'c7 32 ':xd3 .ixd3 33 ll)xc8 'ifxc8 34 'ifxd3 'ifcl+ 35 Φb4 'ifxhl 36 'ifxa6 'ifel + 37 Φb5 'ife2+ 38 Φb6 'iff2+ 39 Φb7 with a winning ending since 39 ... 'ifxf4 fails Ιο 40 'ifa4+. 3) 31 •••.ixhl 32 'ifh3 ':'b8 33 ':'xe6+ fxe6 34 'ifh8+ winning the queen. 32 ':'xc6 The best, since 32ll)d5 'ifxe3 33 ll)f6+ is a draw by perpeιual check and 32 ':'xd3 (32 'ifxd3 'ifxb6 is good for Black) 32....ixhl 33 tα4 is roughlyequal. 32 ••. "'e7+ (D)

Scheveningen

33 "'CS! 33 ':ό .txhl 34 b4 ~xc5 35 'ifxc5 is not so good since Black can avoid the exchange of queens by 35 ... 'ifd8, when White's exposed king means an a1most inevitable perpetual check. After the game Bischoff admitted that he had not foreseen this move ίη tίme. 33 .. ο "'xcS+ Or 33000~ό 34 ':c8+ 'ίi'd8 35 ':xd8+ ~xd8 transposing ίηΙο the game, except for an unimportant change ίη the positίon of Black's king. 34 ':xcS ~xό 35 b4! White is aimίng for a good knίght ν bad bishop ending. He can a1so head for a knight ν knight ending by 35 .txe4 (35 ~M ~d3+ 36 ~c3 ~f2! leads Ιο the same position) 35 ... ~xe4 36 ~M, intending ~a5, but although this is favourable for White, his advantage is less than ίη the game. 35 000 .txhl 36 bxcS .td5? Α blunder ίη severe time-trouble. He should have tried 36000~e7 37

Vαriαtion

~M e5! 38 fxe5 ~e6, but even here White wins by 39 ~c4 .td5 40 a3! .txc4 41 Φχc4 Φχe5 42 ΦΜ! Φd5 43 a4 and Black is ίη a fatal zugzwang. 37 ~d5 exd5 White wins because Black's apawn has moved, while White's can still advance either one or two squares. This extra flexibility means that White can a1ways arrange for the key reciproca1 zugzwang position ιο aπive with Black to play. 38 ~b4 Φd7 39 Φc3 1-0 Black lost οη tίme, but ίη any case 39000ΦΟΟ 40 Φd4 a5 41 a4 and 39000 Φc7 40 Φd4 Φc6 41 a3 a5 42 a4 lead to the same position of reciproca1 zugzwang.

Game6

Nunn-Sax Rotterdαm

le4 2 3 4 5 6 7

00 d4 ~xd4 ω

1989

c5 e6 cxd4 lbf6 d6 h6

g4 h4 lί:X:6 8 ':gl hS This is currently the most popular move. White's best reply is to take οη h5 and since the recapture ... ~xh5 leaves the knight badly placed, Black normally returns ίι ιο f6. The net effect of this is to reach a position similar Ιο that after 6 g4 h6 7 g5 hxg5 8 .txg5, but with White

36 Scheveningen

Vαriαtion

having the two extra tempi h4 and :gl. This might seem Ιο be a great improvement, but ίη fact the disadvantages almost ba1ance the advantages. White has two problems; firstly the h-pawn can become weak without the defence of the rook and secondly Black's ... 'iWb6 effectively pins the f-pawn against the undefended gl-rook, so ίι is harder for Whίte Ιο play f4. Apart from 8 ... h5, Black has: 1) 8.•• dS 9 ~bS .td7 10 exdS ~xdS (1O ... ~xd4 11 ~xd7+ 'iWxd7 12 'iWxd4 ~xd5 l3 ~xd5 'iWxd5 14 'iWxd5 exd5 15 .te3 gave White the better ending ίη Nikolenko-Zakharov, Smolensk 1991) 11 ~xdS exdS 12 ~e3 (D) (this pawn sacrifice is more promising than 12 'iWe2+) and now:

la) The game Karpov-Spassky, Tίlburg 1980 continued 12.. ie7 13 'iWd2 .txh4?! (13 ...0-0 14 ~f5 .txf5 15 gxf5 Φh7 160-0-0 is a1so very

good for White, but Black has better survival chances after 13 ... ~xd4 14 ~xd7+ 'iWxd7 15 'iWxd4 .tf6 16 'iWb4 .te7, when he went οη Ιο draw ίη

Marjanovic-Ceba1o, Yugoslav Ch 1982, a1though he needed Ιο defend accurately υηΙίl move 64 Ιο achieve this) 14 0-0-0 .tf6 (14 ... ~xd4 15 .txd7+ 'iWxd7 16 .txd4 ο-ο 17 f4! followed by g5 and 14 ... 0-015 g5! give White winning attacks) 15 ~f5 ~xf5 16 gxf5 a6 (Black could ηοΙ castle without losing his vital h6pawn, but now his king is permanently pinned down ίη the centre) 17 .txc6+ bxc6 18 .tc5! (with White only needing Ιο ήρ open the d-file by c4 Ιο finish Black off, Spassky launches an ingenious counterattack which fails because of his ίη ability Ιο bήng the h8-rook ίηΙο the game) 18 ...:b8 19 b4 :b5 (Black has Ιο eliminate the deadly bishop) 20 :gel + Φd7 21 c4 :xc5 22 bxc5 .tg5 (after 22 ... 'iWb8 23 cxd5 .tg5 24 :e3 .txe3 25 fxe3 'iWe5 26 dxc6+ Φχc6 27 'iWd7+ White should win easίly enough) 23 f4 'iWf6 24 cxd5! (ηοι 24 fxg5 'iWal + 25 Φc2 'iWxa2+ 26 Φd3 'iWxc4+ 27 Φe3 hxg5 when Black has four pawns and a tremendous attack for the rook) 24 ... 'iWal+ 25 Φc2 'iWxa2+ 26 Φd3 'iWxd2+ (26 ... 'iWb3+ 27 'iWc3 a1so forces the queens off) 27 :xd2 .txf4 (a1though Black has two pawns for the exchange a11 White's pieces are very active and Black is unable Ιο organise himself against the advance of the c-pawn) 28:a2 cxd5 29 :xa6 h5 30 Φd4 h4 31 ΦΧd5 :b8 32 f6 gxf6 33 :xf6 ~g3 34 :xf7+ Φd8 35 :f8+ 1-0. 1b) 12...'ii'xh4 was recommended by Kasparov both ίη ECO and ίη his

Scheveningen book with Nikitin οη the Scheveningen, but of White's responses he only considers 13 'ii'e2, which is curious as 13 'ii'd2 was given ίη Informator30: lbl) 13 'iVe2 ~xd4 14 ~xd4+ 'ii'e7 15 ~xd7+ ΦΧd7 16 ~e3 I:td8 17 0-0-0 Φc8 with an equa1 game according to Makarychev. lb2) 13'iVf3!?a6! 14~xc6bxc6 15 0-0-0 ~d6 16 ~f5 ~xf5 17 gxf5 ~e5 18 ΦbΙ I:tb8 19 b3 'ii'b4! and though White still has compensation for his pawn, Black was able to hold the balance ίη Anand-Sa1ov, Moscow Alekhine mem 1992. lb3) 13 'iVd2! (BTS2's suggestion still seems to be best) and now 13••• ~e7 14 0-0-0 followed by ~f5 is dangerous for Black, whilst if he tries to follow the Salov plan with 13•••86 14 ~xc6 bxc6 15 0-0-0 ~d6 16 ~f5 ~xf5 17 gxf5 ~e5, then the fact that White has not blocked his f-pawn makes all the difference. After 18 f4! ~f6 19 ~c5 Black's queen is hοπίblΥ out of the game and his king extremely exposed. 2) 8... ~7 9 g5 hxg5 (D) and now:

Vαriαtion

2a) 10 hxg5 g6 (lO ... ~de5 11 ~e3 ~d7 12 ~e2 a6 13 f4;t Matu10vit-Simit, Smederevo 1981) 11 ~e3 a6 12 'ii'e2 ~a5 130-0-0 b5 14 f4 'ii'c7?! (14 ... ~c4 is better) 15 ΦbΙ ~b6 16 ~b3! ~ac4 17 ~d4 I:tg8? (17 ... l:th7) 18 'ii'h2! ~e7 19 'iVh7 I:tf8 20 f5 with a winning attack for White, Morovit-J.Polgar, Las Palmas 1994. 2b) 10 ~xg5 'ii'b6 11 ~b3 a6 12 h5 'ii'c7 13 'ii'e2 b5 14 0-0-0 b4 (Tseshkovsky-~bin,SochiI981)

15 ~a4! a5 16 f4 with a good game for White, according to Tseshkovsky. ΜΥ (JN) personal preference is for 10 ~xg5, since Black's knight is badly placed at d7. 3) 8...g6 9 g5 hxg5 10 ~xg5 (10 hxg5 ~d7 transposes to line '2a') 10... ~e7 11 'ii'd2 a6 120-0-0 'ii'b6 13 ~b3 'iVc7 14 f4 ~d7 15 f5 0-0-0 16 ~h3 was better for White ίη Sideif-Zade-~bin, USSR 1982. 9 gxh5 ~xh5 Black has been known to try 9••• l:txh5 1Ο ~g5 I:th8, but this is quite point1ess. Black reaches the same position as ίη the main line, but having forfeited the ήght to cast1e kingside. 10 ~g5 (D) 10 ••. ~Ι6 The alternative is the immediate 10••.'iVb6, which attempts Ιο avoid the 10ss oftime inherent ίη 10...~f6. The problem is that the knight is genuinely badly placed at h5, so Black gains nothing by keeping it there. After 11 ~b3 86 12 ~e2 g6 (12 ... ~f6 transposes ιο the analysis

38 Scheveningen

Variαtion

of 11 ... 'W'b6 ίη the main line) 13 -ΙΜ2 Black may play: 1) 13... j.d7 14 :g2! (White is ηοΙ forced Ιο sacrifice his f-pawn by cast1ing straight away) 14..•'ifc715 0-0-0 b516 a3 (D) and now:

la) 16••• ω 17 "d4! :h7 18 f4 lL!c4 19 j.xh5 :xh5 20 :el j.c621 lL!d5 j.xd5 22 exd5 e5 23 "d3 j.e7 24 lL!d4! and White stands well, Motwani-Roca, Dubai OL 1986. lb) 16•••:b8 (16 ...:c8 should also be met by 17 f4) 17 Ι4 b4 18 axb4lL!xb419 Ι5! (White must open lines before Black can generate a queenside attack) and now: lbl) Gallagher-Chomet, Royan 1988 continued 19•••exf5 20 exf5

j.xf5 21 'ife3+! j.e7 (or21. .. j.e622 lL!d4 "d7 23 j.g4lL!g7 24 j.f6 and Black won't be able to tolerate the pressure much longer) 22 lL!d4 f6 (ηοι the most resi1ient but ίι was lost anyway) 23 lL!xf5 fxg5 24 j.xh5 gxh5 25lL!xd6+ 1-0. lb2) DUΉng the game Ι felt that 19•••85 was Black's only chance, but after 20 fxg6 fxg6 21 :f1! White's attack is much quicker, e.g. 21 ... a4 (if 21 ... j.g7, then 22 j.xh5 and 23 j.f6 is νery strong and οη 21 ... j.c6, 22lL!d4 looks goOO) 22 j.xh5 gxh5 23 "f4 j.c8 (23 ... j.c6 24 "f6 and 23 ... j.b5 24 lL!xb5 lL!d3+ 25 ΦbΙ are ηο better) 24 "f6 :g8 25 j.e3! lL!a2+ (25 .....h7 26 lL!xa4 is sufficient) 26lL!xa2 axb3 27 lL!b4!! (27 lL!c3 "h7 is goOO for White but far less clear than the text) 27 ...:xb4 (27 .....h7 28lL!c6) 28 c3! 'ifh7 29 :xg8 'ifxg8 30 j.g5! and it's time ιο resign. 2) 13..,j.e7 14 :g2! j.d7 15 0-0-0 :c8 16 ΦbΙ "c7 17 a3 b5 18 j.xb5 axb5 19lL!xb5 'ifb8 20 j.xe7 "xb5 21 j.xd6 with a clear adνantage for White, Lobron-Marjanoνic, Reggio Emilia 198516. 3) 13•••'ifc7 (this is eνen worse than the lines aboνe because White need not spend time οη :g2) 14 0-0-0 b5 15 a3 j.d7 16 j.xb5! axb5 17 lL!xb5 'ifb8 18 lL!xd6+ j.xd6 19 "xd6 'ifxd6 20 :xd6 with adνan tage Ιο White, Goνedarica-Mokry, Trnaνa 1987. 11 j.e2 (D) This flexible moνe, which prepares a possible h5, has gained ίη

Scheveningen populaήty, even though White some-

times has to sacrifice his f2-pawn after 'iVd2 and 0-0-0. Ιη reply the immediate ... 'iVb6 turns ουι badly because h5~h6 becomes strong, so Black normally bides his time with ... a6.

11 ... 86 Or: 1) 11••• j.e7 12 'iVd2 lLIxd4 13 'iVxd4 'iVb6 14 j.b5+ Φf8 (14 ... j.d7 15 j.xd7+ ΦΧd7 16 'iVd2 is good for White) 15 'iVxb6 axb6 160-0-0 e5 was unclear ίη Ljubojevic-Adorjan, Linares 1985, but ίι is hard Ιο believe that there is ηο way White can exploit the weak b-pawns. Perhaps 16 1iJa4!? is best. 2) 11•••'ifb6 12 lLIb3 a6 13 h5 'iVc7 (or 13 ... j.d7 14 h6 :th7 15 'Wd2 lLIg8 16 j.e3 'iVc7 17 hxg7 :txg7 18 0-0-0 with a plus for White, AlzateFrias, Dubai OL 1986) 14 h6ltJd7 15 hxg7 j.xg7 16 'iVd2 j.f8 170-0-0 b5 18 a3 j.b7 19 :thl :txhl 20 :txhl lLIce5, Hellers-Sax, New York Open 1987, and now 21 :th8! lLIg6 22 :th7 would have given White a very dangerous attack.

Vαriation

3) 11 •.. j.d7 12 h5 (οη 12 'iVd2, 12... lLIxd4 13 'iVxd4 'iVb6 is interesting, whilst 12lL1db5 'iVb8 is not dangerous for Black) 12••• 86 13 'Wd2 andnow: 3a) 13•••'ifb6 transposes Ιο '2' ίη the note Ιο Black's 12th move . 3b) 13••• j.e7 14 0-0-0 'Wc7? 15 h6! gxh6 16 j.xf6 j.xf6 17 lLIf5!! j.e7 (the brilliant ροίηΙ is 17 ... exf5 18 lLId5 'iVd8 19 'Wxh6!) 18 lLIxe7 Φχe7 (18 ...lLIxe7 was the lesser evil) 19 :tg3! b5 20 'iVf4 :tad8 21 'iVh4+ Φe8 22 j.xb5! with a winning ρο sition for White, Anand-Ye Jiangchuan, Kuala Lumpur 1989. 3c) 13••• b5 14 a3 j.e7 15 j.e3!? (150-0-0) 15 ...lLIxh5 160-0-0 lLIf6 (it seems a strange decision Ιο give the pawn back at once, even if after 16 ... g6 17 f4 lLIf6 18 :thl :tg8 19 j.f3 White has good compensation according Ιο Kasparov) 17 :txg7 'iVb8 18lL1xc6! j.xc6 19 'Wd4 with advantage Ιο White, Kasparov-Sax, Tilburg 1989. 12 'Wd2 j.d7 Playing for ... b5 is a relatively new idea, but the critical continuation is probably the older 12.••'Wb6 13lL1b3 (D) and now: 1) 13••.'Wc7 with a couple of possibilities: la) 14 h5 (the most common continuation but Ι (JG) am ηοΙ totally convinced) 14 ... b5 15 a3 lLIxh5 16 :thl g6 17 j.xh5 gxh5 (17 ...lLIe5 18 'iVe2 gxh5 19 :txh5 :txh5 20 'iVxh5 j.b7 21 0-0-0 was also unclear ίη the game Brunner-Adorjan, Altensteig 1989) 18 'We2 b4 19 axb4 (l9lL1d5 is

40 Scheveningen

Variαtion

extremely speculative) 19 ... tΩxb4 20 0-0-0 J..g7 and Ι don't believe that Black has much to wouy about. Α possible continuation: 21 ':xh5 (21 ΦbΙ is probably the critical line) 21 ... J..xc3 22 bxc3 ':xh5 23 'ifxh5 'ifxc3 24 ':d2 (White now has the surpήsίngΙΥ awkward threat of'ifh7g8 but Black has a spectacular defence) 24 ... J..b7! 25 'ifh7 tΩa2+! 26 ΦbΙ (26 ΦdΙ 'iff3+) 26 ... 'ifxb3+!! 27 cxb3 J..xe4+ 28 Φχa2! J..xh7 29 ':xd6 and a draw is inevitable. lb) 140-0-0 and now: lbl) The game Gallagher-Maxίοη, Bad Worishofen 1991 didn't last vary long: 14••• J..d7 15 h5 J..e7 16 J..f4! tΩxh5?! 17 J..xh5 ':xh5 18 ':xg7 tΩe5 19 ':g8+ J..f8 20 J..xe5 ':xe5 (or 20... dxe5 21 'ifxd7+ 'ifxd7 22 ':xf8+) 21 'ifh6 Φe7 22 'ifh4+ 1-0. 1b2) 14..•b5 is more active. After 15 a3 J..b7 (15 ... ':b8 can also be met by 16 h5) 16 h5 tΩxh5 (16 ... b4 17 axb4 tΩxb4 18 'ifd4 d5 19 h6! ':xh6 20 J..xh6 e5 21 J..f4! won for White ίη Luther-Bonsch, East German Ch 1989) 17 ':h 1 g6, White can transpose into Brunner-Adorjan with 18

J..xh5 gxh5 19 'ife2 tΩe5 20 ':xh5 ':xh5 21 'ifxh5, but he can also try 18 Ι4, which seems to give him good compensation for the pawn. 18.•• b4 19 axb4 tΩxb4 20 'ifd4 is not possible for Black, so he would most likely play 18•• ':c8 when White can has the safe 19 ΦbΙ followed by f5, or the sharper possibility 19 f5 b4 20 axb4 tΩxb4 21 fxe6 (21 'ifd4tΩa2+!) 21 ... fxe6 22 J..g4. 2) 13...J..d714 h5 tΩΠ5 (Black should take everything οη offer; the passive 14... 0-0-0 15 h6 ':h7 16 ο-ο-ο! J..e7 17 J..e3 'ifc7 18 ':xg7 ':xg7 19 hxg7 ':g8 20 ':gl tΩe5 21 J..d4 tΩg6 22 'ifh6 J..c6 23 J..d3 was very good for White ίη Korolev-Agzamov, USSR 1983) 15 ':hl g6 16 0-0-0 'ii'xf'2 (once again the crucial move; 16 ... 'ifc7 17 J..xh5 gxh5 18 'ife2! J..e7 19 J..xe7 tΩxe7 20 ':xh5 ':xh5 21 'ifxh5 tΩg6 was played ίη Tseshkovsky-Mokry, Tmava 1986, and now 22 'ifh2! was promising for White) 17 e5! "Ι5! (D)(17 ... tΩxe5 18 tΩe4 'ii'f5 19 'ife3! J..c6 20 tΩbd2! gives White a crushing attack) and the evaluation of the whole line depends critically οη this position.

Scheveningen White has tried: 2a) 18 .i.xh5 ':xh5 19 ':xh5 gxh5 20 exd6 c!ί)e5 21 c!ί)d4 'ii'g4 22 .i.e7 ltJc4 23 'ii'd3 ~5 24 'ii'e3 ltJc4 25 'ii'd3 c!ί)e5 and there was οηlΥ a draw by repetition, Chandler-Hellers, Thessaloniki OL 1988. 2b) White played 18 exd6 'ii'xg5 19 'ii'xg5 .i.h6 20 'ii'xh6 ':'xh6 21 lΔcs ~5! 22 c!ί)3e4 0-0-0 23 c!ί)xd7! 'it>xd7 24 ':h3 ':e8 ίη A.RodriguezGrooten, Dieren 1987, and although Rodriguez gives 25 b4! as unclear, this line is ηοΙ convincing. It seems Ιο me (JN) that ίι would have been more promising for White Ιο play 21 ~!, intending a combination of c!ί)f6 and c!ί)bc5, while 21 ... 0-0-0 22 c!ί)f6 ':dh8 23 c!ί)xd7 'it>xd7 24 ltJc5+ 'it>c8 25 .i.f3 gives White dangerous threats. 2c) 18 'it>bl!? is the latest try, which rules ουΙ the queen exchange at the cost of another pawn. LauLesiege, Eurodata 1992, continued 18 ... d5 (18 ... 'ii'xe5 19 .i.xh5 ':xh5 20 ':xh5 gxh5 21 ':el 'ii'f5 22 c!ί)d5 is good for White) 19 ':df1!? (19 ':del d4? 20 c!ί)xd4! c!ί)xd4 21 'ii'xd4 'ii'xg5 22 c!ί)e4 'ii'd8 23 .i.xh5 .i.g7 24 .i.xg6 ':xhl 25 ':xhl fxg6 26 c!ί)d6+ won for White ίη HectorMortensen, Grrested 1990, but Hector points ουΙ that 19 ... .i.e7 would have been a much tougher defence) 19 ... 'ii'xe5 (19 ... c!ί)g3 20 ':xh8 c!ί)xf1 21 'ii'f4! 'ii'xf4 22 .i.xf4 g5 23 .i.xg5 c!ί)g3 24 .i.h6! 'it>e7 25 .i.g5+ 'it>e8 26 .i.d3 with excellent compensation for the pawn; an important ροίηι is that 26 ... c!ί)xe5 fails Ιο 27 .i.f4 c!ί)xd3

Vαriαtion

28 .i.d6! as 28 ...0-0-0 is now illegal) 19 ... 'ii'xe5 20 .i.xh5 ':xh5 21 ':xh5 gxh5 22 ':el 'ii'd6 23 c!ί)xd5 .i.e7! 24 c!ί)f6+ 'it>d8 25 'ii't2 and White had compensatίon for the pawns but perhaps οηlΥ enough for an equal game. 13 0-0-0 b5 14 ~c6! The immediate 14 'ii'e3 is met by 14 ... 'ii'b6. 14 .i.xc6 15 'ii'e3 (D)

White has the unpleasant threats of 16 e5 and 16 c!ί)d5, while after 15.••'ii'a5 White can afford Ιο take time ουΙ for 16 'it>bl because 16 ... b4 may be met by 17 ~5 c!ί)xd5 18 exd5 .i.xd5 19 ':xd5 'ii'xd5 20 .i.f3. Black is therefore forced ίηΙο the υη palatable 15 ... 'ii'c7. 15 'ii'c7 16 ~5 .i.xd5 17 exd5 e5 After 17...':c8 White replies 18 c3, and Black has achieved nothing positίve, while he has forfeited the ήght Ιο castle queenside. 17 ... e5 is better, but even so White's lead ίη development and Black's exposed

42 Scheveningen

Variαtion

king give him dangerous attacking chances. 18 <iPbl?! Chess laziness. Of course <iPbl is a desirable rnove, but by giving Black a free ternpo White's attack loses rnuch of its rnornenturn. The irnrnediate 18 Ι4! was cοπect, when White has a clear advantage. Now by accurate defence Black survives the irnrnediate cήsίs. 18 ... ~7 Elirninating the g5-bishop rnakes ίι easier Ιο flee with the king, should that prove necessary, and ultirnately the opposite-coloured bishops rnight provide a drawing rnechanisrn. 19 Ι4 lί)xg5 Νοι 19.••f6 20 fxe5 dxe5 21 ~h5+ <iPd8 22 d6 ~xd6 23 ':'xd6+ 'ii'xd6 24 ':'dl 'ii'xdl+ 25 ~xdl fxg5 26 'ii'xe5 with an excellent position for White. 20 ':'xg5 Α difficult choice, as a1though White may win a pawn by 20 fxe5 (20 hxg5 g6 21 f5!? is probably a1so slightly better for White) 20 ... dxe5 21 ':'xg5 ο-ο-ο! 22 ':'xe5 ~d6, Black completes his development and the opposite-coloured bishops become an irnportant factor. 20 ..• ':'c8 Now 20•••0-0-0 is bad because of 21 a4!, so Black must adopt a different defensive plan. 21 c3 'ίi'cS Of course this is only possible when White has ηοΙ exchanged οη e5. Black gains time Ιο reorganise his defences.

exf4 ':'c7! (D) Αη excellent move. Black's rook covers the vulnerable square ϊ7 and when ίι aπives at e7 the attack οη the bishop will gain more time for Black.

24 a4! White's only chance Ιο make something of his waning initiative is Ιο create a new target οη the queenside. 24 ':'e7 25 ~d3 g6 axb5 26 axb5 Νοι 26••• ~h6? 27 'ii'f6. 27 'ίi'd4 The ending now represents the best winning chance for White. Although White's gain of a pawn is only temporary, the passed b-pawn combined with an exposed Black king gives White a nagging advantage. 'ίi'xd4 27 <iPd8 28 ~xb5+ ':'el+ 29 ':'xd4 ~e7 30 <iPa2 ':'hl 31 ':'gg4

Scheveningen The h-pawn is doomed, so White switches Ιο harassing Black's king. Ιη this the opposite-coloured bishops prove a big help. 32 ':84 ':8xh4 33.:txh4 ':xh4 (D) 33.•• j.xh4 34 ':a7 is worse, since White threatens the f-pawn directly and the d-pawn indirect1y via ':d7+.

Vαriαtion

with his rook. This tempo makes alI the difference and Black can now draw comfortably. j.f6 36 b4 37 ':87+ <Rb8 j.xc3 38 .:txf7 39 b5 ':b4 40 <Ra3 ':bl 41 ':b7+ <Rc8 j.d4 42 b6 43· ':c7+ <Rb8 44 ':b7+

Game7 Kotronias - Kr. Georgiev Kαrditsα 1994

34 :88+? careless check driving Black's king ιο a better square. White should have cut the king off by 34 ':87! (threatening ':d7+) 34 ... (34 ... g5 35 ':d7+ <Re8 36 ':xd6+ <Rf8 37 ':c6 and the d-pawn becomes dangerous) 35 j.c6! with winning chances because Black cannot challenge White ιο a pawn race (35 ... g5 36 b4 g4 37 b5 g3 38 b6 wins because the mate threat gains a tempo). 34 ... <Rc7 35 j.c6 35 ':87+ <Rb6 36 ':xe7 <Rxb5 37 ':xf7 <Rc5 is an easy draw. ':e4! 35 ... White is effectively a tempo down over the above line since after 36 ':87+ <Rb8 White must waste time Α

.:e4

1 e4 c5 2 00 e6 3 d4 cxd4 lΔt'6 4 lΔxd4 5 lΔc3 d6 6 g4 86 Of the sixth move alternatives we only consider 6 ... e5 here; the others may be found ίη Games 8 and 9. 6 ...e5 7 j.b5+ j.d7 8 j.xd7+ 1Wxd7 (8 ... lΔbxd7 9lΔf5 is awful for Black) 9lΔΙ5 h5 10 gxh5 (BTS2 proposed 10 f3, which is also quite a good move) 10...lΔxh5 (10 ... .:xh5 11 lΔd5! and 1O... lΔxe4 11 lΔxg7+ j.xg7 12 lΔxe4 d5 13 h6! are ηοΙ playable for Black) 11 j.h6! (the shock effect of this move is the main reason we have switched systems) 11...lΔc6 (11 ... gxh6 12 'ifxh5 is a disaster, whilst 11 ... g6 12 j.xf8 gxf5 13 j.xd6 is ηο fun either) 121Wxh5 (D) andnow:

44 Scheveningen

Vαriαtion

1) The gaιne Shnιuter-Obukhov, Russia 1993 continued 12•••g6 13 'ffg5 gxf5 14 J.xf8lίXι4 (14 ...:xf8 15 ltJd5) 15 0-0-0 ΦΧf8 16 :hgl 'ffe6 17 :xd4! (this ensures White a permanent attack for a minima1 materia1 investment) 17 ... exd4 18 "g7+ Φe7 19ltJd5+ Φd7 20 'ffxd4 :hc8 (20 ... fxe4 21 'ffa4+ Φc8 {21 ...Φd822'ika5+b623ltJΧb6} 22 'ffc4+ Φd8 23 'ii'c7+ Φe8 24 'ffxb7 and 20 ...:ac8 21 :dl! fxe4 22 'ffxa7 :xh2 23 "xb7+ Φd8 24 "b6+ Φd7 25 "a7+ Φd8 26 "a5+ Φe8 27 ltJc7+ are good for White) 21 'ffb4! :ab8 22 'ffa4+ :c6 23 'ffxa7 "h6+ 24 ΦbΙ "h8 25 ltJb4 :c7 26 e5! :a8 (26 ...b5 a1so meets a nice refutation: 27 e6+! fxe6 28 'ffxb8 'ffxb8 29 :g7+ Φc8 30 :g8+ Φb7 31 :xb8+ ΦΧb8 32ltJa6+ and White wins the resulting king and pawn ending) 27 e6+! Φχe6 (27 ... fxe6 loses to 28 'ffxa8!) 28 "b6 :c5 29 ltJd3 'ffd4 30 'ffxb7 :h8 31ltJxc5+ 1-0. 2) 12•••ltJe7 was not considered by Shnιuter ίη his notes to the game, but this is probably Black's best defence. White now has two possible replies:

2a) 13 "g5ltJxf5 14 exf5 :xh6 (White possesses a far supeήοr minor piece but a1so a weakened kingside which is why Black is not without hope) 15 ltJd5 (15 0-0-0 f6 16 'ffg2 {16 "'g4 g6!} 16 ... 0-0-0 is unclear) 15 ......a4! (Black does best to mix it υρ as 15 ......c6 16 0-0-0 :c8 17 c3 'ffc4 18 f6! is good for White) 16 ltJc7+ (16 0-0-0 :c8 is awkward) 16... Φd7 17ltJxa8 'ii'e4+ 18 Φd2 "d5+ 19 Φe2 (19 Φe3 :h3+) 19 .....c4+ 20 Φf3 'ii'c6+! and Ι can 't see how White can escape the checks as 21 Φg3 J.e7! 22 f6 :xf6! appears bad for White. 2b) 13 "g4 (this seems to be the better choice as the queen controls the fourth rank) 13.••ltJxf5 14 exf5 :xh6 15 0-0-0 and Black has severa1 possibilities which a11 seem insufficient to equa1ise: 2bl) 15.•.0-0-0 16 f4! ;t. 2b2) 15...:g6 16 "e4 :g5 17 ltJd5! :c8 (17 ... 0-0-0 18 'ffe3) 18 ltJe3 'ffc6 19 :d5 is good for White. 2b3) 15...g616OO! gxf517'ii'g5 :g6 (other moves are worse) 18 'ii'xg6 fxg6 19 ltJf6+ Φd8 20 ltJxd7 ΦΧd7 21 Μ! J.h6+ 22 ΦbΙ Φe623 J:thgl :g8 24 c4 and White has some advantage. Black can try to break out with 24 ... b5, but after 25 cxb5 d5 26 :g3! followed by :a3 White may experience some nervous moments but should win ίη the end. 7 g5 lLJrd7 8 h4 bS 8...ltJc6 9 J.e3 will transpose to either Game 8 or Game 9.

9 a3

Scheveningen White should never play this move lightly ίη the Sicilian, especia11y when he plans Ιο castle 10ng. Ιη this case, ίι is imperative that the knight οη c3 has a stable base as ίι holds White's centre togetherwhilst he goes mad οη the kingside. 9 ... .tb7 9 ... lί)b6 is less accurate since after 10 h5 White may meet 10....te7 with the dangerous piece sacήfice 11 'ifg4 e5 12lί)f5 g6 13 hxg6 fxg6 14 .te3! gxf5 15 exf5; ίη Nunn-Walden, Nottingham 1983, the continuation 10...tb8d7 11 :h3 lt)c5 12 g6 f6 13 :g3 e5 14lί)c6 'ifc7 15lί)b4 was good for White. 10 .te3 (D) 10 hS at once is ηοΙ so good because 10... .te7 awkwardly attacks the g5-pawn. 11 g6 doesn't work ίη this position since after 11 ... hxg6 12 hxg6 :xh113 gxf7+ ~xf7 14 'iff3+ .tf6 15 'ifxhl, White's knightond4 is hanging.

la) 11...lί)cS 12 f3 'ifc7 130-0-0 lί)c6 (13 ... lί)bd7 14 .txb5 proved Ιο be better for White ίη Fischer-Najdorf, Leipzig OL 1960) 14 lί)xc6 .txc6 and this position is assessed as slightly better for White by Kasparον ίη

ECO.

lb) 11.Hlί)b6?! 12lί)xe6! fxe613 'ifd4 lt)c4 14 'ifxg7 ~d7 (14 ...:f8 15 .txc4 bxc4 16 'ifxh7 is a1so good for White) 15 .txc4 bxc4 160-0-0 'iff8?! (16 .....g8 would a1so be met by 17 :xd6+!, but the best try was 16... lί)c6, a1though after 17 lί)a4 'iff8 18 'ifc3 White has more than sufficient for a piece) 17 :xd6+! ~xd6 18 'ifd4+ ~c7 19 .b6+ ~c8 20 'ifxe6+ lί)d7 (20 ... ~d8 21 :dl+ ~e8 22 lί)d5 .txd5 23 .xd5 wins for White) 21 :dl 'ifd8 (21 ....e8 is slightly more resilient, but after 22 lί)a4! .td8 23 .xc4+ .tc7 24 .tf4 lί)e5 {24 ... 'ifd8 25lί)b6+!}, White has at the very minimum 25 lί)b6+ q;b8 26 .txe5! 'ifc6 27 .txc7+ 'ifxc7 28 .xc7+ ~xc7 29 lί)xa8+ with ηο less than five pawns for the bishop) 22lί)a4! and Black has ηο defence against :xd7, KarklinsCommons, USA 1972. 2) 10...tα6 11 .e2 lί)de5 12 0-0-0 lί)c4 13 lί)xc6 .txc6 14 (4 'ifa5?! 15lί)d5! gave White acrushing attack ίη Alexander-Lundholm, CΟΠ.

10 ... lί)b6 Or: 1) 10....te7 11 'iνd2 (11 'ife2!?) andnow:

Vαriαtion

1970-1.

3) 10...ω 11 'iνg4!? lί)bd7 12 0-0-0 lί)e5 13 'ifg2 lί)c4 14 .txc4 bxc4 15 h5 intending g6 is good for White according Ιο Boudy. 11 hS tb8d7 12 g6! (D)

46 SCMveningen

Vαriαrion

This discoνery by Adams νirtu ally refutes Black's whole set-up. Preνiously 12 :h3, defending the rook ίη preparation for g6, had been played, but ίι turns ουΙ Ιο be superfluous ίη this parιicular position as White can quite happily sacrifice a whole rook for a raging attack. Neν ertheless, the idea behind :h l-h3 is an important one and has ίη fact led Ιο a reassessment of many lines ίη the Keres Attack. 12 hxg6 13 hxg6 :xhl 14 gxf7+ ~e7? Black should settle for the bad ρο sition that arises after 14...~xΙ715 1Wf3+ 1Wf6 161Wxhl, as he did ίη the prototype game, Adams-C.Hansen, Wijk aan Zee 1991. Play continued 16.•• lί)e5 (Hansen prefers Ιο sacrifice a pawn rather than allow White an excellent attacking position after something like l6 ... lί)c4 Ι7 .txc4 bxc4 Ι8 0-0-0) 17 lί)xe6 1Wxe6 (17 ... lί)f3+ fails Ιο 18 'iWxf3! and Ι7 ... ~xe6 18 .txb6lί)f3+ Ι9 ~e2 is also ηο good for Black) 18 .txb6 d5 19 .td4lί)c6! (19 ... dxe4 20 'iWh5+ lί)g6 2Ι .th3 is crushing) 20 .th3

_e8 21 0-0-0 lί)xd4 22 :xd4 .e5 23 :d3 d4 24 lί)d5 .txd5 25 exd5 .td6 and now: 1) Adams continued Ιο play for mate with 26 :f3+? and was ίη fact lucky Ιο win after 26 ... ~e7 27 :f5 'iWe2 28 'iWgl .ie5 29 f4 .tf6 30 :xf6 gxf6 31 'iWg7+ ~e8 32 'iWg8+ ~e7 33 .h7+ ~e8? (White has nothing after 33 ...~f8 34 ςPbl :e8 35 .ie6 :xe6 36 dxe6 'iWel+) 34 ~bl :d8? (Black would still haνe had some saνing chances after 34 ...d3! 35 cxd3 'iWdl + 36 ~a2 b4 37 1Wb7 {37 .td7+ ~d8 38 axb4 :a7!} 37 ... b3+) 35 .te6 ι-ο. 2) He should haνe headed for the ending with 26 .te6+ ~e7 27 'iWh4+ 'iWf6 28 'iWxf6+ ςPxf6 29 :xd4 when the technical difficuIties are unlikely Ιο proνe insurmountable. 15 .g4! Forced; Black is unable Ιο defend e6 (15 ... lί)c5 Ι6 .ig5+) so he has Ιο make some room for his king, and 15....b8 16 .tg5+ lί)f6 17 'iWxe6+ ~d8 18 'iWe8+ ~c7 19lί)e6 is mate. 16 .xe6+ ~d8 17 .e8+ ~c7 18 lί)e6+ ~c6 Or 18••• ~b8 19 lί)xf8 lί)xf8 20 .ixb6 .tc6 21 'iWe7 and Black is in a hopeless situation. 19 lίJd5! The attack intensifies. 19••••xe8 allows 20 lί)b4# and 19••• lί)xd5 20 exd5 ~xd5 21 0-0-0 ~c6 22 .ig2+ is the end, so Black has Ιο try... 19 :ΧΙΙ+ 20 ~ lί)xd5 21 exd5+ ~xd5 (D)

.cS

Scheveningen

Variαtion

5 ~

6 g4

must confess Ιο have taken some lίbertίes with the move-order of Karpov-Dorfman, which actuaIIy continued 6 ... i..e7. We transpose back ίη a few moves, but this move-order makes ίι easier Ιο explain the proposed repertoίre. 7 g5 ~d7 8 h4 (D)

22 tbxf8! Up υηιil this ροίηΙ Kotronias had simply been following Adams's analysis ίη Informαtor (Georgiev as well but Ι imagine unwittingly) but the text is clearly stronger than Adams's 22 ~c7+ Wc6 23 ~xa8, which, though strong, is ηοΙ immediately decisive. 22 ... ~e5 Useless, but so are 22... ~xf8 23 :d 1+ Wc4 24 :d4+ Wc5 25 :h4+ ~d5 26 'iVe4#, 22.....xe8 23 fxe8'ίW :xe8 24 ~xd7 Wc6 25 ~b6 :xe3 26 fxe3 Wxb6 27 :dl and 22.....c4+ 23 Wel. 23 :dl+ We4 24 :d4+ Φ(5 25 :t'4+ 1-0 25 ... Wg5 26 'iVe7+ and mate next move. Game8 Karpov - Dodman USSR Ch 1976

1 e4 2 00 3 d4 4 ~xd4

cS e6 cxd4 tt)f6

8 ... a6 8...i..e7 is the next game. Here we examine a couple of rarer altematives: 1) 8... ~χd49'iVχd4ω 10i..e2 ~c6 11 'iVd3 i..e7 (or 11 ... a6 12 i..f4 'iVc7 13 0-0-0 ~e5 14 'iί'd4 i..d7 15 h5 with strong pressure for White, Lutikov-Malίch, Leipzig 1977) 12 i..f4 (were it not for this move, exposing the weakness of d6, Black's scheme would be viable - this is one οί the few situatίons ίη the Sicilίan where a direct attack οη d6 works) 12 ... 0-0 130-0-0 e5 14 i..e3 i..e6 15 ~d5 'iVaS 16 a3 i..xd5 17 'iVxd5 'iί'xd5 18 :xd5 and Black has the type of ending Sicilίan players

48 Scheveningen

Vαriαtion

have bad dreams about, Nunn-Jansa, Dortmund 1979. 2) 8...lbb6 9 ~e3 and now: 2a) 9 •••ω 10 f4lbbc4 11 ~xc4 lbxc4 12 "e2 lbxe3 13 "'xe3 followed by 0-0-0 and f5 with good attacking chances according Ιο Glek. 2b) 9•••~e7 10 f4 d5 (1O... h6 11 "'f3 e5 12 li:)xc6 bxc6 13 0-0-0 is good for White) 11 ~b5 ~d7 12 exd5 exd5 13 "'f3.i.b4 14 ο-ο ~xc3 15 bxc3 lbxd4 16 ~xd7+ "'xd7 17 ~xd4 ο-ο 18 f5 with advantage Ιο White, Mark Tseitlin-Lukin, USSR 1987. 2c) 9•••d5 10 .i.b5 ~d7 11 exd5 exd5 12 "'e2 ~e7 130-0-0 ο-ο 14 lL:ιb3.i.e6 15 f4 ':'e8 16 h5 ~b4 17 "'d3! lbc4 (17 ......c8 18lL:ιd4+) 18 ~c5! .i.xc5 19 lL:ιxc5 gave White a clear advantage ίη Glek-Dydyshko, USSR 1991 (ίη fact the game οηlΥ lasted one more move: 19 .....d6?? 20 "xc4). 9 ~e3 "c7(D) 9 •...i.e7 is again the next game.

quickly as possible, while the f1bishop and h 1-rook are left at home because ίι is ηοΙ yet clear which squares are best for these pieces. Αι e2 the queen sets υρ tactical chances down the e-file and avoids attack from a black knight arriving at c4. 10 'itΊι5 can also be considered ίη this position, the one example I've seen being the quickplay game Κhalif man-Machulsky, Moscow Tal mem 1992, which continued 1O ... lL:ιxd4 11 ~xd4 b5 12 0-0-0 .i.b7 (l2 ... b4 looks like a more critical test) 13 ~h3lL:ιc5 14 ':'hel b4 15lL:ιd5 exd5 16 exd5+ Φd8 17 ~xc5 ~c8 (or 17 ... dxc5 18 d6~xd619 g6! with a dangerous attack) 18 ~xc8 ':'xc8 19 ':'e2 dxc5 20 d6 "'d7 21 g6 f6 22 gxh7 Ac6 23 ':'e7 ~xe7 24 dxe7+ Φχe7 25 ':'xd7+ 'iii>xd7 26 ...π+ Φd6 27 "fkxg7 with a won ending for White.

10 ...

~e7

This brings us back ίηΙο KarpovDorfman, but Black has an interesting alternative ίη 1O... b5. The idea is Ιο exploit an obvious weakness of playing "'e2, namely that the c3knight has nowhere Ιο go. The only real answer is Ιο meet the advance of Black's b-pawn by tαJ5, but, provided Black avoids playing ... ~e7, the knight move creates ηο immediate threats and Black is ηοΙ forced Ιο capture ίι. After 10••• b5 11 lL:ιxc6

"xc612 ~d4 (D):

10 "e2 Karpov's move is very logical ίη that ίι prepares queenside castling as

1) Nunn-Howel1, London 1990 continued 12••..i.b7? (stereotyped) 13 0-0-0 ':'c8 14 Ah3! b4 (Black knows what's coming but has ηο

Scheveningen

constructive way of avoiding ίι) 15 lDd5 a5 (15 ... 'i'a4 16lDxb4! a5 17 :a3! 'i'xb4 18 :b3 is very good for White) 16 c4! (cementing the knight ίη οη d5 ensures White oflong-term positional compensation when ίι is captured) 16... Φd8 17 ΦbΙ 1Dc5 (perhaps 17 ... exd5 would have been a better practical chance, a1though if White had found 18 exd5 1i'a6 19 :ο lDe5 20 :e3! he would have maintained a most promising posiιίοη, e.g. 20 ... :c7 21 f4 f6 221i'el! is very unpleasant for Black and 20 ... f6 21 f4 lDd7 22 'ifg4 'ifa8 23 .i.h3 :c7 24 1i'e6 is winning for White) 18 :f3 :c7 19 :e3 :d7 (now 19 ... exd5 10ses at once as after 20 exd5 the black queen won 't be able ιο stay protecting e8 for 10ng) 20 .i.h3! (increasing the pressure) 20...h6 21 g6 fxg6 22 :gl? (this complicates the issue whereas the simple 22 .i.xc5! 'ifxc5 23 lDf4 would have soon led ιο the disintegration of Black's position) 22 ... g5! 23 hxg5 hxg5 24 :xg5 lDxe4? (ίΙ was time ΙΟ capture the knight: 24 ... exd5! 25 exd5 1i'a6 26 :e8+ Φc7 27 .i.xd7 lDxd7 28 .i.xg7 .i.xg7

Vαriαtion

29 1:txg7 ':xe8 30 "xe8 .i.c8 is still good for White but ηοΙ so easy ιο win) 25 .i. b6+ 'itic8 26 :h5! (exchanging off Black's only active piece and exposing his back rank) 26 ... :xh5 27 "xh5 'itib8 (27 ... lDf6 28 :xe6!) 28 :xe4 exd5 (Black finally takes the knight which has been en prise for 13 moves) 29 :e8+ .i.c8 30 .i.xd7 "ii'xd7 31 :xf8 'i'e7 32 "e8 1-0. 2) Black should have played 12••• b4! 13lDds a5 when: 2a) 140-0-0? .i.a6 15 'i'el :c8 16 ':d2 .i.xf1 17 :xf1 'ii'b5 probably threatens Ιο take the knight and White can't retreat ίι Ιο e3 ίη view of 18 ...e5 crapping the bishop. 2b) White should settle for the queenless middlegame that arises after 14 WbS! 'ifxb5 (14 .....xc2? 15 .i.d3) 15 lDc7+ Φd8 16lDxb5 .i.b7 17 .i.g2 where he can still claim an edge. If Black continues 17 ... lDc5, then 18 ο-ο-ο! ίδ good as 18 ... lDxe4 (18 ... .i.xe4 19 .i.xe4lDxe4 20 :hel d5 21 :xe4) 19 :hellDc5 20lDxd6! gives White a clear plus. 11 0-0-0 bS Tactical ideas for White are already ίη the air, for example 12lDf5 exf5 13 lDd5 "d8 14 exf5, but a1though this ίδ quite good for White Black can improve by 12lDf5 ΜΙ 11•••lDxd4 doesn't look like an improvement as after 12 .i.xd4 ο-ο 13 f4 :e8 14 h5 b5 15 g6 White's attack ίδ coming a10ng nicely whilst Black's isn't even ουΙ of the starting blocks. Gufeld-Tilak, Calcutta 1992 concluded 15 ...fxg6 16 hxg6 hxg6

50 Scheveningen Variation 17 'iί'h2 ~f6 18 ~xf6 ~xf6 19 e5 dxe5 20 fxe5 b4 21 ~d3! ~b7 22 ~xg61-0.

1i'xc6 12 tαι:c6 13 ~d4 b4(D) Black forces White to sacήfice οη d5. but this move was itself virtually forced as 13...0·0 14 .1:g1 gives White a crushing attack. e.g. 14 ... b4 15 ~d5 exd5 16 exd5 'iί'xd5 17 'iί'xe7 'iί'xa2 18 g6 hxg6 19 J:txg6 and wins.

exd5 14 ~5 15 ~xg7 .1:g8 16 exd5 "c7 17 ~Ι6 The position ofWhite's bishop at f1 is shown υρ as a defect since 17 ':el ~e5 18 ~xe5 dxe5 19 f4 exf4 achieves nothing when d6 is impossible. If the other rook could come Ιο e1 Black would be finished. 17 ~5 18 ~xe5 dxe5 19 Ι4 Now White wins a third pawn for the piece since 19•••e4 fails to 20 d6 ~xd6 21 'iί'xe4+. Β lack' s king must remain stuck ίη the centre so one

must consider Karpov's sacήfice although ίη the subsequent play Black's resourceful defence almost saves the game. 19 ••. ~Ι5 20 ~h3 White takes time out to neutralise Black's counterpIay as 20 fxe5 at once allows the unclear 20 ... .1:c8 21 .1:h2 "a5. 20 ~xh3 21 .J:Σxh3 ':c8 22 fxeS After this Black activates his queen and Karpov is obliged to play with extreme accuracy ιο maintain his advantage. Ιη his notes Karpov suggested 22 b3 to prevent the following manoeuνre. 22 ••• 1i'c4! 23 ':dd3 "f4+ 23.....xa2 24 d6 (the threat is simply d7+) 24 ... ':c4 (White meets 24 ... .1:c5 by 25 'iί'f2 and 26 dxe7) 25 dxe7 'iί'a1 + 26 Φd2 'iί'xb2 27 .1:d8+ Φχe7 28 .1:d7+ ΦΧd7 29 'iί'xc4 and 23•.. hg5 24 hxg5 'iί'xa2 25 d6 ~xg5+ 26 .1:he3 ':c4 27 'iί'g2 are good for White. 24 ΦbΙ ':c4! The rook follows the queen's path with the aim of causing White some problems οη the back rank. 25 d6 .1:e4 26 .1:he3 .1:xe3 27 .1:xe3 "xh4 (D) IfBlack tήes to save his bishop by 27••. ~d8 (27 ...~f8 28 'ii'xa6 is even worse) he is crushed after 28 ':f3 "g4 29 e6 fxe6 30 d7+ winning the queen, so he quite rightly decides to cοπect,

Scheveningen grab as many pawns as he can while White is takίng his bishop.

28 "f3! Naturally ΩΟΙ 28 dxe7 at once since Black exchanges queens by 28 ... 'ifhl+. White's advantage lίes ίη the insecure black kίng, which causes trouble even when Black restores mateήal equality. "xg5 28 ... 28•••.i.xg5 29 e6 fxe6 30 :xe6+ Φd8 (30 ... Φd7 31 "f7+) 31 'i'c6! and 28...:xg5 29 'ifc6+ Φf8 30 dxe7+ Φχe7 31 a3 win forWhίte. 29 :el 29 "c6+ Φf8 30 dxe7 + "xe7 31 "h6+ :g7 is a lίttle better for White and thίs may ίη fact be hίs best lίne. 29 ... "g2? 29 ...'i'g4 was better, when ίι is far from clear if White can do more thandraw.

"fS :n

30 31 32 dxe7 Μateήal equality is re-establίshed but Black's kίng position makes his defensive task difficult. Detailed analysis of this position would take

Vαriαtion

us Ιοο far afield, but Black does ηοΙ seem Ιο have any real improvements hereafter and the task of defending both his kίng and his pawns SOOD overstretches hίs forces. aS 33 "r4 Φe8 34 'ji'b4+ 35 "xh7 'i'f3 36 'ji'b8+ Φe7 37 'ji'b4+ Φe8 38 "c4 'i'b7 39 b3 One of the most impressive features of this game is the way Karpov managed to conduct all the tactical operations with a vulnerable back rank. Many players, through nervousness or laziness, would have wasted a tempo οη b3 earlίer, and this might well have cost the game (note that although b3 was a good idea at move 22, the ροίηι was Ιο prevent the .....c4-f4 manoeuvre rather than to give the kίng aίr). 39 ... (D)

:e6

• ••• w."B Β'Β• ~ " • _

. .

•

.ΖΒ u

•

"Β Β Β

•

ΒΔ. Β ΔΒΔΒ Β ΒΦΒ Β:Β

•

40 :gl?! Perhaps Karpov assumed that the exposed kίng must succumb quickly to the combined attack of White's

52 Scheveningen Variation

queen and rook, but the task is much more difficult than appears at first. Ι suspect that if Karpov had realised this he would ηοΙ have been so hasty ίη giving back the pawn, because he could have waited for a more favourable moment. 40 1:txe5 41 1:tg8+ Φe7 42 'ifh4+ 'itd7 43 'iVf6! 1:te7 44 'ii'f5+ 'itd6 45 'iVxa5 1:te5 45..•'iVe4 would have held ουι longer, but the result is ηοΙ ίη doubt. 'ite6 46 'ii'd8+ 47 'itb2! Ι6 48 1:tf8 -'g7 49 'ii'c8+ 'itd5 50 'ii'c4+ 1-0

caught οη, perhaps justifiably so. Α possible line for Black is 10.••tDb6 11 ~b3 d5 12 'ife2 ~b4 with unclear complications while 10••• tDde5 is also interesting; Sznapik-Cvitan, Manila OL 1992 continued 11 ~b3 tDxd4 12 ~xd4 ~d7 13 -'e2 tDc6 14 0-0-0 (14 ~e3 is more natural but after 14 ... lϊΊa5 Black has a favourable version of the Velimirovic Attack) 14 ... lϊΊxd4 15 1:txd4 b5 16 1:thdl a5 17 tDxb5 ~xg5+ 18 hxg5 'ifxg5+ 19 f4 'ifxb5 with a good game for Black.

Game9 Zakic - Cvetkovίc Aosta 1989 1 e4 c5 2 tDf3 d6 3 d4 cxd4 tDr6 4 tDxd4 5 tDc3 e6 ~e7 6 g4 7 g5 tDrd7 8 h4 tDc6 9 J.e3 ο-ο If Black plays 9.•.a6 White may again reply 1Ο 'ifh5 when Black has nothing better than 10... 0-0 transposing back ίηΙο the game. 10 'iΊι5 (D) BTS2 concentrated οη the sideline 10 ~c4, but this has never really

Το select another line was ηοΙ so easy as all three possible queen moves (10 'ifd2 and 10 'ife2 are the other two) are quite dangerous for Black. Ι finally opted for 10 'ifh5, ηοΙ so much because of its strength, but because it is the most overtly aggressive and should ensure the right frame of mind for playing the Keres Attack where the slightest dίtheήηg can turn out to be fata1. One ροίηι ίη favour of 'ifh5 is that White is not committed Ιο a pawnstorm and can sometimes carry out a successful attack with just pieces.

Scheveningen Οη the downside the h-pawn is blocked and occasionally the queen will have Ιο retreat Ιο make way for its advance. 10 ••. a6 Jt's rare that Black can get by without this move ίη the Sicilian. Examples here: 1) 10•••dS?!. Α central strike is the recommended reaction Ιο a wing attack. but here ίι just emphasises how active the white pieces are ίη comparison with their counterρarts. Sax-Ehlvest. Reggio Emilia 1988. continued 11 Ο-Ο-ο! dxe4 (11 ...lDxd4 12 ':xd4 .i.c5 13 ':d2 .i.xe3 14 fxe3 is ;!; according Ιο Sax) 12 lDxe4 13lDxc6 bxc6 14 .i.d4! (White goes straight for the jugular) 14 ... e5 (14 .....xa2? 15lDf6+! .i.xf616 gxf6 is winning for White) 15 .i.c3 "xa2 (Black may as well take the pawn as 15 .....c7 16 ':xd7 .i.xd7 17 .i.d3 g6 18 "h6 allows White a winning attack) 16 ':xd7!! (removing one of the defenders of f6) 16 ... .i.xd7 17 lDf6+! .i.xf6 (or 17 ... gxf6 18 gxf6 .i.xf6 19 ':gl+ .i.g7 20 'ifh6) 18 gxf6 'ifal+? (18 ...:fd8 should be met by 19 .i.d3! as after 19 "g5? "al+ 20 Φd2 'ifdl+!! 21 Φe3 "g4 Black has managed Ιο defend himself; relatively best is 18 .....e6!. even though White is well οη Ιορ after 19 .i.h3 'ifxf6 20 .*.xd7) 19 Φd2 "a4 (Black's main idea is Ιο transfer his queen Ιο the kίngside. but White's next move dashes all his hopes) 20 b4! :fd8 21 .i.d3 gxf6 22 :al! (Ι bet you weren't expecting that one) 22 ... 'ifb5 23 "xh7+ Φf8

"a5

Variαtion

24 "h6+ Φe7 25 .i.xb5 cxb5 26 "e31-0. 2) 10••.':e8 11 0-0-0 and now 11 ... a6 transρoses back ίηΙο the main game. but Black has a couple of other trίes: 2a) 11•••.i.f8 12 lDxc6 (12 f4?! lDxd4 13 .i.xd4 e5) 12 ... bxc6 13 f4 d5 14 .i.d3 g6 15 "f3 with good attackίng chances for White according Ιο Sax. 2b) 11.•• lΔxd4 12 .i.xd4 .i.f8 13 .i.d3 a6 14 e5 g6 15 "g4 .i.g7 (15 ...dxe5 16 .i.e3 is dangerous for Black) 16 f4 dxe5 17 fxe5 .i.xe5 18 .i.xe5lDxe5 19 "e2 with a powerful attack Ιο come. 11 0-0-0 The move ... :e8 (which is usually followed by ... .i.f8) is a common prophylactic measure against a whίte attack ίη the Sicίlian. Ιη this particular position. though. ίι has been more common for Black Ιο immediately commence queenside operations with 11.••lDxd4 12 .i.xd4 bS after which White has chooses between 13 .i.d3 or 13 f4. Ι (ΙΟ) am. however. recommending a third possibίlity. 13 eS!? (D).

:e8

54 Scheveningen

Vαriαtion

The idea is simple; Ιο open as many lines as possible for White's aggressively placed pieces, especially the diagonals pointing towards the black king. As far as Ι know 13 e5 has only οccuπed ίη one fairly obscure game, which means that all the analysis given below is οήgίnal and untested. After 13 e5 Ι have examined five possible replies, of which lines '4' and '5' appear Ιο be the most cήtical:

1) 13•••d5?!. Closing the centre allows White Ιο develop an attack free of charge. Α possible continuation: 14 .i.d3 g6 15 'iί'g4 tΔc5 16 .i.xc5!? .i.xc5 17 h5 .i.xf2 18 hxg6 fxg6 19 :xh7 .i.e3+ 20 ~bl 'iί'xg5 21 'iί'h3 with advantage ιο White. 2) 13••. b4? 14 exd6 .i.xd6 15 .i.d3 g6 16 'ii'h6 .i.f4+ (16 ... e5 17 h5! is crushing, e.g. 17 ... exd4 18 hxg6 .i.f4+ 19 'ίtbl tΔf6 20 g7! fol10wed by mate) 17 ~bl e5 18 tΔe2 tΔf6 (18 ... exd4 19 tΔxf4 wins for White) 19 .i.c5! :e8 20 .i.xg6 fxg6 21 :xd8 :xd8 and now 22 f3!, avoiding 22 ... tΔg4, wins for White. 3) 13..•.i.b7 (this also seems to 10se, though the vaήations are much more complicated) 14 exd6 with the lines: 3a) 14••..ixhI15 .i.d3 g6 16 'ii'h6 e5 17 :xh 1 exd4 18 dxe7 'iί'xe7 19 l"Δd5 'ii'e5 20h5! 'iί'xd5 (20 ... 'iί'g7 21 tΔe7+ ~h8 22 hxg6 'iί'xh6 23 :xh6 fxg6 24 tΔxg6+) 21 hxg6 'iί'xhl+ (21 ... :fe8 22 g7! - this g6-g7 move is an important ροίηι, which occurs ίη several variations) 22 'iί'xhl hxg6 23 .i.xg6! fxg6 24 'iί'd5+ ~g7 25

'iί'xd7+ ~g8 26 'iί'd5+ ~h7 27 'iί'xd4 and White has a material ad-

vantage which he should have ηο difficulty converting ίη view of the exposed black king. 3b) 14•••.i.xd6 15 .i.d3! g6 16 'iVh6 (D) and Black can play ... e5 with or without giving a bishop check:

3bl) 16•••e5 17 h5! 'iVe7 (grabbing the material leads to a quick defeat, e.g. 17 ....i.xhI18 :xhl exd4 19 hxg6 .i.f4+ 20 ~bl tΔf6 21 g7!) 18 hxg6 fxg6 19 .txg6! .txhl (or 19 ... hxg6 20 'iί'h8+ ~f7 21 :h7+ ~e8 22 :xe7+ .i.xe7 23 'iί'g7 winning for White) 20 :xhl hxg6 21 'iVh8+ 'ίtf7 22 'iVh7+ (a simpIer line is 22 :h7+ ~e8 23 :xe7+ .i.xe7 24 'iί'hl! with advantage Ιο White, although perhaps less than after 22 'ii'h7+) 22.••~e8 23 'iVxg6+ and now: 3bll) 23•••:f7 24 tΔe4! ~d8 (or 24 ....i.c7 25 :h8+ tΔf8 25 'iί'c6+ +-) 25 tΔxd6 :g7 26 'ii'e4 :b8 27 tΔf5 '6'xg5+ 28 f4 '6'g6 29 :h6! '6'g2 30 :h8+ and White wins. 3b12) 23•••~d8 24 tΔd5 exd4 25 tΔxe7 .i.xe7 26 :h7 and White, with

Scheveningen

his passed pawns and ongoing attack should be winning. 3b13) 23 ... 'ii'f7 24 "xd6 exd4 25 lΔd5! wins. 3b2) 16•••.tf4+ 17 'iPbl e5 (the alternatiνe 17 ... lΔe5 18 .te4! is good for White) 18 .te3! (18lΔe2 .txhl 19 ':xh 1 lΔf6! 20 f3 lΔh5! is much less clear) 18 ....txe3 19 fxe3 .txhl 20 ':xhl and Ι can't see how Black surνiνes. As 2O ••• lΔcs 21 h5 lΔxd3 22 hxg6 is all oνer, he must try 20...f5, when 21 h5 gxh5 (21 .....e8 22lΔd5! and 21 .....e7 22 hxg6 'ii'g7 23 "xg7+ 'iPxg7 24 ':xh7+ 'iPxg6 25 1:Ixd7 are winning for White) 22 g6lΔf6 23 .txf5 ':a7 24 ':gl ':e8 (24 ..."e7 25 g7! wins for White) 25 lΔe4 lΔxe4 26 .te6+! ':xe6 27 gxh7+ 'iPh8 28 ':g8+ "xg8 29 hxg8"+ 'iPxg8 30 "xe6+ and White is winning .. 4) 13...g6 (one of the points behind "h5 ίη the Keres Attack is Ιο tempt Black ίηΙΟ playing this weakening moνe) 14 'ii'e2 (it's ηοΙ clear where the queen should retreat; 14 "g4 dxe5 15 .te3looks quite promising and 14 "f3 also merits some attention as after 14 ... dxe5 15 .txe5 lΔxe5 16 "xa8 'ii'a5 17 'ii'e4 b4, White has the strong rejoinder 18 lΔd5!) 14...dxe5 (14 ....tb7 15 ':'h3 b4 16 exd6 .txd6 17 lΔe4 is better for White) 15 .txe5 (D) and Black has seνeral tries: 4a) 15... b4 16 .tg2! ':'a7 (after 16 ... bxc3 17 .txa8 cxb2+ 18 .txb2 White has the adνantage) 17 lΔe4 (17 .tb8!?) 17 ...'ii'a518.td4':'c719 'iPbl and White is much better.

Vαrίation

4b) 15...'ii'&5 16 .td6! .txd6 (or 16... b417.txe7bxc318.txf8"xa2 19 bxc3 and White can follow υρ with "c4) 17 ':xd6 b418lΔe4! with a clear adνantage for White. 4c) 15....tb7 16 ':h3 b4 17 ':'hd3! .tc6 and now: 4cl) 18 lΔe4 .td5! (18 ... .tb5? 19 ':'xd7! .txd7 20 lΔf6+ .txf6 21 gxf6 and the threat of"d2-h6 is decisiνe) 19 ':'xd5 exd5 20 ':xd5 is νery unclear. 4c2) 18 ':'xd7 .txd7 19lΔe4 appears νery dangerous for Black but Ι haνen't found a way Ιο make the white attack work. 4c3) 18 .tg2! .txg5+(18 ....txg2 19 ':'xd7 "e8 20 lΔe4 .txe4 21 "xe4 and 18 ... bxc3 19 .txc6 "b6 20 .txc3 are good for White) 19 ΦbΙ! (D). Α

most confusing position with of the six minor pieces en prise and the other one caught ίη a fatal ρίη. Ι belieνe that White should emerge νίctοήοus from this chaos. For example: 4c31) 19...'ii'c8 20 ':'xd7 .txd7 21 hxg5 bxc3 22 .txa8! .tb5 23 "f3! "xa8 24 "f6 wins. fiνe

56 Scheveningen Variation

4c32) 19... lΩxe5 20 ':xd8 .1xd8 21 'ii'xe5 .1xg2 22 h5!! is very strong as after 22 ... bxc3 23 h6 f6 24 'ii'xe6+ Φh8 25 'ii'd7 ':g8 26 'ii'f?! White threatens both ':d7 and ':xd8. 5) 13••.dxeS 14 .1xe5 and now: 5a) 14..•g6 15 'ii'h6? (15 'ii'e2 and 15 'ii'f3 transpose ίηΙο lίne '4') 15 ... lt)xe5 16 ':xd8 ':xd8 17 .1g2 ':b8 and with ... lt)g4 and ....1f8 οη the cards White is going Ιο lose his queen and remaίn a rook behind. 5b) 14.••'ir'b6 15 ':xd7! .1xd7 16 .1d3 g6 17 'ii'h6 f6 18 .1xg6! +-. 5c) 14•.•.1b7 15 .1xg7! Φχg7 (15 ....1xh 1 16 .1d3 f5 17 gxf6lt)xf6 18 .1xf6 ':xf6 19 'ii'xh7+ Φf8 20 .1g6 wins) 16 'ii'h6+ Φh8 (16 ...Φg8 17 .1d3 f5 18 gxf6) 17 .1d3 f5 18 g6! ':f? 19 gxf7 .1xhl and now 20 'ii'xe6! +-looks the strongest. 5d) 14•. :iνe8 (the toughest defence) 15 .1g2! (15 .1d3 is met by 15 ... f5 whίlst 15 .1xg7 doesn'tlook quite good enough, e.g. 15 ... Φχg7 16 "h6+ Φh8 17 ':xd7 .1xd7 18 .1d3 f5 19 gxf6 ':f?! 20 fxe7 'ii'xe7 and White has some compensation for the exchange, but nothing more)

15••.lt)xeS (15 ... ':a7 16 .1d4 ':c7 is possible, although White is well centralised after 17 ':hel) 16 .1xa8 b4 (16 ....1d7 17 .1e4 g6 18 'ii'e2 b4 19 lt)bl .1b5 20 "el is good for White) 17 lt)e4! (D) (cuts off the bishop' s retreat, but the tactics seem Ιο be workίng ίη White's favour) and now Black has:

5dl) 17.•..1d7 18 ':hgl! (suddenly White has kingside threats agaίn) 18 ... Φh8 (18 ... 'ii'xa8 19lt)f6+ .1xf6 20 gxf6 lt)g6 21 ':xd7) 19 g6! lt)xg6 20 lt)g5 h6 (20 ... .1xg5+ 21 hxg5 is very good for White) 21 lt)xf?+! with a winning positίon for White . 5d2) 17.....a4 and now White can play: 5d21) 18 Φbι.1d7 19 g6! lt)xg6 20 lt)g5 ':xa8 (20... h6 21lt)xf?! lt)f4 22lt)xh6+ gxh6 23 ':hgl+ wins for White) 21 'ii'xh7+ Φf8 22 h5 .1xg5 23 hxg6 ':c8 24 "h8+ Φe7 25 'ii'xg7 "xc2+ 26 ΦaΙ "xg6 and now 27 "xg6 fxg6 28 ':h7+ Φf6 29 ':hxd7 ':c2 is probably only equal, but 27 "d4 gives White a strong attack.

Scheveningen

5d22) 18:thgl!? (D) with a further branch:

5d221) 18M....xa219lbf6+ ~xί6 20 gxf6 llJg6 (20.....al + 21 'iItd2 :d8+ 22 'iIte2 :txdl 23 :txdl 24 "g51eads ΙΟ mate because after 24 ...•b5+ 25 'iItel the bishop οη a8 prevents Black's ... llJf3#) 21 fxg7 :te8 (21 .....al+ 22 'iItd2 :td8+ 23 'iIte3! and 21 ... Φχg7 22 'it'e5+ 'iItg8 23 b3! are both winning for White) 22 'it'c5! .al+ (22 ... llJf4 23 ~c6 'it'al+ 24 'iItd2 :td8+ 25 'iIte3! :txdl 26 .f8#!) 23 'iItd2 "xb2 24 ~c6 :td8+ 25 'iIte2 :txdl 26 :txdl 'iItxg7 27 h5 'it'e5+ (27 ... llJe5 28 ί4! wins) 28 .xe5+ llJxe5 29 :td8 llJxc6 30 :txc8 llJd4+ 31 'iItd3 llJf3 32 :ta8 and White should win the ending. 5d222) 18...~ι619 'iItbl ~d7 20 ~b7 .b5 21 "f3! :tb8 (this looks strong but White has a tactical way ιο solve his problems; 21 ... llJxh4 22 '6'd3! and 21 ...llJe5 22 "'g3! are also good for White) 22 llJf6+! ~xί6 (22 ... gxf6 23 gxf6 .xb7 24 .xb7 :txb7 25 fxe7 ~e8 26 h5 +-) 23 gxf6 "xb7 24 "xb7 :txb7 25 h5 gxf6 26 hxg6 hxg6 27 :td6 ~b5 28

"a5

Vαriαtion

b3 and White has the edge ίη this ending as Black's queenside pawns are weak. For example, 28 ... Φf8 29 'iItb2 'iIte7 30 :tgdl ί5 31 a4 bxa3+ 32 'iItxa3 with c4 ιο follow is good forWhite. 12 Ι4 The most natural, but 12 llJxc6 bxc6 13 ί4 is also interesting. 12 ... ~f8 (D) The main alternative is 12...llJxd4 13~xd4b5(13 ... e514fxe5dxe515 ~xe5llJxe5 16 .l:xd8 ~g4 17 :txa8 :txa8 18 ~e2 ~xh5 19 ~xh5+) 14 Ι5 ~f8 (14 ... b4 15 fxe6 fxe6 16 ~c4 bxc3 17 ~xe6+ 'iIth8 18 ~xg7+ 'iItxg7 19 "h6+ 'iIth8 20 g6 +- is a variation given by Polugaevsky) 15 ~h3 b4 16 fxe6 fxe6 17 ~5! (an improvement over the 17 g6 h6 18 llJd5 of Sax-Polugaevsky, Haninge 1989 when Black was able ιο defend with 18 ...:tb8) and now: 1) 17...exd5 18 g6 when White threatens both mate οη h7 and "xd5+. 2) 17.. ':b8 now faίls Ιο 18llJc7! g6 19l1Jxe8!, as ίη Hector-Ghitescu, Palma 1989. 3) 17...~b7 is refuted by the sacrifιce 18 llJf6+! according ΙΟ Sax, who ηο doubt had the following variation ίη mind: 18 ... gxf6 19 gxf6 llJxf6 20 :thgl+ ~g7 (20... Φh8 21 'it'xe8 .xe8 22 ~xf6+ ~g7 23 :xg7) 21 :txg7+ 'iItxg7 22 .l:gl+ Φh8 23 "xe8+! "xe8 24 ~xf6#. 4) 17...g6 18 "f3! exd5 19 e5! dxe5 (19 ... ~b7 20 e6llJe5 21 ~xe5 dxe5 22 "f7+ 'iIth8 23 "xb7 d4 24 "e4 is very good for White) 20

58 Scheveningen

Vαriαtion

'Wxd5+ ~h8 21 .te3! and White has a clear advantage (Sax).

13 rs!? This move is very hard Ιο resist when you know what happens after 13 ... exf5. Οη the other hand ίι does concede control of e5 a little prematurely so White might prefer one of the alternatives: 1) 13.td3 g6 (both 13 ... lbxd4 14 .txd4 e5 15 .tc4! g6 16 .txt7+ and 13 ... lbc5 14 lbxc6lbxd3+ 15 ':'xd3 bxc6 16 e5, with the idea of lbe4-f6, are good for White) 14 'Wf3 lbxd4 (14 ....tg7 15 lbde2 followed by h5 is also good for White) 15 .txd4 e5 16 .te3 exf4 (Landenbergue-Magerramov, Bad Wδήshοfeη 1993) and now White should have played 17 'Wxf4, as Black cannot then play 17 ... lbe5 ίη view of 18lbdS J..g7 19 .tb6. 2) The Sax solution, 13 lbxc6 bxc6 14 .td3, was tried ίη WegnerBischoff, Bundesliga 1991. After 14•.•'ii'a5 15 e5 g6 16 'ii'f3 (16 'Wg4) 16•.. dxeS: 2a) White totally lost his way with 17 'ii'xc6 ':'b8 18 'ii'a4 'ii'xa4 19

lbxa4 exf4 when he couldn't even recapture ίη view of20 ...':'b4!. 2b) Instead of grabbing a useless queenside pawn, he could have given Black a torήd time with 17 h5!, for example 17•• .J.g7 (17 ... exf4 18 hxg6 hxg6 {18 ... fxg6 19 ':'xh7!} 19 'Wh3 .tg7 20 .txf4 transposes) 18 hxg6 and now: 2bl) 18•••fxg6 19 ':'xh7! ~xh7 20 'ii'h5+ ~g8 21 .txg6 ':'e7 22 fxe5 and the opening of the f-file will prove fatal. Οη 22 ... lbxe5, 23 .tc5! looks the clearest. 2b2) 18•••hxg6 19 'Wh3 exf4 (or 19 ... Μ 20 lbe4!) 20 'Wh7+ ~f8 21 .txf4 e5 (21 ... lbe5 22 lbe4! looks strong) 22 .te3 and it's hard Ιο see Black surviving for long. 13 ..• exf5? (D) Black must have been counting υροη 14 lbxf5 lbde5 when he has everything under control. Instead 13.••lbde5 would have left the situation unclear.

14 'ii'xf1+!! This thunderbolt had obviously ηοΙ been taken ίηΙο account by Black.

Scheveningen

Vαriation

14 ••• Φxf7 14...~h8 15 ~c4 is completely hopeless. 15 ~c4+ :e6 Both 15 ...~g6 and 15 ...~e7 permit mate ίη one and the only other legal move, 15...d5, leads to a lost position after 16 ~xd5+ :e6 17 lbxe6 'i'a5 18lbc7+ ~e7 19lbxa8.

16 lbxe6 'ii'a5 16...'ii'e7 17lbc7+ ~g6 18 h5#. 17 lbc7+ ~e7 18 lb3d5+ ~d8 19 lbe6+ There is too much action ίη the centre to be distracted by a rook.

19 20 ~d2! 21 ~b3 22 :hel (D)

~e8

'ii'a4 'ii'xe4

22 :b8 As the queen can't go anywhere (22 ... 'i'xh4 fails Ιο 23lbec7++ ~f7 24 lbf4+ d5 25 ~xd5#) Black decides Ιο save his rook.

23 :xe4 24

fxe4

:n!

The attack is far from over.

lbce5 25 ~M W 26 lbdc7+ ~e7 27 :dl! lbc5 Black could have resigned rather than play this move. 28 lbxc5 a5 28... dxc5 29 ~xc5 is mate. 29 lb5a6 axb4 24 ...

30 lbxb8

lbxh4

31 lbb5 ~g4 32 :xd6 e3 33 :b6 e2 34 1%xb7+ ι-ο Finally the black king meets his fate; 34 ... ~e8 35 ~f7+ ~d8 36 lbc6+ ~c8 37 :c7 is mate.

3 Classical Variation Ιη

the first edition of this book Ι (JN) 2 ~f3 5 ~c3 d6, which can also occur with the move order 2 ...d6 and 5 ... lίk:6, with the name 'Classical Vaήation'. This nomenclature seems Ιο have caught οη, so ίι has been maintained for subsequent editions. The line we are recommending against the Classica1 is 6 i.g5, generally ca11ed the Richter-Rauzer Attack even though the treatment used today doesn' t seem Ιο owe anything Ιο Richter. This line is very common ίη practice, so there is a large body of theory. Ιη genera1 we will keep Ιο the main lines ίη the proposed repertoire, but where there are interesting sidelines we will give them a brief mention. The idea of 6 i.g5as ίι is played today is based οη a quick "d2 and 0-0-0, exerting pressure down the dfile and restraining Black from active play ίη the centre. Black's most solid reply is the natural 6 ... e6 7 "d2 i.e7 8 0-0-0 ο-ο, but despite its solid appearance ίι can often lead Ιο sharp tactica1 p1ay. This is covered ίη Game 10. Sometimes Black postpones ... 0-0 so as ιο delay exposing the king Ιο a possible pawn storm. The line 6 ... e6 7 "d2 a6 8 0-0-0 i.d7 is the subject chήsteηed the line 1 e4 c5 ~c6 3 d4 cxd4 4 ~xd4 ~f6

of Game 11 while 6 ...e6 7 "d2 a6 8 0-0-0 h6 appears ίη Game 12. Fina11y, some p1ayers have eΧΡeή mented wίth the omission of 6 ... e6, ηοΙ fearing the doubled pawns resu1ting from i.xf6, and the most popu1ar of these ideas, 6 ... J.d7, forms the basis ofGame 13. Unusua1 lines involving ...e6 are dea1t with ίη Game 10, while the others are ίη Game 13. Game 10 Liss-Leko Budαpest

1993

1 e4

2 ~f3

d6 cxd4 llli6

3 d4 4 ~xd4 5 ~c3 6 i.gS

~c6

e6 Other moves are considered Game 13.

ίη

7 'iί'd2 (D) 7 •.. i.e7 7 •.•a6 appears ίη Games 11 and 12. There are two other alternatives for Black: 1) 7...h6 (7 ... i.d7? is just a mistake: 8 ~db5) 8 i.xf6 gxf6 (after 8 .....xf69~b5 and 100-0-0White wins the d6-pawn) 9 0-0-0 (playing for ο-ο is a1so slight1y better for

Clαssical

White, e.g. 9 :dl a6 10 .te2 h5 11 ο-ο .td7 12 ~b3 'ilc7 13 ΦhΙ 0-0-0 14 f4, or 9 .te2 h5 10 ο-ο a6 11 ΦhΙ! .td7 12 f4 'ilc7 13 :f3! ~xd4 14 'ilxd4.te7 15 :dl h4 16 :fd3 :d8, Van der Wiel-J.Piket, Leiden 1986, and now 17 .tf3 giνes White an edge) 9 ... a6 10 f4 .td7 11 .te2 h5 (11 ...'ilb6 12 .th5 {12 ~b3 is a1so perfectly playable} 12 ... 'ilxd4 13 "xd4 ~xd4 14 :xd4 :g8 15 g3 .te7 16:n .tc6 17 f5 :g5 18.te2 with adνantage, Bondareνsky-Bot vinnik, USSR Ch 1951) 12 ΦbΙ 'ilc7 (or 12... 'ilb6 13 ~b3 0-0-0 14 :hf1 Φb815:f3 .te716:h3 h417 'ilel :hg8 18 .tf3 again with a plus for White, Nunn-Cabrilo, Νονί Sad OL 1990) 13 :hf1 0-0-0 (13 ....te7 14:f3 ~xd4 15 'ilxd4 'ilc5 16 'ild2 .tc6 17 :e3 'ila5 18 a3 :d8 19 .tc4 giνes White an edge, Liberzon-Botvinnik, USSR 1967) 14 ~b3 Φb8 15 :f3 .te7 16 :h3 h4 17 'ilel and Black has ηοι yet equalised, Vasiukov-Shamkoνich, Dubna 1973. Ιι would be fair Ιο conclude that those who belieνe ίι is worth a temρo Ιο encourage White Ιο capture οη f6 are members of a dying breed.

Variation 61

2) 7... ~xd4 8 "xd4 .td7 and now ίη the game Ivanchuk-Saloν, Linares 1992, White produced the powerful noνelty 9 ~b5!. Play continued 9.....b8 (9 ....tc6 10 e5 dxe5 11 'ilxe5 is ± according Ιο Sa1ov) 10 0-0-0 (10 .txf6 gxf6 11 "xf6 :g8 is ηοΙ as promising for White, but as Sa10ν points ουΙ 10 :dl might have been even stronger) 10...8611 ~d6+! (clearer than 11 .txf6.txb5 when 12 .txg7 e5 13 .txe5 .th6+! 14 ~bl dxe5 15 .txb5+ axb5 16 "d7+ Φf8 17 "xb5 is a mess) 11 ....txd6 12 .txf6 .tf4+! 13 ΦbΙ gxf6 14 "xd7+ ΦΙ8 15 g3 .tc7. Sa10ν considers this position to be only slightly better for White (which Ι find hard Ιο belieνe). 2a) He now gives the solid νari ation 16.tc4 b5 17 .tb3 (17 .txe6 fxe6 18 'ilxe6 "b6 is unclear) 17 ...:a7 18 f4.ta5 19 "d6+ "xd6 20 :xd6 ~e7 21 :hdl as~, when the presence of opposite-coloured bishops give Black good drawing chances. 2b) lνanchuk Ρrefeπed a more aggressiνe approach and played 16 .th3, telegraphing his intentions to sacrifice οη e6. After 16 ... b5 17 "c6! :a7 18 .txe6 fxe6 19 "xe6 .td8 (now 19 .....b6 loses to 20 'ilc8+) 20 :d6! .te7 21 :hdl :g8 he opted for 22 :Xa6 :xa6 23 "xa6 Φf7 when White had gained a fourth pawn but lost the initiative. Much better would haνe been 22 :b6! 'ile8 23 f4, maintaining his bind οη the positίon.

8 0-0-0

ο-ο

Clαssical

Variation

Or 8••. lαι:d4 9 "xd4 0-0 and now BTS2 recommended 10 e5 dxe5 11 'iί'xe5, but cuπeηt theory suggests that Black can play 11 ....td7 12 h4 :Ιc8 13 :Ιh3 'ikc7 14 'ikxc7 :Ιχc7 15 lbb5 .txb5 16 .txb5 :Ιfc8, with what Κraιnnik told me (JN) was 'a drawn ending'. So now we are recommending 10 Ι3 (D) with the possibilίtίes:

1) 10.....a5 11 'iί'd2 :Ιd8 12 ΦbΙ 'ikc7 13lbb5 'ikd7 14 g4 a6 15lbd4 b5 16 h4 .tb7 17 .te3 'iί'e8 18 h5 e5 19 lbf5 d5 20 lbxg7! .!bxe4 (White wins after 20 ... Φχg7 21 .th6+ Φh8 22 'iί'g5 'iί'g8 23 'iί'xe5 :Ιe8 24 g5) 21 fxe4 Φχg7 22 .th6+ Φg8 23 'iί'c3 and Black's exposed king gives White an obvious advantage, J .Benjamin-Kogan, US Ch 1984. 2) 10•••e5 11 'iί'd2 .te6 12 ΦbΙ :Ιc8 13 h4 :Ιc5 14 a3 'iί'b8 15 .te3 :Ιcc8 16 .td3 b5 17 .!bd5 .txd5 18 exd5 .td8 19 .tf5 :Ιc4 20 b3 :Ιc7 21 g4 with a clear advantage for White. Chandler-Lobron, Bie11987. 3) 10•••86 was played ίη Kυzmίη ΟΗ, Lvov Ζ 1990, which continued 11 e5 dxe5 12 'ikxe5 "e8 13 .te2

.td7 14 'iί'g3 'iί'c8 15 :Ιd3 :d8 16 'iί'h4 with a satisfactory position for Black. However, having rejected e5 at move 10 there doesn 'ι seem Ιο be any special reason Ιο play ίι at move 11. Starting kingside expansion by 11 h4 is more consistent with playing f3. The positίon after 8... 0-0 is one of the most important ίη the whole Sicilίan Defence and despite decades of practical eχpeήeηce ηο definite assessment can be given. Although 9 lbb3 was often played ίη late fifties and early sixties, ίι feH ίηΙο disuse and 9 f4, which has always been regarded as the main lίne, became virtually universal. However, 9 lbb3 has been regaining popularity and now rivals 9 {4 for the distinction of being considered the 'main lίne'. 9 .!bb3 (D) Notice that 9 .txf6? is bad since Black can play 9 ....txf6 10 lbxc6 bxc6 11 'iί'xd6 'iί'b6 when the threats to b2 and f2 are more than enough compensatίon for the pawn.

The main ροίηι of 9 lbb3 is that an attack against the pawn

ίι unveils

Classicαl Vαriαtion

οη d6 and so prepares J.xf6. Black has four main methods of cοunteήng White's plaπ: The fιrst is Ιο sacrifice the dpawn, for example by 9...h6 10 J.xf6 J.xf6, but a number of games prove that White caπ gain aπ advaπtage. Α second plan is ΙΟ allow J.xf6 and then recapture with the pawn. This leaves Black's king somewhat exposed, but White has ηο immediate method of launching aπ attack. More serious is that Black's central pawns are inflexible and White may have enough time Ιο start a kingside pawn storm. The assessment of this line depends οη the speed of the respective attacks. Black's third plan is Ιο counter White's pressure οη d6 directly by playing 9 ... a5 10 a4 d5 (moreover the immediate 9 ... d5 is just about possible). The final plan is Ιο counterattack f2 by 9 ..."b6, gaining enough time Ιο defend d6 by ...:d8. With the exception of the first, all th~se lines are playable. 9 ••• 86 Apart from the two major alternatives, there are a number of less common ideas: 1) 9 •.•dS 10 J.xf6 J.xf611 exdS J.xc3 (l1 ... ιtJM 12 a3 ιtJxd5 13 ιtJxd5 exd5 14 "xd5 is slightly better for White since aIthough Black has the two bishops ίι is ηοι easy for him Ιο avoid the exchaπge of queens) 12 "xc3 exdS (D) aπd now: la) 13 g3 J.g4 14 f3! (a better choice than 14 :d2 d4! 15 "c5 {15

ιtJxd4 "d5!} 15 ... J.f3 16 :gl :e8

17 J.b5 :e5 18 "c4 "e7 19 "f1! which is unclear, Yudasin-Utemov, USSR 1989) 14 .....g5+ (14 ... J.e6 15 ιtJd4 :c8 16 ιtJxe6 fxe6 17 J.h3! is better for White) 15 ~bl J.f5 16 J.d3 J.g6'? (16 ... J.xd3 17 :xd3 t) 17 f4 "h5 18 g4! (the start of aπ elegant series of moves leading Ιο victory) 18 .....xg4 19 f5 J.h5 (19 ...d4 20 "d2 J.xf5 21 :hgl "h3 22 :g3 "h5 23 :g5 "f3 24 :η and the wriggling is at an end) 20 :dgl d4 21 "d2 22 "g5 g6 23 ιtJd2! (certainly ηοΙ 23 "xh5? "xhl!) 23 .....d5 24 J.e4 "a5 25 "xh5 ιtJb4 26 a3 1-0 Van der Wiel-Mirall~s, Cannes 1990. lb) 13 ιtJd4 should give White a small but safe advaπtage. lc) 13 J.bS!? "g5+ and now both 14 "d2 "xg2 15 :hgl aπd 14 <ϊ&i>bl d4 15 "c5 "xg2 give White aπ attack ίη return for the pawn. 2) 9•••ιtJas 10 ~bl ιtJxb3 11 cxb3 a6 12 f4 b5 13 J.xf6 gxf6 14 J.d3 ~h8 15 f5 b4 16 ιtJe2 e5 17 J.c4, Anand-Mateo, Dubai OL 1986, is worth mentioning because it is a perfect example of what Black should

"f3

Classicαl Vαriαtion

avoid. His kίng position has been weakened without any comρensat ing queenside attack and Black has played ... e5 at a moment when White can reply ~c4 Ιο gain control of d5. 3) 9••• h6 (ίι now seems firmly established that this line is good for White) 10 .i.xf6 ~xf6 11 "xd6 ~xc3 (11 ... 'iWb6 12 'iνc5 'iνc7 13 g3 ~e7 14 'iνe3 a6 15 f4 b5 16 ~g2 lbe5 17 'iνe2 was a1so promising for White ίη Marjanovic-Barlov, Yugoslav Ch 1985) 12 bxc3 'ii'h4 (D) and now:

3a) 13 g3 'iνf6 (ηοι 13 .....xe4? 14 ~d3 and .th7+) 14 "c5 e5 15 ~c4 ~e6 (15 ... ~g4 16 :d6 'iνg5+ 17 ~b2 ~f3 18 :el :ac8 19 h4 'iνg4 20 'ii'e3 :fd8 21 :xd8+ :xd8 22 ~d5 ~g2 23 :gl 'iνf3 24 g4 with advantage Ιο White, Lobron-Kunsztowicz, Bad Neuenahr 1984) 16 ~xe6 (16 ~d5 :ac8 17 'ii'e3 b6 18 ~b2 'iνe7 19 :d3 :c7 20 ~al :fc8 21 :bl 'iνa3 22 'iνcl "a4 23 lbd2 lba5, Gallagher-Van der Poel, San Bernardino 1994, and now instead of 24 'ii'b2? lbc4! with a roughly level game, White could have kept some

advantage with 24 :b4) 16... 'iνxe6 (16...fxe6!? was the reason Ι avoided the exchange οη e6 ίη the above mentioned game; afterwards, Ι discovered that this had never been played although ίι does have the advantage of cοveήng d5 and opening the f-file) 17 :d6 "h3 18 'ii'e3 :fd8 19 :d5 and now both 19..':dc8 20 :hdl :c7 21 f4 exf4 22 'iνxf4 :e7 23 tL1c5, as ίη Chandler- Τοπe, London 1984 and 19.....g2 20 :hdl 'ii'xh2 21lbc5, Κlovan-Ta1, Jurmala 1983, were good for White. 3b) 13 f3!? (this line is a little forgotten but ίι does seem very good forWhite) 13 ...:d814 'iνc7! (maximising the queen's annoyance va1ue) 14 ... Axd1+ 15 ~xdl 'iνf6 (15 ... e5 16.tc4+) 16~cl'ii'xc317 .tb5e5 18 ~xc6 ~e6 (18 ... bxc6 gives White the ορΙίοη of 19 'ii'd8+) 19 :dl bxc6 20 :d3 ± Serper-Reznikov, USSR 1982. We now move οη Ιο the major lines: 4) 9...&5 (quite popular ίη the 1980s, but ίι hasn't been seen so much oflate) 1084 d5 11 ~bS! (D) (first appeared ίη the game Tal-Sisniega, a1though the idea belongs Ιο νίιoliη~) and now Black has: 4a) 11...lbxe4 12lbxe4 dxe4 13 'ii'xd8 ~xd8 14 .txd8lbxd8 15lbc5 f5 (or 15 ... b616lbxe4~b717 :hel ~d5 18 f3 :c8 19lbc3 ~a8 20 ~d7 :c7 21 lbb5 :c5 22 :d6 lbc6 23 lbc3 with a slight plus for White, Rohde-Joshi, USA 1986) 16 :d6 ~f7 17 1ΙΜΙ ~e7 18 .td7 (Tal-Sisnίega, Taxco ΙΖ 1985) and now 18..ixd7

Classical Variation 65

is just slightly better for White according Ιο Tal. Ιη the game 18••. W'7 19 tDxe6 .i.xd7 20 tDc7 .i.xa4 21 tDxa8 led to a qώck win for White. 4b) l l...dxe4 12 "xd8 .i.xd8 13 :hel tDa7 (or 13...h6 14 .i.xf6 .i.xf6 15 tDxe4 with an edge for White) 14 .i.c4 h6 15 .i.xf6 gxf6 16 tDxe4 f5 17 tDd6 .i.c7 18 g3 b6? (18 ...:d8 19 tDb5 tDxb5 20 .i.xb5 would have been sligh~y better for White according Ιο Tal) 19 tDxf5! and White won ίη the famous game Tal-Korchηοί, Montpellier Ct 1985. 4c) ll...tDa7 and now there are two tempting lines for White: 4cl) 12 .i.e2 .i.d7 (12 ... b5 13 exd5 bxa4 14 d6 axb3 15 dxe7 "xe7 16 cxb3 was unclear ίη on-Ryskin, USSR 1987, but 13 .i.xf6 .i.xf6 14 tDxb5 is better according ιο σΗ) 13 .i.xf6.i.xf6 14 exd5 .i.xc3 15 'ifxc3 .i.xa4 16 dxe6 "e7 17 exf7+ Φh8 (Black has some initiative for the two pawns, but ηοΙ nearly enough) 18 .i.c4 :ac8 19 :hel "g5+ 20 'ίi'd2 'ίi'g6 21 .i.e6 :c7 22 'ifd3 "xg2 23 'ίi'd6 and Black's counterattack has collapsed, Gelfand-Ryskin, Minsk 1986.

4c2) 12 .i.xf6 .i.xf6 13 exd5 .i.xc3 14 'i'xc3 tDxb5 15 axb5 a4?! (15 ... exd5 16 tDd4 "b6 is relatively best, but still good for White after 17 ':hel) 16 dxe6! 'ίi'g5+ (the ροίηΙ is that 16... axb3 17 ':xd8 ':al + 18 Φd2 :xd8+ 19 Φe2 :xh 1 fails Ιο 20 "c7:f8 21 e7 :e8 22 'ίi'd8) 17"d2 'iff6 18 tDd4 and Black has very ιίι tle for his minus pawn, HoffmannTimoshchenko, Budapest 1989. 4d) ll....i.b4 (an untested suggestion by Tal) 12 exd5 exd5 13 'i'f4 with an edge for White. 4e) 11...tDb4(D)andWhitehasa choice:

4el) 12 e5 tDd7 13 .i.xe7 "xe7 14 f4 (ίη this French Defence type of position, the exchange of darksquared bishops theoreticaHy favours White, but with the kings castled οη opposite sides of the board the game is more likely Ιο be decided by the speed of the respective attacks rather than by the endgame advantage of the better bishop) 14 ...tDc5 (after 14 ... b6 15 :hel tDc5 16 tDd4 .i.d7 17 ΦbΙ :ac8? 18 g4 :fd8 19 f5 tDe4 20 tDxe4 dxe4 21 c3

Clαssicαl Vαriαtion ι

ll)d3 22 i.xd3 exd3 23 'ii'f4! i.xa4 24 :xd3 White had an excel1ent position ίη Balashov-Κhalifman, Minsk 1986, but 17 ...:fc8 intending ... i.xb5 was better) 15ll)xc5 'ii'xc5 16 h4 (the idea is Ιο bήηg the rook Ιο g3, ηοΙ οηlΥ helping the kingside attack, but also providing useful defence along the third rank) 16 ... b6 and now 17 h5 was unclear ίη Κίη dermann-Felsberger, Vienna 1986. Perhaps 17 :h3 was more accurate because ίη some lines the h-pawn plays ηο important role οη h5, but ίη any case the position is very doubleedged. 4e2) 12 1:thel (D) with a further branch:

4e21) 12••• h6 13 exd5!? (or 13 i.xf6 i.xf6 14 exd5 exd5 15ll)xd5 i.g4 16 f3 i.g5 17ll)e7+! 'ii'xe7 18 :xe7 i.xe7 19 'ίWe2 i.g5+ 20 ΦbΙ i.f5 21ll)d4 i.g6 22 ΦaΙ ! and now White has excellent winning prospects, 011- Κhalifman, USSR 1987) 13 ... exd5 (13 ... hxg5 14 d6) 14 'ii'e3 i.e6 15 i.f4 1Σc8 (Winsnes-Κhalif man, Groningen 1985/6) and now 16 ΦbΙ (intending ll)d4) 16 ... ll)h5 17

i.e5ll)c6 18 i.xc6 bxc6 19 i.d4 is good for White according Ιο Οοη aldson. 4e22) 12.••i.d7 (maybe the most solid move) 13 e5 (13 exd5 i.xb5 14 d6 'ii'xd6 15 'ii'xd6 i.xd6 16 :xd6 i.c6 gives White an edge, as does 13 i.xd7 'ii'xd7 14 e5 ll)e8 15 i.xe7 'ii'xe7 16 f4 ll)c7 17 ll)d4 ll)c6 18 ll)db5 ll)xb5 19 ll)xb5 1:tac8 even though a draw was agreed here ίη Rachels-D.Gurevich, Boston 1988) 13 ... ll)e8 14 h4 ll)c7 15 ll)d4 ll)c6 16 i.xe7 'ii'xe7 17 'ii'g5 and again White has a smal1 advantage, Wang Zili-D.Gurevich, Belgrade 1988. 4e23) 12•••dxe4 13 'ii'xd8 (White has a speculative alternative ίη 13 ll)xe4 ll)xe4 14 'ii'xd8 i.xg5+ 15 'ii'xg5 ll)xg5 16 h4 e5 17 hxg5 i.f5 18 :d2 :fc8 {18 ... f6 was a more cautious move} 19 c3 i.e6 20 ll)al :c5 21 i.d7! with some advantage Ιο White, Vitolin~-Inkiov, Jurmala 1985) 13 ...1:txd8 (13 ...i.xd814ll)xe4 is good for White after 14 ... ll)xe4 15 i.xd8ll)xf2 16 J:[d2 or 14 ... i.e7 15 ll)xf6+ i.xf6 16 i.xf6 gxf6 17 :d6) 14 ll)xe4 ll)bd5 15 c4! ll)c7 (or 15 ... ll)b4 16 1:txd8+ i.xd8 17 :dl i.e7 18 ll)d6 with a clear plus) 16 :xd8+ i.xd8 17 :dl i.e7 18 ll)xf6+ gxf6 19 i.e3ll)xb5 20 axb5, Anand-Inkiov, Calcutta 1986, and the liquidation has left White with a clear advantage. Black stil1 has problems developing his queenside pieces and White's queenside majorίΙΥ is ominously near Ιο creating a passed pawn. True, Black has the two bishops, but White can always

Clαssicαl Vαriαtion

force an exchange (e.g. by .i.c5) ίί they show signs ofbecoming actiνe. After the further moνes 20...f5 21 tDc5 e5 22 lDd7 ί4 23 .i.b6 ί6 24 .i.c7 .i.xd7 25 :xd7 .i.c5 26 .i.d6 .i.xd6 27 :xd6 White had a winning rook ending. 5) 9.....b6 (the counterattack against f2 nullifιes White's threat Ιο take οη ί6, so Black gets time Ιο support his d-pawn by ...:d8; White's usual reaction has been Ιο start a kingside pawn storm, but he must be careful because Ιοο many pawn moνes might encourage Black Ιο open υρ the centre by ... d5) 10 Ι3 (D ) and now:

5a) 10.••a6 11 h4! (this is the reply which has ρυΙ 1O... a6 ουΙ οί faνour) 11 ...:d8 12 h5 'fkc7 (when the pawn reaches h5 Black faces a tricky problem - is he willing ιο let his dark squares be weakened by allowing h6, or should be play ... h6 himself, eνen though this makes the subsequent adνance ofWhite's g-pawn νery strong? After 12 ... h6 13 .i.e3 'fkc7 14 'fk12 tDd7 15 g4 tDce5 16 :gl b5 17 g5 White had deνeloped

a νery dangerous attack ίη HellersJ.Piket, Amsterdam 1985) 13 g4 (or 13 h6 g6 14 'iff4 tDe8 15 .i.xe7 'ifxe7 16 'fke3 b5 17 .i.e2 :b8, Martinoνit-Popoνit, Yugoslaν Ch 1986, and now 18 :d2 followed by :hd 1 is slightly better for White according Ιο Martinoνic) 13 ...b5 14 .i.e3 tDd7 15 g5 tDce5 16 g6! b4 (16 ...fxg617 ί4! lDc4 18 .i.xc4 bxc4 19 tDd5! exd5 20 'fkxd5+ ~h8 21 hxg6 tDf6 22 :xh7+ tDxh7 23 "h5 is mate) 17 gxf7+ ~xf7 18 tDd5! exd5 19 "xd5+ ~ί8 (19 ... ~e8 20 h6 g6 21 ί4 tDg4 22 .i.c4 is also good for White) 20 'ii'xa8! .i.b7 21 "a7 with adνantage Ιο White as 21 ...tDc6 is met by 22 tDd4, Serper-Brodsky, USSR 1986. 5b) 10•••:d8 11 ~bl. Α useful semi-waiting moνe; the reply 11 ...d5 is bad for tactical reasons, so Black plays either 11 ... a6, when Whίte can switch ιο his kingside attack plan without allowing ... d5, or the immediate 11 ... 'ifc7 which has been quite popular recently: 5bl) 11 ••• dS?! 12 .i.xf6 .i.xf6 (Black should ηοΙ play 12... dxe4? because οί the reply 13 .i.xe7! :xd2 14 tDxd2! when 14 ... tDxe7 15 tDc4 'fkc7 16 tDb5 wins the queen; AnandBenjamin, Wijk Μη Zee 1989 continued 14 ... exf3 and now 15 tDc4 fxg2 16 .i.xg2 'fkc7 17 .i.d6 "d8 18 .i.g3 'fke7 19 :hel is crushing for White) 13 exd5 .i.xc3 (l3 ... a5 14 tDa4 'fka7 15 d6 b6 16 "e3! :b8 17 i.b5 was νery good for White ίη Mokry-Conquest, Gausdall989) 14 'ifxc3 exd5 (14...tDb4 is worse since

Clαssicαl Vαriαtion

15 d6! :xd6 16 .tc4 makes ίι hard for Black Ιο complete his development) 15 'iί'c5 followed by .tb5, with advantage Ιο White. 5b2) 11...a6 12 .te3 'iί'c7 13 'iί'f2 (thanks Ιο the threat of .tb6 Black has ηο time for ... d5) 13 ...tbd7 14 h4 b5 with a position similar Ιο line '5a' above. Admittedly White has spent a move οη rbbl, but this is certainly ηοΙ a waste of tempo, and while White's chances are ηοΙ quite as good as ίη line '5a' he has fair attacking prospects. For example, Timman-Salov, Linares 1991 continued IS g4 :'b8 16 g5 tbb6 17 g6!? tba4 18 gxf7+ rbxf7 19 tbe2 .tf6 20 tbed4 tbxd4 21 .txd4 e5 22 .te3 'iti>g8 23 .th3 with just an edge for White. Another possibility is IS .tgS!?, when Govedarica-Thorsteins, Clermοnt-Feπand 1989 continued 15 ... .tb7?! 16 tbd5! exd5 17 exd5.txg5 18 dxc6 with advantage Ιο White. 5c) 11 •• :ii'c7 12 .txf6!? (White parts company with his bishop ίη order Ιο speed up the attack) 12....txf6 13 g4 g6 (Black wishes Ιο keep his bishop οη the long diagonal, even if this means allowing White Ιο open the h-file; 13 ... a6 14 g5 .te7 15 f4 does look quite promising for White) 14 h4 a615 g5 (15 h5 g5!) 15•••.tg7 16 h5 b5 17 hxg6 hxg6 18 f4! (a strong move, threatening Ιο play f5 and clearing the third rank so that the queen's rook has access ιο the hfile) 18..• b4 19 tba4 (19 tbe2 a5 would allow Black counterplay; although the knight can certainly get

trouble οη a4, White hopes that ίι will survive long enough for him Ιο develop his attack) 19•.•:b8 20 'ii'h2 (20 .td3, aiming for f5, is well met by 20 ... tbe7!) 2O.••rbf8 (20....td7 21 :'d3! tbe7 runs ίηΙο trouble: 22 :'h3 <ittf8 23 :h7! tbg8 24 :xg7 rbxg7 25 'iί'h8+ <ittf8 26 :h7 <itte7 27 f5! with a crushing attack) 21 :d3! (D) (preparing ιο triple) and now: ίηΙο

5cl) Benjamin-Anand, Groningen 1993 continued 21 ... e5 22 f5 gxf5 23 :'h3! tbe7! (23 ... fxe4 24 :h7!) 24 :'h8+ tbg8 (24 ....txh8 25 'iί'xh8+ tbg8 26 :h7) 25 :'xg8+ rbxg8 26 'iί'h7+ rbf8 27 exf5 .txf5! (27 ... 'iί'e7 28 f6 .txf6 29 gxf6 'iί'xf6 30 .td3 is better for White) 28 'iί'xf5 'ii'c6 29 g6 :'b7 30 :'h7 'iί'xa4 (D) and now: 5cll) The game continued 31 'iί'g5?! 'iί'e8! 32 .txa6 (Anand's suggestion 32 .tg2 is also insufficient; ηοΙ because of 32 ...:'c7 or 32 ... e4, since both lose Ιο 33 ':xg7!, but 32... d5! when 33 tbc5 can be met by 33 ... 'iί'e7!) 32 ... ':e7! 33 .td3 e4! 34 .tb5?! :e5! 35 gxf7 ':xg5 36 fxe8'iί'+ ':xe8 37 R.xe8 :'gl+ 38

Clαssical

~. ~ -~ '".~ u"

~.~ ~

w.z • ••• : •• •

.Δ8

•.lb. •••••••• ••••

.i...

ΔDΔ.. .Φ.

~c 1 ~xe8 39 a4? (White had Ιο try 39 c3 bxc3 40 bxc3 j,xc3 41 ~c2

j,f6, although Black should still win) 39 ... bxa3 40 bxa3 j,c3! 41 J:th4 d50-1. 5c12) 31 J:txg7 ~xg7 32 gxf7 looks good enough for a draw, e.g. 32••.'itd7 (certainly ηοι 32...1:.xf7? 33 'ίi'g5+ ~h7 34 j,d3+, nor 32 ...1:.f8 33 "g4+ when White picks up the rook οη b7 after both 33 ... ~h7 34 "h3+! and 35 "g2+ and 33 ...~xf7 34 j,c4+ ~f6 35 "f3+) 33 "g5+ 'ifί>xf7 34 ~5! dxc5 (34 .....e7 35 'ίi'h5+) 35 j,c4+ and White has perpetual check, or 32...Φf8 33 "g5 :1bb8 (33 .....d7 34 j,c4 d5 35 ~5 is ΟΚ for White) 34 j,c4 d5 35 'ίi'g8+ ~e7 36 "g5+ and after both 36••. ~d6 37 "f6+ and 36••• ~XΙ7 37 'ίi'f5+ White has at least a draw. 5c13) 31 j,c4!? is an interesting winning attempt. 31 ...d5 can be met by 32 ~c5 and 31 .. :iνd7 by 32 "h5! when 32 ... fxg6? 33 J:th8+! leads Ιο mate. Black should probably respond with 31 .. ':dd7, and ίη this critical position White has various ways Ιο force perpetual check, and one way Ιο continue the fight:

Variation 69

5c131) 32 'iί'h5 fxg6 33 J:th8+ j,xh8 34 "xh8+ ~e7 35 "h4+ is a draw by perpetual check. 5c132) 32 'iνh3 fxg6 33 "e6 (33 1:.h8+ is again a draw, while after 33 j,d5 'ίi'M! ίι is time for White Ιο settle for the draw) 33 ...d5 34 j,xd5 (34 ~5 dxc4 35 J:thl1ooks strong, but Black can reply 35 ... J:te7! 36 'iί'd5 "e8) 34 ... J:txd5 35 "xd5 "d7 and Black has an extra pawn, but his king remains somewhat exposed. 5c133) 32 'iί'Ι3!? is the move ΙΟ play if White wants Ιο win at all costs. Ιι prevents 32 .....c6 and sets up the rather slow threat of"g2 fol10wed by J:txg7. Even ifBlack plays 32 ...1:dc7 33 "g2 "c6, White can still play 34 j,d5 followed by J:txg7. Ι (ΙΝ) don't belίeve that White can be worse ίη this position. 5c2) Benjamin belίeves 21 ..id7 Ιο be the best defence. He gives 22 Ι5 exf5 23 ~ac5 as unclear and Ι imagine this could easily end ίη a draw after 23 ... j,e5 24 'iί'h6+ j,g7 25 "h2 etc. After 22 J:th3 he suggests 22 ... ~e7! so that 23 J:th7 can be met by 23 ...1:h8!. However, Ι lίke the look of 23 "e2! threatening Ιο play f5. One lίne runs 23...~a5 24 f5 j,xa4 25 1:h7!? and now 25 ... J:tg8 26 J:txg7 and 25 ... j,e5 26 f6+ ~e8 27 1:g7! are winning for White. If Black tries 23...m, White plays 24 J:th7 ~e7 25 "h2! and Black has been tricked ίηιο the variation given ίη the notes Ιο Black's 20αι move. Perhaps the best is 23...e5 although after 24 J:th7 Ι prefer White. Another advantage of the queen οη e2 is that

Classicαl Vαriαtion

White may be able Ιο snatch the a6pawn one day. Now we return Ιο the main line after 9 ... a6 (D).

10 ~XΙ6 gxf6 10•••~xf6 is dubious: 11 'ifxd6 ~xc3 (or 11 ... 'ifb6 12 f4 ~e7 13 'ifd2 a5 14 a4 ':'d8 15 i.d3 with advantage Ιο White, ShaposhnikovBoleslavsky, USSR 1950; as usual 12 ... ~xc3 13 bxc3 'ife3+ 14 'iifb2 'ifxe4 loses Ιο 15 ~d3 followed by 16 ~xh7+) 12 bxc3 (the position is the same as after 9 ... h6, except that Black has played ... a6 instead; the verdict is unchanged) 12 ... 'ίi'f6 (12 ... 'ifh413 g3 {13 f3!?} 13 ... 'ίi'f6 14 'ifc5 e5 15 ~c4 ~g4 16 ':'d6 'ίi'g5+ 17 f4 and White was clearly better ίη Ivanovic-Popovic, Νονί Sad 1984) 13 'ifg3 e5 14 ~c4 ~e6 15 ~xe6 'ifxe6 16 ':'d5 and again White had won the opening battle, Benjamin-Chήstiansen, USA 1984.

11 .-ιι6 White plans a general kingside pawn advance by g4, f4, h4 and g5, but first he transports his queen ίηΙο the vicinity of the black king.

11 ••• 'iifh8 12 -.ιι5 Black is now virtually forced Ιο lose time with his queen because he must free the f8-rook Ιο defend h7 by ... ':'g8-g7. 12 ••. 'ίi'e8 12•••':'g8?, Ernst-Chandler, London 1988, 13 'ίi'xf7! ':'g6 14 f4 ~d7 15 ':'d3 e5 16 ':'g3! ~e8 17 'ifd5 exf4 18 ':'xg6 hxg6 19 'fid2 g5 20 lL!d4 lL!xd4 21 'ifxd4 is good for White. 13 Ι4 ':'g8 13••• b5 14 ~d3 ':'g8 would give Whi te the chance Ιο play 15 lL!d5!? Tisdall points out that after 15•••exdS 16 exd5 ':'g7 17 dxc6 ~g4 18 'ifd5 ~xdl 19 ':'xdl White has fantastic compensation for a slight material deficit; of course Black should play 15... ~b7 16lL!xe7 'ifxe7 and ίι is far from certaίn that the exchange of minor pieces favours White, especially as he has spent two tempi achieving ίι

14 g4 b5 15 ~d3 ':'g7 15.•. b4 16 e5 clears e4 for the knight with gain of tempo. 16 h4 b4 17 lL!e2 (D) 17 ••• e5 Ιη view of the hard time Black received ίη this game perhaps his attention will once again turn towards 17•••a5. The analysis runs 18 g5 (18 lL!bd4!?) 18•..a4 19 lL!bd4 and now: 1) 19•••lL!xd4 20 lL!xd4 ~d7 21 gxf6 ~xf6 22 e5 dxe5 23 ':'hgl 'fig8,

CΙassical

Arnason-Inkiov, Plovdiv 1986, and now Inkiov suggests 24 lbf3! ~e8 (24 ... exf4 25 %1xg7 J..xg7 26 lbg5 wins) 25 J..e4 (25 fxe5 also deserves attention, though perhaps Black can defend with 25 ... %1xgl 26 %1xgl ~g7) as good for White. 2) 19... b3!? 20 axb3 (or 20 cxb3 axb3 21 a3, after which 21 ... lbxd4 22 lbxd4 'ii'a4 is unclear) 20 ... axb3 21lbxb3 ~b7 22lbc3lbb4 23 %1hgl 'ii'c6 24 %1g3 'ii'b6 25 'ii'e2 d5 was played ίη Psakhis-Kotronias, Dortmund 1989, and now the best move is 26 'ii'e3!, when Kotronias gives 26 ... lbxd3+ 27 %1xd3 'ii'xe3+ 28 %1dxe3 ~d6 29 e5 fxe5 30 fxe5 ~e7 intending ... h6 as unclear. However after 31 lbb5! Ι doubt that Black has enough for the pawn, e.g. 31 ... h6 32 gxh6 %1xg3 33 %1xg3 ~xh4 34 %1g7 intending lbd6. 18 fS aS 19 %1dgl! Previous games had seen 19 ΦbΙ but it seems that White can do without this precautionary move, thus gaining an important tempo for the attack. 19 .•• ~b7

Variation 71

19••. a4 20 lbd2 b3 is met by 21 a3! and after 21 ... bxc2 22 g5 White is miles ahead ίη the hunt for the opposing king. 20 gS a4?! This natural move is probably a mίstake. Black has insufficient force ίη the vicinity οί the white king to start an attack so there was litt1e point ίη forcing the knight to a more active square. Leko gives the immediate 20..•lbb8 as better, but he still doesn't like Black's position after 21 "h6! lbd7 22 gxf6 %1xgl + 23 %1xgl ~xί6 24lbg3 'ii'f8 25 'ii'e3 a4 26 lbd2 when White has a particularly menacing knight pair. It's worth pointing out that ifWhite had continued as ίη the game, namely 21 gxf6 ~xί6 22 %1xg7 J..xg7 23 ί6 ~xί6 24 %1f1 'ii'e7 25 lbf4, then Black can defend with 25 ... lbd7 26 lbd5 ~xd5 27 exd5 lbf8 as there is ηο lbe4 to finish him off. 21 lbd2 lbb8 As 21 .•• b3 22 a3! achieves nothing, Black makes this defensive manoeuvre ίη order to shore υρ the cήtίcal f6-square. 22 gxf6 22 "h6 is possible but the text seems even stronger. 22 ... J..xf6 23 %1xg7 J..xg7 23••• Φxg7 24 lbf3! looks pretty nasty for Black. 24 {6! White is willing to invest a fair amount οfmateήal to clear the bl-h7 diagonal for his bishop. 24 ••• J..xf6

Clαssical

Variation

25 ':Π 'iie7 (D) 25••• lLJd7 26 ~b5! +- is Leko's comment ίη Informator. True, but White has Ιο take great care as after 26 ... ~g7 the seemingly crushing 27 ':ΧΙ7 gives Black the chance Ιο defend himself by 27 ... ~a6! 28 ~xd7 ~xe2! and it's ηοΙ even clear if White is better; 29 'ii'f5 'ii'g8! with the idea of ... ':f8100ks ΟΚ for Black as does 29lLJf3 'ii'g8! 30 ~e6 ~xB 31 ':xf3 ~h6+ 32 <iPdl 'ii'gl+ with a likely draw. 27 ~xd7 'ii'xd7 28 ':xf7 'ii'c6 is very unclear, and 27 "ilxf7 'fkxf7 28 ':xf7 lLJc5 is ηοΙ recommended but White does have one very strong move though: 27lLJg3!. The knight heads for f5 and ':xf7 is once again threatened. Black will be hard pressed Ιο find any sort of defence, especially as 27 ... lLJf6 fails Ιο 28 ~xe8! lLJxh5 29 lLJxh5 ':xe8 30 ':xf7.

26 lLJr4! exf4 There was ηο choice as 26••• lLJd7 27 lLJd5 ~xd5 (27 ... 'ii'd8 28 lLJxf6 lLJxf6 29 'i'xf7) 28 exd5lLJf8 29lLJe4 ~g7 30 ':xf7 is terminal. 27 e5 Φι8

28 ':gl+? This οηlΥ assists the black king Ιο escape. Much stronger was 28 'iixh7+ Φf8 29 exf6 'i'xf6 30 lLJe4! 'ii'e5 (there is ηο relief for Black ίη 30... 'i'g7 31 'fkxg7+ Wxg7 32lLJxd6 followed by ':xf4) 31 'ii'h6+ We7 32 :'xf4 lLJd7 33 :'xf7+! ΦΧf7 34 lLJg5+ Φe7 35 'ii'h7+ Φd8 (35 ...Φe8 36 ~g6+) 36lLJf7+. 28 Φrs 29 'iVlι6+ Φe8 30 exf6 "ile3! Even ifWhite's 28th move was an eποr it's still nothing short of miraculous that Black managed Ιο save himself from this position. 31 "g7 ~d5! Covering f7. It's now quίte hard for White Ιο get at the black king. 32 ΦdΙ How else can White increase the pressure? 32 .•. b3!! Black finds an incredible saving resource. 33 ':el White cannot be called a coward for avoiding 33 a3 bxc2+ 34 ~xc2 'ii'd4 35 ':el+ <iPd7 36 ':e7+ <iPc6 when his own king is ίη at least as much danger as Black's. 33 bxa2 34 ':xe3+ fxe3 35 ~b5+ Φd8 36 '6'ΙS+ Φc7 37 '6'e7+ Φc8 38 '6'e8+ Φc7 39 '6'e7+ 1/2·1/2 With ηο way Ιο stop the a-pawn White must take the perpetual.

Clαssicαl Vαriαtion

Ζ.

Game 11 Almasi - Gabήel Altensteig 1994

1 e4 cS 2 lΔf3 d6 cxd4 3 d4 4 lt)xd4 lDf6 S lDc3 lDc6 e6 6 .tgS 7 'iWd2 a6 8 0-0-0 .td7 8 ...h6 is examined ίη Game 12. Other ideas are dubious, e.g.: 1) 8 ...'iWb6 9 tLJb3 .td7 10 .te2 'iic7 11 f4 h6 12 .txf6 gxf6 13 .th5 is good for White, Panchenko-Csom, Las Palmas 1978. 2) 8....te7 9 Ι4 'iWc7 (9 ... tLJxd4 10 'iWxd4 'iWa5 11 e5 dxe5 12 fxe5 tLJd5 13 .txe7lDxe7 14 .td31Dc6 15 'iWh4 tLJxe5 16 tLJe4 f6 17 :'hfl with a νery dangerous attack for White, AdlerBannik, USSR 1978) and now: 2a) 10 Φbι.td7 (10...0-0 11 .te2 ':d8 12 .tf3 h6 l3 b4 tLJxd4 14 'iixd4 b5 15 'iWf2 .tb7 16 g4 was good for White ίη Kaνalek-Larsen, Montreall979) IllDf3 ':d8 12 .td3 b5 13 ':hel b4 14 tLJe2 a5 15 lDg3 ο-ο 16 e5 lDd5 17 'iWe2 was unclear ίη Przewoznik-Bielczyk, Katowice 1986. 2b) 10 .te2 lDxd4 11 'iWxd4 b5 12 e5 dxe5 l3 fxe5 lDd5 14 .txe7 tLJxc3 15 .tf3! tLJxdl 16 .td6! and White went οη Ιο win quickly ίη Tal-Larsen, Montreal 1979 after 16 .....c4 17 "b6! tLJf2 18 .tc6+ .td7 19 .txd7+ ΦΧd7 20 "b7+ Φd8 21 "xa8+ 'iWc8 22 "a7 1-0.

2c) 10 .txf6 gxf6 11 g3 .td7 12 f5 lDxd4 l3 'iWxd4 ':c8?! (13 ... 0-0 and l3 ... b5 are possible improνe ments) 14 ΦbΙ b5 15 'iWd2 'iWc5 16 .td3 h5? 17 fxe6 fxe6 18 lDe2 ίη tending tLJf4, Short-Larsen, London 1986, and White has the adνantage. 9 Ι4 (D)

9 h6 Black has two major alternatiνes, 9 ... b5 and 9 ... .te7: 1) 9... bS 10 .txf6 gxf6 11 ΦbΙ. Ιη BTS2 I1lDxc6 .txc6 12 'iWel was recommended, but this moνe order has been abandoned due Ιο 12 ... b4 l3lDd5 a5 14 ':d4 f5! when Black νery quickly creates counterplay οη the long diagonal. For example, 15 :'c4 (15 exf5 .tg7 16 ':dl.txd5 17 ':xd5 'iWf6) 15 ... .txd5 16 exd5 'iWf6 17 dxe6 fxe6 18 ':c7 Φd8! 19 ':b7 (19 ':c6 .tg7 20 ':xd6+ Φe7 21 'iWe5 'iWxe5 22 fxe5 .txe5 is a little better for Black) 19 ... Φc8 20 ':b5 .tg7 21 c3 bxc3 22 .tc4 ':e8 23 'iWe3, which is giνen as unclear by Polulίakhoν, but after 23 .....d4 24 .txe6+ Φc7 White is forced Ιο exchange queens, thereby losίng any chance of gaίnίng

Clαssicαl Variαtion

the advantage. With 11 ~bl White steers clear of this mess and at the same time can aim for similar positions Ιο those analysed ίη BTS2. Black now has: la) 11... b4 (rarely played) 12 lL1ce2 'ii'b6 13 lL1xc6 .i.xc6 14 f5 e5 (14 ....i.xe4 15 fxe6 fxe6 16lL1f4 and now both 16 ...d5 17 .i.d3 and 16 ... b3 17 axb3 Wxb3 18 .i.d3 are good for White) 15lL1g3 Wc5 (15 ... h5 16 h4 gives White a nice target) 16lL1h5!? ~e7 (Damljanovic didn't like the position after 16 ... .i.e7 17 .i.d3 h6 18 :cl) 17 .i.d3 :g8 18 :hgl! (18 "e2 .i.h6) 18 ... :g5 19 g4 .i.h6 20 'ii'e2 (20 lL1xf6 1:Ixg4) 20...:ag8? (better is 20...:gg8 when 21 h4 .i.e3 22 :g3 .i.d4 23 g5 is unclear according Ιο Damljanovic {Ι prefer White} but also possible is 21 :g3, ηοΙ allowing the bishop Ιο get Ιο d4) and now instead of playing 21 :g3? (Vujakovic-Damljanovic, Yugoslavia 1991), White could have won the exchange with 21 h4! :xh5 22 :hl!. Black has insufficient compensatίon, e.g. 22 ... a5 23 gxh5 a4 24 .i.c4 :c8 25 b3 a3 26 :d3 followed by c3 is winning for White. lb) 11.....b6 (D) and now: 1b 1) For those of you ηοΙ content with a positional advantage then Wolff's suggestίon, 12lL1f5!?, might be of interest. His analysis runs 12 ... exf5 13lL1d5 Wd8 (13 .....b7 14 lL1xf6+ ~d8 15 e5!? or 15 g3) 14 exf5 and now 14....i.xf5 loses Ιο 15 "c3 whilstboth 14....i.e715 g4! and 14....i.g7 15 :el + give White good play for the piece.

lb2) 12 lL1xc6 .i.xc6 13 "el! and now Black would like Ιο play ... b4 but for the moment this is met by lL1d5. 13 ... .i.e7 is normal, but as the bishop is often better off οη f8 than e7 Black has tried a couple of other moves recently: lb21) I.Gurevich-Kozul, BielIZ 1993 contίnued 13..•:a7 14 f5 b4 15 lL1e2 e5 16lL1g3 h5 17 .i.e2! :h6 18 .i.c4 a5 19lL1f1 "c5 (19 ... .i.xe4 20 .i.d5 gives excellent positίonal compensation for a pawn) 20 lLJe3 with advantage. lb22) Lau-Grosar, Graz 1993 continued 13...0-0-0 14 .i.d3 h5 15 "h4 .i.e7 16 :hel :de8 17 "h3 ~b8 18 f5 b4 19 lLJe2 e5 20 lL!c 1 a5 21 .i.c4 with a clear advantage for White. lb23) 13....i.e7 14 f5 b4 (White has an edge after 14 .....c5 15 fxe6 fxe6 16 .i.d3) 15 lL1e2 e5 (and ηοΙ 15 ... .i.xe4? 16 fxe6 fxe6 17 lL1g3 .i.d5 18 :xd5! exd5 19lL1f5, which is great for White) 16lL1g3 "c5 17 .i.d3 (17 b3!?) 17 ... a5 and ίη an extremely roundabout way we have transposed ίηΙο the game Ernst-PopονίC, Subotica ΙΖ 1987, which was

Clαssicαl Vαriαtion

featured ίη BTS2. Play continued 18 'fί'e2 (now nothίng can stop .tc4-d5) 18 ... a4 19 .tc4 ο-ο (obviously very ήsky, but there is ηο chance of a successful attack against the white kίng without the other rook) 20 ttJh5 'ίt>h8 21 :hfl a3? (overlookίng that White has an immediate mating threat, but even the best lίne 21 ...:fc8 22 b3 :a7 23 :f3 axb3 24 .txb3 is very (White unpleasant for Black) 22 can afford Ιο abandon the queenside since he has a forced win οη the other flank) 22 ... axb2 23 :h3 (intending 24 ttJxf6 .txf6 25 'ilνh5) 23 ...:g8 24 ttJxf6 1-0. After 24 ... .txf6 25 :xh7+ 'ίt>xh7 26 'fί'h5+ 'ίt>g7 27 'ilνxf7 + 'ίt>h6 (27 ... 'ίt>h8 28 'fί'h5+ 'ίt>g7 29 'fί'g6+ and mate next move) 28 'fί'xf6+ 'ίt>h7 29 'ilνh4+ 'ίt>g7 30 'fί'g5+ the rook is the first of many black pieces Ιο disappear. 2) 9 ....te710 ttJr3 bS (D) (other moves are inconsistent, for example 10... 'ilνc7 11 e5 dxe5 12 fxe5 ttJd5 13 ttJxd5 exd5 14 .txe7 ttJxe7 15 .td3 ο-ο 16 'iIν g5 ttJc6 17 'fί'h5 or ΙO ... h6 11 .txf6 gxf6 12 f5 'fί'c7 13 'ίt>bl 0-0-0 14 .tc4 wίth a clear plus for White ίη both cases) and now:

:f3

2a) 11 83!? (an unusual sidelίne which gives me the chance Ιο demonstrate one of my better efforts; White al10ws his queenside to be weakened but he hopes Ιο break through ίη the centre before Black has time ιο organise an attack) 11 ... b4 12 axb4 ttJxb4 13 .tc4 (ηοι 13 e5?! 'fί'a5 14 'ίt>bl ttJe4) 13 ...'fί'c7 (13 ....tc6 is given as = by ECO, but 14 ttJd4 would be an interesting reply) 14 'ilνe2! (White waίts for ...:c8 before playing .tb3 so that the advance of the a-pawn won't be so dangerous, e.g. 14 .tb3?! a5 15 e5 a4 16 exd6 axb3 17 dxc7 :al+ 18 ttJbl ttJa2#) 14 ... :c8 15 .tb3 ο-ο?! (better is 15 ...e5) 16 e5 dxe5 17 fxe5 (17 ttJxe5!?) 17 ... ttJfd5 18 .txe7 ttJxe7 19 ttJg5 a5 20 :hf1 a4 21 'fί'e4! ttJg6? (D) (Black had ιο play 21 ... ttJf5, even if Whίte can obtain a strong attack after 22 .txa4 .txa4 23 'fί'xb4 .tc6 24 :xf5! exf5 25 e6)

22 :xd7! (White is willing ιο part company with both his rooks if it means landing οη e6) 22 ... 'fί'xd7 23 :xf7! axb3? (the best that Black can do is 23 ...1Ixf7 24 .txe6 'fί'xe6 25

Clαssical Variαtion

lΩxe6 lΩc6 although White should

still be winning) 24 :xd7 :f1 + 25 :dl lΩa2+ 26 lΩxa2 and, ίη GalΙagher-CabήΙο, Royan 1989, Black resigned since he spotted that after 26 ... bxa2, 27 "a4 stops the pawn. 2b) 11 e5 b4 (not 11 ... dxe5 12 fxe5 b4 13 exf6 bxc3 14 'i'xd7+ and White wins) 12 exf6 bxc313 "'xc3 gxf614 .th4 (D) and now:

2bl) 14•••35 (14 .....a5 15 ~xf6 16 .tc4 :c8 17 a3 is good for White) 15 ΦbιlΩb4 (15 ...:b8 16 g4 lΩb4 17 a3 :c8 18 'i'b3 lΩd5 19 ':xd5 with an excellent position for White, Gligori6-Conrady, Dublin 1957) 16 a3 :c8 17 'i'b3 lΩd5 18 :xd5 exd5 and now either 19 lΩd4 or 19 'i'xd5, with very good compensation for the exchange. 2b2) 14•••d5 15 ΦbΙ with a further branch: 2b21) 15•••a5 16 ~b5 :c8 17 lΩd4 lΩxd4 18 ~xd7+ 'i'xd7 19 'i'xd4 :g8 20 g3 'i'b5 21 :hel is good for White. 2b22) 15••• lΩb4 16 lΩd4 'i'a5 (16 ...:c8 17 "b3 'iWa5 18 .tel! .ta4 19 'ifa3 with a clear plus for White, lΩb4

Matanovic-Jansa, Lugano 1968) 17 a31Ωc618 'i'g3lΩxd4 19 :xd4 :b8 20 :d3 Φί8 21 .te2 with a small plus for White according to Lukin. 2b23) 15_lΩa5 16 f5 1%c8 17 'ifd2 'i'c7 18 fxe6 fxe6 19 .td3! 1Ωc4 20 .txc4 'i'xc4 21 :hel :g8 (after 21 ... 0-0, the simplestreply is 22 g4!) 22 h3 :b8 23 ΦaΙ h5 24 g3 with a small plus for White, TseshkovskyLu1άn, USSR 1982. 2b24) 15.••:c8!? 16lΩd4 (the apawn is not of great interest to either side) 16...'i'b6 (16 ...lΩe5?! 17 ...g3 lΩg6 18 f5 lΩxh4 19 fxe6 fxe6 20 'i'xh4 ο-ο 21 'i'g4+ Φf7 22 .td3 was very good for White ίη RachelsShirazi, USA Ch 1992) 17lΩxc6 (17 .tf2 is certainly worth consideraιίοη) 17 ......xc618 'i'd2 :b8 19 ΦaΙ (19 c4!? dxc4 20 .txc4 'i'xc4 21 'ifxd7+ Φf8 22 'ifd4 'i'xd4 23 :xd4 f5 24 ~xe7+ Φχe7 25 :cl gives White a slightly better rook ending) 19... 'i'a4 20 b3 'i'a3 21 .td3 a5 22 'ifc3 a4 23 :bl is unclear, AdamsChήstiansen, Biel1991. 10 .th4 (D)

10 •••

Classicαl Vαriαtion

Or: 1) 10...~e7 (10... lbxd4 11 "xd4 ~c6 12 ~c4 is clearly good for White, while 10...:c8 11 tZ:If3 is similar to the note to Black's 9th move) 11 tZ:If3! b5 12 e5 (12 ~xf6 ~xf6 13 "xd6 :a7 14 e5 ~e7 15 "d3! 16 ΦbΙ :c7 17 "e3! 0-0 18 "e4 was also good for White ίη Georgadze-Makarychev, Nikolaev 1983) 12...b4 13 exf6 bxc3 14 "xc3 gxf6 15 f5 and White stands well, Thiemann-Reynolds, Cοπ. 1966. 2) 10•••tZ:Ixe4 ll1i'el and now: 2a) 11...15 12 tZ:Ixe4 tZ:Ixd4 (after 12 ... gxh4 White can reply 13 "c3) 13 fxg5 tZ:If5 14 tZ:If6+ Φe7 15lZXl5+ Φe8 16 tZ:If6+ Φe7 17 ~f2 ~c6 18 ~d3 hxg5 19 ~xf5 ΦΧf6 20 ~d4+ ΦΧf5 21 ~xh8 Φg6 and by now Black has somehow achieved a quite reasonable position, Jansa-Formanek, Gausdal 1991. Ι think White would have done better to play 14 1i'c3!, e.g. 14 ... ~g7 15 tZ:Ixd6+ Φf8 16 gxh6! ~xc3 (16 ... ~xh6+ 17 ΦbΙ +-) 17 ~xd8 :xd8 (17 ... tZ:Ixd6 18 ~c7!) 18 tZ:Ixb7 and White should be winning. 2b) 11•••tZ:If6 12 tZ:If5 1i'a5 (the lines 12.....b8 13 ~xf6 gxf6 14 tZ:Ie4, 12... ~e7 13 tZ:Ixd6+ Φf8 14 tZ:Ixb7 "c7 15 "d2 and 12.....c7 13 ~xf6 gxf6 14 tZ:Id5 "d8 15 "e3! are all very pleasant for White) 13 tZ:Ixd6+ ~xd614 :xd6 (D) with the

"a5

lίnes:

2b1) 14•••1i'c7 15 :d2 is unsatisfactory after 15•••0-0-016 "f2lΔe7 17 ~d3 ~c6 18 f5 e5 19 :hdl or 15•••1i'xf4 16 ~e2 tZ:Ie4 17 tZ:Ixe4

1i'xe418 "f2, Gιίgοήc-Βarden, Bognor Regis 1957. 2b2) 14•.• tZ:Ie7 15 :d1 tZ:Ig6 16 tZ:Ie4! "xe117 tZ:Id6+ Φe7 18 ~xel tZ:Id5 19 tZ:Ixb7 tZ:Ixf4 20 g3 tZ:Ig6 21 ~g2±.

2b3) 14•••tZ:Ib4 15 a3 tZ:Ibd5 16 1i'e5 ~c6 17 ~c4! is also very good forWhite. 2b4) 14•••0-0-0 15 1:[dl! "c7 (l5 ... tZ:Ie7? 16 tZ:Id5 wins, 15 ... g5 16 fxg5 hxg5 17 ~g3 gives White very strong pressure οη the dark squares and 15 ... e5 16 fxe5 :he8 17 ~g3 tZ:Ixe5 18 tZ:Ib5 ~g4 19 :xd8+ :xd8 20 tZ:Ia7+ Φb8 21 tZ:Ic6+ bxc6 22 1i'xe5+ is a very good ending for White) 16 "f2lΔe7 17 ~d3 ~c618 f5 e5 19 :he1 tZ:Ied5 20 tZ:Ixd5 :xd5 (20 ... ~xd5 21 "a7) 21 "g3 e4 22 "xc7+ Φχc7 23 ~xf6 exd3 24 ~xg7 :hd8 25 ~e5 with good winning chances for White, SpasskyRabar, Gothenburg 1955. 11 fxg5 tZ:Ig4 12 tZ:Ixc6 Ιη BTS2 both this and 12 tZ:If3 were featured, but this time we have opted for just one vaήation, but covered it ίη greater depth. The move

Classicαl Vαriαtion

order 12 ~e2lDge5 13 lLJxc6 ~xc6 is also possible. ~xώ 12 13 ~e2 ω 14 g3 lLJg6 (D)

his pawns but ίι still doesn't seem Ιο be enough for a significant plus) 21 ... ~e7 22 h4 f6 23 ':'gl ':'ag8 24 ~f3 ~f4 25 ':'g4 fxg5 26 hxg5 ~e3 27 ':'el (perhaps 27lLJd5+, but Black could haνe aνoided this with 26... ~xg5) 27 ... ~xg5 28 ':'egl':'f8! 29 ':'xg5 ':'xf3 30 ':'h5 ':'f6 31 ':'gh 1 ':'f4 32 b3 ':'h7 33 ':'el ':'f6 34 ':'ehl ':'f4 35 ':'5h4 ':'xh4 36 ':'xh4 ~f6 37 ':'hl ~e5 with equalίty, AkopianKozul, Moscow OL 1994. 2) 15...h5 (D) and now:

w Αη unusual position with the white bishop incarcerated οη h4. 15 ':'hfi More common is 15 ~bl, with the following possibilίties: 1) 15.• ~e716 gxh6 and now: la) 16...~xh4 17 gxh4 "xh4 18 'ίIi'xd6 ':'xh6? (18 .....e7 is better although White stands well) 19 ~xa6! ':'d8 (19 ...':'xa6 20 'ίIi'b8+ and mate οη d8) 20 "c5 ':'xdl+ 21 ':'xdl 'ίIi'e7 22 "e3 ':'h4 (22 ....:.xh2 23 "g3!) 23 ~b5 with a winning adνanιage for White, Brodsky-Neνednichy, Βυ charest 1995. lb) 16...lDxh4 17 gxh4 ~f6!? (Black clearly has some positional compensation for the sacrificed pawns) 18 ':'hf1 'ίIi'e7 19 ~g4 (with the idea 19 ... 0-0-0 20 lLJd5) 19 ...~e5 20 'ίIi'g5 (20 ~h5 _xh4 21 ~xf7+ ~e7 is fine for Black, as 22 ~g6 is met by 22 ... .:.ag8) 20 .....xg5 21 hxg5 (White has straightened ουι

2a) 16 _e3 ~e7 (this is SUΡeήοr Ιο 16... ~g7 17lLJd5! exd5 18 exd5+ "e7 19 _xe7+ ~xe7 20 dxc6 bxc6 21 c3 with an edge for White, as ίη Chandler-Bellin, Commonwealth Ch 1985,or 16 ... _e7 17 ':'hf1 ~g7 18 ':'f2 ~f8 19 ':'df1, Maίnka-popoνic, Dortmund 1988 and White is better) 17 ':'hf1 "c7 18 "f2 lLJe5 19 h3 0-0-0 20 g4 hxg4 21 ~xg4 lLJxg4 22 hxg4 ':'h7 with an unclear posiιίοη, Jansa-Banas, Czechosloνak Ch 1986. 2b) 16 ':'hfi "c7 (16 ... ~g7!? is better) 17 e5! 0-0-0 18 ~d3lLJxh4 19 gxh4 ~g7 20 "f2! and White

Clαssicαl Vαriαtion

stands well, Tseshkovsky-Fahnenschmίdt, Baden-Baden 1988. 2c) 16:dfi!? intending 17 'i'dl attacking h5 is an interesting idea. 15 ••• J..e7 The alternative is IS ••• hS, when 16 ΦbΙ transposes ίηΙο variation '2b' above. Αη important question is whether White can progress without playing ΦbΙ. One try is 16 e5!?, for example 16... lbxe5 (16 ...dxe5 17 'i'xd8+ :xd8 18 :xd8+ ΦΧd8 19 :xf7 is good for White) 17 g6 (17 ΦbΙ 'fIc7 18 g6 lbxg6 19 J..f6 :h7 20 'i'g5 is good for White, but better is 17 ... lbg6 18 J..d3 lbxh4 19 gxh4 J..g7 with an unclear game) 17 ... J..h6 18 gxf7+ lbxf7 19 J..xd8 J..xd2+ 20 1Ixd2 1Ixd8 21 1Id4 (the threat is 11h4) 21 ... h4! 22 1Ixh4 :xh4 23 gxh4 Φe7 with an unclear ending. Another try is 16 'fIe3, ίη tending 16 ...'i'c7 17 lbd5!. However, Black should play 16 ... J..e7 so that his knight is ready Ιο retum Ιο e5 ifneed be. 16 gxh6 lbxh4 16•••J..xh4 17 gxh4 'i'xh4 18 "xd6 is good for White. 17 gxh4(D)

17 ••• J..f8 17.••J..xh4 18 1Igl! is given as ± by Bonsch but matters are ηοΙ so clear after 18••• J..f6 (the immediate 17 ...J..f6 would 10se Ιο 18 'fIf4), with similar ideas Ιο the Akopian-Kozul game given above. Ι have examined this position ίη some detail and my analysis runs: 19 h7 'fIe7 (D) (forced as 19 ...Φe7 20:df1 J..xc3 21 'ffg5+ is crushing)

20 lbd5!? (20 :df1 0-0-021 'i'h6 is also possible, but the text is full of dangerous pitfaHs for Black; another way Ιο introduce the sacrifice is 20 :g8+ Φd7 21 lbd5, but after 21 ...exd5 22 exd5 :axg8! 23 hxg8'f1 :xg8 24 dxc6+ bxc6 25 1Id3 the situation is far from clear) 20•••exd5 21 exd5 and now we have: 1) 21 .•.J..b5 (21 ... J..xd5 22 :g8+ Φd7 23 'fIxd5) 22 :g8+ Φd7 23 J..g4+ Φc7 24 :el and now: 1a) 24•.•J..e5? (this allows White Ιο play a stunning combination) 25 'iνa5+! b6 26 'iνc3+!! Φb7 27 :xe5! (eliminating the defender of h8) 27 ...dxe5 (27 ...1Ixe5 28 J..c8+ Φb8 29 1Ixe5 dxe5 30 :xh8) 28 J..c8+

CΙαssicαl Vαriαtion

Φa7 (28 ... Φb8 29 d6! also wins) 29 d6! (clearer than 29 ':xh8 "g5+ 30 "d21i'gl+ 31 "dl "g7) 29...1i'xd6 (forced) 30 ':xh8 and the h-pawn will decide the issue. lb) 24...J..xb2+! 25 ΦΧb2 'ft'f6+ 26 "c3+! "xc3+ 27 Φχc3 ':axg8 28 hxg8" ':xg8 29 h3 and White's centralised king and outside passed pawn will make life very unpleasant for Black. Αη important point is that 29 ... Φd8, intending ...J..d7, is met by 30 ':bl Φc7 31 ':f1. 2) 21 •••1i'e5! 22 c3 J..d7! (after 22 ... J..xd5 23 'Wxd5 'Wxd5 24 :xd5 Black can't take οη h7, so he will suffer ίη this ending; 22 ... 'Wxd5 and White has at least the same as above; 22 ... J..a4 23 ':del Φd8 24 ':g8+! Φc7 25 J..g4':xh7 26':xa8! {White must be careful, e.g. 26 ':xe5 :xg8 27 ':ί5 ':xg4 28 b3 J..g7! 29 bxa4 J..h6 30 :xf7+ ':xf7 31 'Wxh6 :g2 and Black wins} 26 ... J..g5 27 :c8+! Φb6 28 ':xe5 J..xd2+ 29 ΦΧd2 dxe5 30 h3 +- with a winning endgame) 23 :del ΦdS! (D) and now:

2a) 24 ':g8+ (24 J..xa6 "h5 25 J..xb7 ':xa2 is ηοΙ recommended)

24 ... Φc7 25 J..g4 ':xh7! 26 ':xa8 (forced, as 26 ':xe5 ':xg8 is winning for Black) 26...J..g5 27 ':xe5 J..xd2+ 28 ΦΧd2 dxe5 29 d6+ Φc6 30 J..f3+ 'iPxd6 31 J..xb7 :xh2+ with an υη clear endgame. 2b) 24 J..dl!? 'Wf5 25 J..c2 (ifhe so wishes, White can take a draw by means οί 25 ':ef1 'ίi'e5 26 ':el) 25 ... "ii'h5 26 ':ef1 (D) and Ι believe White has good compensation for the piece. Here are a few sample varίations:

2b1) 26•••"ii'h4 27 :g8+ Φe7 (27 ... Φc7 28 ':ί4 ':axg8 29 hxg8'W ':xg8 30 ':xh4 ':g 1+ 31 J..d 1 J..xh4 32 Φc2 should be good for White) 28 ':ί4 'Wh5 29 :xf6! :hxg8 30 hxg8tb+ ':xg8 31 ':f1 with advantage to White. 2b2) 26••.J..e7 27 ':g7 J..e8 28 'We3! J..f8 29 'Wb6+ Φc8 30 J..f5+ Φb8 31 ':g8 +-. 2b3) 26••.'ii'e5 27 'Wh6 (27 ':el 'Wh5 is a draw) 27 ... J..e7 28 ':xf7 'iPc7 (28 ...J..e8 29 ':fg7) 29 ':gg7 ':ae8 (29 .....el+ 30 J..dl IΣae8 31 "f4 is even better for White) 30 b3!? White attempts ιο get his king

Clαssicαl Vαriαtion

Ιο

safety whίlst Black is tied down as after 30...'ii'xc3 31 J:xe7 'ii'al + 32 ~d2 'ir'd4+ 33 i.d3 'ir'b2+ 34 ΦdΙ the next check is blocked with 'ir'cl +. 2b4) 26...i.e5 27 J:g5 (another idea is 27 'it'e3!?) 27 ... 'ir'xh2 28 'ir'xh2 i.xh2 29 J:xf7 i.e5 30 J:g8+ (30 i.f5 i.f4+) 30... Φc7 31 i.f5 J:axg8 (31 ...:d8 will only delay the inevitable since Black is ίη zugzwang) 32 hxg8'ii' J:xg8 33 J:xd7+ with White having slίghtly the better of a drawn ending. Black certainly has Ιο go through a minefield Ιο reach this drawn ending, but objectίvely speaking White might be better off with the simple 20 'iνh6 0-0-021 i.d3, followed by J:df1, maintaining the h7-pawn for the moment. If Black exchanges οη c3 White's king becomes slightly exposed, but a more relevant feature is that White's h-pawn becomes far more dangerous once the darksquared bishop disappears. 18 'ir'd4! :xJι6 19 h5 19 .!bd5 deserves investίgation. 19 ..• 'iνg5+ 20 ΦbΙ "e5?! Better is 20•. :ilc5, although after 21 'ir'd2 J:h7 22 i.g4! White stίll has good chances of obtaining the advantage. 22•. .i.g7 can be met by 23 J:f3 and 22•••0-0-0 23 lΔd5 f5?! 24 b4! 'ir'a7 25lΔf6:f7 26 exf5 J:xf6 27 fxe6 is winning for Whiιe. 21 'iib6! Νοι 21 'iνxe5? dxe5, which gives Black excellent compensation. The

text ensures that the black king will ηοΙ be able ιο escape ιο the queenside. 21 .•• J:c8 Bonsch gives 21 .....xh2 22 'ir'c7 and then 22••• J:h7 23 e5! 'ir'xe5 24 i.d3, but points out that 22••.i.e7 would be a tougher nut ιο crack. 22 '6'1'2 J:h7 23 i.g4 1i'cS 23••.i.xe4? loses after 24 lΔxe4 'ii'xe4 25 i.xe6!. 24 '6'g2 i.d7 Defending against i.xe6 and lΔd5.

"f3

J:c7 (D)

26 e5! This thematic pawn sacήfice places Black ίη seήοus trouble. d5 26 •.• 26.••'6'xe5 27 lΔd5 and 26•••dxe5 27 :xd7! J:xd7 (or 27 ... ΦΧd7 28 'ir'd3+) 28 lΔe4 are out of the quesΙίοη, but the blocked centre also favours White, ηοΙ least because Black ηο longer has control over e5. :b6 27 '6'd3 28 '6'd2 J:h8 29 '6'Ι4 :h7

Clαssical Varίαtion

30 .te2! Another ροίηι behind 26 e5 was the opening of the bl-h7 diagonal. 30 .ih6 31 'iWh4 1ke7 32 :t'6! .tg7 33 1kg3 'i't'S 33....txf6 34 exf6 1:txc3 35 bxc3 'ίWίB 36 'ίWb8+. 34 .td3 :Jι6 34•••Wι8 also loses: 35 :xf7! 9;xf7 36 :f1 + Φg8 37 :xf8+ ΦΧίΒ 38lL\xd5 exd5 39 e6 +-. 35 :xh6 .txh6 36 lL\xd5! 1-0 36 ... exd5 37 e6 :tc5 38 exd7+ ΦΧd7 39 b4 :c7 40 .tf5+ Φc6 41 'ίWc3+ 'iPb6 42 'i'd4+ wins for White. Game 12 Short - Ljubojevic Amsterdam Euwe mem 1988 1 2 3 4 5 6 7

e4 lL\f3 d4 lL\xd4

cS d6 cxd4 lL\f6

lL\c3

lL\c6

.tg5 e6 'i'd2 a6 h6 8 0-0-0 Black once again aims Ιο further his queenside ambitions by ....td7 and ...b5 before committing his king, but ίη this line he first of all forces White Ιο decide where Ιο ρυΙ his bishop. 9 .te3 (D) This move is cuπentlΥ thought best. White intends Ιο play f4 and develop his king's bishop Ιο e2 or d3. If

Black castles kingside then White can launch a direct attack by h3 and g4-g5.

9 •.• .td7 Black has a number of possible moves, and although plans with f3 have become quite common for White, we are sticking Ιο the recipe recommended ίη BTS2, which is 'f4 against anything'. The alternatives are: 1) 9.••lL\g4? is a mistake since 10 lL\xc6 bxc6 11 .tc5 gives White an advantage. 2) 9.••'i'c7 is quite often played, but usually transposes Ιο lίnes considered later. Αη independent example is 10 f4 .te7 (10....td7 11 .td3 b5 transposes Ιο the note Ιο Black's 11th move) 11.te2!? (11.td3 .td7 12 Φb1 b5 transposes Ιο line 3 ίη the note Ιο Black's 12th move) 11 ....td7 (or 11 ... lL\a5 12 e5! dxe5 13 fxe5 'ίWxe5 14 .tf4 'ίWc5 15lL\a4 'ίWd5 16 Φb1! with a very strong attack for the sacrificed pawn) 12lL\b3lL\a5 13 lL\xa5 'ίWxa5 14 Φb1 .tc6 15 .tf3 'ίWc7 16 :hel (the immediate 16 g4 may be even better) 16 ...:c8 17 g4

Classicαl Vαriαtion

with advantage Ιο White, Chiburdanidze-Lanka, USSR 1980. 3) 9 ...J.e7 10 Ι4 ~xd4 11 J.xd4 b5 (D) and now we have:

3a) 12 J.e2!? b4 (12 ... J.b7 13 J.f3 b4 14 J.xf6 J.xf6 15 ~e2 'ifa5 16 a3 is good for White) 13 ~a4 ~xe4 14 "iVe3 ~Ι6 15 J.f3 and now: 3al) 15•••dS 16 'iPbl (this move has been played ίη practice, but ίη fact 16 g4 ο-ο 17 ~b6 :'b8 18 'iPbl may be more accurate, transposing ίηΙο Short-A.Rodήguez below but without allowing Black so much choice) 16 ... 0-0 (16 ... J.d7 17 ~b6 :b8 18 g4 J.b5 19 h4 'iPf8 20 g5 ~e8 21 f5 was also dangerous for Black ίη Chandler-Kosten, Hastings 1988/9) 17 ~b6 :'b8 18 g4 J.d6?! (18 ... ~d719~xd7 J.xd7 20 h4 J.f6 21 g5 J.xd4 22 :'xd4 was better, even though White still has a dangerous attack) 19 g5 hxg5 20 fxg5 ~7 21 ~xc8 :'xc8 22 g6 with a clear advantage for White, Short-A.Rodriguez, Subotica 1987. 3a2) 15••':'b816 J.a7 J.d7! (ηοι 16 ... :'b5 17 J.b6 'ifd7 18 ~c5 and White wins, Balashov-Tukmakov,

Sverdlovsk 1987) 17 ~b6 J.b5 18 ΦbΙ!? (after 18 g4 ο-ο 19 J.xb8 'ifxb8 20 g5 hxg5 21 fxg5 J.d8! 22 gxf6 'ifb6 23 'ifxb6 J.xb6 Black had sufficient compensation ίη Κhalif man-Ionov, USSR 1988) 18.•. 0-019 Ι5 flc7 (the only move) 20 J.xb8 ':'xb8 21 ~a8! "iVd8 (ηοι 21 .. :ifc8? 22 'ifa7 J.f8 23 fxe6 fxe6 24 ~c7 and the knight escapes, nor 21 .. :ifd7 22 fxe6 fxe6 23 :'hel d5 24 ~b6, followed by takίng οη e6) 22 fla7! (after 22 fxe6 :'xa8 23 exf7 + 'iPxf7 24 J.xe4 fxe4 25 :'xe4 J.f6 26 ~c7 Black's active minor pieces are at least as valuable as White's rooks) 22...dS! (22 ... exf5 23 :'hel ~e4 24 J.xe4 fxe4 25 ':'xe4 J.f6 26 ~c7 and 22 ... e5 23 :'hel! J.d7 24 ~c7 J.xf5 25 ~xa6 are good for White) 23 ~ε7 J.c6 (threatening 24 ...:'b7 and if 24 ~xa6 then 24 ...:'a8 wins) 24 fxe6 :'b7 25 ~xd5! :xa7 (not 25 ... ~xd5 26 exf7+ 'iPf8 27 'ifxa6 :'c7 28 'ifd3 :d7 29 'ifh7 ~f6 30 'ifh8+ 'iPxf7 31 'ifxd8 and wins) 26 ~xΙ6+ J.xf6 27 :'xd8+ J.xd8 28 exf7+ ΦχΙ7 29 J.xc6 and now: 3a21) Ιη the original game wίth 18 'iPbl !1, Νυηη- Van der WieI, Lucerne 1989, BIack now played the somewhat inaccurate 29 ...J.f6 and after 30 :'d 1 :c7 31 J.f3 J.e7 White couId, have obtained reasonabIe chances with 32 :'d5! :'c5 33 :'d3. 3a22) After 29...:'f2!, however, ίt is very difficult for the white rook Ιο actίvate itseIf. 30 :'el J.f6 31 a3 (obviously White wouId prefer ηοι to split his queenside but there's ηο other way Ιο try ιο make progress)

Clαssicαl Vαriαtion

31 ... bxa3 32 bxa3 and even though White has an extra pawn his winning ~hances are minimal. One example is Adams-Serper, Manila OL 1992 which concluded 32 ... a5 33 h3 1:f5 34 Φa2 1:c5 35 .i.a4 Φf8 36 .i.b3 h5 37 1:e4 1:e5 38 1:a4 1:b5 39 1:c4 Φe7 40 1:e4+ 1:e5 41 1:xe5+ .i.xe5 42 .i.c4 Φd6 43 Φb3 Φc5 44 .i.e2 1/'1.-1/2. Even so, this iS a dangerous line for Black ιο negotiate and there are few players willing ιο include such an ending ίη their repertoire. 3b) 12 'ii'e3 (D) and now:

3b2) 12•••.i.b7 13 .i.xf6 .i.xf6 (13 ... gxf6 14 f5!? 'ili'a5 15 a3 1:c8 16 1:d31:c5 17 .i.e2 1:e5 18 1:n 'ii'c7 19 "g3 .i.xe4 20 lbxe4 1:xe4 21 1:ιc3 gave White good play ίη TalTukmakov, USSR Ch 1983) 14 e5 .i.e7 15 exd6 .i.xd6 16lbe4 .i.xe4 17 'ili'xe4 ο-ο 18 .i.d3 g6 19 h4 and White has good attacking chances, Kovalev-Sjoberg, Ostende 1991. 4) 9 ...lbxd4 10 .i.xd4 b5 11 Ι4 .i.b7 (11 ... b412 .i.xf6 'ili'xf613lbe2 1:b8 14lbd4 1:b6 15 .i.c4 was better for White ίη Tal-Radulov, Malta OL 1980) 12 'ii'e3 (D) and now:

Β Β

3bl) 12.•.'ii'c7 13 e5 dxe5 14 fxe5 lbd7 15 lbe4 0-0 16lbf6+! lbxf6 17 exf6 .i.xf6 18 .i.xf6 gxf6 19 'ii'xh6 (19 1:d4!? is interesting) 19 .....e5 20 h4 .i.b7 21 .i.d3 f5 22 1:h3 1:fd8 23 g4 'ii'g7 24 'ili'e3 fxg4 25 1:g3 (ΚΙο vans-Makarychev, USSR 1983) and now 25 ...1:ac8 26 1:dgl 1:xd3! 27 .xd3 f5 would have been good value for an exchange. However, Ι Uιink White could have obtained the advantage with 22 1:hel 'ili'g7 23 .f4, intending 1:e3-g3. 23 ...1:fd8, with the idea of ...1:d4, is well met by 24 'ili'c7.

4a) 12....i.e7 transposes Ιο '3b2'. 4b) 12... b4 13 .i.xf6 gxf6 (the alternative 13 ... 'iIi'xf6 14lbb5! 'ili'd8 15 lbxd6+ .i.xd6 16 e5 .i.d5 17 1:xd5 exd5 18 exd6+ Φf8 19 'ili'c5 is very good for White) 14 lba4 'ili'a5 15 lbb61:b8 16 .i.c4 is promising for White as 16....i.xe4 17 'ili'xe4 'ii'xb6 18 1:hel and 16...d5 17 exd5! .i.c5 18 'ii'g3! are both very dangerous for Black. 4c) 12...'ii'c7 13 .i.b6 'ii'c8? (better is 13 ... 'ii'c6, although after 14 lbd5 1:c8 15 lbxf6+ gxf6 16 .i.d3

Clαssicαl Vαriαtion

:g8 17 g3 White still has an edge, Nunn-Lobron, Munich 1991) 14 e5 ι!LJd5 15 ι!LJxd5 ~xd5 16 exd6 'ίi'c6 17 :xd5! (this gives White a murderous attack) 17 ... 'ίi'xd5 18 ~e2 'ίi'xa2 (18 ... 'ίi'xd6 also loses after 19 :dl, for example 19 ... 'ίi'e7 20 'ίi'c3 'ίi'b7 21 ~ί3 'ίi'xb6 22 i.c6+! or 19... 'ίi'b8 20 i.f3 i.e7 21 i.c6+ ~f8 22 :d7 with an overwhelming ρο sition) 19 'ίi'f3! i.xd6 (19 ... 'ίi'd5 20 d7+! ~e7 21 :dl 'ίi'xf3 22 d8'ίi'+ and 19 ... :c820d7+~xd7 21 'ίi'b7+ are both the end) 20 'ίi'xa8+ ~e7 21 'ίi'b7+ ~f8 22 i.c5! 'ίi'al+ 23 ~d2 'ίi'xhl 24 'ίi'b8+ 1-0 Nunn-Fedorowicz, Wijk aan Zee 1991. 10 Ι4 b5 Against other moves White can adopt the same general plan οί i.d3, ~bl, followed by a kingside pawn advance, but he has to be careful against 10...i.e7, because 11 i.d3 allows the awkward 11 ... ι!LJg4! and 11 h3 b5 12 i.d3 transposes into the note to White's 12th move. Therefore the best answer to 1O... ~e7 is 11 ~bl b5 12 i.d3 and we are back ίη the main game. i.e7 (D) 11 i.d3 Or: 1) 11...:c812 ~bl ι!LJa5?! 13 e5! b4! 14 ι!LJce2 dxe5 15 fxe5 ι!LJd5 16 :hf1 ι!LJc4 17 i.xc4 :xc4 18 ι!LJί4 ι!LJxe3 19 'ίi'xe3 i.c5 20 ι!LJg6! with a clear advantage for White, HazaiSzabo, Hungary 1983. 2) 11 •••ι!LJxd4 12 i.xd4 b4 13 ι!LJe2 'ίi'a5 14 ~xf6 gxf6 15 ~bl, with advantage to White, PsakhisIvanovit, Sochi 1979.

3) 11 •.• 'ίi'c7 12 ~bl (the piece sacrifice 12 ~xb5 also deserves con~ sideration; Yudasin-Greenfeld, Haifa 1993 continued 12 ... axb5 13 ι!LJdxb5 'ίi'b8 14 ι!LJxd6+ ~xd6 15 'ίi'xd6 'ίi'xd6 16 :xd6 ι!LJa5 17 b3 and such positions often favour White ίη prac· tice) 12••• ι!LJa5 13 'ίi'el (White plans to keep his bishops) and now: 3a) 13•.. b4 14 ι!LJce2 ι!LJc4 15 i.cl i.e7 (15 ... a5 can also be met by 16 ι!LJg3) 16 ι!LJg3 (White plans to drίνe the knight back with 'ίi'e2) 16...:b8 17 'ίi'e2 ι!LJa3+ 18 bxa3 bxa3+ 19 ι!LJb3 a5 20 i.xa3 a4 21 i.c4 axb3 22 cxb3 and Black had ηο compensaιίοη for his pawn, Gallagher-Andreescu, NeucMtell993. 3b) 13.•• ι!LJc4 14 i.cl i.e7 15 i.xc4 bxc4 16 e5 dxe5 17 fxe5 ι!LJh'7 18 'ίi'g3 ι!LJί8 19 'ίi'xg7 :h7 20 'ίi'g3 ι!LJg6 21 ι!LJδ! 'ίi'xe5 (21 ... exf5 22 e6) 22 'ίi'o :c8 23 ι!LJxe7 ι!LJxe7 24 ι!LJe4 ι!LJί5 25 g4 i.c6 26 :hel i.xe4 27 :xe4 1-0 Smirin-Greenfeld, Elenite 1994. 4) 11.•. 'ίi'&5 can be met by 12 ~bl followed by the usual plan.

12 ~bl

86 Classical Variation Ιη

BTS2 a strong case was made for starting the kingside offensive at once with 12 b3 but Black may be able to exploit the omission οfΦbl by playing 12•••lbxd4 13 ~xd4 b4 14lbe2 e5!, the point being that 15 fxe5 dxe5 16 ~xe5? 'ii'a5 17 ~xf6 is met by 17 ... 'ii'xa2. After 14... e5, severaI games have continued 15 ~e3 'ii'a5 16 ΦbΙ ο-ο 17 g4 (D) and now:

1) 17•••1%fd8 18 fxe5 (18 g5!?, intending to meet 18 ...exf4 by 19 lbxf4! hxg5 20 lbd5, aIso deserνes attention) 18 ... dxe5 19 lbg3 ~e6 20 b3 lbe8 and now Adams-Shirov, Groningen 1990, continued 21 84 with unclear consequences. Ιη his notes Ιο the game, Shirov awarded 21 a4 an exclamation mark, claiming that the alternative, 21 'ii'f2 is well met by '21 ... lbd6! 22 ~b6 'ii'a3 23 ~xd8 ~xd8, with ... lZ)b5 to folIow.' However, ίη Dvoirys-Lanka, GausdaI 1991 White played 24 'ii'c5! and obtained a winning position after 24... lZ)b5 25 ~xb5 axb5 26 1%xd8+! :xd8 27 'ii'xe5 1%a8 (27 ...1%d2 28 lZ)h5!) 28 'ii'b2. If instead Black had played 24...~e7 then at the very

least White has 25 'ii'c6 1%a7 (or eIse 'iixa6) 26 'ii'a4 'iixa4 27 bxa4 and with lbf5 to folIow ίι is unlikely that Black has enough compensation for the exchange. 2) 17•••exf4 and White has two recaptures: 2a) 18 ~XΙ4 ~oo and now: 2al) 19 b3 d5 20 e5 lbe4 21 ~xe4 dxe4 22lbd4 'ii'c5 23 'ii'e2 a5 24 :hgl a425 g5 axb3 26 cxb3 hxg5 27 ~xg5 1%fd8 28 ~xe7 'iixe7 29 'ii'xe4 and White won quick1y ίη WahIs-Lutz, BundesIίga 1995. 2a2) 19 lZ)cl d5! (Iess good is 19 ... lZ)d7 20lbb3 'ii'a4?! {20 ... 'ii'b6 21 g5!} 21 g5 h5 22 ~e2 with advantage to White, Gallagher-Sher, Antwerp 1992) 20 e5 lZ)e4 21 'iie2 'ii'b6! 22 ~xe4 dxe4 23 'iixe4 a5 with good play for the pawn, Chandler-Lutz, BundesIiga 1994. 2b) 18lbxr4 ~c6 19 ~d4lbxe4 20 'iie3 (20 ~xe4 ~xe4 21 :hel d5 22 ~xg7 :fc8! 23 1%xe4 dxe4 24 ~xh6 ~g5! with advantage to Black, Nunn-Greenfeld, Pardubice 1993) 20 ... d5 21 ~xe4 dxe4 22 h4 and White has some attacking chances but he is a pawn down, ShabaIov-Sher, BieI 1993. 12 ••• b4 This move has been cήtίcίsed but ίι isn 't clear that the alternatives are better. They are: 1) 12•••0-0 13 g4!? (so far this has only been played by members of the PoIgar family but it is an attractive idea ίη serious danger of spreading; the aIternative, 13 h3, is Iikely Ιο transpose ιο other Iines, for example

Clαssicαl Vαriαtion

13 ... li)xd4 14 J.xd4 J.c6 15 'ii'e3 'ii'c7 is line '2') 13•••li)xg414:hgl li)xe3 IS "xe3 ~h8 (D) (this seems to be forced as 15 ... li)xd4 16 1fxd4 ~f6 loses after 17 e5 dxe5 18 1fe4 g6 19 :xg6+ J.g7 20 :xg7+ ~xg7 21 :gl+ ~f6 22 li)d5+! exd5 23 1fxe5#) and now:

la) 16 eS li)xd4 17 1fxd4 J.c6 18 f5 (18 1fe3 :g8 19li)e4 J.xe4 20 1fxe4 g6 21 1fe3 gave White some pressure for the pawn ίη Zso.PolgarGroszpeter, Hungarian Ch 1995) 18 ... dxe5 19 "g4 J.f6 20 fxe6 1fe7 21 exf7 1fxf7, J.Polgar-Salov, Madrid 1994, and now 22 ~g6 1fe7 23 li)d5 would give White good compensation for the pawn. lb) 16li)f3!? is Salov's suggesιίοη, which does look quite dangerous for Black. For example, 16 ... b4 17 :xg7 ~xg7 (17 ... bxc3 18 :dgl! and the threats of e5 and f5 are hard to meet, e.g. 18 ... J.f610ses to 19 e5! ~xg7 20 'ii'e4) 18 :gl+ ~h8 19 f5 J.g5 20 li)xg5 hxg5 (20 ... "f6 21 e5!) 21 :xg5 and the attack crashes through. 2) 12.....c713 h3 (D) and now:

2a) 13.••lbaS 14 g4 and now both 14.••b4 15 ~e2li)c4 16 ~xc4 1fxc4 17 li)g3 a5 18 :hf1 "a6 19 g5, Hodgson-Csom, Tel Ανίν 1988 and 14••. ~4 15 ~xc4 'ii'xc4 16 "g2 b4 17 e5 :b8! 18 li)e4 li)d5 19 J.cl dxe5 20 fxe5, Gallagher-Wells, NeucMte11995, hold promise for White. 2b) 13••• ~d4 14 J.xd4 J.c6 15 'ife3 0-0 16 :hf1 (16 e5 is also playable) 16...b417li)e2e518fxe5dxe5 19 1fg3 li)xe4 20 1fxe5 'ii'xe5 21 ~xe5 and White has a significant advantage, Aseev-Shirov, Daugavpils 1989. 3) 12•••li)xd4 13 J.xd4 J.c6 (or 13 ... b4 14li)e2 'ii'b8, Hazai-Lobron, Rotterdam 1988, 15 h3 and οη with the game) 14 'ii'e3 'ii'c7 15 e5!? (15 h3 is line '2b') 15 ... dxe5 16 fxe5 li)d5 17 'ii'g3 0-0-0 18 li)xd5 J.xd5 19 'ii'f2 with an edge for White, Adams-D.Garcia, Τeπassa 1991. 13 ~e2 ο-ο 14 h3 Black's problem is that whilst White has an automatic attack by h3, g4, lbg3 and g5, he will be struggling Ιο create any counterchances at all. Ιι is cUΉous that Black's troubles

Clαssicαl Vαriαtion

stem from the apparently innocuous 8 ... h6, which ίη this type of position can easily tum ίηιο a deadly kingside weakness. 'fIc7 14 'ίi'b7 15 g4 ~d4 16 ~g3 i.c6 17 i.xd4 ΑΙΙ Black has achieved is Ιο create a threat ιο e4, which delays White's attack by precisely one move. 18 :hel :Ιe8? This doesn't help the beleaguered kingside. The last chance Was Ιο play 18•.• ~7 intending ...e5 to block the deadly long diagonal, but ίη this case Short gives the line 19 g5 hxg5 20 :gl (threatening ~h5) 20 ... e5 21 ~f5 :fe8 22 fxe5 ~xe5 23 i.xe5 dxe5 24 :xg5! i.xg5 25 'ίi'xg5 f6 26 i.c4+ Φf8 27 'ii'h5 followed by mate ath8. 19 g5 hxg5 20 fxg5 ~7 (D) After this we are treated to a kinghunt ίη the style of the 19th century. 20••• ~h7 is objectively better (anything is better than being mated), but 21 h4 locking the knight out of play is very good for White.

21 i.xg7! Φxg7 22 ~h5+ Φμ There is ηο choice as 22.••Φg8 23 g6 fxg6 24 'ii'h6, 22.•• Φf8 23 g6 i.f6 24 ~xf6 ~xf6 25:f1 Φe7 26 'ii'g5 and 22••• Φh8 23 g6 i.f8 24 ':'gl fxg6 25 ':'xg6 ~e5 26 :h6+! alllead Ιο a disaster. 23 e5+ Φxh5 i.xg5 24 "Ι4 Φh4 25 "xf7+ 26 'iVh7+ Φg3(D)

The main problem when playing such positions with White is trying Ιο keep a broad grin off your face. The main problem when playing such positions with Black is to avoid looking at the broad grin οη your ορ ponent's face. 27 'ίiΊι5 Φh2 Or 27.. ':g8 28 'ίi'g4+ Φh2 29 'ii'gl+ Φxh3 30 i.f1+ Φh4 (30...i.g2 31 :d3+ and 32 i.xg2) 31 'ίi'h2+ Φg4 32 :d4+ Φf5 33 'i'h7+ :g6 34 i.h3# is a nice line given by Short. 28 "xg5 Ιη fact White could have forced mate by 28 'ii'e2+ ΦΧh3 (28 ...i.g2 29 1:ΣhΙ+ Φg3 30 'i'g4+ ΦΩ 31

CΙαssicαl Vαriαtion

':hf1 + .txf1 32 ':xfl + ~e3 33 ':el + ΦΩ 34 .gl+ ~f3 35 ':f1#) 29 '6'h5+.t.h4 30 ':e3+ ':f3 (30...~g2 31 '6'xh4 or 30....tf3 31 ':h 1+ Φg2 32 '6'xh4) 31 J:thl+ ~g2 32 "xh4 1:Ixe3 33 .tf1+ Φf3 34 J:th3#, but there is nothing wrong with winning Black's queen (and having his king οη h2). ':g8 28 .tg2 29 J:td2+ 30 'iνΙ4+ J:tg3 31 .te4 'iνxe4 32 'iνxe4 ι-ο Game 13 Kuzmin Minsk 1982

Kupreichίk -

1 2 3 4 5 6

lbc6

d4 lbxd4 lbc3 .tgS (D)

cxd4 tbr6 d6

6 ••. .td7 This is the most popular alternative Ιο 6 ... e6, but there are other moves:

1) 6•.•g6 7 .txf6 exf6 8 .tc4 (8 .tb5 .td7 9 ο-ο .tg7 10 lΩde2 is a more solίd alternative) 8....tg7 (after this a pawn must be sacήficed, but the alternative, 8 ....t.e7 9 "d2 followed by 0-0-0, is depressing for Black) 9 lΩdb5 ο-ο 10 "xd6 f5 11 0-0-0"a5 (11 .. :6'g5+ 12 f4 .xg2 13 e5 is good for White as Stoica's suggestion, 13 .. :6'g4, can simply be met by 14 1:.hf1! when 14 ... lΩxe5 10ses Ιο 15 .t.e2! and 14 ... .txe5 15 .txf7+! is ηο fun either) 12 .c7 a6 (the lines 12 ....txc3 13 bxc3 .a4 14 lΩd6, 12 ...•b4 13 lΩd6 .txc3 14 bxc3.xc3 15 .txf7+ ~h8 16 ~bl and 12 ... fxe4 13 .xa5 lΩxa5 14 .td5 .th6+ 15 ΦbΙ are all good for White) 13 'ii'xa5lΩxa5 141Ωc7 J:ta7 15 .tb3 .t.xc3 16 bxc3 fxe4 (Kholmov-Chernikov, USSR 1982) and now 17 lΩd5! is good for White according Ιο Κholmov. 2) 6•.•'6'aS 7 .t.xf6 gxf6 8 .t.b5 .td7 9 lΩb3 'ii'c7 10 lΩd5 .d8 11 .h5 e6 12lΩe3 a6 13 .te2 .c7 14 0-0-0.te7 15 ΦbΙ 0-0-0 16 f4 was good for White ίη S.Nikolic-Gufeld, Kislovodsk 1968. 3) 6.•. a6 7 .d2 lΩxd4 (alternatives transpose ίηΙο the main variations) 8 .xd4 e5 9 .a4+ (White has ηο trouble keeping a slight advantage by 9 'ii'd3 .te6 100-0-0 J:tc8 11 lΩd5 .txd5 12 .txf6, but with 9 .a4+ he is playing for' more) 9 ... .td7 10 .txf6 .xf6?! (10 ... gxf6 11 .b3 b5 12 .te2 .te6 13 lΩd5 .th6 14 a4, Marjanovic-Stoica, Istanbul 1988, is good for White, but Black has more hope than after

90 Classical Variation 1O... 1Wxf6) 11 i.b5! 1Wd8 12 J..xd7 1Wxd7 13 1Wb3 ο-ο 14 lίJd5 with a dream position for White, Kotronias-Kova1ev, Debrecen 1992. 4) 6.....b6 7lίJb3 (7 i.e3!? is ίη teresting, e.g. 7 ...1Wxb2 8lίJdb51Wb4 9 i.d21Wc5! 1Oi.e2!?1Wb611 :'bl lίJe5? 12 i.e3 1Wa5 13 ':'b3 g6 14 :'a31Wd8 15 i.xa7 ~ed7 16 f4! and White stands well, Ba1ashov-Petrί enko, Voronezh 1987, but 11 ...1Wd8 was the critica1 test) 7 ... e6 8 J..f4!? lίJeS (after 8 ... e5 the bishop returns Ιο g5) 9 i.e3 'ikc7 10 Ι4 lίJc6 (1O ...lίJg6 ll1Wf3 i.d7 12 i.d3 i.e7 13 0-0-0 i.c6 14 lίJd4 was good for White ίη Greenfeld-Schrenzel, Israel 1983) 11 g4!? (D) (there are anumber of reasonable alternatives, such as 11 i.e20r 11 i.d3, but the direct approach often pays off for White ίη the Sicilίan and also avoids drifting ίηΙο a bad posίtion) and Black has trίed several moves:

4a) 11 •..dS 12 e5 lίJd7 13 lίJb5 1Wd8 (13 ... 1Wb8 14 g5 a6 15 lίJ5d4 1Wc7 161Wf3lίJxd4 17lίJxd4 i.b4 18 'ίti>f2! i.e7 19 h4 was better for White ίη Topalov-Smirin, Burgas 1994)

14 h4 f6, Speelman-Gulko, Hastings 1989-90, and now 15 exf6lίJxf6 16 g5 lίJe4 17 i.g2 would have been very good for White. 4b) 11•.. h6 121Wf3 i.e7 13 h4 h5 14 gxh5 (14 g5!? is another idea) 14 ... lίJxh5 15 0-0-0 with a pleasant position for White, Ga11agher-Gross, Eupen 1991. 4c) 11•..86 12 g5 lίJd7 13 h4 b5 14 1Wd2 lίJb6 15 'ilif2 :'b8 16 i.d3 lίJc4 17 0-0-0 lίJxe3 18 1Wxe3 (perhaps this is a positional gain for Black but he is lagging ίη development and has a long-term king safety problem) 18 ... i.e7 19 h5 b4 20 lίJe2 a5 21 lίJbd4 lίJxd4 22lίJxd4 1Wc5 23 e5! dxe5 (23 ... d5 24 'ίti>bl+) 24 fxe5 h6 25 gxh6 gxh6 26 ~bl i.g5 27 1We2 i.d7 28 :'hf1 :'f8 29 i.c4! (a good move, vacating d3 for the queen and lίning υρ the bishop against e6, where something is liable Ιο explode ίη the ηοΙ Ιοο distant future) 29 ... a4 301Wd3 :'d8 (30 ... 'ilixe5 31 lίJxe6! is one example) 31 lίJb5! i.xb5 32 i.xb5+ ~e7 33 1Wd7+ (strong, even very strong, but ηοΙ as strong as 33 :'xf7+! and mate next move) 33 ...:'xd7 34 ':'xd7+ ~e8 35 :'d5+ 1-0 Kova1ev-Giffard, Clίchy 1991. 7 i.e2 (D) The main lίne is 7 'ikd2 when Black continues 7 ... lίJxd4 8 1Wxd4 'ilia5. The move i.e2 is well motivated since if Black exchanges οη d4 he has lost a tempo over the usua1lίne, so he has Ιο change his plan. 7 •.. 86

CΙαssicαl Vαriαtion

Flexible, but 7 ...'ii'a5 may be more This and other options: 1) 7 •• :ii'b6 8 tΔdb5 (threatening 9 .i.xf6 and 10 tΔd5, while at the same time preventing ... e6) 8 ... 1:.c8 9 ο-ο a6 1Ο .i.xf6 gxf6 11 00 'ii'd8 12 tΔbc3 e6 13 tΔe3 and White's knights proved well-placed ίη VogtMascaήnas, Polanica Zdroj 1977. 2) 7 •••00 8 .i.xf6 (also possible is 8 tΔdb5 'ii'b8 9 a4.i.e7 10 "'d2 a611 tΔa3 "'c7 12 1:.dl1:.d8 13 tΔc4.i.c8 14 .i.e3 "'b8 15 tΔb6 with a slight advantage for White, Spassky-Hort, Moscow 1971) 8 ••.gxf6 (8 ... 'ii'xf6? 9 tΔdb5) 9 0-0 (D) and now: cήtίcal.

2a) 9 ...b5 10 tΔdb5 "'b8 11 a4 a6 12 tΔa3 "'c7 13 "'d2 f5? (Black

didn't like the look of 13 ... 0-0-0 14 tΔc4 when White has dangerous attacking chances, but the move chosen seems rather suicidal) 14 tΔc4 tΔa5 (14 ... 0-0-0 15 tΔb5! axb5 16 axb5 tΔb8 17 'ii'c3! is decisive and 14 ... fxe4 15 tΔxe4 d5 16 tΔf6+ ~d8 17 'i!.fg5! .i.h6 18 'ii'h4 ~c8 19 1:.ad 1 is very good for White) 15 'ii'd4! 1:.h6 (Black would prefer Ιο play 15 ...1:.g8 but this is refuted by 16 tΔb61:tb8 17 tΔbd5! exd5 18 tΔxd5 with tΔf6+ to follow) 16 tΔb6 1:.b8 17 exf5 e5 18 'ii'e3 .i.xf5 19 tΔcd5 'ii'd8 20 'ii'c3! tΔc6 21 .i.xa6 and White had an overwhelming posiιίοη, Gallagher-Dubeck, San Bernardino 1994. 2b) 9 ...a610 ~hl (following the above game with Dubeck Ι amused myself with 1Ο .i.h5 'ii'b6 11 tΔf5!?, but Ι wouldn't dare Ιο suggest such reckless behaviour Ιο readers of BTS3) 10...'ii'c7 (10....i.e711 f4tΔxd4 12 'ii'xd4 'ii'a5 13 1:.adl was better for White ίη Hort-Geller, Palma de Mallorca 1970) 11 .i.h5!? .i.g7 12 f4 f5 (now or never) 13 tΔxc6 bxc6 14 exf5 exf5 15 1:.el + 'itf8 16 'ii'd3 .i.e6 (Black does have a couple of nice bishops but the more salient factors ίη assessing thiS position are his misplaced king and weak pawn structure) 17 1:tadl d5 18 tΔe2 c5 19 tΔg3 .i.xb2 20 tΔxf5 1:.b8 21 tΔg3 (getting ουΙ ofthe f-pawn's way) 21 ... c4 22 'i!.ff3 d4 23 f5 .i.c8 24 f6 .i.e6 25 tΔe2 d3 26 cxd3 c3 27 tΔd4! .i.xa2 (de la Villa-Ubilava, Palma 1992) and now White missed the brilliant 28 tΔb5!! after which Black could have

92 Classical Variation resigned; 28••.%lxbS loses after 29 'ίWg4! as now 29 ...:g8 is hit by 30 'ίWxg8+ and 31 %le8#, and 29 ... .:e5 30 %lxe5 'ίWxe5 31 'ii'c8+! leads Ιο mate ίη a few moves. The main ΡοίηΙ behind 28 lDb5!! is revealed after 28•••axb5 29 'ίWg4!, when ηοΙ οηlΥ is White threatening 'ii'g7+, but also the devastating 'ίWb4+. Nor is salvatίoη Ιο be found ίη refusing the offer as after 28.•. 'ίWb6, 29 'ίWf4! is decisive. 3) 7•.. 'ίWa5 8 ~xΙ6 gxf6 9 ο-ο!? (D) and now:

strong initiative ίη return for the pawn; this was recommended ίη BTS2, but has yet Ιο be tested at the highest level) 11 'ii'xc5 dxc5 12 lDb5! (even more promising than the 12 lDd5 of BTS2) 12••• 0-0-0 (12 ... ~d8 13 %lfdl is very unpleasant for Black as White is threatening Ιο double οη the d-file, and even 13 ... a6 can be met by 14 %ld3!) 13 lDxa7+ ~b8 14lDb5 ~g7 and now: 3cl) 15 Ι4?! f5 16 e5 f6 17 exf6 ~xf6 18 c3 ~c6 19 %ladl %lhg8 and Black's active bishops gave him reasonable compensation for the pawn, Short-Anand, Amsterdam 1992. 3c2) Anand proposes 15 ~d3! because now 15 ... f5 16 exf5 ~xb2 17 %labl is clearly good for White and 15 ... ~xb5 16 ~xb5 %ld4 17 c3 %ld2 (17 ... .:xe4 18 ':fdl with the idea of a4-a5-a6 is terminal according Ιο Anand) 18 %labl simply leaves Black a pawn down. 8 ~XΙ6 gxf6

3a) 9 ..:6'g5 10 lDf5! ':g8 (White has the advanιage after 1O... ~xf5 11 f4 'ίWg6 12 ~h5) 11 lDg3 with an edge for White. 3b) 9 ..•'ίWe5 (too ambitious) 10 lDf3 'ίWc5 (10 ... 'ίWa5 is better even though this is an admission that Black's last move was a mistake) 11 lDd5 %lc8 12 c3 a6 13 lDd4 ~g7 14 b4 'ίWa7 15 ~h5 with advantage Ιο White, Stoica-Kotronias, Istanbul 1988. 3c) 9 ... lDxd4 10 'ίWxd4 'ίWc5?! (1O...%lc8 11 οο! 'ίWc5 12 'ίWd2 'ίWxc2 13 'ίWe3 'ίWc5 14 'ίWf4 gives White a

There is a second possibility which, like the main line, is based οη the move lDf5, namely 9 lDf5 'ίWa5 10 ο-ο %lc8 l1lDd5 'ίWd8 12lDde3! lDe5 13 f4lDg614 'ίWd3! h5 15 %ladl b5 16 a4 and White is better, Sznapik-Hawelko, Poland 1984. 9 ••• 'ίWb6 9 ..•e6 transposes Ιο '2b' ίη the note Ιο Black's 7th move.

ο-ο

10 1Df5! (D)

Many players would have automatically retreated the knight Ιο b3 but Kupreichik realises that ίη this situatίon he can afford Ιο give υρ his

Clαssicαl Vαriαtion

b-pawn. Although Black can organise ... e6 ιο expel the knight we have already seen ίη Vogt-Mascarifias above that knights οη c3 and e3 can be well placed.

10 ... ο-ο-ο?! Black prepares ... e6 but ίι turns ουΙ that this is ιοο slow. He should have tried 10•••e6 (1O ... 'ffxb2? at once fails Ιο 11 tbd5 and 12 :bl) 11 lbxd6+ ~xd6 12 1fxd6 1fxb2 13 1Udl and now: 1) 13••• lbe5 14 lbd5! exd5 15 'Wxf6 lbc4 (15 ...:g8 16 :abl 'ffc3 17 :xd5 10ses at once) 16 c3 :f8 (16 ...:g817 ~h5) 17 :abl 'ffxe218 :el followed by exd5+ forcing Black Ιο play ... ~e6, when White continues dxe6 with a winning attack. 2) 13..•0-0-0 14 lbd5! exd5 15 exd5 'ffc3 (15 ... 'ii'e5 16 dxc6 'ffxd6 17 :xd6 ~xc6 18 :xf6 wins a pawn) 16 dxc6 'ffxc6 17 'ii'f4 fol10wed by ~f3 when Black has problems with his exposed king and his weak f-pawns. 3) 13••':d8! 14 'ffg3 (the sacή fice 14 :abl 'ffxc3 15 :xb7 flops

after the response 15 ... ~c8 16 :e7+ ι;Pf8) 14.. :it'a3 15 ~h5 and White still has some pressure, although far less than ίη the game.

11 a4 11 lbd5 'ffa7 only leaves White with the probIem of meeting ...e6. 11 .•. lbb4 Ιη the Sicίlian, BIack can normaIIy only contemplate castling queenside when White has also pIayed 0-0-0, since ίη a race between attacks οη opposite wings the missing bIack c-pawn gives White a Iarge head start. 12 a5 'ii'c7 13 lba4 Black's ... tLΊb4 has stopped lbd5 by White but there are other ways ιο reach b6. 13 ••. e6 (D) 13••.'ii'xa5 14 c3lbc6 15 b4 'ffc7 16 'ffb3 gives White two free tempi and an open a-fiIe for his attack.

~xB4 14 lbd4 White could ηοΙ prevent this exchange by pIaying 14lbb6+ Φb8 15 lbd4 as then 15 ... e5 followed by ...'ffxc2 wouId confuse the issue.

Clαssicαl Vαriαtion

15 :'xa4 d5 Black's only chance is Ιο find some counterplay quickly, or else he will be crushed by c3 followed by b4-b5. 16 c3 tbc6 Νοι 16•.•dxe4 17 cxb4 e5 when 18 'iί'bl threatening :cl wins. 17 exd5 :xd5 18 ~f3 :d6 18...1txaS 19lL1xc6 :xa4 20 'iί'xa4 bxc6 21 ~xc6 is one ofthose positions ίη which the opposite-coloured bishops increase the strength of an attack Ιο alarming proportions.

19 :c4 :g8? Black overlooks the threat. 19...e5 was necessary but even then White has the choice of two lίnes: 20 lLIxc6 Axdl 21 :xdl bxc6 22 Axc6 gives White an ending with aπ extra pawn and the better position, while some players might prefer 20 ~xc6 bxc6 21 'iί'g4+ aπd 22lL1f5. 20 'iWa4 So simple; c6 collapses and with ίι Black's position. :g5 20 21 ~xc6 bxc6 22 lLIxc6 ι-ο

4 Pelikan Variation This line arises after 1 e4 c5 2 lbf3 lbc6 3 d4 cxd4 4 lbxd4 lbf6 5 lbc3 e5. Black is willing Ιο accept a backward d-pawn ίη return for active piece play and, ίη some variations, the two bishops. The hίstοήcal background to this line is rather obscure since many players have adopted ίι over the years with different ideas ίη mind. The names ofLasker and Pelikan are associated with ίι, but the modern handling probably owes most Ιο the Russian grandmaster Sveshnikov. We have given Pelikan's name Ιο the whole system with 5 ... e5, reserving that of Sveshnikov for the 8... b5 variation, today considered the main line. It has gained many other adherents ίη recent years and is regarded as an excellent way ιο play for a win with Black, since unbalanced positions arise ίη almost every line. There is a second move order by which the Pelikan can arise, namely 1 e4 c5 2 lbf3 lbc6 3 d4 cxd4 4 lbxd4 lbf6 5 lbc3 e6 (οτ 2 ... e6 and 5... lbc6) 6lbdb5 d67 j.f4 e5 8 j.g5, reaching the same position as after 1 e4 c5 2lbf3 lbc6 3 d4 cxd4 4lbxd4 lbf6 5 lbc3 e5 6lbdb5 d6 7 j.g5, but ίη one extra move. Το avoid the confusion of having two different move numbers ίη each position we will take the 5 ... e5 order as standard,

even though ίη practice the two move orders are equally common. Since Black is incuning strategic weaknesses White's most logical (and best) lines are those ίη which he limits his immediate ambitions Ιο nullifying Black's piece play and only later turns his mind Ιο the exploitation of his long-term advantage. Our recommendation for White runs 6lbdb5 d6 7 j.g5 a6 8 lba3 b5 (the less common line 8 ... j.e6 is considered ίη Game 14) 9 j.xf6 gxf6 10 lbd5 when Β lack chooses between 10... f5 and 1O... j.g7. Both lines lead Ιο unbalanced position where Black pits his two bishops and central pawn majority against White's control of d5 and superior pawn structure. The offside knight οη a3 can be an important factor, and White usual1y aims Ιο bring knight back ίηΙο the game by playing c3 οτ c4. ΑΙΙ the lines of the Sveshnikov (8 ... b5) are covered ίη Game 15. Game 14 Karpov - Νυππ London 1982

1 e4 2 lbf3

3 d4 4 lbxd4 5 lbc3

cxd4 lbr6 e5

lbc6

Pelikαn Vαriαtion

6 ~b5(D)

6 •.. d6 BIack's 6th move aItematives give White a cIear pIus: 1) 6...a67 tί)d6+ j"xd6 8 'i'xd6 'i'e7 9 'i'xe7+ tί)xe7 (9 ... Φχe7 10 j"g5 tί)b4 11 0-0-0 is similar) 10 j"g5 and White has undisputed controI of d5. 2) 6... j"c5 7 .te3!? (7 tί)d6+ is a1so good) 7 ... j"xe3 8 tbd6+ Φf8 9 fxe3 'ii'b6 1Ο tί)c4 'i'c5 11 'i'd6+ 'i'xd6 12 tΔxd6 and f7 is about Ιο come under heaνy attack by j"c4. 3) 6....tb4 7 a3 j"xc3+ 8 tΔxc3 d6 9 j"g5 h6 (9 ... a6 10 tbd5 is also very pleasant for White) 10 j"xf6 'i'xf6 11 tΔb5 and Black willlose his d-pawn for insufficient compensaΑΙΙ

ιίοη.

4) 6...h6 (by preventing .tg5 BIack avoids the loss of control of d5 as ίη Iίne 1. but the move is really just ιοο slow) 7 tΔd6+ j"xd6 8 'i'xd6 'i'e7 9 tΔb5 (Spassky introduced this pawn sacrifice - the older lines 9 'i'xe7+ Φχe7 10 b3 and 10 j"e3 also give White a favourable ending) 9 ... 'it'xd6 (9 ... 0-0 10 'i'xe7 tΔxe7 11

tΔd60r 10 b3 tΔxe4 11 'i'xe7 tΔxe7 12 j"a3) 10 tΔxd6+ Φe7 11 tΔf5+ Φf8 12 b3 d5 (12 ... tΔxe4 13 j"a3+ Φg8 and now 14 f3 or 14 tΔd6) 13 .ta3+ Φg8 14 exd5 tί)xd5 15 tbd6

1:b8 16 j"c4 j"e6 17 0-0-0 and White has a very pleasant position. Spassky-Gheorghiu. Bath 1973. 7 .tg5 a6 Black must meet the threat of tΔd5 so the οηlΥ other move is 7....te6. but then White does ηο! have ιο retreat his b5-knight to the bad square a3 and can gain the advantage by 8 tί)d5 :c8 (8 ... j"xd5 9 exd5 tΔe7 is good for White after 10 'i'f3 or 10 c3 a6 11 'i'a4) 9 c3 a6 10 tΔa3 j"xd5 11 j"xf6 gxf6 12 'ίi'xd5 'ίi'a5 13 .tc4. Jansa-Danek, Czechoslovak Ch 1982. with a gήΡ οη d5. j"e6 8 tΔa3 8...b5 is the Iίne popularised by Sveshnikov and is examined ίη Game 15. Other moves are definitely ίηfeήοr:

1) 8....te7 (Black commits the bishop too soon) 9 tΔc4 j"e6 (after 9 ... tΔd4 10 j"xf6 j"xf6 11 tΔd5 b5 12 tΔcb6 :b8 13 tΔxc8 :xc8 14 c3 tΔe6 15 a4! White has a clear advantage. Averbakh-Korchnoi. semi-fina1 USSR Ch 1950. 9 ... 0-0 10 j"xf6 j"xf6 11 'i'xd6 doesn't give BIack enough for the pawn and 9... b5 10 j"xf6 gxf6 11 tΔe3 gives White a crushing bind) 1Ο j"xf6 gxf6 11 tΔe3 (thanks ιο Black's ... .te7 he cannot now dislodge the knight by ... j"h6) 11 ...'ίi'd7 12 tΔcd5 followed by j"d3 and 'it'h5. once again with a total Iίght-squared bind.

Pelikαn Vαriαtion

2) 8••.d5 (not cοπect) 9 lί)xd5 ~xa3 1Ο bxa3 '-a5+ 11 'ii'd2 'ii'xd2+

12 ~xd2 lί)xd5 13 exd5 lί)d4 14 ~d3 followed by ο-ο and f4, when White has two bishops ίη an open position and a moderately relevant extrapawn. 9 lί)c4

:c8 (D)

9••..i.e7 transposes Ιο line '1' of the last note, while the altemative 9•••tί)d4 (9 ... b5 10 ~xf6 '-xf6 11 lί)e3 and 12lί)cd5 is very good for White) often leads to the knight being driven back with loss of time, e.g. 10 ~xf6 gxf6 (10 .....xf6 11 lί)b6 :b8 12lί)cd5 'ii'd8 13 c31eaves Black a tempo down οη KarpovΝυηη) 11 lί)e3 :c8 12 ~d3 ~h6 (12 ... h5 13 ο-ο h4 14lί)cd5 ~g7 15 c3 lί)c6 16 'ii'f3 is excellent for White, Bronstein-Pilnik, Moscow 1956) 13 0-0 0-0 14lί)cd5 and again Black's knight is expelled by c3, Dely-Flesch, Hungarian Ch 1965.

10 ~XΙ6 '-ΧΙ6 With this move Black accepts a loss of time to avoid damage to his pawn structure but as a result he is driven into a passive position. The

important a1ternative 10.••gxf6 is met by 11 ~d3 lί)e7 (11 ... lί)d4 12 lί)e3 transρoses to the last note while 11 ...:g8?! 12 0-0 ~h6 1300 f5 14 'ii'h5! ~f8 15 lί)cb6 f4 16 lί)xc8 ~g417 'ii'xh7 :g7 18 'ii'xg7! ~xg7 19 lί)cb6 gave White too much for the queen ίη Matulovic-Arnason, Zeman 1983)12lί)e3~b6(12 ...'ii'b6 13 Ο-Ο! 'ii'xb2 14 lί)cd5 ~xd5 15 lί)xd5 lί)xd5 16 exd5 'ii'd4 17 "f3 Φe7 18 a4! :c7 19 :fdl "c3 20 :abl 'ii'a5 21 'ii'e4 gave White more than enough for the pawn ίη MednisLombardy, USA Ch 1978 - this type of pawn sacrifice ίη return for Iightsquared pressure and attacking chances occurs frequently ίη the 1O ... gxf6Iine) 13 ο-ο ~xe314 fxe3 'δ'Μ and now: 1) 15 '-f3 h5 (15 .....xb2? 16 00 ~xd5 17 exd5 is bad for Black) 16lZXι5 ~xd5 17 exd5 :h6 18 :abl _a5!? (18 ...:c7 19 c4 was good for Whίte ίη Tseshkovsky-Chandler, Minsk 1982) 19 e4 f5 (19 .....xa2? 20 'ii'e3 :h8 21 'ii'b6 wins) 20 exf5 (20 a3 f4 is not as clear) 20...lί)xd5 21 ΦhΙ 'ii'c5 22 :bel ;t Am.Rodriguez-Espinoza, Cali 1990. White has the more effective minor piece and Black's rook is misplaced οη h6. 2) 15 "cl!? was preferred ίη Anand-Morovic, Las Palmas 1993, not with the idea of defending the bpawn but to force Black to defend his f-pawn; after 15 ... lί)g8 16 ΦhΙ 'ii'c5 17 'ii'd2 h5 18 :adl h4 19 h3 Φf8 20 'ii'e2 :h6 21lZXι5 White had the better game. 11 lί)M

98 Pelikan Variation Ιι

is very dangerous for White Ιο take the pawn by 11 lDxd6+ .txd6 12 'ii'xd6 when 12 ... :'d8! 13 'ii'c5 lDd4 14 .td3 "'g5 15 ~f1 :'c8 16 'ii'b4 b5 gives Black adequate compensatίon.

11 ..•

:'b8 12 lDcd5 -'d8 12..:ii'g6 13 'ii'd3 .te7 14lDc7+ ~d8 15 lDcd5 folIowed by 0-0-0 gives White a dangerous attack. 13 c3 .te7 Attempting ΙΟ develop the bishop more actively runs ίηΙο trouble after 13...g614 'ii'a4! .txd5 15lDxd5 .tg7 16 .txa6 ':a8 17 'ii'b5! (even more effective than 17 .txb7, as given ίη BTS2) 17 ... .:xa6 18 'ii'xb7 with a winning positίon for White. 14 .tc4 Karpov conectly steers clear of the complίcations resulting from the win of a pawn by 14 lDc4 ο-ο 15 lDxe7+ 'ii'xe7 16 'ii'xd6 'ii'h4, and quietly consolidates his gήΡ οη d5. White's knights are rather clumsily placed but Black's possibilities for actίve play are very lίmited and aiming for ... f5 is his οηlΥ constructive plan. ο-ο 14 15 ο-ο .tg5 16 a4 ~h8 Chekhov suggests the imaginative 16..:ii'e8 intending 17 ....td8, but after 17 a5 (with the queen οη e8 Black can ηο longer meet this move by ... lDxa5) 17 ....td8 18 b4 ~h8 19 :'a2 f5 20 exf5 .txf5 21 lDe3 White has the better chances. 17 -'e2(D)

This move and the next are evidence ofKarpov's understanding of the position. White would lίke Ιο play 17 a5 defending the b6-knight and thereby freeing the tangle of minor pieces, but at the moment it just allows 17 ... lDxa5! 18 ':xa5 .txd5.1t seems natural, therefore, ιο prepare a5 wίth 17 b4 but ίη Sznapik-Simic, Smederevo 1981 Black obtained active play by 17 ... f5 18 b5lDa5 19 .td3 g6 and equalised cornfortably.

Karpov's first concern is ιο take the sting ουΙ of ... f5 by prepaήng Ιο answer ίι with exf5 and f4. For this purpose 'ii'e2, which pins the e-pawn against the loose bishop οη e6, and his next move ~hl, removing the king from the vulnerable diagonal, are excellent preparatίon. Only when Black's counterplay is completely neutralised does White retum to the exploitation οί his queenside space advantage and d5 control. 17 ... g6 Ι made use οί Karpov's chess lesson five years later ίη the game Nunn-Manor, London Lloyds Bank 1987, which continued 17...aS!? 18

Pelikan :adl j,h6 (after 18 ... g6 19 j,a2! Black should avoid 19.. .tΔe7 20 'ii'b5 and 19 ... j,h6 20 lbc4 f5 21 exf5 gxf5 22lbdb6 with a clear plus for White, but even his best line 19...f5 20 exf5 gxf5 21lbc4 f4 22 f3 is slightly better for White) 19 ΦhΙ lbe7?! (19 ... g6 is more solid, as ίη Karpov-Nunn) 20 lbxe7 "xb6 (the alternative 20 .....xe7 21 j,xe6 fxe6 22 'ifb5 is very bad for Black) 21 lbf5 and Black had ηο compensation for his serious weaknesses. White went οη Ιο win. j,h6 18 ΦhΙ 19 b4?! Afterwards Karpov thought that this was still Ιοο soon and that 19 Aadl or 19 :ael would have been better. 19 Ι5 20 exf5 gxf5 21 Ι4 j,xd5 Black hopes for sa1vation ίη the drawing tendencies of opposite-coloured bishops, but White's bishop has a fine outpost at d5 whereas Black's is rather useless. 22 lbxd5 e4?! 22•••lbe7 at once is more logical, based οη the fact that White cannot win a pawn by 23 fxe5 lbxd5 24 j,xd5 dxe5 25 "xe5+ because of 25 ... j,g7 26 "e6 :f6. 23 a5 j,g7 24 :acl lbe7 25 :fdl lbxd5 26 j,xd5 'ilc7 27 :c2 'ile7 28 'ile3 29 c4 :c7

:bc8

Variαtion

White has two ways of makίng further progress. He can either play b5 Ιο leave Black with an isolated pawn οη the queenside, which will be hard Ιο defend when his bishop is operating only οη the kingside, or he can prepare g4 Ιο attack Black's king. For the moment White is ηοι sure which plan offers the best chances. 30 g3 :e8 31 :g2 'ilf6 (D)

32 g4 This doesn't have the desired effect and ίι would probably have been better Ιο try the other plan. If Black moved his forces Ιο the queenside White could then have contemplated g41ater. 32 fxg4 33 :xg4 'ilc3 34 :g3 'ilxb4 If Black swaps queens Karpov gives the line 34•••'ilxe3 35 :xe3 :ce7 36 b5! j,h6 37 :n 38 :xe4 :xe4 39 j,xe4 :xf4 40 :xf4 j,xf4 41 R.xb7 and wins. 35 :dgl 'irb2 Stopping White's threat of'ifd4.

:f8

100 Pelikan

Vαriαtion

36 :gS 'ii'f6 'ii'81+ 37 :lg4 38 ~g2 'ii'b2+ :ce7 39 Φh3 'ii'f6? 40 fS This was the sealed move (move 41 ίη the garne, which started with the 2 ... e6 move order) and, as so often happens, after a 10ng period of difficult defence a player's relief at reaching the time control results ίη a casual sealed move. White obviously has considerable pressure for the pawn but after 40...:f8 (Κarpoν suggests 40 ... 'iνal threatening ... 'iνf1 +) ίι is likely that Black can draw. Black's passive queen move gives White the freedom of action he needs Ιο mount the decisive assault. 41 :bS :ιs 42 :gh4 h6 White cannot now play 43 :xJι6+ i.xh6 44 :xh6+ due Ιο 44 ... 'iνxh6+ 45 'iνxh6 :h7. 43 :g4 :eS

44 :gg5

:c8 (D)

45 ~g4! ~h7? Going under without a fight. Ι should have tried 45o..:xd5 46 cxd5 :c2 although White is winnίng even after this. 46 :g6 'ii'f8 47 'ii'g5 'ii'xf5+ Otherwise 48 :gxh6+ R.xh6 49 'ii'g6+ is the end. :ΧΙ5 48 'ii'xf5 49 J:xg7+ ~xg7 50 :xf5 1-0 Game 15 Short-Sax Sαint John Ct (1) 1988 1 2 3 4 5 6 7

e4 tLlf3 d4 tLlxd4 tLlc3 tLldb5 i.g5

c5 tLlc6 cxd4 tLlf6 eS d6 86 8 tLla3 b5 9 R.xf6 gxf6 9ooo'ii'xf6 10 tLld5 -'d8 11 c4 tLle7 (after 11 ... b4 12 'iνθ4 R.d7 13 tLlb5! axb5 14 'iνxa8 'iνxa815 tLlc7+ White wins mateήal) 12 cxb5 lDxd5 13 'ii'xd5 (13 exd5 is also an effective move) 13 ... R.e614 'ii'd2d5 15 bxa6 i.xa3 16 i.b5+! gives White a clear advantage. 10 tLIdS (D) 10 fS This is a a1ready an important moment for Black. Αι one time 1O... f5 was the only move played ίη this ρο sition, but recently the altemative 10oooR.g7 has become ρορυlαι. The 000

There isn 't much Black can do Ιο meet the threat of ~g4 followed by :g6.

Pelikαn Vαriαtion

101

14...0·015 lΔc2 (D) we arrive at the basic position of this line:

idea behind this move is Ιο delay Black's thematic break ... f5, giving first ρήοήΙΥ Ιο removing the powerful knight from d5 by ... lΔe7. Αι the time of BTS2 the theory of this line was still very much ίη its infancy, which made ίι difficult Ιο recommend one line ahead of another. This time we have taken the plunge and are suggesting the solid continuation 11 .td3 followed by c3. Before investigating this line ίη detail, it is worth mentioning the possibility 10•••.te6 11 c3 .tg7 12 lΔc2 and now 12...f5 13 exf5 .txf5 14lΔe3leaves Black a tempo down over a normalline, while 12....txd5 13 exd5!? (13 'ifxd5lΔe7 14 'ifd2 f5 15 exf5 d5 16 'ifg5 ο-ο 17 f6lΔg6) 13 ... lΔe7 14 a4! ο-ο 15 axb5 axb5 16 :xa8 'ifxa8 17lΔb4 'ifb7 18 g3! was good for White ίη Gallagher-Korpics, Oberwart 1993. After 10....tg7 11 .td3 lΔe7 12 lΔxe7 "xe7 13 c3 Ι5 14 ο·ο!? (14 lΔc2 is more common, but the move order we are proposing rules ουΙ a number of extra options for Black and reduces considerably the quanιίΙΥ of theory one has Ιο know)

1) 15...fxe4?! 16 .txe4 :b8 17 lΔb4! with a clear plus for White.

2) 15...:b8 16 exf5 e4 17 :el .txf5 18 lΔd4 .txd4 (better than 18 ....tg6, which turned ουΙ very badly for Black ίη Κramnik-Nunn, Monaco 1994 after 19 .txe4! .txe4 20 f3 d5 21 fxe4 dxe4 22 'ifg4! 'ifc5 23 ΦhΙ when 23 ... f5 is impossible because of 24 'ifxg7+!) 19 cxd4 d5 20 'ifd2 :b6, Magem-San Segundo, Madrid 1994, and now Magem gives 21 .tf1 'ifh4 22 :e3 :h6 23 :g3+ Φh8 25 h3 as slightly better for White. 3) 15...d5 16 exf5 e4 17 .te2 :d8 18lΔd4 .txd4 and now ίη Psakhis-Dolmatov, Κlaipeda 1988 White, for some reason, recaptured with the pawn and agreed a draw a couple of moves later. 19 'ifxd4 looks more natural, and after 19 ... .txf5 20 a4 White has a distinct positionaI advantage. 4) 15....tb7 16 exf5 (D) and now: 4a) 16...e4 17 :el d5 18 .tf1 'ifg5 19 lΔd4 .txd4 20 "xd4 'ifxf5

102 Pelikan Variation

21 a4 is even better for White than the simίlar position ίη line 3 just above as Black's bishop is more passively placed. The game de la VilIaAnglada, Ibercaja 1993 finished rather abruptly: 21 ...-tc6 22 axb5 J.xb5 23 c4! (opening the third rank) 23 ...-tc6? 24 :'a3! f6 25 cxd5 and Blackresigned as 25 ...'ir'xd5loses Ιο 26 J.c4 and 25 ...-txd5 Ιο 26:a5 fol10wed by J.c4. 4b) 16.. :iνg5 17 tDe3 (17 f3 deserves consideration) with a further branch: 4bl) 17...dS 18 f4! (destroying the black centre) 18 ... 'ir'h6 (18 ... exf4 19 'ir'g4! 'ir'xg4 20 tDxg4 f6 {20... h5 21 f6} 21 :'xf4 with a winning ending for White, Topalov-Spasov, Βυ dapest 1993) 19 f6 'ir'xf6 20 fxe5 'ir'b6 21 :'el :'fe8 22 ΦhΙ :'xe5 23 tDf5 with advanιage Ιο White, Mithrakanth-Prasad, Indian Ch 1992. 4b2) 17..•b5!? (Black invests a tempo Ιο secure control over g4) 18 J.e2 d5 19 J.f3 :'ad8 20 'ir'b3 (obviously ηοΙ 20 J.xd5 :'xd5! 21 tDxd5 :'d8) 20 ... e4 21 J.e2 d4 22 cxd4 :'xd4 23 :'ad 1 :'fd8 24 :'xd4 :'xd4 25 'ir'c2! and Black has insufficient

compensation for the pawn, Topa10ν- Vaiser, Mesa 1992. 11 J.d3 J.e6 11......g5 (11 ... f4 12 g3! is also good for White) 12 g4! Φd8 13 gxf5 J.xf5 14 tDe3 J.e6 15 'ir'd2 gives White a positional advanιage. 12 ο-ο J.g7 Or: 1) 12..•f4?! 13 c4! :'g8 (13 ...bxc4 14 J.c2! J.g7 15 J.a4 :'c8 16 tDxc4 and 13 ... b4 14 'ir'a4! J.d7 15 tDb5 are very good for White) 14 f3 b4 and now Timman-I.Sokolov, Amsterdam 1994 was agreed drawn after 15 tDc2?! a5 16 b3 J.e7 17 :'f2, but White should have played 15 "'a4 :'c8! 16 tDc2 (thanks Ιο the inclusion off3 and ...:'g8, 16 tDxb4? 10ses a piece) 16 ... a5 17 a3 with some advantage. He wilI always be able Ιο defend his kingside with :'f2 and J.f1. 2) 12..•J.xd5 13 exd5 tDe714 c3. BTS2 concentrated οη 14 tDxb5, which wins a pawn but allows Black some positional compensation. This time we are opting for the quieter continuation. White plans Ιο concentrate his forces against Black's weakest ροίηι, f5, by playing 'ir'h5 folIowed by lί:)c2-e3 and if Black plays ... e4 White wi1l be 100king ιο break υρ Black's centre with f3. 14••• J.g715 'iiΊι5 (D) and now: 2a) 15......d716 :adl :c8 (hoping Ιο generate some counterplay against d5 by playing ... :'c5) 17 lί:)c2 :'c5 (17 ... e4 18 J.e2 followed by f3 is favourable Ιο White; 17 ...0-0 18 g3! e4 19 J.e2 :'c5 20 lί:)e3 f4!?

Pelikan Variation 103

21 gxf4 f5 22 f3 "a7! was unclear ίη Renet-T.Horvath, European Club Ch 1993, as 23 fxe4 is met by 23 ... :c4! but perhaps White should have played 22 ~hl as 22 ... lbg6 23 Ίi'g5 looks ίη his favour) 18 lbe3 e4 19 j.bl ο-ο (19 ... f4 20 lbg4 Ίi'f5 21 Ίi'xf5lbxf5 22 f3! is good for White) 20 g3! (ηοι 20 f3 f4 21 .txe4 f5) 20 ... b4?! (Black should have prepared this with 20...:b8) 21 cxb4 :b5 22 a3 .txb2 23 ΦhΙ! (23 f3 is still too early because of 23 ...Ίi'a7) 23 ...:c8 24 f3 :c3 25 fxe4 :xa3 (not 25 ... :xe3? 26 "g5+) 26lbxf5 lbxf5 27 e5!! lbxg3+ 28 hxg3 .txe5 29 Ίi'xh7+ ~f8 30 :cl! :b7 31.tf5 Ίi'b5 32 .te6 1-0 Dolmatov-Chekhov, Germany 1992. 2b) 15...e416 .tc2 and now: 2bl) 16..."a5?! 17 :ael :a7 (or 17 ... .txc3 18 bxc3 Ίi'xa3 19 .txe4! fxe4 20 :xe4 :a7 21 :ael Ίi'xc3 22 Ίi'g5 Ίi'c7 23 h4!, followed by h5h6, with a strong attack according to Κramnik) 18 ~hl (18 .tb3 is also good) 18 ....txc3 19 bxc3 Ίi'xa3 20 .tb3 Ίi'b2 21 f3 with excellent prospects for White, Zso.Polgar-Κram nik, Guarapuava 1991.

2b2) 16... 0-0 17 :ael :c8 18 .tb3! (a strong move, neutralising ... b4, clearing c2 for the knight and protecting the d-pawn) 18 ...:c5 19 lbc2 Ίi'd7 20 f3! with advantage Ιο White, Kovalev-Palac, Neu Isenburg 1992. 2b3) 16... Ίi'c8!? 17 :ael ο-ο 18 <ithl lbg6! (ίη a previous outing Κrasenkov had got ίηΙο trouble after 18 ... b4?! 19 cxb4 .txb2 20 :e3 f6 21 g4! .txa3 22 gxf5 :f7 23 :gl+ :g7 24 :xg7+ ~xg7, TseshkovskyΚrasenkov, Voskresensk 1992, and now 25 :g3+ <ith8 26 Ίi'f7 Ίi'f8 27 Ίi'xf8+ :xf8 28 :xa3 lbxd5 29 .txe4 lbxb4 30 :a4 would have given White a very good ending) 19 .tbl (19 f3 is met by 19 ... b4 but .tb3 deserves attention, either here or οη the previous move) 19 ...:e8 20 f4 exf3 21 Ίi'xf3 :xel 22 :xel f4, Almasi-Κrasenkov, MalmQ 1994, and now 23 .txg6 hxg6 24 Ίi'xf4 looks a little better for White, as long as he spends the next few moves bringing his knight back ίηΙο the game. 13 Ίi'h5 Ι4 Νοι 13... 0-0? 14 exf5 .txd5 15 f6 e4 16 fxg7 :e8 17 Ίi'xd5, nor 13...h6?! 14 f4! opening the position while Black's king is still ίη the centre. 14 c4 bxc4 After 14...b4 (14 ... .txd5 15 exd5 lbe7 16 :adl b4 17 lbbl lbg6 18 g3 with advantage Ιο White, VogtGeorgadze, Halle 1978) 15lbc2 :b8 White has a variety of promising ideas, since with d5 secured he can

104

Pelikαn Vαriαtion

play οη the queenside with b3 and a3, or οη the kingside with g3. Ρί nally .!l)el-f3-h4-f5 can be awkward. The speculative 14•••0-0 15 cxb5 lΔd4 16 lΔc2 lΔxb5 (16 ... lΔxc2 17 .txc2 axb5 18 .tb3 is a safe positional plus for White) is dubious after 17 a4 (17 lΔcb4lΔd4 18 1Icl is also proιnising) 17 ...lΔa7 18 a51Δc6 19 b4 f5 20 lΔb6 1Ia7 21 exf5 .tf7 22 1Wh3 "f6 23 .te4 .!l)e7 24 1Iad 1 and White keeps control of d5. 15 .txc4 ο-ο 16 1Iac1 (D)

2) 16.. .1Ia7!? with the alternatives: 2a) 17 lΔxΙ4!? exf4 (17 ....txc4? 18 1Ixc4lΔa5 19IΣa4 exf4 20 IΣxa5 .txb2 211Δc4 is good for White) 18 .txe6lΔe7! (ηοι 18 ... fxe6 19 1Ixc6) 19 .tc4! (after 19 .td5 or 19 .tf5 Black can take the bishop and then play 20 ... .txb2, when the sacήfice 21IΣbl .txa3 22 1Ib3 .tc5 23 IΣh3 doesn't work because Black can advance his f-pawn and defend the second rank) 19 ... lΔg6 (ηοι now 19 ... .txb2? 20 1Ibl.tg7 21.J:[b3 ίη tending Ah3) 20 1Ic2 1Ie7 21 .td5 1Ie5 221Wdl 231Δc4 1Ih5 24 h3 1Ig5 25 ~hl .te5 and Black has enough for the pawn, Dvoirys-Ba§agit, Sibenik 1988. 2b) 17 1Ifd11Wb8!? (17 ... ~h8? 18 lΔxf4 exf4 19 .txe6 really does work because f7 hangs after ... lΔe7) 18 b3 (18lΔxf4 exf4 19.txe6lΔe7 is still unclear) 18 ...lΔd4 19 1Δc2 lΔb5 and now 20 lΔcb4! is good for White. 3) 16••.~h8 171ΣfdllΔd4 18lΔc2 is slightly better for White. 4) 16•••1Ib8 (the main line) 17 b3 (D) and now:

"h4

16 ••• l'Δe7 The position after 16 .J:[ac 1 is the key ΙΟ the whole line and one of the most important ίη the Sveshnikov Variation. Black has a wide range of options: 1) 16•••lbd4?! 171Δc2! lΔxc2(the line 17 ... f3 18 lΔxd4 fxg2 19 lΔf5! gxf11W+ 20 ΦΧf1 .txf5 21 exf5 h6 22 IΣc3! "g5 23 1Wxg5 hxg5 24 lΔe7+ ~h7 25 1Ih3+ .th6 26 f6 e4 27 .td5 wins for White) 18 1Ixc2 'iIth8 19 1Ifcl 1Ia7 20 b4 1Ig8 21 a4 with an excellent position for White, Mokry-Vodichka, r>etin 1979.

Pelikαn Vαriation

4a) 17...~h8 18 ~xf4!? exf4 19 i.xe6 ~d4 20 i.f5! ~xf5 21 'ifxf5 i.b2 22 tbc4 i.xc123 :xcl is a very promising exchange sacήfice. 4b) 17...i.xdS 18 i.xd5 ~M 19 1:tfdl ~xa2 20 11c6, followed by tbc4, and Black is very badly placed. 4c) 17.. :ii'a518~b1!? (18 ~c2 'ii'xa2 should lead Ιο a draw after 19 'ii'h4 h6! 20 ~f6+ i.xf6 21 'ifxf6 i.xc4 22 "xh6 i.xf1 23 'it'g5+) l8... ~h8 (ηοΙ 18 ...'ii'xa2? 19 ~bc3 'ifa5 20 11al 'it'c5 21 11xa6) 19 ~bc3 (D) and now:

105

4c2) 19...~d4 20 11fdl (20 ~xf4 i.xc4 21 bxc4 'ifc5 is much less clear) 20 ... f5?! (20 ... f3!?) 21 ~e7! i.xc4 22 bxc4 :f6 23 ~xf5 ~xf5 24 exf5 :g8 (24 ...11bf8 25 ~e4 11xf5 26 ~g5! h6 27 11xd6! is also winning for White) 25 ~e4 11h6 26 ~xd6! 1-0 Varavin-Chemiaev, Moscow 1992. 4d) 17.....d718 11fdl Wh8 (ηοι 18 ... i.g4? 19 'ifg5, while 18 ... ~4 19 tbc2 ~xc2 20 11xc2 wh8 21 :d3 i.g4 22 'it'h4 f5 23 f3 fxe4 24 fxe4 is a lίttle better for White) 19 'ii'h4 (D) and now:

Β Β

4cl) 19...f5 20 ~e7 ~xe7 21 i.xe6 11f6! (ηοι 21. .. fxe4? 22 ~xe4 d5 23 :c5 'ifb4 24 ~g5 h6 25 11c7 breaking through) 22 exf5 e4! 23 'ife2! (23 ~xe4? 1:txe6 24 ~g5 'ii'xf5 and Black wins) 23 ... d5, EstevezTimoshchenko, Managua 1988, and now 24 'it'd2! f3 (24 ... ~xf5 25 "xd5 'ii'xd5 26 i.xd5 e3 27 ~e4 is very good for White) reaches the cήtίcal position. Black has some compensaιίοη for the pawn after 25 ~xe4 'ifxd2 26 ~xd2 fxg2 27 ~xg2 ~xf5, but 25 g4! appears quite unpleasant for Black (25 ...i.h6 26 'it'd4).

4dl) 19...Ι5!? 20 ~xf4 exf4 21 i.xe6 'it'xe6 22 :xc6 fxe4 (ηοι 22 ...:bc8? 23 exf5 'ife2 24 J:tcxd6 11ce8 25 h3 'ifxa2 27 f6 1-0 Dolmatov-Figuero, SeviIle 1993) 23 :cxd6 'ife8 and now both 24 ~c4 e3 25 fxe3 (ηοι 25 11el i.c3) 25 ... fxe3 26 11el 'it'fΊ! 27 ~xe3 :be8 28 h3!, Turzo-Shaked, World U-16 1994 and BTS2's suggestion 24 "g4, with the idea 24 ... e3 25 "f3 (25 'it'e2!?) followed by ~c4, leaνe Black having ιο prove that he has sufficient compensation.

106 Pelikan

Vαriation

4d2) 19.. .JιxdS 20 i.xdS! (better than the old continuation 20 :txd5) 2OM.tbb4 (after 20 ...tDd4 21 :tc4! both 21 ... f3 22 tbc2 tbe2+ 23 Φη tbf424 gxf3 and 21 ... f5 22 tbc2 fxe4 23 tbxd4 exd4 24 i.xe4, Sveshnίkov Vyzhmanavin, Protvino 1987, are good for White) and now: 4d21) 21 :d2 f5 22 tbc4 (22 Wh3 shouId be met by 22 ... tbxd5 23 :txd5 Wb7 24 Wd3 fxe4 25 Wxe4 f3 rather than 22 ... We7 23 exf5 e4 24 tbc4 :tf6 25 :el e3 26 fxe3 %:.h6 27 Wg4 i.c3 28 "xf4 :tf6 29 :tedl i.xd2 30 :txd2 with an advantage for White, Klovans-Vyzhmanavin, USSR 1987) 22 ...tbxd5 (22 ...:tb5 23 :tcdl tbxd5 may be better) 23 ]hd5 fxe4 24 :txd6 "f5 25 h3! h5 26 "e7 e3 27 fxe3 fxe3 28 :td7!, RenetKorchnoi, Lugano 1988, and several sources, including BTS2, consider White to be better here. However, Ι (JG) am ηοΙ so sure; for example Renet gives 28 .....t2+ 29 ΦhΙ "g3 30 tbxe3 :tn 31 :tn as a clear advantage to White, but Black should continue 30•••:tbe8 31 "c5 :t2 32 :tf1 :txg2! 33 :txg2 "xh3+ with fair drawing chances. 4d22) 21 :tc3! (Renet's suggesιίοη ίη his notes Ιο the above game with Korchnoi) 21 ... tbxa2 (after 21 ... f5 22 %:.h3 i.f6 23 i.e6! both 23 ..."e7 24 'fj'b5 fxe4 25 tbc4 d5 26 tbd6 and 23 ... i.xh4 24 i.xd7 are good for White) 22 %:.h3 h6 23 tbc4 %:.bd8 24 g4! and White has dangerous attacking chances since if Black exchanges οη g3 the route is open for White's knight Ιο move Ιο f5.

17 Afdl There is an interesting altemative ίη 17 tDc7!? i.xc4 (17 .....xc7 18 i.xe6 "b7 19 i.b3! is good for White since 19... "xe4 loses Ιο 20 i.c2) 18 :txc4 %:.c8 19 %:.fc 1; perhaps 19..."d7 is the best reply. %:.c8 17 18 tΔxe7+ "xe7 (D)

Up to this ροίηι the game has followed Matanovic-Sax, Buenos Aires OL 1978, which continued 19 'i'e2 Φh8 20 i.xa6 f5! with dangerous counterplay. Black went οη Ιο win and later ίι was suggested that 20 %:.c2 followed by doubling rooks οη the c-file would have been good for White. Short prefers Ιο double rooks οη thed-file with his queen still actively posted οη h5. 19 Ac3! Black has ηο immediate threats, so White can afford Ιο take time ουΙ Ιο prepare b3 followed by %:.cd3. Black must aim for ... f5, since this provides the οηlΥ possible counterplay Ιο offset the backward d-pawn. The immediate 19... f5 fails to 20 exf5.

Pelikan 19 ••• ~h8 Short's innovation caused Sax Ιο use a lοι of time over this and his next few moves, so that before Iong White was an hour ahead οη the clock. 19...1:tc6? 20 ~xe6 1:txc3 21 ~f5 wins. 20 b3 (5 20 ...dS 21 exd5 ~xd5 22 1:txd5 'ifxa3 fails after 23 ~d3 e4 (23 ...h6 24 'iff5 and 23 ... f5 24 ~xf5 are even worse) 24 ~xe4 h6 25 1:txc8 1:txc8 26h3. BIack couId have tried 20...~d7, but then White can keep the advantage by 21 lbbl! (intending 1:tcd3 and lbc3) 21 ... f5 22 1:tcd3 fxe423 1:txd6 followed by lbc3. 21 1:th3 h6 22 ~xe6 "xe6 23 1:thd3 1:tcd8 AIthough this looks passive ίι is the best way ιο defend the d-pawn, for example 23....:c6 24 'ii'f3 or 23...fxe4 24 1:txd6 'ile7 25 lbc4 with a cIear pIus for Whίte ίη both cases. There is IίttIe point Ieaving a rook οη the c-fiIe when it will soon be blocked by lbc4. 24 "e2! It is easy to win a pawn but if this involves allowing BIack to advance his central pawns BIack's powerfuI bishop may well provide enough compensation. One such Iίne is 24 lbc4 (24 'ii'f3? 'ii'g6! 25 exf5 'ii'xf5 followed by ... e4 is aIso dubious) 24 ... fxe4 25 1:txd6 1:txd6 26 1:txd6 'ile7 27 1:txa6 e3 28 fxe3 fxe3 and the threats ιο White's king practically force him ιο pIay 29 1:txh6+ (29

Vαriαtion

107

lbxe3 "c5 30 'iWe2 e4 threatening ... ~d4 is awkward) 29 ... ~xh6 30 'ii'xh6+ ~g8 31 'ii'xe3 but 31 ...'ii'f6 32 h3 'ii'f1 + 33 ~h2 'ii'f4+ is a IikeIy draw because the knight and queenside pawns cannot assume a stabIe defensive configuration. White correctIy prefers ιο keep the bind, and cash ίη later. 24 ... fxe4 24.....g6 25 f3 d5 (25 ... fxe4 26 'ii'xe4 is good for White) 26 1:txd5 :xd5 27 exd5 e4 28 lbc4 e3 may seem uncIear, but White's passed pawn is just as far advanced as Black's and he has an extra pawn ίη hand. f3 25 "xe4 26 lbc4! (D) Much better than 26 1:txf3 d5 27 'iWh4 (27 1:txf8+ ~xf8 attacks a3) 27 ...1:tfe8 followed by the advance of the centra1 pawns, with good compensation for the pawn.

26 ... 1:tf4 26...fxg2 27 1:txd6 1:txd6 28 1:txd6 'ii'n 29 1:td2 and 30 lbe3 is a1so very good for White. "g4 27 1r'dS

108 Pelikan

Vαriαtion

Black cannot aνoid a miserable ending. Although he now succeeds ίη playing ... d5 his bishop remains blocked by the e5-pawn. 28 29'-xf3 30 gxf3 dS 31 ~Ω .tf6 Despite Whίte's doubled pawn the position should be a win since Black's d-pawn is easily blockaded by White's king and the bishop will be impeded by the central pawns. Moreoνer, White possesses a powerful queenside pawn majοήty. 32 lbb6 d4

:xf3

41 b7 42 J:ιbl

:xf3 '-xf3

33 ~e2

33 :cl! was more accurate, actiνating the rook before Black can play ....tg5. Then 34 :c8 is a threat, and after 33 ... ~g7 34 :c7+ ~g6 35 J:ιd7 White plays his rook behind Black's passed pawn and follows υρ with ~e2-d3. Howeνer the moνe played should also be sufficient Ιο win. 33 .tg5 34 lbc4 .tf4 35 h3 :g8 Black will have Ιο play this sooner or later, for otherwise White plays ~d3 and J:ιgl, seizing the open fιle for his own use. 36 b4 1tg2 The rook must moνe υρ the g-fιle Ιο allow the black king Ιο cross, and at g2 ίι delays White's 'i!;>d3. 37 84 <l;g7 38 b5 axb5 39 axb5 ~f6 40 b6 ~e6

The win stiII requires some work. White fιrst secures the adνanced pawn and moνes his king up. Black is paralysed and can only adopt a waίting strategy. 43 J:ιb5 ~d7 44 lba5 ~c7 Whίte's only WΟΠΥ is that Black might try Ιο liquidate Ιο a rook and pawn ending by ....tg5-d8xa5. This might prove awkward Ιο win as White would be left with only {- and h-pawns. The immediate 44....tg5 45 J:ιxe5 .td8 faίls Ιο 46 :d5+ ~e6 47 :xd8. 45 ~d3 ~d6 46 h4 Defιnitely ruling ουΙ ....tg5. 46 ~c7 47 J:ιb2 .th2 48 ~e4 White's plan is Ιο play :c2+ at a moment when Black must reply ... ~d7 ΙΟ preνent J:ιc8. Then White will seize the g-file by J:ιcl-g1. Ρί nally the ρenetration of the rook combined with the adνance of the

Pelikan white king Ιο c4 and b5 wiIl decide the game. The immediate 48 :tc2+ faίls Ιο 48 ... ~b6! 49 :tc8 ~xa5, so White must blockade the e-pawn before startίng his plan. 48 ••• ~f4(D)

Black can only delay the end. Thίs move covers c 1 ίη order Ιο prevent :tcl-gl. 49 :tc2+ Φd7

Vαriαtion

109

Or 49•.•Φb6 50 :c8 :txb7 51 llJxb7 ~xb7 52 :tf8 (threatening nxf4) 52 ...~h2 53 nf6 and wins. 50 Φd3! Black is ίη zugzwang. The rook cannot move because οί lbc6, the king cannot move because οί nc8 and ... h5 doesn't help. 50 ••• ~h2 51 :tcl ~Ι4 The same again! Black has Ιο free gl. 52 :tgl ~d6 53 ~c4 1-0 The finish mίght be 53•••:rs 54 lbb3 (ίί White gets the chance Ιο play llJb3-c5 the winning process is simplίfied) 54 ...:tb8 55 lbc5 :th8 56 :tg6+ ~c7 57 Φb5 d3 58 :tc6+ Φb8 59 Φb6 mating, or 53 .•• ~c7 54 :tg6 :th8 55 ~c5 d3 56 :tg7+ ~b8 57 ~b6 with a simίlar conclusion.

5 TheDragon The Dragon is characterised by the initial moves 1 e4 c5 2 ~f3 d6 3 d4 cxd4 4 ~xd4lΔf6 5 ~c3 g6, intending Ιο develop Black's bishop actively at g7. If White plays quietly and castles kingside the bishop will still be useful ίη supporting Black's minority attack οη the queenside, while if White aims Ιο attack the enemy king and plays 0-0-0 the Dragon bishop comes ίηΙο its own. Hosts of players have seen their queensides disintegrate under the laser-like power of the g7 bishop, often supported by ... 'Wa5 and ...:c8 Ιο step υρ the pressure οη c3. The true Dragon player will analyse six exchange sacrifices οη c3 before breakfast, and White needs ιο be constantly οη the alert for combinations based οη blowing open the long diagonal. The Dragon is one of the most controversiallines ίη the Sicilian. Αι various times over the years ίι has appeared Ιο be ίη its death throes, only Ιο be suddenly revived by the discovery of new ideas for Black. Practitioners regard the variation as their Ρήvate property, and defend ίι with an almost religious fervour against the many players who wish Ιο commit the heresy οί mating Black down the h-file. BTS2 proposed the main line of the Yugoslav

Attack, 9 .tc4 followed by attempting to mate Black down the h-file, but this time we are opting for 9 0-0-0 followed by attempting Ιο mate Black down the h-file. The advantage of this approach is clearest if Black insists οη getting involved ίη a mutual king-hunt, when the tempi White has saved οη .tc4-b3 are very valuable. The slίght catch is that Black can reply 9 ... d5, somewhat changing the character of the game. This pawn sacήfice has a very respectable reputation but thanks Ιο the injection of several interesting new ideas White has been coping quite well with ίι recently. This is examined ίη Game 17, whilst Game 16 deals with 9 ...~xM and all the other alternatives for Black. General principles aren't much help ίη the Dragon, since success or failure is determined mainly by tactical consideratίons. Game 16 - Ward London Uoyds Bαnk 1994 Ι. GureVΊch

Ι e4 2 ffi

3 d4 4 ~xd4 5 Μ 6 .te3

c5 d6 cxd4 ~Ι6

The Drαgon 111 If White intends castling queenside this moνe is the most usual. 6 .te2 only fιts ίη with ο-ο, since the bishop is usually better placed at f1 or c4 ίη the more aggressiνe lines resulting from castling οη opposite wings. 6 ••• .tg7 The Dragon differs from many other Sicilian systems ίη that Black often omits the typical moνe ... a6. Time is of particular importance ίη the Dragon and Black simply cannot afford the tempo spent οη preparing ... b5, which can often be played without ... a6 ίη case White castles queenside. The idea of playing ... a6 and ... b5 before castling has been tried, but after 6•••a6 (ηοι 6... tΔg4?? losing material after 7 .tb5+) 7 f3 lbbd7 81i'd2 b5 9 a4! bxa4 (9 ...b4 10 tΔd5 is alSO very good for White) 10 :xa4.tg7 11 .te2 ο-ο 12 ο-ο tΔc5 13 :a3 .tb7 14 :fal 1i'c8 15 tΔb3 White had strong queenside pressure ίη Kaνalek-Bilek, Sousse ΙΖ 1967. 7 f3 (D) This is more or less forced as 7 .tc4 and 7 1i'd2 can both be met by 7 ...lbg4.

7 .•. ο-ο Or: 1) 7•••a6 (7 ... tΔbd7 puts ηο pressure οη d4 so simply 81i'd2 followed by 0-0-0, .tc4 and .th6 giνes White a strong attack) 8 'i'd2 tΔbd7 (8 ...b5 9 a4! is similar Ιο Kaνalek-Bilek aboνe) and now: la) 9 .th6 (9 0-0-0 is also good) 9 ... .txh6 (9 ... 0-0 is suicidal since White's intended h4-h5 attack is much stronger than usual) 10 Wxh6 b5 11 0-0-0.tb7 12 ~bl, Bastrikoν Khasin, USSR 1961, and Black's king is stuck ίη the centre. lb) 9 g4 tΔb6 10 0-0-0 lbfd7 11 .th6.txh612 Wxh6e513 tΔb3 tΔf6 14 1i'd2 Φe7 15 ~e2 1i'c7 16 f4! exf4?! 17 g5 tΔe8? 181i'd4! 1-0 was the game Arakhamia-Chiburdanidze, Belgrade 1992. White has the double threat of 19 1i'xh8 and 19 'Wxb6. 2) 7••• tΔc6 (this can transpose ίηΙο the main line if Black castles quickly, so we only explore lines ίη which Black plays ....td7 and ...:c8 before ... 0-0) 8 'Wd2 .td7 9 0-0-0 :c8 10 g4 tΔe5 (10 ... 0-0 transposes Ιο line '2' ίη the note Ιο Black's 9th moνe) 11 h4 h5 (11. .. 0-0 is variation '2a4' ίη the note Ιο Black's 9th moνe) 12 g5 tΔh7 13 f4 tΔg4 and now P.Littlewood-Mestel, London 1978, continued 14 .tgl 0-0 15 ΦbΙ e5 16 tΔf3 .te6 17 fxe5 tΔxe5 18 tΔxe5.i.xe5 19 tΔb5 :e8 and Black had ηο problems. Howeνer, White can improνe by playing f5 at some ροίηΙ, for example 15 f5 or, more proνocatiνely, 14 f5!? since eνen

112 The Drαgon though 14 ... lίJxe3 15 'ii'xe3 0-0 exchanges off White's dark-squared bishop, ίι isn't easy Ιο see a constructive plan for Black after 16 ~h3.

ΊΜ2

lbc6

12 ~e2) 12... e6 (12 ... lbc4 13 j,xc4 dxc4 14 'ii'e2+) 13 h4 h5 14 ~c5 lΔe7 15 lbb5 a6 16lbbd4 and now White's position is preferable. 9 0-0-0 (D)

Or 8••• dS (other moves meet with common-sense replies, e.g. 8 ... ~e6 9lbxe6 fxe6 10 e5 or 8 ... a6 9 0-0-0 b5 10 h4) 9 eS (D) and now:

9 ...

lbxd4

9 ...dS is the subject of the next

1) 9...lbe8 10 f4 f6 11 0-0-0 fxe5 (11 ... lbc612lbf3) 12fxe5lbc6 (af-

ter 12 ... ~xe5 13 lΔf3 White either regains the pawn with advantage or, after 13 ... ~xc3 14 'ii'xc3 e6 15 h4, gets an enormous attack) 13 lΔf3 ~g4 (13 ... e6 14 ~h6 leaves Black with a very bad bishop) 14 lΔxd5 :xf3 15 gxf3 j,xf3 16 ~g2 ~xdl 17 ':xdl ~xe5 18 ~c5 e6 19lΔe7+ ~g7 20 lΔxc6 'iixd2+ 21 :xd2 bxc6 22 :d7+ with advantage Ιο White. 2) 9...lΔfd7 10 f4 lbc6 11 0-0-0 (11 lΔb3 e6 12 0-0-0 f6 13 lbxd5! fxe5 14 lbc3 'fIe7 15 ~bl exf4 16 ~xf4 gave White a clear positional plus ίη L.Milov-Afek, Konsumex 1992) 11 ... lbb6 12lΔf3 (this suggesιίοη of Afek's is perhaps even stronger than the usual 12 lΔb3 or

game whilst the other altematives are examined below: 1) 9...~e6 and now: la) 10 ~bl1eaves Black with ηο better continuation than 10...lbxd4 11 ~xd4 'iic7 transposing ιο the main line. If instead 10.. ':c8, then 11lbxe6 fxe6 12 ~c4 'iid7 13 ~b3 lΔe5 14 'ili'e2 a6 15lba4!, threatening lbb6, lbc5 and f4-f5, is excellent forWhite. lb) White can also try 10 lbxe6 fxe6 11 g3! (11 ~c4 'ili'c8! is not so clear) when Romero-Martin, Spanish Ch 1990 continued 11 .. :it'd7 12 ~h3 lΔe5 13 'iie2 b5 14 f4lbc4 15 e5lbe8 16 ~d4lbc7 17lbe4 d5 18 lbc5 'iWc6 19 ~g4 b4 20 h4 a5 21 h5 a4 22 hxg6 h6! 23 :'h2 1ta5 24 :dhl b3 25 a3 :xc5 26 c3! and with the queenside blocked, Black can only await his fate οη the other wing.

The 2) 9•••.td7 (D) is thought Ιο be a lίttle slow when White hasn't wasted

time οη .tc4. White now has:

:Ιχe4

113

lDxe4+ 28 c;i;>e3 .txg2 29 .txg7 Q;xg7 30 .txg21Dc5 was best for Black; the weird ending resulting appears Ιο be better for White) 27 .txg7 "e3+ 28 Φc2 .txf3 29 .tb2! e5 30 "d2 'ίWc5+ 31 ~bllbe3 32 ltcl 'ίWb6 33 "h2lDf5 34 ltc8+ ~g7 35];tb4 1-0. 2a2) 11... 12 ~bl (12 h5 allows 12... lΔb4, with ... 1:txc3 Ιο follow, but 12 lDb3 "c7 13 h5 is an interesting alternative) 12...ltfd8? (12 ... lDe5 transposes Ιο line '2a4' and 12 ... lΔxd4 13 .txd4 ':fd8 is assessed as slightly better for White by Htibner) 13lDb3 "c7 14 h5 .te6 15 hxg6 fxg6, Htibner-Hort, match 1979 and now 16lΔd5! would have given White a clear plus according ΙΟ Htibner. 2a3) 11•..h5 12 .te2 (ίι is by ηο means certain that this is White's best; Ligterink-Sosonko, Dutch Ch 1978 went 12 gxh5 lΔxh5 13 ':gl ~h7 14 .te2lΔxd4 15 .txd4 .th6 16 .te3 .tg7 17 .td4 .th6 1/2_1/2 but 14 ~bl avoiding the draw may be good for White) 12 ... lΔe5 13 gxh5 lΔxh5 14 ':dgl (14 .:thgl is an alternative, but whether better or worse is ηοΙ easy Ιο decide) 14 ... lΔc4 15 .txc4 (a defect of 12 .te2 is that this capture will involve a 10ss of a tempo) 15 ... ':xc416:g5 (threatening 17 ':xh5 gxh5 181:tgl) 16... ':c5 17 lΔd5 e6 18 ':xh5 exd5 19 ':xd5 ':xd5 20 exd5 and White was just a little better ίη Speelman-Liu Wenzhe, China 1981. 2a4) 11 ••• lΔe5 12 ΦbΙ (after 12 h5 "a5 13 ~bl, Seirawan and Jim

"a5

2a) 10 g4 ltc811 h4 (11 ~bl is interesting as after 11 ... lbe5 12 h4 we have transposed Ιο lίne '2a4' cutting ουΙ lines '2a', '2a2' and '2a3', while 12 .th6 is recommended by Miles and Moskow who give 12 .th6lDc4 13 .txc4 ltxc4 14lDde2 followed by h4 and h5 as good for White) and now Black has: 2al) 11••• b5 12 lDcxb5 lDe5 13 h5 lDxf3 (once Β lack has started sacrificing he must continue) 14lDxf3 .txg4 15 'ίWg2 (15 .te2 lΔxe4 16 'ίWel! would be a safe answer Ιο Black's aggression) 15 .....a5 16 a3 ltxc2+! 17 ~xc2 'ίWa4+ 18 ~d2 'ίWb3 19lΔc3 'ίWxb2+ 20 ~d3 'ίWxa3 with hοπendοus complications ίη the game Mestel-Christiansen, Hastings 1978/9. This remarkable game concluded 21 .tcl (21 ':cl was probably better) 21 ... 'ίWb4 22 .td2 ltc8 23 hxg6 hxg6 (23 ... h5!? reserving h6 for the bishop was possible) 24 lth4 ltxc3+ 25 .txc3 "xe4+ 26 ~d2 lΔd5? (the line 26 ... .txf3 27

Drαgon

114 The Dragon Gallagher {probably a long-Iost cousin Ιο judge from his chess style} give 13 ... ':'xc3! 14 'iνxc3 'iνxc3 15 bxc3 lDxf3! 16lDxf3 .txg4 17 .tg2 lDxe4 with a lοt of play for the rook) 12...'ίi'a5 (12 ... b5 13 lDcxb5lDxf3 is ηοΙ so effective with the king οη bl) 13 lDd5 'ίi'xd2 14 lDxe7+ Φh8 15 .txd2! ':'ce8 (D)

2a41) ThegameTimoshenko-Jim GaIlagher, J acksonviIle 1990 continued 16 .tb4?! (16 lDd5 lDxd5 17 exd5 lDxg4) 16 ... lDxf3! 17 lDxf3 lDxe4 18lDd5 .txg4 19 .tg2lDf2 20 .txd6 ':'g8 21 lDg5 h6 22 lDxf7+ Φh7 23 ':'df1lDxhI24 ':'xhl 1Ie2 25 lDf4 ':'xg2! 26 lDxg2 .tf3 27 ':'h2 (27 :gl .td4 28 lDel .txgl 29 lDxf3 .tf2 is ΟΚ for Black) 27 ... ':'e8 28 .tg3 ':'f8! 29 lDg5+ hxg5 30 hxg5+ Φg8 and the bishop pair compensated for the missing pawn. 2a42) However, Ι belίeve that 16 h5! would refute Black's interesting concept. If now 16•••:Xe7, then 17 h6! traps the bishop and after 16•••gxh5 White can simply recapture or play 17lDef5! as 17 ... hxg4 18 lDxg7 costs Black the exchange. Nor

can Black save himselfby means of 16•..lDxe4 17 fxe4 ':'xe7 (17 ....txg4 18 h6) as 18 h6 .tf6 19 g5 stίll traps the bishop. There remains the attempt Ιο block the h-pawn's further advance with 16•.. h6, but then 17 g5! looks crushing, e.g. 17 ...lDxh5 18 gxh6 ':'xe7 (18 ....tf6 19lDd5) 19 hxg7+ Φχg7 20.tb4!. 2b) 10 h4 ':'c8 11 h5 lDxh5 12 'ii'f2!? (12lDxc6 ':'xc6 13 .th6 .txc3 14 bxc3 'ii'a5 looks good for Black) 12 ... lDe5 13 g4lDf6 14lDd5 with a strong attack for White. Ιη ShakedLeonard, Philadelphia 1993 Black sacήficed unsoundly by 14... lDfxg4 and lost quickly after 15 fxg4lDxg4 16 'ii'h4 h5 17lDxe7+ Φh8 18 .tg5 .th6 19 .txh6lDxh6 20 "f6+ Φh7 21 lDf3 lDg4 22 lDg5+ Φh6 23 lDxf7+ 1-0. 3) 9•••lDe5 and now: 3a) 10 g4 e6 11 .te2 d5 12 g5 lDh5 13 f4 1Dc4 14 .txc4 dxc4 15 e5 ± Z.Almasi-Szalanczy, Kecskemet 1993. 3b) 10 .th6 .txh6 11 'ii'xh6 Φh8 12 'ii'd2 .td7 13 f4lDc6 14 i.e2 ;!; Luther-Szalanczy, Kecskemet 1993. 10 i.xd4 .te6 11 ΦbΙ! This prevents l1 ... 'ίi'a5, which would now be met by the standard trick 12lDd5 'iνxd2 13lDxe7+, and White wins a pawn. However, Black doesn't rea1ly have an a1ternative plan so he is forced fιrst to play ... flc7 and ... ':'fc8 and only then .....a5 (with f8 avaiIable for the king the trick doesn't work). 11 ••• 'ίi'c7 (D)

The

Drαgon

115

Mιujanovic-Mesing,

Bela Crkva 1984.17 j.xc4 and 18 j.xf6 was ηοΙ a very cήtίcal test, but for the moment sharper continuations such as 17 (4 or 17 g4 remain unexplored. Against either Black will start his counterattack with 17 ...b5. 12 :fc8 13 h4 14 .g5! First played ίη a quickplay game by Ivanchuk, this move virtually forces an exchange of queens οη g5, which is very favourable for White. Previously 14 a3 had been played and after 14 ...:ab8 15 h4 b5 a sharp struggle arose with chances for both sides. 14 ... 'iixg5 If Black plays 14..:ir'c7, then the simplest would be Ιο play 15 j.d3 quickly and awaίt with cuήοsίtΥ Black's next move. 15 hxg5 ~7 16 j.xg7 Φχg7(D)

.&5

12 g4

.&5

Or 12 h4 :fc8 13 h5 (after 13 ... lbxh5 14 j.xg7 Φχg7 15 g4 lbf616e5dχe51ΤW'h6+Φg818g5

lbh5 19 j.d3! White had a decisive attack ίη Evans-Zuckerman, New York 1987, whichconcluded 19 ...e4 20 lbxe4 'ίi'f4 21 :xh5! gxh5 22 lbf6+ exf6 23 j.xh7+ Φh8 24 j.f5+ 'iPg8 25 'ίi'h7 + Φf8 26 "h8+ 1-0) 14 hxg6 hxg6 15 a3 :ab8 16 j.d3 and now: 1) 16.•• b5?! 17 'ίi'g5! (a strong move which both slows down the ορ posing attack and brings the queen ίηΙο the vicinity of the black king) 17 ... j.c4 (17 .....c7 18 e5 dxe5 19 j.xe5 'ίi'c5 20 f4 :b7?! 21 j.xg6 fxg6 22 'ίi'xg6! j.f7 23 :h8+! 1-0 VanderWiel-Sax, Plovdiv 1983) 18 f4 e5 19 fxe5 b4 20 exf6! .xg5 21 fxg7 f6 22 j.xc4+ Φχg7 (22 ...:xc4 23 :h8+ Φχg7 24 :xb8 bxc3 25 :b7+ Φf8 26 j.xc3 is good for White as the black king is simply Ιοο exposed) 23 lbd5! ± CuipersBerendse, Holland 1984. 2) 16••• j.c4! 17 j.xc4 :xc4 18 j.xf6 j.xf6 19 ~5 "xd2 20 lbxf6+ Φg7 led Ιο equality ίη the game

17 :b4! the original game with 14 "g5!, Ivanchuk-Kir.Georgiev, Tilburg 1993, White played the less Ιη

116 The Drαgon accurate 17 1Le2. Gurevich's move, which protects g4 so that the f-pawn can advance and gives White the ορ ιίοη οί doublίng οη the h-file, is more powerful. 17 ••. Ι6 Black must grab some space before it's Ιοο late. 18 gxf6+ tΩxΙ6 19 g5 ~7 Or 19••• lί:Ίh5 20 ί4 -ZΣf8 21 f5lί:Ίg3 22 fxe6lί:Ίxf1 23 b3! -ZΣf2 24 e5! with a clear advantage for White according Ιο Gurevich. 20 Ι4 ];[cS 21 tΩb5! The knight heads for its ideal square, d4. lbf8 21 22 lDd4 i.d7 23 i.d3 (D)

Whίte has an extremely pleasant position. Αll his pieces are more active than their counterparts and he also has a SUΡeήοr pawn structure. Black now takes some drastίc actίon, reasoning that ίί White is going Ιο get ί5 ίη then he might as well have a big centre ίη return. Objectively

speaking though, Black's nextmove probably changes the assessment οί his positίon from bad Ιο 10st. 23 •.• e5 Ι can sympathise with someone who's ηοΙ willing Ιο defend the posiιίοη after 23••• lt:'ιe6 24 -ZΣdhl -ZΣh8 25 tΩxe6+ i.xe6 26 b3, when White can slowly build υρ his position υηΙίΙ he feels like strikίng, safe ίη the knowledge that Black has nothing Ιο do but wait.

24 tΩb3

-ZΣc7

25 Ι5 gxf5 26 exf5 d5 27 c3 a5 Οί course 27•••e4, which is met by 28 f6+, concedes control οί the central dark squares and facilitates the knight's entry ίηιο the game. a4 28 i.c2 29 lί:Ίcl :aS 30 tΩe2! Once the knight reaches the kingside Black will be ίη seήοus danger οί getting mated, notwithstanding the reduced mateήal. 30 i.e8 31 a3 Φg8 32 -ZΣgl :a6 33 tΩg3 -ZΣ81 34 Μ! Black has done the best he can over the last few moves but a knight οη g4 will be Ιοο much Ιο cope with. Ν ote how White is ίη ηο hurry Ιο push his kingside pawns. Οη f5 and g5 they severely restήct the black minor pieces. 34 b5 35 tΩe3 -ZΣd6

The Drαgon 117 36 o!tJg4 e4 36...:xg5 loses Ιο 37lΔh6+. 37 :th2?! The immediate 37 1:h6! would have been more Ιο the Ροίηι. 37 ... lί)d7 38 :h6! ]:tb6 38...:xh6 39 gxh6 1:e7 40 lΔf6+ Wf8 41 lΔxd5 1:e5 42 lΔe3 is also hopeless. 39 ':'xb6 lΔxb6 40 lί)f6+ Φr7 40...~ 41 :hl and Black can't take οη g5. 4Ι

lΔxh7

lί)d7

42 43 44 45 46

.i.dl g6 gxh7+

Wg8 ]:txh7

ΦcΙ Φd2

Φxh7 lΔΙ6 ι-ο

After 10 exd5 lΔxd5 11 lΔxc6 bxc6 12 lΔxd5 (12 .i.d4!) 12 ...cxd5 13 'ifxd5 'ifc7! Black has Ιοο much for apawn. White does have one interesting alternative though, 10 ΦbΙ!?, which has recently been introduced ίηΙο practice by L.Mίlov. 1) The far from obvious idea is Ιο meet 10...lΔxd4 with 11 e5! (D).

Game 17 Luther - Μ. Hoffmann Lippstαdt 1994 Ι

2 3 4 5 6 7 8 9

00 d4 lΔxd4 ω

cS d6 cxd4 lΔΙ6

g6 .i.g7

.i.e3 f3 ο-ο 1i'd2 lΔc6 0-0-0 d5 (D) ΙΟ 1i'eI Α logical move hoping Ιο take advantage of the fact that the black queen is οη the same line as the white rook. Although Black's last move is theoretically a pawn sacήfice, White should ηοΙ even consider taking ίι.

Εχpeήence is still very limited, but what we've seen so far suggests that White is doing well here. la) For example, L.Mίlov-Golu bev, Biel 1994 continued 11...lί)d7 12 .i.xd4lΔxe5 (12 ... e6 13 f4 f6 14 exf6.i.xf6 15 'ife3lΔb6 16 h4 was distinctly better for White ίη the

118 The Drαgon game Milov-CΊrkven~ic, Nagykanizsa Open 1993) 13 "e3! lbc6 14 ~χg7Φχg715lDΧd5"a5 (15 ... ~e6 16 "c3+ f6 17 ~c4 is good for White) 16 b4! 1i'a4 17 b5 1Σb8! 18 "b3 (18 Φc1 ι?) 18 .....xb3 19 axb3 e6 20 bxc6 exd5 and now instead of the inaccurate 21 c7?, which a1lowed Black Ιο equa1ise after 21 ...:ta8 22 :xd5 ~e6 23 :tc5 :tfe8! 24 ~d3 :ac8 25 1Σa5 a6, the straίghtforward 21 :txd5 bxc6 22 J:ta5 would have left Black with a difficult defensive task. Alternatives Ιο 11 ...lDd7 also look good for Whίte, e.g.: lb) 11...lDxc2? 12 exf6 +-. lc) 11•••lM5 12 exf6 exf6 13 ~c5±.

ld) Perhaps 11•••lDxf3 is most cήtίcal. After 12 gxf3 1Dd7 White

can play either 13 f4 followed by the advance of the h-pawn, ΟΓ 13lDxd5 as 13 ... lDxe5 a110ws 14lDf6+ and after 13 ... ~xe5 14 ~h6 White has a dangerous initiative. 2) Golubev assesses 10•••e6 as ;1;. Ιι is Ιοο soon for any definitive judgement of 10 ΦbΙ, especia1ly without any examples of 10... e6. 10 .. e6 Or: 1) 10...:te8?! 11 ~b5 ~d7 12 ~xc6! bxc6 (12 ... ~xc6 13 e5lDd7 14 e6 ±) 13 e5 c5 14 lDb3! c4!? (14 ...d4 15 lDxc5 dxc3 16 exf6 "b6 17 "xc3 "xf6 18lDd7 +-) 15lbc5 (151Dd4) 15 ... ~c6 16 exf6 ~xf6 17 'iff2 and Black didn't have enough for the piece, Jansa-W.Watson, Prague 1992. 0

2) 10...e5 (thίs has been ουΙ of favour recently but ίι seems Ιο cause White at least as many problems as the αιaίη lίne) 11 lDxc6 bxc6 12 exd5lDxd5 (12 ... cxd5 looks more natural, but after 13 ~g5 ~e6 14 ~c4! White has the advantage; the best that Black can do is 14.....c7 15 ~xf6 dxc4 16 J.xg7 Φχg7 17 "e3 :tab8 18 :thel f6 19 lDe4 with an edge for White, Kuporosov-A.Kova~evic, νrnja~ka Banja 1992) 13 J.c4 ~e6 14 lDe4 'ii'c7 (Black can also try 14 ...h6 with the ideaofmeeting 15 J.c5 by 15 ... f5; Dragon players norma1ly don't even hesitate for a moment before playing such exchange sacrifices, but ίη this case White seems Ιο be able Ιο gain the advantage after 16 J.xf8 'iWxf8 17 lDf2 'iWe7 18 1Dd3! when 18 ... lDb6 19 W'b4! is good for White, as are 18 .....f6 19 :td2 :te8 20 :e2 ~f7 21 'i'f2! and 18 ... J.f7 19 'i'a5 e4 20 :thel e3 21 'i'c5 'i'xc5 22lDxc5 f4 23 g3 g5 24 h4 ~f6 25 hxg5 hxg5 26 gxf4 gxf4 27 %lh 1 ±; analysis by Perez Cruz) 15 J.c5 1U'd8 (D) and ίη this position White has at least three reasonable tries:

The Dragon 119 2a) 16 'ii'h4 h6 17 g4 lL!f4! (better than 17 ...:ab8 18 g5 h5 19"f2 "b7 20 b3 with an edge for White, Κramnik-Rosselli, Maringa 1991; or 17 ...:d7 18 g5 h5 19 lL!f6+!? lL!xf6 20 ~xe6 :xdl+ 21 ':xdl lL!d5 22 ~xd5 cxd5 23 :xd5 "a5 24 a3 "b5 25 "e4! and Black had ηο compensation for the pawn ίη Mainka-Lindemann, Oberwart 1991) 18 ~d6? (18 ':xd8+ :xd8 19 ~xe6lL!xe6 20 ~e7 might be better for White as 20 ... 11d4 is met by 21 lL!f6+ Φh8 22 lL!e8, but Black should prefer 18 .....xd8 with a roughly balanced position) 18 ...:xd6! 19lL!xd6 ~d5! 20 ~xd5 cxd5 21lL!b5 "c5 22lL!c3 ':c8 with an excellent game for Black, Baron-Komljenovic, Ibercaja 1993. 2b) 16 g4lL!f4 17 "c3 :d5! 18 ΦbΙ ':ad8 19 ~xd5 cxd5 20 ~e3 (20 lL!g5 d4! 21 ~xd4 "d6! is good for Black) 20 .....e7 21 'ii'c5 'i'b7 22 "a3 "c6 23 'i'c3 "a8! 24 ~xf4 :c8! 25 "a5 exf4 and Black's strong bishops give him ample compensation for the exchange, Xie Jun-Gufeld, Kuala Lumpur 1994. 2c) 16 lL!g5 ~c8 17 g4 (D) (a suggestion ofDvoirys) with the possibilities: 2cl) After 17...h6, the sensible continuation is 18 tbe4 ~e6 19 'ii'f2, and the pressure against a7 makes ίι hard for Black Ιο allow an exchange οη d8. The crazy continuation is 18 h4 hxg5 (18 ... :b8 19 lL!e4 f5 20 lbc3!? ~e6 21 gxf5 gxf5 22 11g1 looks dangerous for Black) 19 hxg5, but after 19 ... ~b7 (19 ... lL!e7 20

:xd8 'i'xd8 21 'i'h4 "c7 22 'i'h7+ Φf8 23 ':h6!, with the threat of 24 ':xg6, is much more dangerous) 20 'i'h4lbe7 21 'ii'h7+ Φf8 White may ηοΙ have enough compenastion for the piece since Black is already preparing Ιο liquidate rooks. 2c2) 17...:b818 ~a3 ~h6!? (or 18 ...h6 19 lL!e4 ~e6 20 lbc5 t; Greenfeld suggests 18 .....b6) 19 'ii'h4 (19 h4? f6) 19... Φg7 20 Ι4! and now: 2c21) 20...f6 (20... lL!xf4 21 :xd8 "xd8 22 "xh6+! wins) 21 :d3! exf4 (21 ... lLIxf4 22 :xd8 "xd8 23 "xh6+ and wins) 22 :h3! fxg5 23 "xh6+ Φf6 24 b3! with dangerous threats. 2c22) 20...':e8 21 fxe5!? ':xe5 22 ~f8+! is very good for White. 2c23) 20...~xg5 21 fxg5 ~e6 22 ':de Ι! lL!f4 23 ~xe6 lLIxe6 24 :e3! 'i'a5!? (24 .....d7 25 :d3! lLId4 26 :el±) 25 "f2! :d4 (25 .....d5! t) 26 :hel ! ':b5 27 "f6+ 'iPg8 28 c3 ± Svidler-Alterman, Haifa 1995. 11 h4! Until recently 11 g4 was the normal move here, but then it was discovered that ... e5 is much stronger

120 The Drαgon once White has weakened his kingside. Por exarnple, the game Morozevich-Savchenko, Moscow 1991 continued 11 ... e5! 12lΔxc6 bxc6 13 exd5 cxd5 14 .tg5 .tb7 (with the pawn οη g2, 15 .tc4! would give White the advantage, and ίι would still bave been better than the course Morozevich took) 15 'ii'xe5? h6 16 .th4 (16 .txf6 .txf6 is tremendous forBlack) 16 ... g517 .tellΔxg418 'ii'g3lΔe3 19 :d2 'ii'a5 20 'ii'f2 .txc3 21 ΦbΙ 'ir'b4 0-1. After 11 h4, which has the meήt of ηοι weakening the kingside so much, White plans Ιο start an attack οη the h-file which his queen οη el is particularly well placed Ιο join ίη. Ιη many lίnes fih4 will prove difficult for Black Ιο meet. 11 ... fic7 (D) Altematives are: 1) 11 •. :ti'e7 12 lΔb3 %1d8 (οη 12... b6, thereply 13 h5100ks strong) 13 exd5 lΔxd5 14 .tg5! lΔf6 15 %1xd8+ 'iνxd8 16lΔe4 a5 17 a3 a4 18 lΔbc5 h6 19 lΔxf6+ .txf6 20 .txf6 'iνxf6 21lΔe4 fid4 22 .tb5 Φg7 23 'iνc3! and the weakness of the apawn proved decisive ίη Sax-Piket, Tilburg 1989. 2) 11 ...:e8 12 lΔb3 lΔa5 13 h5 lΔxb3+ 14 axb3 'ii'a5 (14 ... lΔxh5 15 g4 lΔf6 16 'ii'h4 is extremely dangerous for Black) 15 hxg6 fxg6 16 'iti>bl dxe4 17 fxe4lΔg4 18 .td4 e5 19 .tgl 'iic7?! 20 lΔb5 'ii'e7 21 .tc4+ 'iti>h8 22 lΔd6 (Smagin-Rajkovic, Bundesliga 1991) with a winning advantage for White as 22...1:tf8 fails Ιο stop 23lΔf7+! οη account of

23 ...:xf7 24 .txf7 "ikxf7 25 :d8+ .tf8 26 .tc5.

12 hS case can also be made ουΙ for continuing ίη positional vein. Por example, 12 exd5 exd5 13 'ii'd2 :e8 14 .tf4 "ika5 15 lΔb3 "ikb4 16 a3 "ike7 17 .tg5 .te6 18 lΔb5 was quite good for White ίη Dvoirys-Los, Groningen 1993. lΔxhS 12 ... 12•.. lΔxd4 13 .txd4 e5 14 lΔb5 'iνe7 15 .tc3lΔxh5 16 exd5 was better for White ίη Sίήgοs-Αlterman, Komotini 1992. 13 exdS Knaak's proposal,13 g4, is very interesting. He gives 13.•• lΔxd4 14 .txd4 .txd4 15 :xd4 'ii'g3 (after 15 ... 'ii'e5 16 fid2lΔg3 17 :gl dxe4 18 f4lΔxf1 19 :'xf1 'ii'g7 20 g5 e5 21 1:[d6 e3 22 'ii'xe3 exf4 23 'iνxf4 .th3 24 1:tf2 :ae8 25lΔd5 :el + 26 Φd2 "ike5 27 :h2! White gained material ίη the game R.Mainka-Alterman, London Lloyds Bank 1994) 16 'iνd2 lΔf4 17 exd5 e5 18 1:[e4 'iνxf3 19 'iνh2 h5 20 gxh5 as being clearly better for White, but perhaps Α

The Dragon 121 13•••lbg3 would be a more cήtίcal test of Κnaak:'s idea. 13 ••• lbxd4?! 13•••exd5! is better, as 14lbxd5 "e5 is uncIear and Luther's suggesιίοη 14 g4 faίIs Ιο 14 ... lbxd4! 15 .txd4 "f4+ 16 .te3 "xf3. Therefore ίη AImasi-Kir.Georgiev, Groningen 1994 White trίed 14lbdb5 and instead of 14.....e7? 15lbxd5 "e5 Ι 6 lbbc3 with advantage ιο White, Black should have played 14.....g3!, which AImasi assesses as equal. So unless an improvement is found for White here, the altematίves Ιο the 12th and 13th moves given above look more promίsing. 14 .txd4 .txd4 14.....g3 15 "d2 .txd4 16 "xd4 e5 (16 .....f4+ 17 "'xf4lbxf4 18 d6 is good for White) 17 "'gl! .tf5 18 'iitbl :fc8 19 d6 is cIearly better for White according Ιο Luther. 15 1%xd4 "'g3 (D)

"e3

16 exd5 After 16•••e5 17 :dh4 the black queen is ίη trouble. Luther points ουΙ 17 ... .tf5 18 ll)e4 .txe4 19 1:g4! .txd5 201:xh5.

17 lbxd5 17 1:txh5 was temptίng, but White cοπectlΥ fιgures that he can gaίn a strong attack without parting with any mateήal. 17 ••• 'it>h8 BIack can' t have enjoyed playing this ugly move, but the probIem is that the naturaI17•...te6 would present White with the bήΙΙίaηCΥ ΡήΖe after 18 1:xh5! gxh5 19 1:g4+!! hxg4 20 "'g5+ Φh8 21 "f6+ Φg8 22 lbe7#. 18 1:e4! ΒΥ opening the long diagonal for the queen White prevents 18•••i.e6 which would now Iose Ιο 19 "'c3+. 18 ••• .tf5 19 1:e5 Threatenίng 1:xf5 or 1:xh5. 19 ••. Ι6! (D)

The οηlΥ chance. 20 1:e7? White is bIuffed out. 20 1:txf5! was stiIl very strong because after 20••• μΙ5 21 1:xh5 1:tae8, 22lbxf6! is immediately decisive. 20•••1:ae8 at once is a better try although after 21 1:fxh5! gxh5 (after 21 ...1:xe3 22

122 The Dragon :xh7+ Φg8 23 lbe7+! 1Σχe7 24 .tc4+ White forces mate) 22 'ii'd4 1ΣeΙ + 23 Φd2 1ΣaΙ 24llx:3 'ii'el + 25 Φd3 the black attack has run ουΙ of momentum and White remains with his mateήal plus. 20 ••• 1Σad8 21 Ι4 White sensibly opts for an exchange of queens, avoiding 21 c4 b5! and 21 'ii'h6 1Σf7!. 21 ••• "ίi'xe3+ Νοι 21 •.. :xd5?? 22 'ii'xg3 tDxg3 23 1Σhχh7+ Φg8 24 :eg7#. The alternative 21 •••"ίi'g4Ioses, but ίη a less obvious fashion: 22 g3! and now 22..JιxdS 23 .te2 wins as 23 ...'ii'xg3 24 'ii'xg3 leads Ιο the same mate as above, and 22••..txc2 23 tDc3! defends against Black's threats and prepares .te2. 22 tDxe3 .tc8 23 g4 tDg7 24 .td3?! Luther prefers 24 .tb5, retreating the bishop Ιο d3 only after Black has weakened his queenside with ... a6. 24 Ι5 25 μΙ5 μΙ5 26 tDc4 :re8 27 tDe5 Ι believe this is where White let his remaining advantage slip. He should have played 27 1Σχe8+ as Black cannot recapture with the rook: 27.••:xe8 28 tDd61Σf8 29 :h6 and Black is ίη virtual zugzwang, ηοΙ Ιο mention the concrete threat of28 .tc4. Therefore Black has Ιο play

27••• tDxe8 and after 28 tDe5 White retains a nagging edge. His pawn structure is superior and his king is ready for a quick march Ιο the centre. 1Σχe7 27 28 tDg6+ Φg8 29 tDxe7+ Φh8 30 1Σh6 .te6 31 :r6 :e8 32 tDxr5 tDh5 33 1Σh6 tDxr4 34 .te4 .txf5 35 .txf5 1::te7 36 Φd2 The game is now level. The simplest would have been 36•••Φg7 37 1::txh7+ Φf6 38 :xe7 9;xe7 with a dead drawn position. The remainder of the game is given ίη brief with Black making error after eποr and somehow contriving Ιο lose an ending that was almost impossible ιο lose: 36••• Φg8 37 c4 :r7 38 .te4 :e7 39.tc2 Φf8? 40 :xh7? (40 .txh7 would give some chances) 40••• 1Σχh7 41 .txh7 9;e7 42 Φe3 tDh5 43 Φd4 tDr6 (43 ...b6!) 44 .tr5 'iii'd6 4! cS+ Φc6 46 b4 b6 47 ΦeS tDe8 48 .te4+ Φd7 49 b5 tDc7?! (49 ... bxc5 50 .tc6+ Φd8 51 .txe8 Φχe8 52 'iii'd5 Φd7 53 Φχc5 Φc7 54 a4 'iii'b7 55 a5 a6 =) 50 αι,6 axb6 51 84 Φe7?? (Black could still have saved himself with a sta1emate trick: 51 ...9;c8 52 Φd6 Φb8 53 'iii'd7 Φa7! and 54 .ta8 is met by 54 ... 'iii'b8) 52 .tr5 1-0. Nothing can be done Ιο stop the white king penetrating.

6 Kan Variation This line, which starts 1 e4 c5 2lbf3 e6 3 d4 cxd4 4 lbxd4 a6, is notable for the f1exibility afforded Ιο Black, since by delaying his piece development he keeps the maximum possible range of options open. Because there are many reasonable choices at each move it is pointless Ιο give precise lines against all possible moveorders, so ίη this chapter there will be a greater emphasis οη general principles. The continuation recommended ίη this chapter, 5 J.d3, is the most common line ίη practice. Αι the moment Black's most popular reply is Ιο set υρ a 'hedgehog' posiιίοη by 5 ... lbf6 6 ο-ο d6 (or 6 ... 'ίi'c7 7 'ίi'e2 d6). After 7 c4 Black may choose Ιο develop his bishop οη e7 immediately, but he sometimes brings the queenside ουΙ fιrst ίη order Ιο retain the possibility of ... g6 and ... J.g7. The 'hedgehog' name is deήved from the way Black curls υρ οη the fιrst two ranks, moves like ... lbbd7, ... b6, ...J.b7, ...1:te8, ... 'ίi'c7, ... J.f8, ...1:tac8 and ... 'ίi'b8 being typical. Black's slow development invites White Ιο attack, but experience has shown that the unwary attacker can easily impale himself οη Black's spines, and such attacks have Ιο be well-organised ίf they are Ιο stand much chance of success. Moreover White has Ιο watch ουΙ for

Black's ... b5 and ... d5 breaks. Game 19 deals with the lines arising after 5... lbf6, including the 'hedgehog'. ΑΙΙ Black's other 5th moves, such as 5 ...J.c5, 5... lbc6, 5 ... lbe7 and 5 ... g6, are ίη Game 18. Game 18 YakoVΊch

- Emms

Cαppelle Ια Grαnde

1 e4 2

3 d4 4 lbxd4 5 J.d3 (D)

1993

c5 e6 cxd4 a6

5 ...

J.c5

Black's alternatives are arranged ίη descending order of importance: 1) S.••lbc6 (ίΙ is surpήsίηg that this solid line is not played more frequently; although the symmetrical position gives Black few winning

124

Kαn Vαriαtion

chances, ίι is quite hard for White Ιο prove any advantage) 6lbxc6 and now: la) 6.••dxc6 7 ll)d2 e5 8 "h5 i.d6 9lbc4 ~c7 (9 ... ll)f6 10 lΔxd6+ 'Wxd6 11 'ife2 i.e6 12 ο-ο gives White the opportunity Ιο make his dark-squared bishop a potent force, as ίη Jansa-Ceba1o, Smederevska Pa1anka 1978 after 12... ~7 13 1:dl 'We7 14 b3 ο-ο 15 a4 a5 16 i.a3lbc5 17 .tc4!) 1O.tg5lΔf611"e2h612 i.h4 'We7 130-0-0 (13 ο-ο?! allowed Black Ιο stir υρ trouble by 13 ... g5 14 i.g3 h5 15 f3 h4 ίη LigterinkMiles, Lone Pine 1979) 13 ... i.e6 14 ί4 i.xc4 (14 ... i.g4 15 i.xf6 i.xe2 16 i.xe7 i.xdl 17 i.d6 i.xd6 18 ll)xd6+ cJi>d7 19l1)xf7 wins materia1 for White) 15 i.xc4 b5 16 i.b3 ο-ο 17 i.xf6 "xf6 18 :'d7 and White's pressure against f7 gives him an advantage, Tseshkovsky-Miles, BledPortoroz 1979. lb) 6... bxc6 (ουι οί favour ever since the famous Fischer-Petrosian game mentioned below) 7 ο-ο d5 (7 ... e5 8 ί4 i.c5+ {8 ... d6 is more solid} 9 cJi>hllΔe7 10 'Wh5lΔg6 11 ί5lΔί4 12 i.xf4 exf4 13 ί6 i.d4 14 fxg7 .txg7 15 :'xf4 'We7 16lΔc3 is very good for White, Ravinsky-Vorotnikov, USSR 1963, while 7 ... g6?! 8 e5 i.g7 9 ί4 d6, Van der WielAnand, Thessaloniki OL 1984, should have been met by 10 exd6 "xd611lΔd2 'Wd4+ 12 ΦhΙll)f613 lΔc4ll)d5 14lΔd6+ cJi>e7 15 c3 with a fine position for White) 8 c4lΔf6 9 cxd5 cxd5 10 exd5lΔxd5 (1O...exd5 11ll)c3 i.e7 12 'Wa4+ "d7 13 :'el!

"xa4 14lΔxa4 i.e6 15 .te3 ο-ο 16 .tc5 is Fischer-Petrosian, Buenos Aires Ct (7) 1971, which was won by Fischer, while 10.....xd5 11 ll)c3 "d7 12 i.g5 i.e7 13 "e2 i.b7 14 1:acl ο-ο 15 ':fdl, although keeping Black's pawns intact, gave White a dangerous initiative ίη Mikh~ chishin-Gorchakov, USSR 1972) 11 i.e4 (D) and now:

lbl) 11 ..':b8 12 "f3 ί5 13 .txd5 "xd5 14 'Wxd5 exd5 15 ':'dl .te6 16 lbc3 ':'d8 17 .tg5 ':'d7 18 lΔe2 gives White the better ending. lb2) 11....te7 12 lΔc3 .tb7 13 'Wa4+ 'Wd7 14 "xd7+cJi>xd7 15 ':'dl is also promising after 15....:.ad8 16 lΔxd5 i.xd5 17 i.xd5 exd5 18 ':'xd5+ cJi>e6 19 ':xd8 ':'xd8 20 i.e3, Matanovic-Roos, Le Havre 1966, or alternatively 15 ... i.f6 16 lΔxd5 i.xd5 17 i.xd5 exd5 18 ':'xd5+ cJi>e6 19 ':'d2 ':'hd8 20 ':'e2+, AverbakhTaimanov, USSR Ch 1960. lb3) 11..':a7 12 'ii'd4 ':'d7 13 lbc3lΔxc3 (13 ....tb7 14lbxd5 .txd5 15 i.xd5 ':'xd5 16 'ii'a4+ wins the apawn) 14 'Wxc3 and Black's uncastled king gives him plenty οί

Kαn Vαriαtion

problems, Beliavsky-Kurajica, Sarajevo 1982. 2) 5 •••g6 6 c4 i.g7 (this is an attempt Ιο reach a kind of hedgehog position, but with the bishop more actively deployed at g7) 7lL1b3 (D) and now:

2a) 7 •••lLIe7 8 lLIc3 ο-ο (8 ... d5 9 cxd5 exd5 10 i.g5 h6 11 i.xe7 i.xc3+ 12 bxc3 'fIxe7 13 ο-ο dxe4 14 i.xe4 0-0 15:el andWhitehasa big lead ίn development, DonchevPri6, Τουlοn 1988) 9 ο-ο lLIbc6 10 i.e2 b6 (1O ... f5 11 c5 b6 12 cxb6 'fIxb6 13 i.e3 1i'd8 was unclear ίη Psakhis-Oratovsky, Tel Ανίν 1993 but perhaps 11 exf5 is better) 11 i.g5 h6 12 i.e3 f5 13 'ifd2 ~h7 14 f3 with an edge for White, AnandBologan, Calcutta 1992. 2b) 7•••d6 8 lLIc3 lLIf6 9 i.f4 (9 ο-ο followed by i.f4 transposes to lines given ίη Game 19, but White can do better here because Black has ηο time for ... lLIc6 and ... lLIe8) 9... 0-0 10 i.e2 e5 (10 ...lLIe8 11 c5 is particularly unpleasant when Black's queen is undefended) 11 i.e3 i.e6 12 ο-ο lLIc613 f3 :c8 14 :cl, with a

125

favourable position for White ίn the game Ljubojevic-Rajkovic, Yugoslavia 1980. The plan of directly attacking the d-pawn by i.f4 and i.e2 is a logical way Ιο exploit Black's ... i.g7, and ίη this case ίι gives White the advantage. 3) 5•••lLIe7 (Black aims Ιο play ... lLIc6, but only when he can recapture with a piece) and now: 3a) 6 ο-ο lLIec6 (or 6 ... g6 7 c4 i.g7 8 i.e3lL1bc6 9lL1xc6 bxc6 {9 ... lLIxc6 t is better} 10 c5! i.xb2 11 lLId2 ο-ο, Topalov-J .Polgar, Las Palmas 1994, and now 12 i.h6 ~g7 13 ..txg7 ~xg7 14 lLIc4 would have given White a permanent positional bind) 7 c3 i.e7 8 i.e3 ο-ο 9 f4 d6 1Ο lLIf3 lLId7 11lL1bd2 gave White a small but enduring plus ίη GeorgievPeev, Bulgarian Ch 1980/1. 3b) 6 lLIc3 lLIec6 7 lLIb3 i.e7 8 'ifh5! d6 9 ~e3 lLId7 10 f4 b5 11 0-0-0 b4?! 12lL1a4! e5 13 ί5 ο-ο 14 g4! gave White an automatic attack, Mikhalchishin-Dorfman, Lvov 1983. 4) 5 ••:fkc7 6 ο-ο and now Black can transpose ίnΙο Game 19 by 6 •••lLIf6, or try: 4a) 6•••ω?! 7 lLIxc6 1i'xc6 (or else Black has an inferior version οί variation 1 above) 8 c4 (8 lLId2 is probably also good) 8... g6 9 lLIc3 i.g710:ellL1e711 i.g5 d612 'fkd2 with a very good position for White. 4b) 6 ••• ~c5 7 lLIb3 i.e7 8 'fkg4 (White could also develop normally) 8 •••..tf6 and now: 4bl) Nunn-I.Gurevich, Hastings 1992/3 continued 9 lLIc3 h5 1Ο 1i'e2

126

Kαn Variαtion

~c6 11 f4 d6 12 .te3 (12 .td2 may be better) 12 ... .txc3 13 bxc3 ~f6 with an unclear game. 4b2) Ιη his notes Ιο the game Gurevich gives 9 .tr4!? h5 (forced) 10 "g3 h4 11 "g4 .te5 12 .txe5 "xe5 13 ι!bc3 ~f6 14 'ikxg7 :g8 15 "h6 "xh2+ 16 ~xh2 ~g4+ 17 ~hl ~xh6 as equa1, but Ι (JG) don't see what's wrong with 14 We2 when White has a healthy lead ίη development and will be able to gain further time with f4. 5) 5 ••• b6 6 ο-ο .tb7 7 ~3 (7 c4 also led Ιο a good game for White ίη Benjamin-Dizdarevic, Manila OL 1992 after 7 ...d6 8 f4 ~d7 9 ~c3 g6 10 f5 .tg7 11.tc2 'fie7 12 fxe6 fxe6 13 ~f3! ~gf6 14 .tf4 e5 15 .tg5) 7...d6 8 Ι4 (D) and now:

5a) 8...lDd7? 9 f5 e5 10 ~e6! (the main drawback of the early queenside fianchetto is that it weakense6) 10... fxe6(1O .....c811 ~d5! is very strong as 11 ... fxe6 loses to 12 "h5+ Φd8 13 fxe6 ~f6 14 'ikf7 and 11 ... .txd5 12 exd5 fxe6 13 "h5+ ~d8 14 fxe6 ~df6 15 "f7 ~e7 16 .tg5 is a1so winning for

White) 11 "h5+ g6 (11 ... ~e7 12 fxe6 g6 13 .tg5+ ~f6 14 "f3 .tg7 15 ~d5+ .txd5 16 exd5 leaves Black ίη a terrible tangle) 12 fxg6 ~gf6 13 g7+ ~xh5 14 gxh8" "e7 15 "g8 and Black had ηο rea1 compensation for the exchange in Fogarasi-Portisch, Hungary 1994. 5b) Οη 8...~6, Fogarasi recommends 9 g4! and after 9 ... d5 10 e5 ~e4 11 'fif3 ~xc3 12 bxc3 ~7 13 .td2 .te7 14 :ael White has a clear advantage. 6) 5...b5 6 ο-ο .tb7 7 a4 b4 8 ~d2 ~e7 9 f4 ~c6 10 ~f3 d6 11 'ikel with advantage Ιο White, Wedberg-Dzindzihashvili, New York 1991. One of the main advantages of 5 .td3 as opposed to 5 ι!bc3 is that an early ... b5 by Black is hardly ever a WΟΠΥ, since White may undermine Black's queenside pawns by a4 without fearίng a loss of time after ...b4. 7) 5.....b6 6 c3! d6 (6 ... ~6 7 ο-ο ~xd4 8 cxd4 "xd4 9 ~3 is dangerous for Black) 7 0-0 ~f6 8 a4 .te7 9 ~d2 "c7 1Ο a5 ο-ο 11 ~c4 ~bd7 12 'ii'e2:e813 .tg5! .tf8?! (13 ...h6 14 .th4 b5 is just slightly better for White) 14 .th4 with a good game for White, Lazic-Martinovic, Yugoslavia 1987. .ta7 6 ~b3 6....tb6 is also possible but then Black will have Iίttle choice but to exchange οη e3. 6....te7 appears from time to time, the idea being akίη to that of the .....b6, .....c7 ploy seen ίη many Iίnes of the Sicilian, namely to force the knight ιο retreat from its active

Κaπ

central ρost. White can try 7 "14 but the simplest seems Ιο develop normally. For example 7 ο-ο d6 8 c4 followed by 1Dc3, f4, ΦhΙ, etc. 7 "e2 7 ο-ο ~c6 8 "14 (Ο), hoping Ιο take advantage of the fact that the dark-squared bishop has deserted the kίngside, is worthy of attentίon.

Blackhas: 1) 8.....f6 9 ~c3 ~ge7 10 .tg5 "iVg6 11 'ίi'h4 (ECO assesses this position as ;1;) and now ίη PopovicSchlosser, Bmo 1992 Black produced a real howler, 11 ... ~e5??, with the obvious idea of 12 .txe7?? ~f3+ but after 12 .te2! he felt obliged Ιο resign as he couldn't deal with the double threat of 13 .txe7 and 13 .th5. 2) 8 •.• ~Ι6! 9 "xg7 (9 'ίIfg3 or 9 'ίIfM are reasonable alternatives) 9•••:18 10 "h6 and now: 2a) 10•••tbeSll.te2! b5 (11 ...d5 12 ~c3 dxe4 13 .tg5 :g6 14 'ίIfM .td7 15 ~xe4 .tc6 16 .txf6! and White was already winning ίη Hellers-Sjoberg, MalmQ 1994) 12.tg5 (Ljubojevic gives 12 .te3 :g6 13

Variation 127

'ίIfh3 as a clear advantage for White, but perhaps Black can try 13 ... ~xe4 since 14 .txa7 ':xa7 15 'ίi'e3 is met by 'ίIfg5! when 16 'ίi'xe4 loses after 16 ....tb7! and 16 'ίi'xg5 ~xg5 17 f4? ~h3+ 18 Φhι.tb7! is alSO immediately decisive) 12 ... :g6 13 'ίi'M .tb7 14 ~ld2 h6 (or 14 ...:c8 15 c3 ;Ι; Oll-Stangl, Tίlburg 1994) 15 .txf6 lLxf6, Ljubojevic-Lobron, Plovdiv 1983, when Ljubojevic gives 16 c3 as ;1;, but Ι lίke 16 a4 as 16 ... b4 can be met by 17lba5!. 2b) 10•••:g6 11 'ίIfh3 e5! led Ιο a draw by repetition ίη J.Horνath Farago, Budapest 1987 after 12 'ίi'M :g4 13 'ίi'h6 :g614 ~4, etc. and ίι has ιο be said that ίι would be extremely risky for White Ιο try Ιο continue the game. For example, if he continues 14 "d2, then 14 ... d6 (threatening ....th3) 15 ΦhΙ ~g4 100ks very strong as 16 g3 is met by 16... ~xh2! 17 ΦΧh2 'ίIfd7! and 16 h3 'ifh4! looks crushing. Το sum up, this could be an interesting lίne Ιο have ίη your repertoire if you are willing Ιο take the ήsk of a quick draw. 7 ~c6 8 .te3 (Ο) 8 ~6 Up until about four years ago Black regularly exchanged bishops οη e3, but since then he has almost invariably invited White Ιο exchange bishops οη a7, ίη the belief that his rook will be more active οη the second rank. Perhaps this is just a quirk of fashion so we still have ιο examine the older lines ίη some detail:

128 Kan

Variαtion

1) 8•••.i.xe3 9 'ifxe3 d6 (9 ... e5 10 ltιc3 ltιge7 11 0-0-0 d6, Wahls-Far-

ago, Altensteig 1987, and now 12 :d2 followed by :hdl and .i.e2 creates awkward pressure οη the dfile and 9 ... ltιge7 10 ltιc3 ο-ο 11 0-0-0 'fIic7 12 f4 d6 13 'fIih3ltιb4 14 :hf1! e5 15 f5 d5 16 exd5lί)exd5 17 ltιxd5 ltιxd5 18 'ifh4 ltιf6 19 g4 e4 20 g5 exd3 21 :xd3 'fIie5 22 gxf6 'fIixf6 23 "xf6 gxf6 24 :d6 gave White a very good ending ίη KorlovBatakov, Cou 1984) 10 ltιc3lί)f611 0-0-0 ο-ο (moves such as 11 ... b5 and 11 .....c7 are well met by 12 g4 and 11 ... e5 led Ιο a promising position for White ίη Lekander-Schoneberg, Cou 1980 after 12 :d2 .te6 13 :hdl 'fIic7 14 .te2 :d8 15 g4! 0-0 16'ii'g3.txb317cxb3!lί)d418~bl

b5 19 .td3 "c6 20 f3 ':fe8 21 "g2 b4 22 g5!) 12 Ι4 'ii'c7 (D) and White has a number of attacking ideas: la) 13 'ifh3ltιb4 14 g4 b5 15 g5 ltιxd3+ (15 ... ltιe8 16'ii'h4 f6 17 a3 fxg5 18 fxg5 ltιxd3+ 19 :xd3 :b8 was unclear ίη Vogt-Velikov, E.Germany-Bulgaria 1987) 16 :xd3ltιe8 17 f5 b41ed Ιο a sharp finish ίη Bronstein-Suetin, Moscow Ch 1982 after

18 'ii'h4 bxc3 19 :h3 cxb2+ 20ΦbΙ f6 21 'fIixh7+ Φπ 22 :gl :g8 23 :h6 exf5 24 :xf6+ ltιxf6 25 gxf6 ΦΧf6 26 'ii'xg8 .tb7 27 'ii'h7 .i.xe4 28 'fIig6+ Φe5 29 'fIig3+ 1/2-1/2. lb) 13 :hgl (this is probably good enough for a slight advantage) 13 ... b5 14 g4 b4 15 g5ltιe8 16lί)e2 (16 ltιbl a5 17 :g4!? a4 18 ltι3d2 .ta6 19 .txa6 :xa6 20:h4 g6 21 'ii'h3 f5 22 gxf6 ':xf6 23 lί)c4 may be slightly better for White, Wedberg-Spraggett, New York Open 1987) 16... 'fIia7 (16 ...a5 17 ~bd4 ltιxd4 18 'ii'xd4 .ta6 19 ΦbΙ was a little better for White ίη ArnasonSuetin, Sochi 1980) 17 'ii'h3! g6 18 f5 exf5 19 exf5 ltιe7 20 ltιg3! 'ii'e3+?! (20 ... a5 is better, but still favours White) 21 ΦbιJ.χf5 22.txf5ltιxf5 23 :del 'ii'f4 24 :gf1 'ii'h4 25 'ii'xh4 ltιxh4 26 :e4 with an excellent ending for White, Arnason-Κirov, Plovdiv 1986. lc) 13 g4 with a further branch: 1c 1) Accepting the offer must be a cήtίcaΙ test ofWhite's willingness Ιο play g4 without the preparatory :hgl. Ιη Short-Velikov, European Club Ch 1987 the continuation was

Καπ Vαriation

13 ... llJxg4 14 'iί'g3 llJf6 15 :hgl llJe8 16 ~bl (16 f5!? is natural) I6 ... llJe7?! (16 ... b5 and 16 ... f6 have been suggested as possible improvements) 17 llJd4 Wc5 18llJf3 ί6 19 c5! with a very strong attack for White. Velikov must have found an improvement because he repeated this line ίη a later game IvanovicVelikov, SaintJohn Open 1988. υη t"ortunately Ivanovic varied by 16 'ii'h4, so we don't know what Velikov's intention was. Despite this hint, Ι (ΙΝ) believe White has good compensation for the pawn and ίι would require a brave player ΙΟ take this line οη as Black. lc2) 13... b5 14 g5 llJd7 15 Ι5!? b4 16 llJe2 a5 (positions ίη which the players are attacking οη opposite wings are extremely difficult ΙΟ assess; unless one οί the players is well ίη front ίι is likely that a single tempo will decide the race and obscure tactical points will often have a crucial influence οη the play) 17 'iWh3 (attacking e6 directly, and generating a concealed threat ΙΟ h7) 17...exf5 18exf5 llJde5 19l1Jf4 (the ρίη a10ng the c8-h3 diagona1 is awkward for White and bringing the knight Ιο d5 is the οηlΥ way Ιο make progress) 19...a4 20 llJd5 ~d8 (now that b3 and g5 are under attack, White is committed Ιο the sacήficίa1 path) 21 :hgl! (D) and now: lc21) Ιη the game Kengis-Nevednichy, USSR 1979 Black continued 21...llJxd3+? (eliminating one of the attacking pieces, but ίη doing so activating the dl-rook) 22 :xd3

129

llJe5 (now White wins by force) 23 llJf6+ gxf6 24 'it'h6! (24 gxf6+ llJg6 repulses the attack) 24 ... llJxd3+ (now 24 ... llJg610ses Ιο 25 :h3 :e8 26 fxg6 fxg6 27 'it'xh7+ Φf8 28 'iί'h8+ whilst 24... Φh8 25 :h3 ~xί5 26 g6! ~xg6 27 :'xg6 is a1so mate) 25 ~bl! (Black was hoping for 25 cxd3 ~h8 when 26 gxf6 :g8 27 :g7 ~xί5 defends, as does 26 g6 fxg6 27 fxg6 'it'c7+ and 28 ... 'it'g7) 25 ... fxg5 (25 ...Φh8 26 g6 fxg6 27 fxg6 threatens both 28 'it'xh7# and 28 g7+, while 25 ... 'it'b6 26 :'g3 οηlΥ makes matters worse) 26 f6 'it'xf6 27 'it'xf6 g4 (although rook, bishop and two pawns amount Ιο enough material Ιο ba1ance a queen, Black still suffers from his bad king position) 28 'it'g5+ Φh8 29 'it'f6+ Φg8 30 llJd4 llJe5 31 h3 (intending 32 hxg4 followed by llJf5) 31 ... h5 32 'it'g5+ llJg6 33 'it'xh5 gxh3 34 'it'd5 (attacking a8 and g8) 34...~e6 35 llJxe6 h2 36 :Ιχg6+ 1-0 (36... fxg6 37 llJxf8+ and 'it'xa8 wins all the black pieces). Αη extremely energetic performance fromWhite. lc22) Black's best defence is 21 ...axb3! 22 llJf6+ μΙ6 23 Wh6!

130 Kan Variation (23 gxf6+ tLJg6 leads nowhere as the f5-pawn is pinned while 24 'ilh6 'ilxf6 25 fxg6 fxg6 26 .txg6 'ilf4+ 27 'ilxf4 :xf4 28 .te8+ ~f8 29 .txc6 bxa2 30 ~d2 :d4+ wins for Black) 23...lbg6! (23 ... ~h8 24 gxf6 :g8 25 :g7 lbxd3+ 26 :xd3 .txf5 27 :h3! results ίη mate at h7) 24 fxg6 fxg6 2S .txg6 :a7! 26 gxf6 hxg6 (26 ... 'ilxf6 27 .txh7+ ~h8 28 :'g8+! :'xg8 29 .tg6+ mates, or 27 ... Wf7 28 :'df1) 27 :xg6+ (27 'ilxg6+ ~h8 leads Ιο nothing as the d 1 rook cannot reach the h-file, ':d5 being met by ... lbe5) 27...Φf7 (D) reaching a remarkable position ίη which ίι appears that White must mate, but ίι isn't certain that he can doso.

1c221)

Ιη

the second edition

(JN) commented that Ι couldn't see a

mate after 28 :g7+ ~e6 (28 ... ~e8 29 'ile3+ lbe5 30 'ilxa7 'ilxf6 31 axb3 is very good for White, with material equality but weak black pawns and an exposed black king) 29 'ilh3+ ~xf6 30 "h6+ ~e5 (after 30... ~f5 31 :g2! Whitedoesmate), for example 31 :'g3 .tf5 32 c4

.te4!. Ι still don't see a mate, (nor can Fritz - JG) and nobody wrote Ιο me suggesting one. lc222) If we return Ιο the position after 27 ... ~f7 White can gain a massive advantage by 28 ':el lbe5 29 :'g7+ We6 (29 ... ~e8 30 :'xa7 'ii'xf6 31 'ii'h5+ ~d8 32 axb3 is similar) 30 :'xa7 bxa2 (30 .....xf6 31 'ilh3+ and 'ilxb3+) 31 'ii'h3+ Φd5 32 "b3+ ~c6 33 "xa2, with a slight material plus for White together with a raging attack. 2) Black can also develop his knight Ιο e7 rather than f6. Against this plan we are recommending that White castles short, ηοΙ getting ίη volved ίη the habitual pawn storming competition as his attack won't have as much momentum without a knight οη f6 for the g-pawn Ιο sink its teeth ίηΙο. Moreover with the knight οη e7 Black won't have as much control over squares like h5 and g4, both ίη the vicinity of his castled king, so White may be able Ιο develop an attack merely by piece play; and if he's ηοΙ going Ιο pawnstorm, his own king wil1 be much safer οη the kingside, well ουΙ of the way of Black's queenside counterplay. Α good example of this strategy is the game Κindermann-Zso.Polgar, ΜίίηsΙer 1994: 8... lbge7 (if Black refuses Ιο commit this knight then White can also wait, playing moves like 9lbc3 and 10 f4 before deciding οη which side Ιο castle) 9 lbc3 "c7 (9 ... d6 10 f4 ο-ο 11 .txa7 :'xa7 12 0-0 b5 13 :'ael1eads Ιο very similar play) 10 f4 d6 11.txa7 (perhaps this

Kαn

exchange could have been rnade earlier) 11. ..:xa7 120-0 b5 13 :ael b4 14 tLJdl ο-ο 15 tLJe3 d5 16 e5 ί5 17 exf6 :xf6 18 "h5 g6 19 'fί'g5 :f7 20 tLJg4 and White had a clear positional advantage as well as good attackίng chances against the black king. d6(D) 9 ω

10 0-0-0 b5 10•••0-0 11 ί4 e5 12 i.xa7 :xa7 13 Ι5 b5 transposes back ίηΙο the rnain garne, but ίη Spraggett-111escas, Spain 1994 White tried 13 'i!νf2!? and achieved sorne advantage after 13••• tLJg4?! 14 'fί'g3 exf4 15 'iί'xf4 tLJge5 16 i.e2! b5 17 tLJd4 tLJxd4 18 :xd4 i.e6 19 :hd 1, but Β lack would have done better ιο play 13•••exf4. 11 i.xa7 White should rnake this exchange sooner rather then later, for exarnple 11 (4 b4 12 tLJa4 i.d7 13 i.xa7? loses a piece after 13 ... tLJxa7!. This is precisely what happened ιο Μί chael Adarns ίη his garne agaίnst Hjartarson frorn the Paris leg οί the 1994 Intel Rapid Grand Prix, with

Variation 131

the slight difference that the black knight was οη e7 instead οί f6. 11 ••• :Xa7 12 Ι4 b4 When Black has exchanged οη e3 he usually rneets the threat οί e5 by playing .. :iic7, but with the rook οη a7 this would seriously harnper the co-ordination οί his forces; therefore he has ιο play ...e5 hirnself, but first he should dήve the white knight away frorn its control οί d5. Neglecting ιο do so leads ιο a good garne for White, for exarnple: 12•.•e5 13 f5 b4 (or 13 ... 0-0 14 g4 tLJd4?, Yudasin-Nikolaev, Podolsk 1991 and now White should have played 15 tLJxd4 exd4 16 "f2! when 16... 'iWb6 17 :hgl! followed by tLJe2 and 16 ... tLJxg4 17 "xd4 are both very good for hirn) 14 tLJd5 tLJxd5 15 exd5 and now Yudasin considers 15.....g5+ 16 <itbl tLJe7 17 h4!, 15•.•lΔa5 16 tLJxa5 'fί'xa5 17 ΦbΙ "xd5 18 i.e4 "c5 19 :d5 "b6 20 :hdl :d7 21 ί6! and 15••. tLJe7 16 ί6! gxf6 17"f2 all ιο be ίη White's favour. 13 tLJa4 This is οί course a very doubleedged square for the knight, but the alternative, 15 tLJbl, allows Black an autornatic attack by ... a5-a4 etc. 13 ..• e5 14 f5 ο-ο The immediate 14•••i.d7 would be rnet by 15 "e3. 15 g4 This doesn'trequire any explanaιίοη.

15 •••

i.d7

132 Kan Variation 16 g5 lΔe8 17 'ife3! White is now threatening Ιο solve all his problems by playing 18 tDb6, so Black has ηο choice but Ιο give up his a-pawn. 17 Ahg1, ignoring the plight of the knight οη a4 ίη favour of an immediate kingside attack has also been tried. Icelandic Grandmaster Stefansson analysed the posiιίοη Ιο a draw: 17 ... tΔb8! 18 'iί'h5 .txa4 19 :g3 .tb5 (19 ... ~7 20 .tc4) 20 1:th3 h6 21 1:1g1! (ίη PrieStefansson, Reykjavik 1993 White played the awful 21 f6?? and was 10st after 21 ... tDxf6 22 gxf6 'iί'xf6) 21 ... 'iί'b6 22 1:tg2 f6 23 gxh6 .txd3 24 cxd3 Ac7+ 25 'i1i>bl ~6 26 hxg7 Axg7 27 'iί'h7+ q;f7 28 1:1xg7+ tDxg7 29 'iί'ι6+ Wg8 30 'iί'h7+ with perpetual check. 17 Ab7 18 .txa6 (D)

After 19 .tb5 (not 19 'iί'b6? 'iί'a8) 19•••tDd4 20 .txd7 "xd7 21 tDb6 'iνc6!, White is forced to play 22 Axd4!, although this does seem Ιο be good enough for some advantage, e.g.: 1) 22••.'iνxb6 23 %:td3 'iί'a6 (the lίne 23 .....xe3+ 24 1:txe3 Axa2 25 Adl is excellentfor White) 241Wel! so that moves such as ...1Wc4 or ...1:tc7 can be met by ~bl followed by ~1 if Black attacks a2 agaίn . 2) 22••.exd4 23 1Wxd4 :'c7 (after 23 ... :'xa2 24 <ϊPbl followed by ~5 White has tremendous play for the exchange) and now 24 c3 bxc3 25 bxc3 'iί'xc3+ 26 'ii'xc3 1:txc3+ 27 q;d2 is unclear, so White should probably swallow his Ρήde and play 24 tΔa1. Of course this is ηοΙ an ideal square for the knight but Black's kingside pieces are also ίη a bit of a huddle and White is hoping Ιο relieve the pressure οη the c-file by ~5, thereby enablίng him Ιο return the knight οη al Ιο a more desirable square. There is also ηο need Ιο fear 24 ... b3 as after 25 axb3 1:1a7 White can block the a-file with 26 tΔa4. 19 .tc4! (D)

18 ... Ab8 Although this position has been reached οη several occasions nobody has played (or considered ίη theίr notes) 18•••Aa7, which appears Ιο me at least as critical as the text.

Kαn Vαriαtion

Much strooger than the previous1y p1ayed 19 'Wd2 or 19 ~bl. White wastes 00 time p1aciog his bishop 00 its most effective diagona1, thereby creating dangerous kingside threats based οη g5-g6. 19 ... ~h8 B1ack removes his king from the sensitive diagona1 as after 19••• lDd4 20 g6! lDf6 (20 ....txa4 21 gxf7+ is good for White) 21 gxf7+ ~h8 22 lDxd4 exd4 23 :xd4 .txa4 24 e5 White's attack crashes through the midd1e. Nor can he b10ck the gpawn's advance with 19••. g6 as the opening of the h-fι1e by 20 fxg6 hxg6 21 h4! and h5 will prove far more important than any piece that may be 10st οη the queenside. 20 'il'd2! Α very good move, introducing ideas of lDac5 and defusing the threat of ... lDd4, which can now be met by 21lDxd4 exd4 22 b3 as the queen wiIl ηοι be en prise as ίι wou1d have been οη e3. 20 •.• lba5 The alternative is 20••.:a8 when 21lDac5! :xa2 22 ~bl :a7 23lDxd7 'iί'xd7 24 'Wf2! followed by :d3-h3 should be very good for White. 21 lDxa5 'iνxB5 22 b3 (D) After 22••..txa4 23 bxa4 'Wxa4 24 1:thgl! White's attack will triumph as he is ab1e Ιο combine attack and defence. Α possib1e continuation: 24 ...lDc7 25 :g3 lDb5 26 .txb5 1:txb5 27 'Wxd6 :g8 28 :h3 'Wxa2 29 g6! and White wins because 29•••fxg610ses Ιο 30 :xh7+!, 29•••h6

:c8

:e8 Ιο

Ιο 30 1:txh6+! and 29.•. 1t'd7.

133 30

23 'iVd5! Now B1ack will even be denied the p1easure of dreaming about an attack οη the white king. As the ending is comp1ete1y hope1ess he is forced Ιο retreat 1eaving White with an extra pawn and a positiona1 bind. The game is essential1y decided. 23 1ta7 24 <ιtbl 'fIe3 25 :hel! White temporari1y returns the pawn ίη order Ιο simplify ίηΙΟ a technically won position. 'fIxg5 25 26 lDb6 .iOO 27 'ii'a5 'iνd8 28 .id5! .ixd5 Οη 28 ••• lDf6, White wins by 29 'Wxb4 1:[b8 30 .ixc6 :xb6 31.tb5! followed by a4. 29 1hd5 30 1Dc4 'iνh4 The queen returns Ιο the kingside ίη search of counterp1ay as the ending after 3O•••lDf6 31 'ii'xd8 :xd8 32 lDxe5! is 10st.

:00

134 Kan Variation 31 "'xb4

'ifxh2

32 :Ιc1!

~ι8

42 43 fxg6 44 a8'iW 45 "xf7+!

Leaving aside the fact that 32••• h5 a1lows 33lL!xe5, ίι would be dangerous for Black Ιο push his h-pawn with all four rooks οη the board as White may be able Ιο start an attack with 1:ΣddΙ and 1:ΣhΙ. 33 lαιd6 1:Σxd6 34 J:lxd6 lΔxd6 h5 35 "'xd6 36 a4(D)

g5 ~g7 1:ΣΧΒ8

1-0

Game 19 - Gheorghiu Hamburg 1984

Νυnn

1e4

2 3 4 5 6

e6 cxd4 a6 llJf6

llJf3 d4 llJxd4 J.d3 ο-ο

(D)

This sort of pawn race is ηοΙ uncommon ίη the Sicilian, usua1ly resulting ίη both players having Ιο give υρ their rook for the opposing passed pawn. Here, this will mean that White's b-pawn will be left Ιο carry the day. 36 ... h4 37 aS h3 38 a6 "'C4 39 'i'dS h2 40 a7! 'iWxc1+ 41 ~xc1 h1'iW+ 42 ~b2 Black won the race but loses the game. His next move slightly accelerates his resignation.

6 d6 6 ...d57 e5 is very bad for Black since the natura1 7 ... llJfd7 loses Ιο 8 llJxe6!. The surprising move 6...eS was unveiled ίη Fedorowicz-Dorfman, New York 1989, and now 7 J.g5! exd4 (7 ... h6 8 J.xf6 "xf6 9llJe2 d6 10 llJbc3 J.e6 11 f4 is good for White) 8 e5 J.e7 (8 .....a5 9 J.d2! and 8 ... h6 9 exf6 hxg5 10 :el+ are very bad for Black) 9 exf6 J.xf6 10 J.xf6 "xf6 11 1:ΣeΙ+ ~f8 12 J.e4 is a little better for White since Black

Kan Variation 135 will still have an isolated pawn after White regaίns the front d-pawn. Black has a major alternatίve ίη 6..:fic7 7 'ife2 (D) (7 c4lbc6 8lίΊxc6 dxc6 is now considered satisfactory Ι'οι Black because White has spent u move οη c4, which ίη this position only serves Ιο weaken d4) and now we examine severa1 possibilities which are distinct Ιο an early .....c7:

1) 7 .•• lίΊc6 is bad because of 8 lίΊxc6 and 9 e5. 2) 7...J..c5 is (οι was) a specia1ity of Eingorn and I'Gurevich but ίι has now been discovered that the simple 8 lίΊb3 J..a7 (8 ... J..e7 is also met by 9 e5; forexample, 9 ... lbd51O c4lίΊb4 11 J..f4lίΊxd3 12 "xd3 with a clear advantage for White) 9 e5! gives White an edge. Ulybin-I'Gurevich, Santiago jr W ch 1990 continued 9 ... lίΊd5 10 c4 lίΊb4 11 J..f4 lίΊxd3 (11 ... lίΊ8c612lίΊld2lίΊxd313 'ifxd3 f5 14 "g3 ο-ο 15 c5! was very good for White ίη Κudήn-I.Gure vich, Philadelphia 1990) 12 'ifxd3 f6 (otherwise Black will suffocate) 13 "f3! (better than 13 "g3 g5!) 13 ...fxe5 14 ~+ g6 15 "xe5 'ifxe5

16 J..xe5 0-0, and now Whiιe should have played 17 lίΊc3 as 17...b5 18 cxb5 axb5 19 lίΊxb5 J..a6 20 a4 doesn't give Black enough for the pawn and 17...lίΊc6 18 J..d6 J:tf5 19 lίΊe4 leaves him struggling Ιο complete his development. 3) 7...J..d6 8 ΦhΙ lί'Ic6 9 c3 (9 lίΊxc6 dxc6 10 f4 e5 11 f5 a1so gives White an edge according Ιο Anand) 9 ... lίΊe5 10 f4 lίΊxd3 11 "xd3 J..e7 12 c4! d6 13 lίΊc3 ο-ο 14 b3 and White's extra space gave him some advantage ίη Anand-Ljubojevic, Linares 1993. 4) 7 ...d6 8 c4 g6 (8 ...J..e7 will be similar Ιο the maίn line) has become quite a popular way for Black Ιο play. The kingside fιanchetto reduces White's prospects of a direct kingside attack but οη the other hand d6 is weakened. 9 lίΊc3 (ίΙ is also worth considering b3 and J..b2) 9 ...J..g7 10 J:tdl (White's idea is Ιο attack d6 by lίΊf3, J..f4, J:tacl and J..b 1) 10...0-0 11 lίΊΙ3 (D ) (11 J..c2 is less accurate; the bishop is better placed οη b1 so White should ηοΙ retreat ίι until he has played J:tac1) and now:

136

Kαn Vαriαtion

4a) Η ••• ω and now there is a further branch: 4al) 12 i.f4 e5!? (12 ... llJg4 13 l:ιaclllJge5 14 'it'e3!? 'it'e7 15 i.e2 i.d7 16llJe1! l:ιad8 17 i.h6 was a1ready very pleasant for White ίη Mίlos-Motwani, Manίla OL 1992; Black should probably have played 14 ... llJxd3 with just an edge for White) 13 i.e3 i.g4 is a new idea from Black who hopes Ιο equalise the game by gaining control of d4. After 14 h3 llJd4 (14 ... ..txf3 15 'it'xf3llJd7 16llJd5 'it'd8 17 Μ! {preventing the d7-knight reaching e6 and d4 via c5} 17 ... ~d4 18 'it'g3 gave White an edge ίη KotroniasStefansson, Komotini 1993) 15 i.xd4 exd4 16llJd5 ..txf3 17 'it'xf3 a draw was agreed ίη Almasi-Farago, Ηυη gary 1992, aIthough White still has at least a sma1l edge. 4a2) 12 h3!? avoids the above simplifications and is ίη fact quite a useful waiting move. For example, 12 ... ~d7 13 i.e3 (now that Black has committed himselfto the ... llJd7e5 manoeuvre there is ηο need for White Ιο play ..tf4) 13 ... b6 14 l:ιacl i.b7 15 i.bl llJce5 16llJxe5 llJxe5 17 b3 l:ιfe8 18 'it'd2 i.f8 19 ~a4 llJd7 20 ~xb6! (an important tactical ροίηι which is also relevant Ιο the very simίlar lines considered ίη '4b') 20 ... llJxb6 21 'it'a5 llJd5 22 1:ιΧd5! gave White a clear advantage ίη Luther-Von Gleich, Βοηη 1993. 4b) H •••llJbd712 ..tf4llJg4 (another way Ιο relieve the pressure οη d6 is 12 ... llJh5, one example being Gallagher-Goldstern, Bad Zurzach

1995 which continued 13 i.e3 b6 14 'it'd2 J:[e8 15 ..te2 ..tf8 16 ..th6 ..te7 17 ..tg5llJhf6 18 :tacl..tb7 19 'it'f4 rj;g7 20 'it'h4 with an edge for White) 13 1tacl ~geS 14 b3 (14 ~xe5!? ~xe5 15 ..te3 ..td7 16 f4 ~xd3 17 'it'xd3 :t"d8 18 'it'xd6 'it'xc4 19 i.b6 i.f8 20 'it'd4! 'it'c8 21 '.tί>hl won for White ίη Moutousis-Van Wely, Dortrnund 1992 but Black's play was poor; 14 ... dxe5 seems the most obvious altemative) 14••. b61S i.bl (D) and now:

4bl) IS••...tb7? (if Black wishes avoid the exchange οη f3 he should play 15 ... l:ιe8) 16 'ίWd2! (this move unexpectedly wins a pawn) 16...1tfc8 (after 16... ~xf3+ 17 gxf3 ~5 18 'it'xd6 ~xf3+ 19 '.tί>g2llJh4+ 20 rj;g3!, 20 ...'iWxd6 loses the exchange and οη 20 ... 'it'c8 White can play 21 rj;xh4 with impunity) 17 'it'xd6 'it'xd6 18 1:Ixd6 ~xf3+ 19 gxf3 ..tc6 20 ~e21eft Black with ηο compensation for the pawn ίη Armas-Bosboom, Groningen 1991. 4b2) IS... ~xf3+ 16 "xf3 lbes 17 'iVe2 ..tb7 (17 ...1:Ib8 is similar) 18 'iVd2 l:ιfd8 and now: Ιο

Καπ

4b21) 19 .tg5 f6 (better than 19 ... 1:ιd7 20 h3, intending f4, with advantage Ιο White ίη the game Armas-Ionescu, Romanian Ch 1988) 20 .te3 :b8 (ηοι 20 ... g5, 21 ~a4!) 21 h3 g5 22 lΔe2 was unclear ίη Wolff-Sadler, London 1991. 4b22) 19 .te3 was answered by 19...:ab8 20 h3 .ta8 21 f4lΔc6 22 'ίi'f2, which gave White an edge ίη Armas-Gheorghiu, Romanian Ch 1987. Perhaps Black could have tried 19..•lΔg4. He did well, however, Ιο avoid 19....tf8, which runs ίηΙο 20 lΔa4 ~d7 21 lΔxb6! lΔxb6 22 'iWa5 ~d5 23 :xd5!. 4b23) 19 h3 .tf8 20 .te3 :ac8 21 .tg5!? .te7 22 j,xe7 'iixe7 23 f4 ~c6 24 'iif2 with some advantage tΌrWhίte, Velίcka-Bosboom, Groningen 1991. 7 c4 (D)

7 ... .te7 Other moves: 1) 7... g6 8lΔc3 .tg7 9lΔb3! ο-ο lO .te2 ~c6 11 .tf4 lΔe8 (Black has little choice as 11 ... lΔe5 12 c5 is very awkward) and after 121i'd2 b6 13 1:ιfdΙ lΔe5 14 :acl 'ilc7 Black

Variation 137

equalίsed ίη Nunn-Gheorghiu, Υί enna 1986. However 12 c5! is unpleasant for Black, for example 12...dxc5 13 .xd8 ~xd8 14lΔa4! or 12...e5 13 .te3 .te6 14 ~5. 2) 7....td7 8 ~c3 ~c6 9 lΔxc6 .txc6 10 "e2 .te7 11 b3 0-0 12 .tb2 and now: 2a) 12...1:ιc8 13 ΦhΙ ~7 14 f4 .tf6 15 1:ιadΙ 'iic7 16 .tbl b5 17 cxb5 axb5 18 e5 dxe5 19 lΔxb5 .txb5 20 'iixb5 :fd8 21 f5! was very good for White ίη HowellAldama, Capablanca mem 1993. 2b) After 12..:ir'b8 13 a4 :e8 14 1:ιaeΙ ~d7 15 f4 .th4! 16 :dl .tf6 17 b4!? the posi ιίοη was unclear ίη Nunn-Bischoff, Dortmund 1987; 14 Ι4! was more accurate, ηοΙ committing the al-rook for the moment. 2c) 12...1:ιe8 13 f4 d5!? 14 cxd5 exd5 15 e5lΔe4 16lΔxe4! dxe4 17 .txe4 .tb5 and now 18 '6'g4?! 'ΙΜ2! 19 .txh7+!? was unclear ίη Arnason-Toshkov, Jurmala 1987. White can draw by 18 '6'c2 .txf1 19 .txh7+ 'itth8 20 ΦΧf1 1:ιc8 21 'it'd3! g6 22 j,xg6 hxg6 23 'it'xg6 'it'd2, but 18 "f3!? is the most promising, ηοΙ οηlΥ playing for a possible attack by 'ii'h3, but also lίning up against the b7-pawn. 3) 7... b6 normally transposes Ιο lines considered below. 8 ~c3 White has two main attackίng plans, which are distinguished by the development of his queen's bishop. Firstly he may buίld up a slow kίng side attack by b3, .tb2, lΔc3, 'it'e2, f4, 1:ιaeΙ and so οη, with the ultίmate

138

Kαn Vαriαtion

aim of a breakthrough by f5. The other plan is Ιο prepare for e5 by lΔc3, 'ife2. f4, ~d2 and J:[ael. The important ροίηι is that with the bishop οη b2 the e5 plan is much less effective, because White ends υρ with a pawn οη e5 and this would block the bishop οη b2.

8 ...

ο-ο

9 'fIe2 Moves such as 'ife2, ~hl and f4 are logical because they do ηοΙ commit White ιο one plan or the other. ΜΥ (JN) view is that the e5 plan is most effective against ... ι!LJbd7 by Black, because then the queen's knight blocks the retreat of the one οη f6. Therefore ίι is often useful Ιο delay committing the c l-bishop υηιΗ Black has moved his b8-knight. Against ... 1Δc6 White will take οη c6, then play b3 and ~b2, and against ... lΔbd7 White will continue with ~d2 and J:[ael. 9 ... b6 Black also delays for as 10ng as possible. Αη example where he didn't is Vehi-Robovic, Biel 1993, which continued 9...ι!LJbd7 10 f4 'ifc7 11 ΦhΙ J:[e8 12 ~d2 (12 οο!?) 12... ~f8 13 J:[ael e5 14 fxe5 dxe5 15 lΔf5 Φh8 16 ~g5 with a strong ίη itiative for White. 10 Ι4 Despite the above (rather subjective) coιnments ίι is quite reasonable ιο play b3 straight away, the advantage being that White can sometimes manage without ΦhΙ. After 10 b3 ~b7 11 .tb2 (D) there are two variations:

1) 11 ..•ι!LJc6 (as the earlier explanation makes clear. this move plays ίηιο White's hands since we reach positions similar Ιο the maίn line be10w, but with White having saved about half a tempo by missing ουΙ Φh1) 12lΔxc6~xc6andBlackhas

been highly unsuccessful from this position: la) 13 :ad1 'ίi'b8?! 14 a3 (what οη earth is this for?) 14 ... J:[d8? (ίι doesn't matter about the tempo spent οη a3 ίη view of the way Black plays) 15 f4 ι!LJd7 16lΔd5! ~f8 17 J:[f3! J:[e8 18 J:[h3! g6 19 'ifg4 'ifd8 20 J:[f1 ~B7 21 ~xg7 Φχg7 22 f5! with a massive attack, Ivanovic-Ermenkov, Plovdiv Echt 1983. lb) 13 f4lΔd7 14 J:[adl b5? 15 cxb5! axb5 16 ~xb5 'ifb6+ 17 J:[f2! ~xb5 18 'ifxb5 'ifxb5 19 lΔxb5 J:[xa2 20 ~xB7 J:[fa8 21 J:[xa2 J:[xa2 22 ~d4 e5, Hellers-Adamski, Eek10 1985, and now 23 ~c3 gives White a wonending. lc) 13 J:[ae1 J:[e8 14 f4 g6 15 e5! dxe5 16 fxe5 ~c5+ 17 Φh1lΔg4 18 ~e4! ι!LJxe5 19 ~xc6 ι!LJxc6 20 'iff3 with a tremendous attack for White, Ermenkov-Gheorghiu, Prague 1985.

Kαn Vαriαtion

2) 11...tΔbd7 12 :adl (12 f4 is ulso playable) 12...:e8 (12 ...'ii'c7 13 .1bl :fe8 14 f4 :ac8 15 tΔf3 ~f8 16 <iIth 1 ~c6 is less accurate, and 17 c5! ~xf3 18:Xf3 dxe5 19 fxe5 tΔg4 20 :xf7! gave White a very dangerous attack ίη Plachetka-Ravikumar, Copenhagen 1980) 13 ~bl 'ii'b8 14 1"4 ~f8 15 <iIthl :a7 16 tΔf3 ~a8 with a double-edged position, λkes ~on-Mestel, Copenhagen 1980, although Ι (JN) still favour White. 10 ~b7 11 <i'hl (D)

11 .•.

tΔc6

Or: 1) 11 ...tΔbd7 12 ~d2 'ii'c7 13 .l:r.acl (13 :ael :fe8 14 :f3 g6 15 .l:r.g3! <iIth8 16 :h3 e5 17 tΔf3 exf418 i.xf4 ~f8 19 'ii'f2 tΔc5 20 ~c2 was also good for White ίη IvanovicPeev, Balasiha 1977) 13 ... g6 14 b4 .l:r.ac8 15 a3 'ii'b8 (the advantage of playing f4 is that the weakening of c4 created by White playing b4 cannot be exploited by ... tΔe5) 16 tΔf3 :fe8 17 :cel ~f8 18 tΔg5 h6 (18 ... e5 19 f5 gave White a dangerous attack at ηο material cost ίη the

139

game Commons-Najdorf, Lone Pine 1976) 19 tΔxf7! <iItxf7 20 e5 tΔg8 21 'ii'g4 tΔe7 22 ~xg6+! tΔxg6 23 f5 tΔdxe5 24 fxe6+! (24 fxg6+ <iItg8 is unclear) 24 ...Φe7 (24...Φg7 25 :xe5 dxe5 26 :f7+ mates) 25 'ii'xg6! (the climax of a magnificent combinaΙίοη) 25 ... <i'd8 26 .l:r.xe5 dxe5 27 ~xh6 :xc4 (27 ... ~xh6 28 :dl+ mates) 28 'ii'xe8+ (White gives υρ his queen after all!) 28 ... <i'xe8 29 :xf8+ <i'e7 30 :xb8 ~c6 31 tΔdl b5 32 <iItgl <i'xe6 33 :b6 1-0 Commons-Peev, Plovdiv 1976. 2) 11 ...:e8 (dubious as ίι allows White Ιο play for e5 without delay) 12 tΔf3 g6 (12 ... tΔbd7 13 e5 dxe5 14 fxe5 tΔg4 15 J.f4 J.xf3 16 :xf3 and 12 ... tΔc6 13 e5 dxe5 14 fxe5 tΔg4 15 ~f4 are good forWhite) 13 e5! tΔh5, Sax-Bellon, Dubai OL 1986, and now Sax recommends 14 J.e4 tΔc6 15 g4 tΔg7 16 f5!, when 16...exf517 gxf5 dxe5 18 fxg6 hxg6 19 tΔxe5, 16...gxfS 17 gxf5 tΔxf5 (or 17 ...exf5 18 ~d5) 18 J.xf5 exf5 19 tΔd5 and 16...dxe517fxg6f5 (or 17 ... hxg618 tΔxe5) 18 gxf5 exf5 19 ~d5+ <i'h8 20 tΔxe5 are all good for White. 12 tΔxc6 ~xc6 13 b3 tΔd7 Or 13.. :'-c7 14 J.b2 :ad8 and now: 1) IS .l:r.adl g6 16 J.bl tΔh5 17 :d3 (17 g4 tΔg7 18 f5! exf5 19 gxf5 was probably better) 17 ...e5! 18 fxe5 dxe5 19 :xd8 :xd8 20 tΔd5 with just an edge for White, ΜatuΙοvίό Τήηgοv, Vmja~ka Banja 1986. 2) IS :ael (threatening tΔd5) 15 ... ~b7 16 ~bl tΔd7? (16 ... g6 is

140 Kan Variation

probably better, when 17 "d3 is οηlΥ a slίght plus for Whiιe) and now 17 'ii'h5 1:tfe8 18 :e3 ~ί6 19 'iί'h3 g6 20 ί5! gave White a decisive attack ίη Nunn-Gheorghiu, Bie11983, which is annotated ίη detaίl ίη Secrets ofGrandmaster Ρ/αΥ by John Νυηη and Peter Griffiths. For some reason Gheorghiu repeated the entire lίne ίη the game Mokry-Gheorghiu, Prague 1985. That game continued 17 'ji'g4 Jl.f6 18 :le3 g6 19 :h3 and then 19••• ~h8?? 20 ~d5! winning, as 20 ... exd5 is met by 21 "h4. Οί course 19...:lfe8 is better, but 20 'iί'g3 intending ί5 gives White a dangerous attack ίη any case.

chances) 17 :d3 g6 18 :fdl tΩc5 19 ~d5! exd5 20 .txf6 ~xd3 21 exd5! ~xί4 22 "f3 ~h5 23 .txd8 :xd8 24 dxc6 White had a clear advantage ίη the game Marjanovic-Rajkovic, Yugoslav Ch 1983. 15 :adl The purpose οί this (rather than :ae 1) is Ιο prevent the development ofBlack's e7-bishop ιο the 10ng diagonal. After 15 ... .tf6, forexample, 16.tbl "c7 17 "d2 attacks d6. 15 16 .tbl (D)

:e8

14 iιb2 (D)

14 ... g6 14....tf6 leaves d6 weak and after 15 :adl "c7 (15 .. .'fί'b8 16 .tbl b5 17 cxb5 axb5 was Maίnka-Espig, Bundeslίga 1991, and now ίι seems ιο me (JG) that the sacrificialline 18 e5! dxe5 19 :xd7 .txd7 20 "d3 :d8 {20... e4 21 ~xe4!} 21 'iί'xh7+ ~f8 22 lDe4 would have left White with a very powerful attack) 16 .tbl :fd8?! (16 ... :ad8 offered better

16 ... "fIc7 The altematίves are: 1) 16...:a7 17 a4 (the dίrect 17 :d3 was also tempting) 17 ... .tf8 18 .tc2 "a8 19 ~c5? (19 ...:c7 was better) 20 ~d5! exd5 21 cxd5 .txa4 22 bxa4 .tg7 23 .txg7 Wxg7 24 e5! and White stands well, Ρορ ovic-Kotronias, Pucarevo 1987. 2) 16...'ii'b8 17 ί5 (17 a4!? is possible, but Ι (JN) like 17 "d3 b5 18 cxb5 axb5 19 tl)e2! b4 20 tl)d4 .tb7 21 "h3 with dangerous threats οη the kίngside) 17 ... b5 18 fxe6 fxe6 19 cxb5 axb5 20 ~d5!? .txd5 (and

"f2

Kan ηοΙ 20 ... exd5 21 exd5 j.f6 22 "d2 .txb2 23 dxc6 with advantage Ιο White) 21 exd5 e5 22 "g4 ll:Jf6 23 "e6+ ~g7 was unc1ear ίη PrasadGheorghiu, Bie11985. 3) 16••.J.f8 17 e5! dxe5 18 j.e4! gives White a dangerous attack. The game Vogt-Gheorghiu, E.GermanyRomania 1984 continued 18 .....c7 (l8 ... j.xe4 19l1:Jxe4 j.g7 20 j.xe5 .txe5 21 fxe5 rl;g7 22 'ίi'Ω wins after 22 .....e7 23 'ίi'f6+! rl;g8 24 "xe7 ':xe7 25 ':xd7 or 22 ....:e7 23 ll:Jg5 "g8 24 ll:Jxf7 ':f8 25 ':xd7!) 19 'ii'f3 j.xe4 20 ll:Jxe4 f5 (20... j.g7 21 fxe5 ll:Jxe5 22 ll:Jf6+ rl;h8 23 j.xe5 wins mateήal) 21':xd7 fxe4 22 "dl 'ii'c6 23 .i.xe5 ':ac8 24 "d4 j.e7 (24 ... .:e7 25 ':d6 drops the b-pawn, while 24 ... b5 25 "a7! and 24 ... j.c5 25 ':g7+ rl;f8 26 j.d6+ 1ead Ιο mate) 25 f5! (the immediate 25 j.h8 .i.f8 26 ':g7+ rl;xh8 27 ':xg6+ is met by ...e5, but if B1ack now p1ays 25 ... exf5 this line wins) 25 ....:cd8 (25 ... gxf5 26 j.h8 j.f8 27 ':xf5! exf5 28 ':g7+ mates) 26 ':xe7! (26 fxg6! ':xd7 27 gxh7+ rl;xh7 28 ':f7+ rl;h6 29 'iνe3+ j.g5 30 "h3+ a1so wins) 26 ... exf5 (26 ... .:xd4 27 ':g7+ is mate next move) 27 ':xe8+ "xe8 28 "xb61-0. 17 "d3 Α flexible move attacking d6 and preparing 'ίi'd4 or 'iνh3 according Ιο circumstance. 17 .i.f8 (D) This move cost B1ack forty minutes, presumab1y checking that the line 18 'ίi'd4 e5 (18 ....i.g7 19 'iνxg7+ wins) 19l1:Jd5 j.xd5 20 "xd5 exf4 0.0

Varίation

141

21 1:xf4ll:Je5 22 1:df1 1:a7 presented dangers.

ηο

18 ll:Jb5!? shock for B1ack. The ροίηι is that after 18 ... axb5 19 "c3 e5 20 cxb5 1:ac8 21 bxc6 "xc6 22 'ίi'f3 B1ack cannot exchange at f4 since he has ηο satisfactory way Ιο cover f7, so White gets Ιο p1ay 23 f5, when B1ack's light squares look very sickly. 18 ... .i.xb5 19 cxb5 axb5 19...85 20 :cl ll:Jc5 21 'iνc3 is much worse as B1ack has Ιο p1ay the weakening ... e5. 20 "xb5 'ίi'b7 The upshot ofWhite's mini-combination is that he has the two bishops and a queenside pawn majοήtΥ. Now 21 a4! would have been 10gica1, relieving the bl-bishop of its defensive duty and pinning down the b6-pawn. If B1ack remains passive the bl-bishop can eventually move Ιο b5. Duήng the game Ι was concemed about the rep1y 21 ... ll:Jc5 22 "c4 1:ιac8, but 23 f5 e5 24 :f3 gives B1ack ηο way Ιο exp10it the Α

142

Kαn Vαrίαtion

position ofWhite's queen (24 ... lL!e6 25 fxe6!). and consolidation by 25 'ife2 will be good for White. b5! 21 1i'e2?! Black takes the chance Ιο prevent a4. Now ... b4 would permanently cripple White's queenside pawns, so ... 22 b4 Black cannot transfer his knight ιο c4 because b5 is weak, for example 22•••"a6 (threatening ... lL!b6) 23 .td3 'ifxa2 24 .txb5 'ifa7 (24 ...:ed8 25 .txd7 and :al) 25 .tc6 followed by .td4 and b5, with an excellent ρο sition for White. :a6(D) 22 ••• Another useful defensive move, cοveήng d6 ίη preparation for ....tg7. Here Ι thought for a Iong time trying ιο find a way ιο keep the advantage.

23 :f3! 24 .txg7 25 .J:[fd3

.tg7 Φχg7

1i'c7

:Jι3

The ροίηι οί White's manoeuvre is that he gains a tempo by attacking b5 to transfer his rook ιο the h-fiIe,

thereby reviving his threats against BIack's king. 26 ... :b6 Νοι 26...1i'c4? 27 'ifb2+ followed by 28 J:tc3, when BIack's queen is suddenIy trapped. 27 1i'd2 The threat is 28 e5 d5 29 ί5 and the queen gets ιο h6. 27 ••• lL!f6 Once again Black finds a good defence, relieving the ρίη down the dfile. 28 f5 exf5 28•••e5 29 'ifh6+ 'iPg8 30 g4! is very unpleasant. d5! 29 exf5 ΑIl these difficuIt moves were very time-consuming, so that BIack had only a couple οί minutes left ιο reach move 40. Αι first sight White can win by 30 fxg6 fxg6 311i'h6+ 'iitg8 32 .txg6 hxg6 33 'ifh8+ (33 'ifxg6+ 'ifg7 34 'iff5 :ce6 defends) 33 ...Φf7 34 :h7+ Φe6 35 :el+ (35 'ifxf6+ ΦΧί6 36 :xc7 :be6 gives BIack enough for the pawn ίη view οί his active rooks and king), but then comes 35 ...Φί5! 36 :f1+ Φg5! 37 h4+ Φg4 and White's queen is trapped. 30 fxg6?! The immediate 30:rt! was much better, when the defence BIack pIays ίη the game would have been prevented. 30 ••. fxg6 31 :η Now truly threatening 32 'ifh6+ Φg8 33 .txg6. 31 ••• lL!g4!

Kαn Vαriαtion

Suddenly Black exploits White's vulnerable back rank. Ιί 32 'ii'xd5? Ihen 32 ... it)f2+ wins. 32 'ii'd4+ :16 (D)

White cannot take the rook or the knight, nor can he play :hf3. The move played supports the weak back rank and simultaneously attacks the knight. 33 .td3 h5 Forced, but effective. The draw would now be ίη sight ίί ίι were ηοΙ for Black's time-trouble. 34 :xf6 ~Ι6 35 :13 :el+ 36 .tn 'ii'e7?! 36•••'ii'e5 37 'ii'xe5 :xe5 38 .txb5 gives White good winning chances as the d-pawn cannot pass over the d3-square. However, 36•••'ii'd6! was more accurate, for example 37 ~gl (37 h3 :e4 38 'ii'c3 :xb4 39 :xf6 'iί'xf6) 37 ...:e4 38 'ii'c5 (38 'ii'c3 d4 39 'ii'd3 it)g4! 40 g3 :el threatening ... it)e5 is unpleasant) 38 ... 'ii'xc5 39

143

bxc5 :el and the draw is inevitable (40 ':c3 d4). 37 ~gl :e4 Black can also choose ιο wait, since progress isn't easy for White, but ίη time-trouble ίι is very natural Ιο break the ρίη. 38 'ii'c3 d4 Νοι 38•••:Xb4 39 :xf6 d4 because of 40 :e6!. 39 'ii'd3 With two moves still Ιο make Black suddenly finds himself ίη trouble over the b5-pawn. 39.••'ii'xb4 loses Ιο 40 :xf6, so the best chance is 39••• it)d5 40 'ii'xb5 it)xb4. Then 41 .tc4 :el+ 42 'ittf2 ~h6 isn'tdangerous, so White's winning prospects are very slight. it)g4? 39 ••• 40 h3 and Black's flag dropped before he could make his 40th move. After 40... it)e3 41 '6'xb5 it)xf1 42 :xf1 :e2 43 '6'd3 '6'e3+ 44 '6'xe3 dxe3, suggested by Black after the game, White can win by 45 a4! (but ηοΙ 45 b5 :xa2 46 :bl :d2! 47 ~f1 :f2+ 48 ~el :xg2 49 :b4 :d2 50 b6 :d8 51 b7 :b8 when Black has drawing chances) 45 ...:b2 46 b5 :a2 47 b6 :xa4 48 :bl :d4 49 ~f1! (49 b7 e2) 49 ...:d8 50 b7 :b8 51 ~e2, etc. Obviously Black has other ways Ιο play, but the two connected passed pawns give White good winning chances ίη any case.

7 Maroczy Bind This most commonly arises ifBlack plays an early ... g6, for example 1 e4 c5 2 lLJf3 lLJc6 3 d4 cxd4 4lLJxd4 g6, aiming to reach a Dragon position without having played ... d6. This restricts White's options quite severely, since he has Ιο be carefuI ηοΙ Ιο allow Black Ιο play ... d5 ίη one go, saving a cruciaI tempo. Αη effective way ουΙ of this diIemma is Ιο continue 5 c4, setting υρ the formation of pawns οη c4 and e4 known as the Maroczy Bind. The asset ofthis formation is the automatic restraint of ... b5 and ... d5, Black's basic freeing thrusts. Black does sometimes succeed ίη organising ... b5, but this is normally only good when White has made a mistake. White's main asset is hίs space advantage, Ieading Ιο the corolIary that he should avoid exchanges which would relieve the cramp ίη Black's position. If Black does succeed ίη liquidating Ιο an ending, White's c4-pawn and darksquared weaknesses can become a ιi ability. Play often becomes a matter of slow manoeuvring as White tries Ιο increase his space advantage and force weaknesses ίη the black posiιίοη while his opponent remains crouched οη the fιrst couple of ranks waiting for the first sign of over-extension ιο Iaunch a counterattack. The Maroczy Bind can also occur

if Black adopts an unusual move order, for example 1 e4 c5 2 lLJf3 lLJc6 3 d4 cxd4 4lLJxd4 d6 and now 5 c4 wiII probably transpose Ιο this chapter after 5 ... lLJf6 6lLJc3 g6. ΑΙ though the Maroczy Bind ί:. sIίghtly passive for Black, players such as Larsen, Petursson and VeIίmirovic have shown that by patiently waiting for a lapse of concentration from White this line can offer winning chances for Black. The theoreticaI ορίηίοη is that White should maintain a sIίght advantage, but ηο one should beIίeve that this is a line ίη White cannot lose. Game20

Karpov - Kavalek Nice ΟΙ 1974 cS

1 e4 2 lLJf3

lί)c6

3 d4

cxd4

4 lLJxd4

5 c4

Here Black has two possibilities. He may play 5 ... ~g7 ίη order ιο force White's ~e3, but ίη doing this he forfeits the chance Ιο take οη d4 at a moment when White must recapture with the queen. The alternative is 5 ... lLJf6, which wilI often transpose ιο 5 ... ~g7 if Black does ηοΙ take υρ the chance Ιο play ... lLJxd4.

Mαroc'Y ΒίΜ

The 5 ... Jιg7 systems are examined ίπ Game 21, whίle ίη this game we look at 5... tΔί6. 5 tΔt'6 6 tΔc3 (D)

6 ... d6 If Black adopts the move order 6 ... tΔxd4 7 'iWxd4 d6 (possibly Ιο rule ουΙ tΔc2) then White has an ίη teresting alternative based οη playing Jιd3 rather than Jιe2, providing the e4-pawn with useful extra protection, for example 6 ••• tΔxd4 7 'iWxd4 d6 8 Jιg5 i.g7 9 'i'd2 ο-ο (9 ... Jιe6 10 :tcl :tc8 11 b3 'iWa5 12 [3 h6 13 Jιe3 ο-ο 14 Jιd3 Φh7 15 ο-ο a6 16 h3 tΔd7 17 ί4 ί5 18 exf5 Jιxf5 19 Jιe2! was good for Whίte ίη Polugaevsky-BeIίavsky, USSR Ch 1975) and now: 1) 10 Ι3 is interesting, for example after 1O ... Jιe6 11 :tcl 'ila5 12 b3 and now rather than 12•• 0861 13 tΔd5! 'iWxd2+ 14 ΦΧd2 Jιxd5 15 cxd5 :tfc8 16 :txc8! :txc8 17 g3 'ίtί>f8 18 Jιh3 :tc7 19 :tcl :txcl 20 'ίtί>xc 1 which left White winning ίη Byrne-GarciaPadron, Tonemolinos 1977, 12..•:tfc8 was better, when 13

145

Jιe2 a6 14 tΔa4 transposes ιο Kar-

pov-Kavalek. 2) 10 Jιd3 Jιe6 (10 ... a5 11 ο-ο a4 12 I1acl Jιe6 13 'iWc2 gave White his usual comfortable space advantage ίη Portisch-Reshevsky, Petropolis ΙΖ 1973) 11 ο-ο (more accurate than 11 I1c 1 as ... 'iWa5 would now allow tΔd5) 11 ...:c8 12 b3 a6 13 I1acl tΔd7 (perhaps 13 ...:e8 is better, with the idea of meeting 14 ί4 with 14... b5, although even here it's unclear whether Black has sufficient compensation for the pawn after 15 cxb5 axb5 16 Jιxf6 Jιxf6 17 tΔxb5; if this position is ηοΙ Ιο White's liking then he can try 14 ΦhΙ or 14 :fel) 14 ί4 [6 15 Jιh4 Jιh6 16 tΔd5 tΔc5 17 Jιbl b5 18 cxb5 axb5 19 b4 tΔa4 20 :xc8 Jιxc8 21 'iWe2 e6? (21 ... Jιd7 is preferable, when White should ηοΙ play 22 e5 dxe5 23 fxe5 οη account οί 23 ... i.c6!, but 22 [5 with advanιage) 22 'ilxb5 Jιd7 23 tΔxf6+! 'iWxf6 24 W'xd7 'iWxh4 25 g3 W'g4 26 'iWxa4 Jιxf4 27 W'b3 with a clear advantage for White, PsakhisKagan, Tel Ανίν 1992. 7 Jιe2 tΔxd4 This is Black's last chance Ιο force White Ιο recapture οη d4 with hίs queen. If he plays 7 •••Jιg7 White should transpose Ιο Game 24 by 8 i.e3. Αι one time 8 tΔc2 was thought the best reply Ιο 7 ... i.g7, but after 8 ... tΔd7 9 i.d2 a5! (ηοΙ 9...tΔc5?! 10 b4 tΔe6 11 :c 1 ο-ο 12 ο-ο ί5 13 exf5 gxf5 14 ί4 tΔed4 15 tΔxd4lbxd4 16 Jιe3 with a positional advantage for White, Nunn-Rind, Manchester 1980) 10 ο-ο ο-ο 11 :cl tΔc5 Black

146

Mαroczy

Bind

has a much better version of the Maroczy Bind than ίη other lines. Ιη Nunn-Petursson, Wijk aan Zee 1990 Ι made matters worse by playing 12 b3? lt:Jb4!, when Black had a clear advantage. .tg7 8 "xd4 9 .tg5 (D)

9 •••

ο-ο

There is little Ιο be gained by delaying castling: 1) 9•••.td7 10 "d2 h6 (1O ....tc6 should be met by 11 f3) 11 .tf4 a6 12 ο-ο .tc6 13 f3 ο-ο 14 a4lt:Jd7 15 a5! lt:Jc5 16 ':a3 with a slight advantage for White, Pomar-Cordovil, Malaga 1972. 2) 9...h6 10 .te3 ο-ο 11 "d2 Φh7 (11 .....a5 12 ο-ο forces ... Φh7 ίη any case) 12 ο-ο .te6 13 Ι4 (13 .td4 ':c8 14 b3 a6 15 "e3 lt:Jd7 16 .txg7 Φχg7 17 f4 "b6 18 "xb6 lt:Jxb6 19 f5 .td7 20 ':adl οηlΥ gave White a slight edge ίη TimmanRibli, Amsterdam 1973) 13••..:c8 (after 13 .....a5 14 f5 forces 14 ....td7 since 14 ... gxf5 15 exf5 .txf5 allows 16 ':xf5, so White avoids wasting a tempo οη b3) 14 b3 (D) and now:

2a) 14•.•a6 15 :adl 'iί'a5 16 ~d4 is good for White after both 16••. b5 17 f5 .td7 18 .txf6! exf6 19 lt:Jd5 'iί'xd2 20 :Ιχd2, Nunn-van der Sterren, Groningen 1974/5 and 16•••.td7 17 .txf6 exf6 18lt:Jd5 'iί'c5+ 19 ΦhΙ a5 20 f5, Nunn-I.Ivanov, London 1987, so it makes sense for Black to try Ιο force White's rook ιο the less active square c 1. 2b) 14••."a5 15 a3 (15 ':acl is also possible, aIthough 15 ... a6 16 f5 .td7 17 h3 .tc6 18 .td3 lt:Jd7 was unclear ίη Gulko-Petrosian, Biel ΙΖ 1976) 15 ... a6 16 f5 ~d7 17 b4 "e5 (Larsen-Fischer, Denver Ct (2) 1971) and now, according to vaήous analyses of this famous match, 18 ':ad 1 would have been good for White. 3) 9•.. ~e6 10 ':cl 'iνaS 11 "d2 ':c8?! (11 ... 0-0 is more accurate) 12 Ι3! leaves Black ίη a rather awkward situation since he has committed the wrong rook to c8. The point is that after ... 0-0 and ... ':fc8 White is never threatening Ιο play lt:Jd5, because the sequence .....xd2, lt:Jxe7+ Φf8 costs White his enant knight. With the a8-rook οη c8, however,

Mαroczy ΒίΜ

Black wilI sooner or later have Ιο waste time meeting this threat. 3a) So, 12... 0·0 13 b3 a6 14 ο-ο is good for White. 3b) It is also Ιοο dangerous Ιο snatch the c-pawn, e.g. 12....i.xc4 13 ι!t)d5! Wxa214 0·0 ι!t)xd5 15 ':'xc4! ':'xc4 16 Wxd5 ':'a4 17 .i.b5+ Φf8 18 ':'cl! and ίη GelIer-Stean, Teesside 1975 Black resigned because of 18... .i.d4+ 19 "xd4! ':'xd4 20.i.h6+ mating. Ιη fact White had an even more convincing win by 14 ι!t)b4, since 14 .....b3 15 .i.dl traps the queen. 3c) IfBlack doesn't cast1e he will 500η run ουΙ of things Ιο do, e.g. 12... a613 b3 b5 14 οο! "xd2+ 15 Φχd2.i.χd5 16 cxd5 Φd7 17 a4 h6 18 .i.e3 ':'xc 1 19 ':'xc 1 ':'b8 20 ':'c6 with a very good ending for White, Nunn-Reuben, London 1978. 10 "d2 10 "e3 is an important alternative which has the advantage of firmly defending the e-pawn; this makes the Iines with a quick ... a6 and ... b5, which can easily cause problems after 10 "d2, much less effective. After 10....i.e6 11 ':'εΙ!? (D) Blackhas: 1) 11...a612 0-0 b5 13 cxb5 axb5 14 a3! Wb8?! (14 ... is suggested by Petursson, but he hasn' t repeated this line) 15 .i.xb5 and Black didn't have enough for the pawn ίη Arnason-Petursson, Reykjavik 1990. 2) 11 .....b6 12 'ίWd2 "b4 13 Ι3 and now: 2a) 13....i.xc4 14 a3 Wb3 15 .i.xc4 Wxc4 16llJd5 'ii'b3 17 ι!t)xe7+

.:.a5

147

Φh8 is given as unclear by Petursson, but Ι (JG) prefer White. Α sample variation: 18 ο-ο Wb6+ 19 ':'f2 ':'fe8 20 .i.xf6 .i.xf6 21 00 'ii'd4 22 ':'c7±. 2b) 13....:.ιε8 14 b3 a6 15 ι!t)a4 'ii'xd2+ 16 ΦΧd2 and we are back ίη Karpov-Kava1ek, a1though judging from the notes ίη various sources nobody else seems Ιο have noticed. Dona1dson and Silman, for example, ίη Accelerαted Drαgons (Cadogan 1993) give 11 ... Wb6 and 12.. .'ίi'b4 exclamation marks and consider the position after 16 ΦΧd2 Ιο be equal, whilst two pages earlier ίη their book they give the same position as better for White (albeit after 15 ΦΧd2). 3) 11.....a5 12 ο-ο and now: 3a) 12....:.fc8 13 b3 a6 14 f4 'ii'c5 15 'ii'xc5 ':'xc5 16 .i.f3 ':'b8 17 e5 (17 tba4 followed by c5 is also interesting) 17 ... dxe5 18 fxe5 ~8 19 .i.xe7 ':'xe5 20.i.h4 with a clear advantage for White, Ciocaltea-Spiridonov, Tίmisoara 1982. 3b) 12...1Ue813 b3lbd714 .i.h6 (White's problem is that 14 f4 'ii'c5 15 Wxc5 ι!t)xc5 is very awkward for him and after 14 ':'fdl he has Ιο

148

Mαroczy ΒίΜ

take ίηΙο account 14....txc3 and 15 .. :ii'xa2) 14...'iic5 (14 ....th8 15 J:lfdl) 15 "-d2 'iid4 and now: 3bl) 16 'iVxd4 .txd4 17 ':fdl .tc5 18 ~d5 .txd5 19 exd5 a5 20 .td2 .tb4! with an equal game, Stangl-Espig, Dortmund 1991. 3b2) Ι (JG) prefer 16 ~b5! as after 16 .. :ii'xd2 17 .txd2 ':ec8 18 .te3 White retaίns an edge. 10 .•• .te6 Black cannot do without this move so 1O ... a6, for example, is of ηο independent significance and will transpose to lines considered below. 11 J:lcl Perhaps this move is ηοΙ the most accurate as ίη several variations when Black plays ... a6 and ...b5 he is able Ιο follow υρ with ....txa2. The main alternative is 11 ο-ο (D), with the following possibίlities:

1) 11 ...:c8 12 :adl!? 'iia5 (ηοΙ 12....txc4 13 e5 ~8 14 .txc4 ':xc4 15 'it'e2! ':c7 16 exd6 ~xd6 17 ~5 +-) 13 b3 a6 14 ~d5 'iVxd2 15 ~xe7+ Φh8 16 .txd2 1:tce8 17 .tb4 1:txe7 18 J.xd6 ':ee8 19 e5 ~g4 20 .txf8 ':xf8 21 .tf3 ':b8 22 h3 ~xe5

23 .txb7 a5 24 .td5 with a clear advantage for White, J .Polgar-Dzindzihashvili, New York 1992. 2) 11 ..•a6 12 ':adl!? (12 ':cl transposes ίηΙο the note Ιο Black's 11th move) 12... 'ii'a5 (12 ...b5 13 e5!? dxe5 14 'ii'xd8 ':fxd8 15 ':xd8 1txd8 16 cxb5 axb5 17 .txb5 should be better for White) 13 .txf6!? .txf6 (perhaps Black can try 13 ... exf6 14 'ii'xd6 f5) 14 ~d5 'ii'xa2 15 ~xf6+ exf6 16 'iic3! 'iia4 17 ':xd6 1tac8 18 b3 'iia3 19 ':fdl with advantage Ιο White, Tiviakov-Polak, Oakham 1992. Ιη fact, the game only lasted anotherthreemoves: 19 ...1tc720h4 Φg7 21 h5 gxh5? 22 1:txe6 1-0. 3) 11 .••'iVa5 12 [3 (12 :adl a6 transposes Ιο '2') 12 ... a6 13 ':fdl (13 tί)d5?! 'it'xd2 14 ~xe7+ Φh8 15 .txd2 ':fe8 16 tί)d5 tί)xd5 17 cxd5 .txd5 18 .tc3 1/2-1/2 Short-Andersson, Νονί Sad OL 1990) 13 ... ':fc8 14 ~d5 'iixd2 15 ':xd2 .txd5 (after 15 ... tί)xd5? 16exd5 .td7 17 .txe7! .th6 White should play ηοΙ 18 1:tc2 .tf5 and 19 ...1te8, but 18 1tadl! .txd2 19 1txd2 :e8 20 .txd6 with a clear plus for White, Βήttοn-Dοn aldson, Rhodes 1980) 16 cxd5 ~f8 and White' can only claim a very slίght plus. Ιη general these endings with an open c-file are very drawish unless Black has weakened his queenside by playing ...b5, which both allows a4 and gives White an entry ροίηΙ at c6 (see Nunn-Reuben, lίne '3' ίη the note Ιο Black's 9th move for a demonstration). Another way for White to obtaίn some pressure is Ιο seize the c-file by playing

Mαroczy ΒίΜ

his bishop ιο the h3-c8 diagonal. As neither situation exists here, White's advantage is insignificant and ίη practice BIack wouId have few problems reaching the draw. Now we return Ιο the main line after 11 1%cl (D):

11 ... 'iί'a5 11•.•a6 is more cήtical; White has severaI sensibIe Iooking replies, but against each one of them Black seems Ιο be abIe Ιο gain reasonable play with 12...b5, e.g. 1) 12 0·0 b5 13 cxb5 axb5 14 .ixb5 (against 14 a3, Andres proposes 14 ... 'it'a5 15 .ixb5 ~xe4 16 ~xe4 'iί'xb5 17 .ixe7 1%fb8 with enough compensation for the pawn) 14 .....txa2 15 .ic6 1%a6 16 ~xa2 :xa2 17 1%c2 'it'b8 with an equal game, Ortega-Andres, Cuba 1984. 2) 12 (3 b5 13 cxb5 axb5 14 a3 ~d7! 15 b4?! (15 ~b5 and .ixb5 are met by the awkward 15 ... ~c5, but are still probably better than this odd move) 15 ...1%xa3 16 tbxb5 1%a2 with a very active game for Black, Topalov-Antunes, Candas 1992. 3) 12 b3 and now:

149

:c8

3a) 12... with two possibilities for White: 3al) 130·0 b5 14 cxb5 axb5 15 ..txf6 (BTS2 recommended 15 .ixb5 "a5 16 .id3 Axc3 17 'ifxc3 {17 1%xc3 ~xe4! is good for Black} , unfortunately overlooking 17 .....xg5) 15 ....ixf616 ~xb5 'it'b6. This posiιίοη has been reached οη several occasions and ίι is unclear whether Black's bishop pair provides sufficient compensation for the pawn. The latest example, Gufeld-Konguvel, Calcutta 1994 suggests ηοι: 17 Μ!? 'ifb7 (perhaps Black should be looking Ιο improve υροη this) 18 a4! 'ii'xe4 19 ~7 Ab8 20 ~xe6 fxe6 21 Ac4 'ifd5 22 'it'c2 and White's passed pawns give him the advantage. 3a2) 13 13 b5 14 cxb5 axb5 15 ~xb5 (15 .ixb5 1fa5) 15...1%xcl+ 16 'iνxcl 'ifa5+ 17 "d2 1%a8 18 ~c3! (D) (this is an improvement over 18 'ii'xa5 Axa5 19 a3 .ixb3 =) and now:

3a21) Frolov-Tangborn, Kobanya 1992 continued 18...h6 19 .ie3 ~g4 20 .id4 .ixd4 21 "xd4 :c8

150 Maroczy ΒίΜ 22 Φd2 :xc3 (22 ... lΔe5 23 :cl is very good for White) 23 'iί'xc3 'i'xa2+ 24 ΦeΙ lΔe5 25 cJtf2 'ifa7+ 26 'ife3 when Black's compensation was obviously insufficient. 3a22) 18... is a more natura1 try, but this also seems good for White after 19lΔa4 :cl+ 20.tdl 'ifxd2+ 21 ΦΧd2 :al 22 lΔc3 h6 (22 ...Φf8!?) 23 .te3lΔd7 24 :gl!. 3b) 12••• b5! 13 cxb5 axb5 14 .txb5 (14 .txf6 .txf615lΔxb5 'i'a5 100ks fine for Black) 14 ...'ifa5 150-0 :ac8 16 .td3 :xc3 17 :xc3 lΔxe4 18 .txe4 .txc3 19 'i'e3 d5 with an excel1ent game for Black, LukovBrendel, Κrumbach 1991. If the above variations are really true then White should try either 10 We3 or 11 ο-ο. 12 f3 :fc8 13 b3 ΒΥ securely defending c4 and e4 White has prevented any tricks based οη an immediate ... b5 so Black has nothing better than Ιο prepare this thrust with ... a6. 13 ..• a6 14 lΔa4 White chooses a favourable moment Ιο exchange queens. Black's last move weakened b6 and he must waste a tempo preventing White's knight fork. 14 .•• 'iVxd2+ After 14.• :iνd8 White may play: 1) 15 c5!? (as suggested by Karρον) 15 ...:c6 (the main idea is that after 15 ... dxc5 16 'iί'xd8+ :xd8 17 lΔb6 :ab8, 18 .tf4 wins mateήa1) 16 Φf2! 'iί'f8 17 'ifb4lΔd7 18 cxd6

:c8

:xd6 19 'i'xb7 .td4+ 20 .te3 :b8 21 'iVc7 with advantage Ιο White, Κrasenkov- Hernandez, Palma 1989 as after 21 ... .txe3+ 22 Φχe3 Black has ηο way Ιο exploit the advanced position of the white king. 2) 15 .te3 is a quieter continuation which promises White an edge, e.g. 15 ... :ab8?! (15 ... lΔd7) 16 .ta7 :a8 17 .tb6 "f8 18 .te3 :ab8 19 lΔb6 :c7 20 ο-ο lΔd7 21 lΔd5 with an excel1ent position, Nunn-Blum, London 1979. 15 ΦΧd2 (D)

:c6

15 ... After Black 10st with this move ίη Karpov-Kava1ek attention turned Ιο 15•••lΔd7, but this doesn't seem Ιο be much better, for example 16 g4! (16 h4 f6 17 .te3 f5 al10ws Black more freedom) and now: 1) 16•••f617 .te3 f5 18 exf5 gxf5 19 h3! :f8 20 f4 d5 21 cxd5 .txd5 22 :hd 1 :ac8 23 gxf5 b5 24 :xc8 :xc8 25 ΦeΙ! with advantage Ιο White, Beliavsky-Tiviakov, Groningen 1993. 2) 16•••Φf'8 17 h4 :c6 18 :c2 1Δc5 191Δc3 a5 20 lΔd5 :e8 21 .te3

Maroczy ΒίΜ 151 i.c8 22 h5 e6 23 lΔc3 f5 24 hxg6 hxg6 25 exf5 exf5 26 gxf5 i.xf5 27 :cc 1 with a favourable ending for White, Averbakh-Popov, Polanica Zdroj 1976. Ιι is cuήοus that the exchange of queens is just what White needs Ιο start a kingside attack. The cxplanation is that Black's counterplay by ... b5 would be very dangerous with queens οη the board, since ίι would lead ιο an attack against the centralised white king. With queens off this counterplay is relatively harmless. 16 lΔc3 :8c8 Karpov suggested 16...:e8 as a possible improvement, so as Ιο trap White's bishop ίη case of 17 lΔd5 lΔd7 18lΔxe7+?! :xe7 19 i.xe7 f6 20 i.d8 b6. 16•••:cc8 was played ίη Sakharov-Pereira, Cοπ. 1976, which fίnished ίη a draw after 17 lΔa4 :c6!. Ι don't suppose Karpov would have agreed a draw if Kava1ek had 'found' 16...:cc8!. 17 lΔd5 ~f8 18 i.e3 lΔd7 Defending such an ending is an unpleasant task at the best of times, doubly so against Karpov. White has the choice of expanding οη the queenside by a tίmely Μ, or of gaining space οη the other flank by g4 and h4, as ίη the note Ιο Black's 15th move. Until White shows his hand Black can only wait. i.xd5 19 h4 Black resolves ΙΟ do away with the dangerous knight. 19•••h5 would be well met by 20 lΔf4 and Karpov's suggestion of 19•••f5 would require

strong nerves ίη view ofBlack's king position. 20 exd5 :6c7

21 h5

Φg8?! (D)

This move is probably a mistake. 21•..%te8 followed by ... e6 would have opened the position υρ for White's two bishops, but by activating his rooks οη the central fιles Black would have developed counterplay against White's king.

Ι4!

Most players would have rejected this as ίι allows Black's knight Ιο settle at e4 (supported by ... f5). Karρον, however, is actually aiming Ιο provoke ... f5, which gives him the lever g4 by which he can prise open Black's kingside. 22 lΔc5 23 i.g4 lΔe4+ 24 Wd3 Ι5 25 i.f3 b5 It looks as though Black's counterplay has got off the ground at last but White defuses ίι adroitly.

26 g4 27 1:Ixc4 28 bxc4

bxc4+ :xc4 lΔc5+

152

Mαroczy ΒίΜ

Black had little choice as he could allow the white rook Ιο occupy the b-file, nor could he play ...:b8 without losing a pawn at e4. 29 .txc5! If there are rooks οη the board ορ posite-coloured bishops tend Ιο lose their drawish influence. Here Black runs ίηΙο trouble because his king is badly placed and he will have two or even three pawns stuck οη light squares, where they cannot be defended by his bishop. :Ιχc5 29 .•. Β lack plays for a counterattack by ....:a5. 29 ...dxc5 30 h6 .td4 31 ':bl is also unpleasant, the a6-pawn being parιicularly weak. 30 h6 .tf8 It 100ks awful Ιο bury the bishop but Black lacked a reasonable alternative, for example 3O•••.tf6 31 :bl threatens 32 gxf5 gxf5 33 ':b8+ Φϊι 34 .th5#, or 3O•••fxg4 31 .txg4 .tf8 (31 ....tf6 32 .te6+ ~f8 33 :bl) 32 .te6+ Φh8 33 f5 :a5 34 :bl :a3+ 35 ~e2 ':xa2+ 36 ~f1 .txh6 37 f6 and the pawn slips through (37 ...exf6 38 :b8+ wins a piece). 31 Φc3 Karpov also analyses 31 g5 :a5 32 ':bl as good for White, but the variations are by ηο means simple and ίη practice it is ηοΙ surprising that he chose ιο prevent 31 ...':a5 by simple means (32 Φb3, and the rook has Ιο go back). 31 ••. fxg4 32 .txg4 Φt7?! (D) Black resolves Ιο extract his kίng, even at the cost of the h7 -pawn. Ιη ηοΙ

spite of its dangerous appearance, he would probably have done better ιο try 32•••:c7 33 .te6+ 'ίth8 34 f5 :b7! (preventing 35 :bl .txh6 36 f6 exf6 37 ':b8+), when White finds ίι hard Ιο make progress because of Black's attack οη the h6-pawn.

33.te6+ 'ίtι6 34 .tg8 :c7 34•.•.txh6 35 :xh6 ~g7 leads Ιο a 10st rook and pawn ending after 36 :xh7+ and 37 :xe7. 35 .txh7 e6 35••• ~7 is refuted by 36 f5 g5 37 f6! exf6 38 .tf5 ~g8 39 h7+ ~h8 40 :bl and 41 :b8. 36 .tg8 exd5 37 h7 Νοι 37 .txd5? :h7. 37 ..• .tg7? Loses by force. After37•••:xc4+ 38 Φd3 .tg7 39 .txd5 (39 h8'i1ί' .txh8 40 :xh8 ':c8 and 41 ... Φg7) 39 ...:c8 we reach the same position as ίn the game but with White having a pawn less. Black would still be worse, but he would have chances of adraw. 38 .txd5 .th8

Mαroczy ΒίΜ

39 <itd3 ΦΙ5 40 <ite3 :e7+ 41 Φf3 aS 42 a4 :'c7 43 .i.e4+ ΦΙ6 44 :Jι6 ':g7 44...Φg7 45 l1xg6+ <itxh7 46 :gl+ <ith6 47 :'hl+ and 48 :h7+ wins the rook. 45 <itg4 1-0 BIack is totalIy paralysed. Garne 21 Vaganian - Ivkov Moscow 1985

1 2 3 4 5 6

tΩf3

tΩc6

cxd4 g6 .i.g7

tΩxd4

c4 .i.e3 (D)

6 ••• tΩf6 BIack can aIso pIay ... tΩh6 foIIowed by ... f5, either with or without ... d6, but this idea does ηοΙ equalise: 1) 6•.. d6 7 tΩc3 tΩh6 8 .i.e2 ο-ο 9 ο-ο [5 10 exf5 gxf5 (10 ... tΩxd4 11 .i.xd4.i.xd4 12 'ίWxd4 tΩxf5 13 'ίWd2

153

.i.d7 was good for White after both 14 .i.f3, Tal-Kupreichik, Sochi 1970 and 14 .i.g4, VileIa-Estevez, Cienfuegos 1980 aπd 10... .i.xd4 11 .i.xh6 l1xf5 traπsposes Ιο '2') 11 [4 'ίWb6 (11 ... .i.d7 12 'ίWd2 tΩg4 13 .i.xg4 fxg4 14 tΩd5 gave White aπ edge ίη the game Szabo-Larsen, Vinkovci 1970) 12 tΩxf5 'ίWxb2 13 tΩxh6+ .i.xh6 14 :'cl (or 14 tΩd5 .i.f5 15 .i.d3 'ίWg7 16 <ithl with advaπtage for White, Yermolinsky-Chepukaitis, USSR 1980) 14 ... .i.g7 15 ':c2 (15 tΩd5 sacήficίng the a-pawn was possibIe) 15 ... 'ίWa3 16 'ίWd2 wίth an edge for White, Spassov-Nicevski, Sofia 1976. 2) 6 ••• tΩh6 7 tΩc3 ο-ο 8 .i.e2 Ι5 9 exf5 .i.xd410 .i.xh6 :'xf5 11 ο-ο d6 (D) (11 ... 'ίWb6 is answered by 12 tΩd5 !). This position was reached πο Iess than four times by Afifi ίπ the 1990 Maπίla Interzonal.

2a) Two games continued 12 .i.f3: 2al) The first of the two games, against Ηϋbner, was a bit of a disaster. There followed 12..•.i.g7 13 .i.xg7 (13 .i.e3 is aIso quite good)

154

Mαroczy

Bind

13 ... Φχg7 14.te4':fl 15 'iί'd2 'iί'b6 16.td5 e617 .txc6! "xc618 ':adl 'iί'xc4 19 "xd6 White had a positiona11y won garne. 2a2) Later ίη the tournarnent Afifi Ρrefeπed 12....txc3 and after 13 bxc3 lΔe5 14 .te4 ':h5 15 .te3 (Rachels-Afifi, Manila ΙΖ 1990) 15 ... 'iί'c7 would haνe left the situation unclear. 2b) White can aνoid the doubling of his pawns with 12 jt'd2 and this should be good enough Ιο giνe him an edge. For exarnple, 12•••'iί'&5 13 'iii>hl ':Π 14 f4 was promising ίη Sharnkoνich- Vasiukoν, USSR 1965 and 12••..tg7 13 .txg7 Φχg7 (Haklά Afifι, Bahrain 1990) 14 .td3!?':fl 15 .te4 would transpose ιο Ηϋbηer Afifi giνen aboνe. 7 lΔc3 ο-ο Black has a major aIternatiνe ίη 7 ... lΔg4, which has recently become more popular. This new respectability has been based partly οη a new idea for Black inνolνing kingside pawn expansion, and partly οη a realisation that the older lines are ηοΙ so bad for Black as had been thought. White has his typica1 space adνantage, but Black's position is solid and Larsen ίη particular has achieνed quite good results for Black. After 7..• lΔg4 8 'ifxg4lΔxd4 (or 8 ... i.xd4 9 .txd4lΔxd4 10 0-0-0 e5 11 'ίi'g3 d6 12 f4 f6 13 f5! Φπ 14 lΔb5 lΔxb5 15 cxb5 with an excellent position for White, MestelKarlsson, Las Palmas 1982) 9 'ii'dl (D) Black has three possibilities:

1) 9•••e5 and now White has the choice between a complicated and a safe line, both of which promise him some adνantage: la) 10 lΔbS (the sharp continuation) 1Ο ... 0-0 11 'iί'd2! (11 lΔxd4? exd4 12 .txd4 'iί'a5+ is what Black is hoping for; a recent exarnple of this old trap being Brunner-EkstrDm, Suhr 1990 which continued 13 'iii>e2 :e8 14 f3 d5! 15 .txg7 ':xe4+! 16 'iii>d3 ':xc4 17 'iii>e3 'iί'c5+ 18 'iii>d2 'iii>xg7 19 .txc4 dxc4 with a winning attack for Black) 11 ... 'iί'h4 (11 ...'iί'e7 12 f3 f5 13 .td3 d6 14 .tg5 .tf6 15 .txf6 'iί'xf6 16 lΔxd4 exd4 17 ο-ο left Black's d-pawn fatalIy weak ίη Yebelin-Sίlman, Budapest 1994) 12 .td3 d5 13 cxd5 lΔxb5 14 .txb5 'iί'xe4 15 ο-ο :d8 16 ':fdl and White is clearly better. Note that 16 ....te6 fails ιο 17 f3 'ίi'xd5 18 "e2 trapping Black's queen. lb) 10 .td3 (the solid continuation) 10 ... 0-0 11 0-0 d6 12 "d2 (12 a4!? deserνes attention; SpraggettGarcia Ilundain, Candas 1992 continued 12 ....te6 13lΔb5 a614lΔxd4 exd4 15 .td2 :c8 16 b3 f5 17 exf5 .txf5 18 "f3 with adνantage Ιο

Mαroczy ΒίΜ

White) 12 ....te6 (or 12 ... f5 13 exf5 gxf5 14 f4 lDc6 15 :adl 'fIe7 16 .te2 lDd4 17 .txd4 exd4 18 lDd5 with a clear plus for White, Anders1ion-Rogers, Malta 1980) 13 :acl a6 14 b3 :c8 15 f3 and now both 15.••f5 16 exf5 gxf5 17 f4 'ί1i'f6 18 lDe2 :cd8 19 lDxd4 exd4 20 .tf2, Ghitcscu-Radovici, Romania Ch 1977, and 15...'fIa5 16 :fd 1 f5 17 exf5 lbxf5 18 .te4, Tal-Partos, Nice 1974, were very good for White. 2) 9•••lDc6 10 'ί1i'd2 'ί1i'a5 11 :c 1 ο-ο 12 .te2 d6 13 ο-ο .te6 14 b3 :ac8 15 f4 with a good position for White, Polugaevsky-Suetin, Κislo νodsk 1972. 3) 9...lbe6 (the main line) 10 :cl (D) and now:

3a) 10...'fIa5 and ίι is unclear whether White should develop his bishop at d3 or e2: 3al) 11 .td3 leads Ισ a further branch: 3al1) 11....txc3+ 12 :xc3 'fIxa2 13 'ficl 'ί1i'a5 14 c5 is extremely dangerous for Black. 3a12) 11...d6 12 ο-ο (better than 12 'fid2 .td7 13 ο-ο .tc6 14 :fel

155

ο-ο 15 .th6 'ί1i'e5! 16 .txg7 'ί1i'xg7 17 :cdllbc5 18.tf1 a5 with equality, Ν ogueiras-Korchnoi, MontpeIIier Ct 1985) 12... 0-0 13 .tbl.td7 14 f4 .tc6?! (14 ... lbc5 15 lDd5 is better for White, but ησΙ as bad as the game) 15 f5lDc5 16 f6! (the start of an extremely surpήsίηg combinaιίοη) 16... .txf617 :xf6! exf618 b4! 'fixb4 19 'fixd6lba6 20 'ί1i'xf6 (this is the ρσίηΙ; now Black is getting mated ση the dark squares) 20 ...:fe8 (20 ...:fd8 21 .th6 'fic5+ 22 ~hl 'ί1i'd4 23 e5 wins) 21 .td4! (ηοι 21 .th6?? 'fic5+ 22 ~hl 'fie5) 21 ... ~f8 22 a3! 'ί1i'a5 23 :f1 'ί1i'c7 24 'fig7+ 1-0 Filίpenko-ΚIiukίn, Bie11993. 3a13) 11...b6 12 ο-ο .tb7 (if Black plays 12 ... g5, then White can continue as ίη '3a2' with f3,:f2 and .tf1) 13 f4!? (probably better than 13 '6'd2 g5 14 :fdl d6 when 15 f3 .te5 16 ΦhΙ .tf4 17 .txf4lbxf4 18 .tf1lbe6 19 a3 'fie5 20 lbd5 h5 21 b4 ~f8 22 :e 1 :c8 was equal ίη Popovic-Cebalo, Yugoslavia 1988, and 15 a3 h5 16 :c2 .td4 17 b4 'fie5 18lbd5 .txe3 19 fxe3 :c8 20 :f1lDg7 21 'fif2 f6 was unclear ίη Ljubojeνic-Korchnoi, Tίlburg 1987) 13 ... 0-0 14 .tbl d6 15:f2 :ac8 16 lbd5 .txd5 17 exd5 lbc5 18 a3 and White is better, Α.RσdήgueΖ-Ηer nandez, Cuban Ch 1988. 3a2) 11 .te2 b6 12 ο-ο .tb7 13 Ι3 g5 (Larsen's plan increases the dark-squared pressure and reserves e5 for the queen, but the obvious danger is that Black's king has Ισ stay ίη the centre) 14:t'2! (a number of other games had continued with

156

Mαrocz:y ΒίΜ

'fi'd2, but the rook transfer Ιο d2 appears ιο be the best way of meeting Black's double-edged plan) 14•••h5 15 j.n (D) and now:

3a21) 15.•. j.xc3?! 16 1:txc3 g4 backfired hοπίblΥ ίη Chandler-Larsen, Hastings 1990 after 17 f4! j.xe4 18 f5! g3 (18 ... j.xf5 19 j.d4! ~xd4 20 'ii'xd4 e5 21 1:.e3! wins mateήal) 19 hxg3 ~c5 (perhaps Larsen had intended 19 ... j.g4, oνerlooking 20 'ii'f3) 20 1:.a3 'iί'b4 21 'ii'd4 :g8 22 j.d2 and Black's queen is trapped. 3a22) 15... 'fi'e5 16 .1:[d2 with a couple of examples: 3a221) 16...d6 17 ~d5 Φf8 18 b4 j.h6 19 'fi'b3 g4 20 j.xh6+ :xh6 21 'ίi'e3 fI g7 22 f4 with a distinct adνantage for White, Short-Larsen, Hastings 1987/8. 3a222) 16... j.c6 (Black refrains from ... d6 as he hopes Ιο build υρ pressure οη the b8-h2 diagonal by playing ... flb8 and ... j.e5; an interesting, if somewhat optimistic idea) 17 b4 :d8 18 ~d5 flb8 19 c5 bxc5 20 bxc5 .te5 21 h3 :g8 22 'ίi'b3 g4 23 'fi'xb8 j.xb8 24 hxg4 hxg4 25 f4! j.xd5 26 exd5 j.xf4 27 j.xf4 ~xf4

28 1:.d4 ~h5 29 1:ta4 and notwithstanding his pawn minus White has much the better of this ending, Stangl-Becker, Bundesliga 1991/2. 3b) 10... b6 11 j.d3 (11 b4 is also good, for example 11. .. j.b7 12 j.d3 ο-ο 13 ο-ο ~d4 14 j.bl ~6 15 a3 d6 16 'fi'd3 ':c8 17 f4 and White has consolidated his space adνantage, Suba- Taimanoν, Bucharest 1979, or 13 ...1:tc8 14 f4 with attacking chances for White) 11 ... j.b7 12 ο-ο flb8 (Black intends a νariant of Larsen's plan Ιο dominate the dark squares οη the kingside; normal deνelopment would lead Ιο positions similar Ιο Suba-Taimanoν aboνe) 13 'ii'd2 'ii'd6 14 ~d5 g5 15 b4 h5 16 ':fdl j.e5 17 h3 j.f4 18 j.f1! and Black's attack has become bogged down while White has all sorts of threats against Black's king and queen, Mochaloν-Kapengut, USSR First League 1976. 3c) 10...d611 Μ! (when Black's knight is οη e6 White should ίη genera1 aim Ιο play b4 as quickly as possible, preνenting Black cementing his knight οη c5 by ...a5) 11 ...0-0 12 .te2 (D) and now:

Maroczy ΒίΜ 157

3cl) 12•••b613 ο-ο ~b7 14lM5 (14 ... 'ii'd7 15 ~g4! f5 16 ~h3 17lbxc7 'ii'xc7 18 exf5 gxf5 19 c5! with advantage Ιο White, Ador.iun-Larsen, Hastίngs 1986Π) 15~g5 1'6 16 ~e3 Φh8 17 'ii'b3 'iWd7 18 :tfdl and White enjoyed a large ~pace advantage, C.Hansen-Larsen, Esbjerg 1988. 3c2) 12.•.a5 13 a3 axb4 14 axb4 .id7 (14 ...1Ia3 15lbd5) 15 ο-ο ~c6 16 'ii'd2 1Ia3 (16 ... ~xc3 17 _xc3 .ixe4 faίls Ιο 18 ~h6 1:te8 19 lIcel I'ollowed by ~g4 with a catastrophe nΙ g7) 17lbd5 'iith8 18 ~ b6 'ii'd7 19 1'4 with a fine position for White, I>ortisch-Pfleger, Manila 1974. 3d) 10•.. 0-0 11 b4 will quickly Iranspose ίηΙο '3b' or '3c'. 8 ~e2(D) fΔc7 tΔc7

'ii'a5 14 f4 e5 15 fxe5lbxe5 161:i.xf6! ~xf6 17lM5 ~d8? (Black should have tried 17 ...:a6! 18 1Ic 1! ~d7 19 :c5 'ii'd8, although White has an ominous initiative) 18 ~d4! f6 19 ~c3 'ii'a6 20 lbbc7 'ii'a7+ 21 ~d4 'ii'b8 22 tιJxa8 with a clear plus for White, Nunn-Haίk, Paήs 1983. 2) 8...b6 (an important alternative) 9 ο-ο ~b7 10 Ι3 (D) (when Black develops his bishop at b7 the extra protection of the e-pawn afforded by f3 is usually a good idea) and now Black has an extensive range of possibilitίes:

8 •.. d6 Black may try Ιο do without this move: Ι) 8••• a5 9 ο-ο a4 10 c5!? (an attempt at outright refutatίon; Ι Ο 1Mb5 would be similar Ιο 9 ... a5 below) 1O... d5 11 cxd6 'ii'xd6 12lbdb5 'ii'b4 (12 ... 'ii'xdl 13 1:i.axdl gives White some endgame advantage) 13 a3

2a) 10... tιJh5 11 tιJxc6!? with a choice of recaptures: 2al) 11•••~xOO 12 :cl f5 13 exf5 gxf5 14 f4lbf6 15 ~f3 1:tc8 16 b3 'ii'e8 17 tιJd5 'ii'f7 18 tιJxf6+ 'ii'xf6 19 1:i.c2 ~xf3?! 20 ':xf3 d6 21 ':d2 ι;i;>h8 22 ~d4 gave White his standard favourable position ίη Nunn-Ristoja, Malta OL 1980. 2a2) 11•••dxc6 12 c5 b5 13 g4!? (13 f4 b4 14lba4lbf6 is considered as harmless for Black by Donaldson and Sίlman, but they οηlΥ take 15 ~f3 and 15 ~d3 ίηιο account

158

Mαroczy ΒίΜ

when the natural15 'i'c2 should give White an edge) 13 ... lbf6 14 'i'c2 (14 e5!?) 14 ... e5 with οηlΥ a small plus for White, Khuzman-Hergott, Biel 1993. 2b) 10••• d6 (this is inconsistent with the choice of ... b6) 11 'i'd2 'i'd7 12 a4!? e6 13 1tfdl 1tfd8 14lbxc6 'i'xc6 15 a5 bxa5 16 lbb5 with threats Ιο d6, a7 and a5, GheorghiuBellon, Las Palmas 1976. 2c) 10•••1tc8 11 'i'd2 lbh5 (the alternative 11 ...1te8121tacl 'i'c713 b4! lbh5 14 lbxc6 .i.xc6 15 lbd5 'i'b8 16 f4lbf6 17 .i.f3 d6 18 .i.d4 was very good for White ίη Νυηη Karlsson, Helsinkί 1981) 12 1tfd1 lbe513 b3 Ι514 exf5 gxf515lbd5! (D) and now:

2cl) 15..•'i'e8?! (Κir.Georgiev Saint John 1988) 16 f4! .i.xd5 (16 ... lbg4 17 lbxf5lbxe3 18 lbxg7lbxg7 19 'i'xe3 wins) 17 cxd5 lbg4 18lbxf5! .i.xa119 .i.xg4lbf6 (19 ... .i.f6 20 d6! e6 21lbe7+ .i.xe7 22 dxe7 1tf6 23 'i'xd7 wins and 19 ... .i.c3 20 'i'd3lbg7 21 d6 are ηο better) 20 ll:\h6+ <l;g7 21 1txal ~xh6 22 f5+ <l;g7 23 .i.h6+ <l;h8 24 .i.xf8 Κristensen,

lbxg4 25 'i'd4+ lbf6 26 .i.h6 wins forWhite. 2c2) 15•• ..txd5 16 cxd5 f4 17 .i.f2 'i'e8 and now 18 .i.d3?! lbxd3 19 'i'xd3 'i'g6 20 'i'xg6 hxg6 21 1tacl 1txcl 22 1txcl .i.xd4! 23 .i.xd4 1tf5 left White fighting for a draw ίη Sherzer-Edelman, New York 1993. However, Whίte's 18th move was a bit panicky. Much better would have been 18 d6 when both 18..•exd6 19 lbb5 'i' g6 20 1tac 1 and 18••. e6 19 lbb5 'i'g6 20 J:lacl are good for White. 2d) lo...lbxd4 11 .i.xd4 d6 12 'i'd2 lbd7 13 .i.xg7 (13 .i.e3 is a promising alternative) 13 ... ~xg7 14 f4 1tc8 15 1tadllbf6 16 e5 dxe5 17 fxe5lbg8 18 'i'e3! with some advantage for White, Cvetkovic-Cebalo, Yugoslavia 1985. 2e) 10••• lbe8 11 'i'd2 lbc7 12 1:Iadllbe6 13lbdb5 d6 14lbd5 1tb8 15 f4lbc5 16.i.f3 a6 17lbd4lbxd4 18 .i.xd4 .i.xd4+ 19 'i'xd4 b5, Agapov-Kimelfeld, USSR 1985, and now 20 e5 would have kept some advantage for White. 2f) 10•••e6 11 'i'd2 (11 lbdb5 is probably also good, e.g. 11 ... d5 12 cxd5 exd5 13 exd5lbe7 14 d6lbf5 15 .i.f2lbe8 16 d7lbf6 17 g4lbe7 18 .i.h4 with advantage, A.Rodriguez-Pinal, Sagua la Grande 1984) 11. .. d5 (11. .. 'i'e7 is more solίd, as ίη Panno-Bellon, Buenos Aires 1994, which continued 12 1tfel 1tfd8 13 .i.f1 lbxd4 14 .i.xd4 lbe8 15 .i.f2 lbc7 16 1tadl d6 17 a4 with an edge for White) 12lbxc6 .i.xc6 13 cxd5 exd5 14 e5lbd7 15 f4lbc5 16 1tadl !

MαroczY ΒίΜ

f6(16 ... tDe417tDxe4fxe418'i'd6! i.b7 19 i.c4 is a1so unp1easant) 17 tDxd5 fxe5 18 i.c4 ~h8 19 fxe5 :xf1+ (19 ... i.xe5 20 i.d4 'i'd6 21 i.xe5+ 'i'xe5 22 :fellΔe4 23 he4! 1-0 Kuporosov-Yakovich, USSR 1984) 20 :xf1 'i'h4 21 :f4! 'i'h5 22 tDe7 i.b7 23 b4 tDe4 24 'i'd7 :b8 25 tDc8! :xc8 26 'i'xb7 J:d8 27 'i'xe4 i.xe5 28 i.e2 1-0 Kuporosov-Ma1ishauskas, USSR 1985. 2g) 10•.•'i'b8 11 'i'd2 :d8 12 :fdl d6 13 tDb3!? (13 'i'el :d7 14 'i'h4 is another possibility) 13 ...:d7 14 a4 'i'd8 15 a5 bxa5 16 tDb5 i.a6 17 tDxa5 i.xb5 18 cxb5 tDxa5 19 'ifxa5 'i'xa5 20 :xa5 with a winning endgame for White, ArakhamiaCane1a, Νονί Sad OL 1990. 9 ο-ο (D)

9 ..• i.d7 Or: 1) 9••• tDxd4 (and ηοΙ 9 ... tDg4? 10sing a piece after 10 i.xg4 i.xg4 11 tDxc6) 10 i.xd4 i.e6 11 ί4 :c8 (11 .. :ii'c8 12 b3 :d8 {12 ... i.g4 13 e5!} and now 13 h3, keeping B1ack's pieces ουΙ οί g4 gives White a good game) 12 b3 i.d7 13 e5 dxe5 14 fxe5

159

tDe8 15 'ii'd2 .tc6 16 'ii'e3 with advantage Ιο White, A.Soko10v-Nemet, Bem 1992. 2) 9••• tDd7 10 :bl tDc5 11 'ii'd2 tDxd4 12 i.xd4 a5 13 b3 i.xd4 14 'ii'xd4 b6 15 :fel i.b7 16 :bdl ί6 17 i.g4 i.c6 18 h4! with a slight advantage for White, Anand-Morovic, Νονί Sad OL 1990. 3) 9•••&5 10 ί3 tDd7 11 tDdb5 tDc5 12 'ii'd2 a4 13 :fdl 'i'a5 14 :acl i.e6 15 tDd5 'ii'xd2 16 :xd2 i.xd5? (16 ...:fd8 is on1y slight1y better for White) 17 cxd5 tDb4 18 :xc5! dxc5 19 i.xc5 tDxa2 20 i.xe7 and B1ack is ίη troub1e, AnderssonLarsen, Linares 1983. 4) 9•••:e8, another Larsen idea, is the most important a1temative Ιο 9... i.d7. White has several tries: 4a) 10 1:tcl tDxd4 11 i.xd4 i.h6! 12 :c2 b6 (12 ... e5 13 .te3 .txe3 14 fxe3 is unc1ear) 13 ί3 .tb7 14 'ii'el tDh5 was fine for B1ack ίη PyhiiliiRantanen, Pori 1986. 4b) 10:bla611f3.td712tDc2 tDe5 13 a4 'i'c7 14 tDa3 e6 15 ί4 with some advantage for White, Wolff-Larsen, New York 1990. 4c) 10 tDb3 .te6 11 f3 tDe5 12 tDd5 .td7 13 'i'd2 tDxd5 14 cxd5 b5 15 :acl a5 16 tDd4 and White stood very well ίη Wahls-Vanscura, Budapest 1988, but Black played the opening honib1y. 4d) 10 tDc2, also avoiding the knight exchange and preparing solid development by :cl was recommended ίη BTS2 but has still ηοΙ been tested. 10 'ii'd2

160

Mαroczy ΒίΜ

Ιι is also possible Ιο move the knight οη d4, thereby frustrating Black's plan οί ... ltJxd4 and ... i.c6. Ιη the gaιne Korchnoi-Soos, Rome 1982, White continued 10 ltJb31tJa5 (1O ... a5 11 a4 1tJb4 12 f3 i.c6 13 'ιi'd21::ιc8 14 clthlltJd7 151tJd5 was also better for White, Tarjan-Strauss, USA 1982) 11 f31tJxb3 12 axb3 a6 13 b4 i.e6 14 'li'd2 :c8 15 b3 ~7 16 :a2, while Schmidt-Kagan, Lucerne 1982, went 10 ~2 i.e6 11 'li'd2 a5 12 ί4 a4 13 :abl i.g4 14 i.d3 i.c8 15 h31tJd7 16 i.e2 and ίη both cases White had a good posiιίοη. It could well be that these lines are as strong as the traditional continuations 1O:c 1 and 10 'it'd2, but they do ηοΙ seem Ιο have been tested ίη recent years. Α separate question is whether White should play 10 1::ιc 1 or 10 'A'd2. We are concentrating οη the latter move for two reasons. Firstly, the rook οη cl is sometimes vulnerable ιο unwelcome attacks by ... ltJxd4 and ... i.h6, and secondly White often starts a queenside advance by a3 and b4 later οη. Ιί Black plays ... a5 White needs a rook οη bl Ιο force through b4, but playing the rook Ιο c1 and then Ιο b1 loses a tempo. 10 ... ltJxd4 10... 11 f3 a6 12 1:taclltJxd4 13 i.xd4 .te6 14 b3 is ιοο passive and White has a very comfortable position, Smejkal-Diez del Coual, Skopje OL 1972. 11 .txd4 i.c6 12 f3 a5

:c8

The move-order is flexible, but White always answers ... ltJd7 by i.e3 (to avoid freeing exchanges) and ... a5 by b3 (or else Black plays ... a4 followed by ... 'ίi'a5, with an active position). 13 b3 ltJd7 (D) Or 13... (l3 ... ltJh5?! 14 i.e3! ί5 15 exf5 gxf5 16 ί4 was promising for White ίη Kavalek-Larsen, USANordic match 1986) 14 :fdlltJd7 15 i.e3 ltJc5 16 :acl (16 1::ιab1 appears more consistent) 16 ... 'ίi'b6 17 ltJb5 :ec8 18 'ίi'e1 (l81tJd4 is better) 18 ... i.xb5! 19 cxb5 i.h6 with an equal position, Arnason-Karlsson, Helsinki 1986.

:e8

14 i.e3 14 i.f2!? is an interesting alternative. After 14... ltJcS 15 :abl Black has: 1) 15... 'iί'b6 16 a3! 'iί'd8 (as 16 ... ltJxb3, which would win with the white bishop οη e3, now fails ΙΟ 17 'li'd 1! Β lack has ηο choice but Ιο retreat his queen without having accomplished the main point behind ... 'it'b6, naιnely Ιο play ... %:tfc8) and now:

Mαroczy ΒίΜ

la) 17 ':fcl a4!. lb) 17 b4 (premature) 17 ... axb4 18 axb4lDa4 19lDxa4 ~xa4 20 ':fc 1 'iνa8 with a very comfortable game for Black, Arnason-Larsen, Ostersund Ζ 1992. lc) 17 .i.dl (often a useful move ίη this variation as ίι allows White ιο play b4 under more favourable circumstances) 17 ... b6 18 b4lDd7 19 .i. b3 ;j; Wells-McElligott, Dublin 1993. 2) 15••.Wc7 16 IIfcl ':fc8 17 .i.dl 'iνd8 18 a3 b6 19 b4 axb4 20 axb4 lDd7 21 .i.b3 'iνf8 22 .i.e3 is simίlar Ιο 'lc' and gave White an edge ίη Renet-Spangenberg, Buenos Aires 1994. 3) 15..•e6 16 ι!Z:Ib5!? .i.e5 17 ι!Z:Id4 'ii'f6 was played ίη Hellers-Cebalo, Debrecen Echt 1992, and now 18 g3! would threaten Ιο push back the black pieces. Note that 18 ... .i.xe4 19 fxe4 ι!Z:Ixe4 20 'ii'e3 ι!Z:Ixf2 faί1s Ιο 21

161

by the fact that with the a1ternative 15 :Scl (thίs position can also arise via 10 ~cl), White is struggling Ιο gain any advantage. After 15•••'iVb6 (15 ... f5 16 exf5 gxf5 17 ι!Z:Id5':f7 18 ':fdl b6! 19 .i.g5 ':a7 was unclear ίη Vaganian-Yudasin, USSR Ch 1988) 16lDb511fc8 17 ':fdl Wd8 there is: 1) 18 .i.n 'ii'f8 19 1Dc3 b6 20 ι!Z:Id5 ':ab8 21 ':bl.i.e5 22 .i.h6 .i.g7 23 .i.g5 :b7 24 ':el ι!Z:Ie6 25 .i.e3 1Dc5! with a level game, Sax-Petursson, Reykjavik 1988. 2) 18 ι!Z:Id4 was answered by 18..:ii'f8?! 19 ':bl .i.f6 20 a3 'ii'g7 21 b4 axb4 22 axb4 lDe6 23 lDxe6 fxe6 24 b5 .i.e8 25 f4 with a clear plus for White ίη Νυηη- Velimίrovic, Dubai OL 1986. However, Black should have played 18....i.d7 19 ':bl lDe6 20 lDxe6 .i.xe6 with equa1ity.

ι!Z:Ic2!.

4) 15•••Ι5 16 exf5 gxf5 17 a3 ςi(h8 18 b4 axb4 19 axb4lDa4 20 lDxa4 1/2-1{2 Bagίrov-Yudasin, USSR 1991. Perhaps 18 b4 was a Iittle hasty, but even so Ι would have thought that the final position was ίη White's favour, e.g. 20 ... ':xa4 21 b5!? .i.d7 22 c5 dxc5 23 .i.xc5 leaves him with the better pawn structure. 14 ••• lDc5 15 lιtabl (D) Ιη order Ιο make progress White must expel the knight from c5, and so he needs ιο play a3 and b4. 15 lIabl appears the most 10gica1, and this intuitίve assessment is supported

15 ... 'iνb6 After 15...e6 16 .i.dl! intending lDe2-d4, a3 and b4 Black's panic reaction 16... f5 17 exf5 ':xf5 18 ι!Z:Ie2 was good for White ίη Τήηgοv-Ηaik, Vrnja~ka Banja 1986. 15...b6 16 .i.dl (White's problem is that the immediate a3 may be

162 Maroczy Bind met by ... a4; the idea οί -i,dl is Ιο be able ΙΟ take the black knight when ίι aπives at b3, but White has other ways ιο ηυΙΙίίΥ ... a4, for example with the slow preparatory plan οί ':fc l-c2, .i.f1 and "12 Ιο line up against Black's weak pawn οη b6) 16...~8 17 a3 ':c8 18 ~5 -i,xd5?! (probably bad, but White is slightly better ίη any case) 19 exd5 a4 20 b4 lbb3 21 "e2 gave White an excellent position ίη Anand-Larsen, Cannes 1989. 16 ':fcl Or 16lbb5 (ηοΙ 16 a3? lbxb3! 17 "d3 -i,d4) 16 ... .:fc8 17 ':fdl "d8 18lbd4 (18 ~hl 19lbc3 b6 20 a3 ':ab8 21 -i,f1 h5 22lbd5 ':b7 23 b4 axb4 24 axb4lbd7 25 c5! gave White a good game ίη GreenfeldBrulc, Israel1992, but it's hard Ιο believe that lbc3-b5, followed shortly by lbb5-c3 is White's best course οί action) 18 ...'iff8 19 a3 (19 -i,f1 may be better, but Ι doubt that White has more than a ιίηΥ edge) 19 ... -i,d7! (ηοι 19 ... -i,f6? 20 b4 axb4 21 axb4 lbe6 22 lbxe6 fxe6 23 ί4 with the same White advantage as ίη NunnVelimίrovic above) 20 b4 axb4 21 axb4 lbe6 22 ':al ':xal 23 ':xal ':a8 24 ':xa8 "xa8 25 lbxe6 -i,xe6 with complete equality, Jansa-Petursson, Nrestved 1988. 16 ... ':fc8 17 Ac2 Again this is useful preparation for a3 and b4. Now Black cannot meet 18 a3 by 18 ... lbxb3 because 19 ~1! wins material. 17 ... 'ίWd8

"f8

There is ηο ροίηι Ιο 17...'ίWb4 because after 18 "cl Black will be driven back by a3 with great 10ss οί tίme.

18 -i,n -i,e5 Whίte intends "12, followed by a3 and b4, so Black has Ιο organise some counterplay. With ... -i,e5 he hopes Ιο become active with ... e6 and .. :ti'h4, but the exposed bishop οη e5 ίδ a target which causes White Ιο switch plans away from his queenside pawn advance Ιο a more aggressive idea.

19 lbdl!? The knight transfer Ιο h6, which gaίns time along the way when lbg4 hits the bishop, is ηοΙ a very thematic approach, but chess cannot always be played according ιο the recipe book. 19 ... "e8 (D) Preparing ... b5 is the most natural way to counter White's slow buίld up towards a kingside attack. Moreover the queen defends the f7-pawn which might come under fire after ~f2-g4-h6+.

b5?!

MαroC7;Y ΒίΜ

It would have been better Ιο play 20...J.g7, so as ΙΟ meet 21 tΩg4 by 21. .. h5. Then White could hardly venture ίηΙο the lion's den by 22 tΩh6+ ~h7, so the knight would have Ιο return Ιο f2. After 20 ...J.g7 White should switch plans agaίn by 21 tΩd3 b6 22 'iff2, followed by a3 und b4. 21 tΩg4 J.g7 J.xb5 22 cxb5 ~f8 (D) 23 tΩh6+ Unfortunately Black has Ιο go Ιο 1"8 because 23.•• ~h810ses a pawn afler 24 tΩxf7+, whίle 23••• J.xh6 24 .txh6 gives White a permanent advantage.

163

24 J.xb5 1Vxb5 25 1Vd5 1Ve8 26 e5! ':d8?! Ιη a difficult position Black faίls Ιο offer the most resistance. He should try Ιο give υρ a pawn Ιο reach an ending ίη which the offside knight οη h6 gives White problems. Therefore 26...ltJe6 was best (26 ... ':a6 27 exd6 is bad since 27 ... ':xd6 28 'ifxf7+! wins a clear pawn), with the idea 27 ':xc8 ':xc8 28 exd6 exd6 29 1Vxd6+ 'ife7 30 'ifxe7+ ~xe7 with excellent counterplay. White should prefer 29 tΩg4!, whereupon the weak pawns οη d6 and a5 make Black's position unattractive, but hίs chances would certainly be much better than ίη the game. 27 exd6 exd6 28 ':el ':ac8? The final collapse. 28...ltJe6 was necessary Ιο meet the threat οί 29 ':xc5, when 29 tΩg4 gives White a positional advantage but nothing decisive. 29 ':xcS! :XcS 30 1Vxc5 ι-ο

8 Taimanov Variation The first moves of αιίδ system run 1 e4 c5 2 lbf3 e6 3 d4 cxd4 4 lbxd4 lbc6 (or2 ...lbc6and4 ... e6). Ina way this resembles the Kan Variation, since Black keeps the f8-b4 diagonal open for his bishop, but here White cannot play .i.d3 δΟ there are fewer options for the first player. BTS2 recommended the Maroczy set-up starting with 5 lbb5, but this time we are opting for 5lbc3. There are several reasons for this change but ΓΗ just give one example from my own practice. After 5 lbb5 d6 6 c4ltJf6 7lί:1ι c3 a6 SltJa3 .i.e7 9 .i.e2 0-0 10 ο-ο b6 11 .i.e3lbe5 12 f4lbed7 13 .i.f3 .i.b7 14 We2 White's main idea, perhaps after some further preparatory moves such as 1%ac 1 and 1%fd 1, is to play g4-g5 gaining space οη the kingside and relieving some ofthe pressure against e4. However, ίη GaHagher-Cramling, Bern 1991, this idea was nipped ίη the bud by 14... h6!. Now any subsequent g4 can be met by ... lbh7 and ...g5 causing White some positional difficulties as Black will gain control of e5. Deρήved of my on1y active plan Ι floundered around for a few moves before settling οη a draw. The new system, 5 lbc3, is also not without its drawbacks mainly because the CΙassical Scheveningen is ηοΙ part of the repertoire we are

proposing ίη αιίδ book. Therefore we have Ιο rely οη some slightly less common systems, but as you will see they are also quite dangerous for Black. Game 22 deals with 5 ... a6, against whίch we are recommending 6 lbxc6, and 5 ... d6, against which we suggest 6 g4 inviting transposiιίοη ιο the Keres Attack ofChapter 2. Game 23 deals with 5 ...Wc7; Ιο avoid transposition Ιο the Scheveningen we are suggesting 6 f4. Game22 Gallagher - Rufener Bern 1994

1 2 3 4

e4 tbf3 d4 lbxd4

c5 e6 cxd4

lbc6

5 tα3 (D)

5 •••

Tαimαnov Vαriαtion

After 5•••d6 we suggest 6 g4!? Of course there are many other good moves such as 6 .te3 or 6 .te2 but these involve being ready ιο play lines ίη the Scheveningen that do ηοΙ form part of our repertoire. 6 g4 caused astonishment when ίι was first adopted by ηο less than Anatoly Karpov during the 1985 World Championship match. Black has: 1) 6••• lίJI6 transposing Ιο Chapter 2. 2) 6 ...h67 h4 (consistent) 7 .... a6 8 .tg2 (8 .te3 is quite playable but after 8...lίJf6 we have transposed Ιο a line ίη the Keres Attack that is ηοΙ discussed ίη Chapter 2) 8 ....te7 9 .te3lίJxd4 (9 ... .txh4 10 lίJxc6 bxc6 11 J:txh4! "xh4 12 "xd6 .td7 13 g5! is much Ιοο risky for Black) 10 "xd4 e5 11 "dl .te6 (11 ....txh4? 12 lίJd5!, threatening .tb6 is very strong and 11 ... lίJί6 12 .tf3! is also good for White) 12lίJd5 :'c8 13 c3 lίJf6 14lίJxe7 "xe7 15 g5 (15 .tf3, maintaining the tension, deserves consideration) 15 ... hxg5 16 hxg5 :'xh 1+ 17 .txh IlίJg4 18 .td2 "f8! 19 "f3 (19 .tf3 would have been met by 19..."h8!) 19..."h8 20.tg2 'ii'h4 21 b3! (threatening 'iPe2 and then J:thl) 21 ... d5! 22 "g3 "xg3 23 fxg3, Karpov-Kasparov, Moscow Wch (14) 1985, with arather strange ending ίη which Black was able Ιο maintain the balance despite having an awkwardly placed knight οη g4. 3) 6 ...lίJxd47 "xd4 tΔe7 8 .te3 lίJc6 9 'ii'd2 .te7 100-0-0 a611 'iPbl ο-ο 12 g5 b5 13 f4 leads Ιο a fairly normal Sicilian position with White

165

attacking οη the kingside and Black advancing οη the queenside. Ιη the game Lobron-Schulze, Bundesliga 1993 White gained the upper hand after 13 ... b4 14lίJa4 J:tb8 15 h4 .a5 16 b3 d5 17 exd5 exd5 18 .td3 .tg4 19 J:tdg 1 .tf3 20 J:th3 .te4 21 ί5 J:tfe8 22 ί6 .txd3 23 "xd3 .td6 24 h5. 4) 6 ... a6 7 .te3lίJge7 (this is the most popular way for Black Ιο play; alternatively 7 ....td7 8 g5 lίJge7 9 lίJde2!? is slightly better for White according Ιο Ehlvest) 8 lίJb3 (8 g5 !? is also possible, with a probable transposition Ιο line '3') 8 ...b5 (D) (or 8 ... lίJg6 9 "e2 .te7 100-0-0 b5 11 ί4 h6 12 'i'f2 .td7 13 'iPbl J:tb8 14 .te2lίJa5 15lίJxa5 .xa5 16.td3 with an edge for White, BologanFrolov, USSR Ch 1991) and now:

4a) 914 .tb7 10 "f3?! (10 .tg2 lίJa5 11lίJxa5 'ii'xa5 12 "e2 trans-

poses Ιο line '4b', but ifWhite employs this move order Black may try 1O... lίJg6) 10...g5! (Blackgivesupa pawn Ιο gain control of e5, ηοΙ an uncommon idea ίη the Sicilian) 11 fxg5 lίJe5 12 'ii'g2 b4 13 tΔe2 h5! (to give you an idea οί modern preparation,

166

Tαimanov Vαriαtion

Judit Polgar says that she had this position οη the board ίη her room οη the morning of the game) 14 gxh5? (Shίrov heads straίght into the trap; 14 gxh6 J.xh6 15 J.xh6 ':xh6 gives Black good play, but 14 0-0-0 would have been unclear) 14 ... lDf5! 15 J.f2?! (a better chance was Ιο sacήfice the queen by 15 exf5 J.xg2 16 J.xg2 a1though Black must be doing well after 16 ....:xh5) 15 .....xg5! 16 tba5 (Shίrov had been relying οη this counterstroke) 16... tbe3! (played ίη stantly) 17 "g3 (17 "xg5lDf3# and 17 J.xe3 'ii'xe3 18 tbxb7 tbf3+ 19 ΦdΙ "d2# are the attractive points Ιο Black's combination) 17 .....xg3 18 lDxg3 tbxc2+ 19 ΦdΙlDχaΙ 20 lDxb7 b3! and the knight escaped, leaving Black a clear exchange υρ, Shίrov-J.Polgar, Buenos Aίres 1994. 4b) 9 "e2 lDa5 10 lDxa5 (this 100ks better than 10 0-0-0 lDc4, as played ίη Shirov-Salov, Linares 1993) 1O..... xa5 11 J.g2 J.b7 12 f4lDc6 13 ο-ο J.e7 14 g5 'ii'c7 15 "f2 J.d8 16lDe2! ο-ο 17 h4lDe7, Lanka- Yermolίnsky, Lucerne Wcht 1993, and now 18 tbd4! d5 (18 ...e5 19 fxe5 dxe5 20 lDe6!) 19 e5 would have been very promising for White according Ιο Lanka. bxc6 6 lDxc6 6...dxc6 is just bad for Black and is never seen ίη practίce. 7 J.d3 (D) 7 ••. d5 Black has a couple of important alternatives; he can set υρ a sma11 centre with 7 ... d6 or he can delay any centra1 actίon υηtίl he sees what

set-up White has adopted, by playing 7 .....c7. 1) 7••• d6 8 ο-ο lDf6 9 Ι4 with the branch: la) 9.. .J.e7 (9 ... e5 10 ΦhΙ J.e7 11 fxe5 dxe5 12 J.e3 ο-ο 13 tba4 gave White an edge ίη Van der Wiel Zapata, Pa1ma 1989) 10 "e2 (10 should be considered) 10..•lDd7 11 ΦhΙ "c7 and now: lal) The game Reeh-J.Horvath, Mitropa Cup 1990 continued 12 J.d2 ο-ο 13 ':ael tbc5 14 e5 d5 15 J.xh7+!? (good enough for a draw) 15 ... Φχh716 "h5+Φg817 1:Ie3 f5 18 ':h3 J.d7 19':ff3 J.e8 20 ~7+ Φf7 21 "h5+ Φg8 22 "h7+ Φf7 23 "h5+ Φg8 1/2-1J2. la2) 12lDa4!? (often a good idea when Black has played an early ... lDd7 as ίι prevents ...lDc5) 12...0-0 13 c4 and after a continuation such as 13 ... c5 14 b3 J.b7 15 J.b2 White has an edge. lb) 9.....c7 10 "e2 (10 "el is also quite a promίsing move and ίη Ga11agher-Landenbergue, Martigny 1993 Ι tήed 10 "f3, achieving a clear advantage after 10...J.b7 11 b3 c5 12 J.b2 J.e7 13 ':ael lDd7 14

"f3

Taimanov Variation 167 lΔdl! i.f615 i.xf6, when 15 ... lΔxf6

16 "g3 0-017 e5 gives White aformidable attack, and my opponent's choice, 15 ... gxf6 left him with a miserable positίon after 16 "h5) 1O ... d5 (it may seem strange ιο play ... d5 so soon after having played ... d6 but Black is arguing that f2-f4 doesn't help White's position as it lίmits the scope of his queen's bishop; 10 ... e5 can be met ίη simίlar fashion to the Van der Wiel-Zapata game ίη '1a' above) 11 i.d2 i.e7 (it would require a brave man ιο take the b-pawn with .....b6+, but if White doesn't feellίke risking this he could first play ΦhΙ) 12 :ael ο-ο 13 ΦhΙ :e8 14 e5 lΔd7 15 lΔa4 (this, followed by c4, is very common ίη these lίnes and ensures that Whίte can at least hold his own οη the queenside) 15 ... lΔc5 16 lΔxc5 i.xc5 17 c4 a5 18 :cl "b6 19 b3 (the last solίd move) 19 ... i.d7 20 g6 21 i.el (heading for h4 and f6) 21 ... i.e7 22 :h3 f5 (this fatally weakens g6, but it's unlίkely that Black could have held the kίngside ίη the long αιη; he really needs a knight οη f8) 23 exf6 i.xf6 24 i.f2! c5 (24 .. :6'c7 loses after 25 :xh7! ΦΧh7 26 'iί'h5+ Φg8 27 'iί'xg6+ i.g7 28 'iί'h7+ 1;f7 29 i.g6+ Φe7 30 i.c5+ Φf6 31 i.h5! and mate ίη a few moves) 24 ... c5 25 'iί'g4! :e7 (White was threatening :xh7!) 26 i.xg6! hxg6 (of course 26...:g7 allows 27 i.xh7+) 27 'iί'xg6+ i.g7 28 cxd5! :c8 29 f5! and White soon won, Kosten-Collinson, British Ch (Plymouth) 1989.

2) 7 •• :flc7 8 ο-ο lΔΙ6 (8 ...c5 9 'iί'e2 i.b7 10 ο-ο lΔe7 11 b3lΔg6 12 i.d2 i.e7 13 '-h5 0-0-0 14lΔa4 d5 15 exd5 :xd5 16 'iί'e2 'iί'c617 i.c3 :h5?! 18 :ael i.f6 19 i.xf6 gxf6 20 .te4 'iί'xe4 21 'iί'xe4 i.xe4 22 :xe4 :d8 23 :c4 gave White a very good ending ίη Gallagher-Zviagintsev, Loosdorf 1993, but Ι have to admit that i.xa6!, either οη move 18 or οη move 20, sίmply never occuπed to me) 9 Ι4 (9 'iί'e2!?) 9...d5 (9 ... d6 is '1b') 10"(3 .te7 (D) with a couple of examples:

:f3

2a) 11 b3 ο-ο 12 i.b2 :d8 13 e5 lΔd7 14 "h3 g6 15 lΔdl! (having forced a weakening οη the dark squares around the black kίng, White sends his knight off οη a long journey Ιο try ιο exploit them) 15 ... c5 16 c4 lΔb6 17 lΔf2 dxc4 18 bxc4 i.b7 19 lί)g4 h5 20 lί)h6+ Φg7 21 f5! 'iί'c6 (οι 21 ... Φχh6 22 fxg6 fxg6 23 1:[f7 with a crushing attack) 22 fxg6 ..tί'xh6 23 :xf7 :xd3 24 :h7+ 1-0 Rohde-Spraggett, San Francisco 1987. Α model attack from White. 2b) 11 ΦhΙ i.b7 (11 ... 0-0 couId be met by 12 e5lί)d7 13 'iiΊι3) 12 f5!?

168

Tαimαnov Vαriαtion

e5 13 ii'g3 d4 14lDbl c5 15lDd2 0-0 16lDc4! Φh8 (16 ... .txe4 is met by 17 .th6 and 16 ... lDxe4 by 17 .txe4 .txe418 f6! .txf619:xf6) 17 :el (now White has a big positional plus) 17 ... lt)d7 18 .td2 a5 (Black is worήed about White playing .ta5 at some ροίηι) 19 a4! (fιxing the weakness οη a5 and preparing Ιο swing the queen's rook ιο the kingside) 19 ....tf6 20:a3! lDb8 21 ii'g41Dc6 22 .te2 (White's last two moves have cleared the third rank for the rook) 22 ... :fd8 23 :h3! Φg8 24 1:th6! g6 25 'ίi'h3lDe7 26 .td3 :a6 (now it's Black's turn Ιο try Ιο swing his rook ίηΙο the defence - but Ιο ηο avail) 27 :xh7 .tg7 28 1:txg7+! Φχg7 29 ii'h6+ Φg8 30 .txa5! (a nice ροίηι) 30...'ίi'c6 (30...1:txa5 31 f6 mates) 31 .txd8 lDxf5 32 ii'g5 1-0. Undoubtedly Black's play ίη these two games can be improved upon but they illustrate well White's ideas ίη this system. The lines where Black plays ...ilic7 are of particular importance as they can equally be reached via Game 23.

8 ο-ο

lDf6 (D)

9 :el 9 'iνe2 is much less explored than the text but ίι has made a few successful appearances recently. For example, Smyslov-C.Hansen, Biel ΙΖ 1993 continued 9....te7 10 b3 0-0 11 .tb2 a5 (11 ... lDd7 12lDa4 .tf6 was an incredibly bad idea from the system's originator, and after 13 e5 .te7 14 c4 a5 15 cxd5 cxd5 161:tacl White had a clear plus ίη the game David-Taimanov, Prague 1993; perhaps Black can try 11 ... c5) 12 :adl ii'c7 13 .!Da4!? (White wishes Ιο play c4 even if ίι costs him the bishop pair) 13 ....tb7 14 c4 :fd8 (14 ... dxe4 15 .txe4lDxe4 16 ii'xe4 c5 17 ii'e3 :fd8 18 .te5 'ii'c6 19 f3 is what Black should have played according Ιο Smyslov, even if White is still slightly better) 15 e5 lDd7 16 cxd5 cxd5 17 :cl ii'b8 18 .td4 with a very promising position for White. It's worth seeing a bit more of this game as the veteran Smyslov soon transformed his positional advantage ίηΙο a devastating kingside attack: 18 ....ta3 19 :c3! (threatening .txh7+) 19... h6 20 .tb5lDf8 21 :g3 lDg6 22 lDb6 :a7 23 lDd7 1fa8 24 1:txg6! fxg6 25 1fg4 .ta6 26 1fxe6+ Φh7 27 .txa6 1:txa6 28 1ff7 1:txd7 29 'ii'xd7 1fc6 30 1ff7 1fe6 31 'ii'xe6 1:txe6 32 f4 and White won. 9 ... .te7 9••..tb7 (a major alternative) 10 eS lDd7 11 lDa4 (D) and now Black has a choice: 1) 11...'iνh4 (11 ...ii'a5 12 b3.tb4 13 :e2 is slight1y better for White according Ιο Sax, who also believes

Tαimanov Vαriαtion

169

ο-ο 16 :Ιχc4 .txe4 17 l1cxe4 :fd8

that 12 c3 is worthy of attention) 12 b3 lbc5 13 :bl! lbxd3 (Black is ίη 00 position ιο open the b-file by lbxa4) 14 cxd3 (ifBlack now continues passively White can aim for a good knight ν bad bishop position by .te3-c5) 14 ....tb4? (a sιrange decision forcing White ιο activate his rook; better is 14 ... c5, υροη which Sax recoιnmends 15 1:te3) 15 :e3 g6 (15 ... 0-0 loses Ιο 16 :g3) 16 :g3 .te7 17 .te3 f5 (17 ... d418 :h3) 18 exf6 'ii'xf6 19 :f3 'ii'h4 20 'ii'e2 .tc8 (20 ... e5 21 .tc5) 21 'iί'b2, Sax-Zapata, Subotica ΙΖ 1987, and ηοΙ only has White a sιrategically won game but he also collected some material after 21 ... :f8 22 :xf8+ .txf8 23 lbb6 :b8 24 'ii'e5. 2) 11•••.te7 12 'ii'g4 g5 13 b3 h5 14 'ii'e2 c5 15 c4 d4 16 .te4 'iic7 17 .txb7 'iixb7 18lbb2 gave White an edge ίη Sίeπο Gonzalez-Nogueiras, St C1ara 1987. White has a safer king and should eventually aim Ιο open the position by playing b4. 3) 11 ••. 'ii'c7! 12 .tf4 c5 13 c4 dxc4 and now: 3a) Adams-Andersson, Cannes 1989 continued 14.te4 .te7 15 :ΙcΙ

18 'iic2lbb6 19lbxb6 'it'xb6 20 .te3 'iVa5 21 1:tal 1td5 22 h3 1tad8 23 :c1 h6 24 b3 'it'b5 with an equal game. 3b) Perhaps White should simply have played 14 .txc4, which gives him some chances of an edge. For example 14 .....c6 (14 ... Jιe7 should ηοΙ be met by 15 'ii'g4 οη account of 15 ... g5!, but by 15 lbc3 when 15 ... lbb6 16 .tb3100ks a little better for White) 15 Jιf1 (better than 15 f3 which would make ίι virtually impossible for White Ιο develop any sort of kingside attack) 15 ... Jιe7 16 lbc3 ο-ο 17 'iVg4 :fd8 18 Jιh6 and now 18..•Jιf8 can be met by 19 :e3, threatening 20 Jιxg7 and 18••• g6 by 19 Jιg5! when White can hope ιο exploit the weakened dark squares around the black king. 10 e5 lbd7 11 'iί'g4 (D)

11 g6 Black can also play 11 .• /~)f8. After 12 Jιd2 f5?! (12 ...lbc5 13lba4 h5 14 'iid4lbxa4 15 "xa4 c5 16 c4 was slightly better for White ίη Rogers-Razuvaev, Tίlburg 1993 and

170

Tαimanov Vαriαtion

12... h5 13 'iff3 <Rg8 14 ~a4 ~g5 15 'W'e2 ~xd2 16 'W'xd2 g6 17 1:ιacΙ also gave White an edge ίη J.HorvathVan ΜΗ, Budapest 1989 but both these continuations are more solid than 12 ... f5) 13 exf6 ~xί6 14 'W'g3. Black has an impressive centre but it's not going anywhere and his king is not ideally placed. Ιη MestelDaly, British Ch 1992 White eventually triumphed ίη a cascade οί sacrifices: 14 .. 5~ίΊ 15 ~a4 ~d616 f4 c5 17 c4 ~d7 18lbc3 1:ιe8 191:ιe2 ~c6 20 1:ιaeΙ 'W'd7 21 'W'h3 g6 22 ί5! exf5 23 cxd5 ~b5 24 ~xb5 axb5 25 1:ιe6! c4 26 ~c3! ~c5+ 27 ~hl ~h5 28 ~xί5 'W'xd5 29 ~xg6+! hxg6 30 1:ιf1 + ~g8 31 1:ιχg6+ ~h7 32 1:ιg7+ <Rh6 33 1Η6+! ~xg7 34 'W'g4+ 1-0. 12 ~h6(D) Ιί White wishes Ιο avoid the complications οί the folIowing note then he should play 12 b3. One example is the game Spraggett-A.Sokolov, Saint John Ct (10) 1988, which continued 12 ... a5 13 ~a4 ~c5 14 ~h6 ~xa4 15 'iWxa4 ~d7 16 'ίi'ί4 ~ί8 and now 17 h4 would have given White an edge.

12 ••• ~f8?! This tamely accepts a worse posiιίοη. The cήtίcalιine is 12•• .1:ιb8: 1) White has usually sacήficed his b-pawn with 13 'ii'h3!?, giving ήse Ιο the following possibilities: la) 13•••1:ιg8 14 ~d2!? (14 b3 is also quite good) 14 ... h5 15 ~a4 ~c5 16 ~xc5 -*.xc5 17 b3 a5 18 "f3 'ίi'b6 19 c4 ~b4 20 ~xb4 "xb4 21 1:ιec 1 and White had his customary edge ίη Rogers-Small, WellingΙοη 1988. lb) 13••. 1:ιb4 14 -*.g7 1th4 15 "g3 1tg8 16 ~ί6! with a clear plus for White according Ιο Sax. Α possible continuation is 16... -*.xf6 17 exf6 1tb4 18 "d6 :xb2 19 1tabl!? with excelIent attacking prospects for White. lc) 13•••1txb2! 14 ~g7 1tg8 15 'ii'xh71txg7 16 "i'xg7 ~b4! (D) is perhaps going Ιο force a reassessment οί this line. White has:

lcl) 17tί)a4~xeI18~xb2~c3 (18 ...~xf2+!?) 19 1tbl (19 ~xg6 'ίi'e7!) 19 ..... b6 and now Black is going Ιο win the knight οη b2, and Ι can't see how White will be able to

Tαimanov Vαriation

exploit the resulting ρίη along the bfile. lc2) 17 J.xg6 and now: lc21) 17...fxg6? 18 'ii'xg6+ ιJ;e7 19 'ii'g5+ ιJ;e8 20 'ii'h5+ ιJ;e7 21 %Σe3 is very good for White as 21 •.• Μ 22lba4! %Σχc2 23 'ii'h4+ picks υρ the bishop and 21 ••• J.xc3 22 %Σχc3 'ii'b6 23 'ii'g5+ ιJ;e8 24 h4 willleave Black struggling Ιο stop the h-pawn. lc22) But 17•• :ii'e7! gives Black a good game after both 18 %Σe3 fxg6 19 'ii'xg6+ 'ikf7 and 18 'ikg8+ tbf8 19 tba4 J.xel 20 tbxb2 (20 %ΣχeΙ %Σb4) 20 ... J.c3. lc3) Socοπο-Lebredο, Cabaigun 1990 continued 17 %Σe3 J.xc3 18 J.xg6 tbxe5 19 %Σχe5 (perhaps 19 J.xf7+ tbxf7 20 'ii'xc3 with a mess is the best that White can do) 19...fxg6 20 'ii'xg6+ ιJ;e7 with a ηοΙ ιοο dissimilar position Ιο certain lines of the famous Poisoned Pawn Variation where the black king also gets booted around, only for White Ιο see ίι find sanctuary amongst a central pawn mass (21 %Σh5 'ii'f8 22 %Σh7+ ιJ;d6 23 'ii'g3+ J.e5 24 'fί'a3 c5, for example). Ι have examined this posiιίοη ίη some detail and can't see any way for White ιο get at the black king but the passed h-pawn did enable him Ιο maintain the balance οη a number of occasions. Ι think it's best if Ι leave you Ιο form your own conclusions. 2) Ιη view of the above lines ίι is worthwhile re-examining the solid 13 b3. ECO dismisses this as equal after 13 ... %Σb4 14 'fί'e2 J.g5 (this is ηοι forced) IS a3 %Σb8 16 J.xg5

171

'fί'xg5, but does Black really have enough for the pawn after IS J.xgS 'fί'xg5 16 J.xa6 J.xa6 17 .xa6 ίη this line? Ι doubt ίι ιJ;xf8 13 J.xf8 ιJ;g7 14 lba4 IS c4 %Σb8 16 b3 .e7 (D)

17 %ΣacΙ tbb6 18 tbxb6 %Σχb6 19 h4 The last few moves are fairly ΙΥΡί cal for this variation and require little explanation (ίfΥοu'νe looked at the variations above). White stands better as Black is restricted by the weak squares οη his kingside and must also take care that White can't open the queenside at a moment favourable Ιο him. These factors, combined with White's extra space and more active minor piece, mean that ίι will take an extremeIy resolute performance from Black ΙΟ avoid being overstretched at some ροίηΙ 19 ••• hS Weakens g6, but Black can't allow White Ιο play h5. 20 .r4 aS

172 Taimanov Variation 21

1:te3

Black plans Ιο activate his bishop by ... J.a6. but this couldn't be done at once ίη view of 22 c5. 22 .rΣK3 J.a6? (D) Black should have reinforced the g6 square with 22.••:h6, although this does of course leave him very passively placed. The move chosen meets a tactical refutation.

23 :c3! The fιrst ροίηΙ behind this move is that ίι defends b3 so White is now threatening 24 a3 :bb8 25 cxd5 J.xd3 26 d6!. That is what would happen οη 23 .. ':h6, for example. So Black played... 23 ... dxc4 ... and discovered the second Ροίηι behind :c3... 24 .txg6! cxb3 Taking the bishop leads Ιο mate: 24...fxg6 25 :xg6+ Φχκ6 26 :g3+ Φh7 27 'We4+ <lth6 28 'ikg6#. 2S .te4+ Φf8 26 axb3 Black is quite lucky Ιο be οη the board at all but with his king posiιίοη destroyed and several pawn

weaknesses he will still need a miracle Ιο avoid his fate. cS 26 ••• 27

"e3

Breaking the the c5-pawn.

ρίη

and eyeing

υρ

:d4 28 J.f3! With dl protected and his bishop οη a more secure square White is now ready for positive action. If Black takes οη h4, White simply takes οη c5 opening υρ the queenside. 28 Φe8 29 'ii'cl c4 30 bxc4 "xh4 If Black had tried Ιο blockade with 30..:iνcs, White wins with 31 'We3!; the threat is :d3 and since 31.. :i'a7 loses after 32 :d3 J:ιd7 33 J.c6, Black must try 31 ...J.xc4 when there follows 32 J.e2 'Wd5 33 'Wcl! J.a6 (33 ...J.xe2 34 :c8+ <J;e7 35 'Wc7+ 'Wd7 36 'Wc5+) 34 :c8+! J.xc8 35 'ffxc8+ 'ffd8 36 J.b5+ <J;f8 37 'i'c5+ 'fie7 38 'fixd4 and White wins. 31 "a3! :Xc4 32 J.c6+ 1-0 27 ...

Game23 Gallagher - Cramling Biel1990

1 2 3 4 S

00 d4 tαιd4

lbc3 Ι4

(D)

cS e6 cxd4 lbc6 "c7

Tαimanov Vαriαtion

173

23 b3 lbxe3 24 "xe3 (White has a positionally won game) 24 ...b5 25 :hf1 J.b7 26 "b6 :ac8 27 1Wd6! "xd6 28 exd6 :cd8 29lbd7! :xf1 30 :xf1 ~g8 31lbf6+ ~h8 32 g5 e5 33 lί)d7! (White returns Ιο d7 for his fourth knight sacήfice of the game) 33 ... ~g8 34 lbxe5 :xd6 35 :e6 36 :g7+ 1-0 GallagherJansa, Royan 1989. 6••• a6 is the main alternative Ιο the text; 7 lbxc6! 1Wxc6 (7 ... bxc6 is played with about equal frequency and transposes Ιο lines considered ίη Game 22) 8 J.d3 (D) (8 J.e2 b5 9 J.f3 is another reasonable system which doesn 't form part of our reρer toire) and now:

:f7 6 ... fud4 6•••d6 is an attempt Ιο reach a Scheveningen, but White has some dangerous plans based οη castling queenside. For example 7 .te3 (7 lbxc6 is ηοΙ so good when Black hasn't wasted a tempo οη ... a6) 7... lί)f6 8 J.e2 (perhaps 8 "f3 is more natural when White achieved a good game ίη Lau-Andonov, St John 1988 after 8 ... J.d7 9 0-0-0 lbxd4 10 J.xd4 a6 11 g4 e5 12 J.e3 exf4 13 J.xf4 J.c6 14 g5 ~7 15 J.c4 lbb6 16 J.b3, but ίfI were Ιο recommend only that, then Ι wouldn't be able Ιο show off the following game) 8...J.e7 9 "d2 ο-ο 10 0-0-0 a611 g4 d5 (11 ... lbxd4 12 .txd4 e5 13 J.e3 J.xg414 J.xg4lbxg415lί)d5 1Wd8 16 J.b6 'ii'd7 17 :hgl! is winning for White) 12 e5lbd7 (the position now resembles a French Defence) 13 h4lί)b6 14 h5lba5 15lί)f5!? lί)bc4 (15 ... gxf5 16 J.xb6 'ii'xb617lί)xd5 and White will regain his piece with interest) 16lί)xe7+ "xe7 17 J.xc4 lί)xc4 18 "d4 f6 (18 .....b4 19lbe2 and everything is under control) 19 h6 fxe5 20 fxe5 g6? 21 lί)e4! b6 (21 ... lbxe3 22lbf6+) 22lί)f6+ ~h8

1) 8 ...J.c5?! (a dubious move order) 9 "e2 lbe7 10 J.d2 b5 11 e5!? d5? (better is 11 ... J.b7, when 12 J.e4 "c7 13 J.xb7 "xb7 14lbe4 "c6 15 b4 is unclear but 12 lbe4 should be good enough for an edge) 12 exd6 (perhaps Black expected 12lbxb5, which is ηοΙ so clear after 12...:b8) 12.....xd613lί)e4 1Wc614 lί)xc5 "xc5 15 J.c3 ~f8 (15 ...0-0 160-0-0 and Black will be subjected

174

Tαimanov Vαriαtion

Ιο a massive attack) 16 .1e4 :a7 17 0-0-0 :d7 18 :xd7 .1xd7 19 :d 1 .1e8 20 1Id2 and White had a winning position ίη Ga1lagher-Meulders, Eupen 1992. Ιη fact after only seven more moves Black was ίη tota1 zugzwang: 20 ... f5 21 .1f3 q;f7 22 1Id6! 1Ixd6 23 :xd6 h5 24 :xa6 :h7 25 :a7 g6 26 .1b4 q;f6 27 h4 and any move Black makes loses at least the exchange, including 27 ...e5 28 fxe5+ ~e6 29 .1xe7 q;xe7 30 .1d5+. 2) 8 ••• b5 9 'ii'e2 .1b7 (9 ... b4 is met by 10 lL!d5 and after 1O... 'ii'd6 White can choose between 11 e5 1Ic6 12lL!e3 with a nice square οη c4 awaiting the knight, or IIlL!e3 1Ixf4 12lL!c411c7 13 e5 with a lοt of pressure for a pawn) 10 .td2 (D). Of course White doesn't play 10 .1e3, which would deprive him of the answer lL!d5 ίη reply Ιο ... b4. White is now ready Ιο castle either side depending οη developments. There haven't been many practica1 examples from this position, but what we 've seen so far suggests that White is doing quite well. Black's main moves are:

2a) 10••• lL!e7 (10 ... .1e7 is very passive and after 11 a4!? b4 12lL!a2 a5 13 c3 bxc3 14lL!xc3lL!f615 :cl, White was already much better ίη Hector-Gonotaxtegi, San Sebastian 1991) 11 a3 (11 a4 and 11 ο-ο come ίηΙο consideration, but now that ... b4 can't be met by lL!d5 it's ηοΙ a bad idea Ιο secure the knight οη c3) 11 ... lL!g6 (11 ... lL!c8 should also be met by 12 Ο-Ο) 12 ο-ο .te7 (after 12....1c5+ 13 q;hl White follows the same plan as ίη the game) 13 :f3! 1Ic7 14 :af1 h5 (14 ... 0-0 15 :h3 gives White a very strong attack as long as he remembers Ιο play q;hl before 1Ih5 ίη view of ...1Ic5+ exchanging queens) 15 q;hl h4 16 :h3 e5? (a mistake ίη a poor position) 17 f5 lL!f4 (17 ... lL!f8 18 1Ig4 .tf6 19 .tg5 is also very bad) 18 .txf4 exf4 19 e5 f3 20 :hxf3! (of course) 20 ... .1xf3 (flicking ίη 20 ... h3 would make ηο difference after 21 g3) 21 1Ixf3 :b8 22 :el with an overwhelming position for White, Yudasin-Balashov, Minsk 1982. 2b) 10....tc5 11 0-0-0 lL!e7 (playing 11. .. b4 is still ήsky; after 12lL!d5 1Id6 13lL!e311xf4 14lL!c4 1Ic7 15 1Ig4 White has a very strong initiative) 1211h5! b4?! (12 ...%Σc8 13 q;bl ο-ο 14 %Σhf1 "b6 15 g4 with an edge for White, Tolnai-C.Horvath, Ηυη garian Ch 1989) 13 lL!a4! d5 (the only move as both 13 ...1Ixa4 14 1Ixc511xa2 15 .1xb4 and 13 ....1d6 1411a5! are very bad for Black) 14 lL!xc5 "xc5 15 :hel :c8 16 ~bl a5 (Black is ίη difficulties as he can't castle) 17 .te311c7 18 exd5lL!xd5

Tαimanov Vαriαtion

19 f5! ο-ο (forced) 20 fxe6 g6 21 'it'h6 and White had a clear advantage ίη Gallagher-Illescas, French League 1990. 7 'it'xd4 a6 8 R.e2 bS 8"0~7 is well met by 9 'Wf2. One example is the game Am.Rodrίguez Zapata, Bayamo 1987, which continued 9 ... b5 (more accurate than 9...lίk6 as ίη that case White will be able to omίΙ a3, for example 1Ο R.e3 b5 11 ο-ο R.e7 12 e5 ο-ο 13 lbe4 R.b7 14 R.b6 'it'c8 15 R.c5 With a clear advantage for White according Ιο Boleslavsky) 10 R.e3 R.b7 (10 ... b4 11lL1a4 'Wxc2 12lL1b6 :b8 13 lίk4! lLIc6 14 e5 followed by 0-0 gives White tremendous play for the pawn) 11 a3 lLIc6 12 ο-ο R.e7 13 R.b6 'Wc8 14 :adl ο-ο 15 R.f3 f6 16 :fel with a pleasant position for White. 9 R.e3 R.b7 Black can try 9ooolLle7 here as well. After 10 ο-ο lLIc6 11 'i'd3 R.b7 12 e5 R.e7 13 :adl :c8 14 a3lL1a5 15 ΦhΙlL1c4 16 R.cl R.c6 17lL1e4 ο-ο 18 R.f3 White had a slight advantage ίη Chandler-Zapata, Amsterdam 1987. 10 0-0-0 Of course this is quite risky but the results with 10 ο-ο have been uninspiring. Black is thought Ιο be able Ιο equalize by 1O...:c8 11 J:ιfdl (11 :adllL1f6 12 R.f3 h5 is also ΟΚ for Black) 11 ... R.c5 12 'Wxc5 'it'xc5 13 R.xc5 :Xc5 14 :d4lL1e7! 15 :adl and now both 15 ... R.c6 and 15 ... d5 are roughly level. This ending is ηοΙ

175

as promίsing for White as the one we consider later (note to Black's 12th move). There White castled queenside, which ίη effect gains a tempo as the rook reaches the d-file ίη one go and the black knight didn 't have the opportunity Ιο develop ιο the superior e7-square. One may also argue that the white king is better off οη c 1 than g1. 10 000 11 :d2 (D) This may appear artificial, but White must protect the c2-square.

:c8

11 000 lLIf6 l1oooR.cS 12 'Wxc5 'Wxc5 13 R.xc5 :xc5 14 :hdl R.c6 (14 ... :c7 15 R.xb5! and 14...lLIf6 15 e5 lLId5 16 lLIe4 are very good for White) 15 a3 lLIf6 16 R.f3 is ;j; according Ιο Kupreichik. 12 J.f3 'it'a5 12oooR.cS 13 'it'xc5 'Wxc5 14 R.xc5 :xc5 15 :el is considered Ιο be a little better for White. 12ooo'it'c4 is well met by 13 e5. 12o.ob4 13lL1a4 'Wc4 14 b3! 'Wxd4 15 :xd4 gave White an edge ίη Kupreichik-Kotronias, Lvov 1988.

176

Tαimanov Vαriαtion

That leaves 12...J.e7 (D) as the main altemative. White then has:

1) 13 :hdl ο-ο 14 e5 and now: la) 14•••J.xf3? 15 gxf3 b4? 16 exf6 bxc3 17 1::tg2! 'i'b7 18 1::txg7 + Φh8 19 :'g8+! and Black resigned ίη Ehlvest-Andersson, Belfort 1988 ίη view of 19 ...:txg8 20 fxe7+ 1::tg7 21 1::tgl 'i'xb2+ 22 ΦdΙ "bl+ 23 J.cl when mate follows shortly. lb) 14••. b4! 15 exf6? (15ltJa4 is better, when 15 ... J.xf3 16 gxf3ltJd5 17ltJb6ltJxb6 18 "xb6 "c4 19 b3 'iWc3 20 "d4 leads Ιο a roughly level endgame) 15 ... bxc3 16 'i'xc3 (16 fxe7 cxd2+ 17 1::txd2 :'fe8 18 J.xb7 "xb7 19 "xd7 :'xe7 is good for Black) 16.....xc3! 17 fxe7 (17 bxc3 J.a3+ 18 'iitbl J.xf3 followed by ... :'b8+ wins for Black) 17 ... 'i'xe3 18 exf8"+ ΦΧf8 19 J.xb7 :tb8 and Black soon won, Hector-P.Cramling, Valby 1991. Α word of advice: if you find yourself facing a male Swedish grandmaster then you should head for this varίation, regardless of colour, as they seem Ιο find a way Ιο lose all their pieces or get mated within 20 moves.

2) 13 g4!? is thought Ιο allow Black to equalίse with 13••• J.c5 14 "ii'xc5 "ii'xc5 15 J.xc5 :Xc5, the idea being that the tempo White has gained, g2-g4, actually harms his position. However, I'm ηοΙ so sure about this theory; admittedly the ρίη οη the e-pawn can prove annoying but White does have new possibilίties based οη playing g5. Ι (JG) have looked at this positίon ίη some depth and found some fascinatίng variations. ΜΥ analysis runs 16 :'hdl! (D) (White must be careful ηοΙ Ιο play g5 Ιοο early as Black can sometimes develop counterplay by ... ltJg8 and ... h6, while ίη this particular position the exchange sacrifice ση c3 is also interestίng) and now:

2a) 16..• b4 17 ltJa4 :'c7 18 :td4 looks very good for White, for example 18 ... a5 19 g5 ltJg8 20 ltJb6 J.c6 21ltJc4 with a clear advantage. 2b) 16••• :'c7 17 :td4 J.c6 18 a4! also promises White a slίght advantage. 2c) 16•••J.c6100ks lίke the cήtί calline as 17 :'d4 can now be met by 17 ... e5 and 17 a3 a5 doesn't solve

Taimanov Variation 177 White's problems with his vulnerable e-pawn. Therefore 17 g5 and now: 2cl) AfterI7••• b418lbd5! both 18•••exd5 19 exd5 lbxd5 20 .*.xd5 and 18••• lbxd5 19 exd5 .*.b7 20 .*.g2! are better for White, an important ροίηι ίη the latter variation being the tactic 20...0-0 21 dxe6! .*.xg222 exd7. 2c2) 17•••lbg818lt)e2! (D) (now that the pressure οη e4 has been relieved the knight is free to find a square with more potential) with a further branch:

2c21) 18••. h6? (what could be more natural?) 19lbd4 when Black has three moves: 2c211) 19.. ib7 20 lbxe6!! fxe6 (obviously 20 ... dxe6 al10ws mate ίη two) 21 ':xd7 and Black is defenceless. 21 .••.*.c6 gets mated ίη a few moves after 22 .*.h5+ and 21 •••.*.d5 22 J:ιxg7 only delays the inevitable. The best he can do is 21 •••hxg5 (preventing .*.h5+) 22 ];[xb7 J:[xh2 23 c3 when 23••. gxf4 loses a piece after 24 J:[b8+ and 23••.lbf6 24 e5 lbd5 25 .*.xd5 gives White a won ending.

2c212) 19••ia820lbxe6!! fxe6 21 ':xd7 (21 i.h5+ g6! 22 i.xg6+ Φf8 23 :xd7 lt)e7 is less clear-cut) 21 ... hxg5 22 :d8+ </;;f7 23 :xa8 (23 :d7+ is also promising) and now 23•••:xh2 24 c3 is winning ίοι White as Black is liable ιο lose his knight and 23••• gxf4 is wel1 met by 24 h4! with a fatal ρίη οη the eighth rank. 2c213) 19•••hxg5 (Black may try this if he spots 20 lbxe6 ίη time) 20 lbxc6 :xc6 21 :xd7 :c8 22 e5! (the most important thing is ιο activate the bishop; now 22...:xh210ses ιο 23 i.c6!) 22 ... lbe7 23 i.b7 :b8 24 fxg5 and now both 24••• lbg6 25 i.xa6 lbxe5 26 :c7! with the idea of :c5 and 24•••:xh2 25 .*.xa6lΩd5 26 :b7! are good for White. 2c22) 18•..f5?! (another case of Black looking ίοι active play before he's ready for it) 19l2Jd4 and now: 2c221) 19•.ixe4 20 lbxe6! dxe6 21 :d8+ </;;f7 22 i.h5+! g6 23 ':ld7+ lbe7 24 :xh8 :xc2+ 25 </;;dl and White wins. 2c222) 19..•fxe4 when White can gaίn the advanιage ίη prosaίc fashion by 20 lbxc6 exf3 21 lbe5 d5 22 lbxf3 fol1owed by lbd4, οι ίη more spectacular fashion by 20 i.h5+ g6 21 lbxc6 :xc6 22 :xd7 gxh5 23 g6!! lbf6 (23 ...hxg6 24 :d8+ Φf7 25 :ld7+ Φί6 {25 ... lbe7 26 ':xh8 +-} 26 :f8#!) 24 :d8+ </;;e7 25 :xh8 hxg6 when Black has some compensation for the exchange. 2c23) 18.••lbe7! 19 lbd4 i.b7 when after a normal continuation like 20 lbb3 :c7 21 lba5 .*.c8 it's

178

Tαimαnov Vαriαtion

difficult Ιο see how White can step υρ the pressure. However, he does have an abnormal continuation at his disposal, 20 lDxe6!?, which is nowhere near as devastating as ίη the above lίnes but still poses Black some problems. After 20••.fxe6 21 ':xd7 .*.dS! 22 :a7 (22 IΣd6 loses Ιο 22 ...1Ic6! and 22 1Ixe7+ rj;xe7 23 exd5 ':c4 looks at least ΟΚ for Black) there are two possibilities: 2c231) 22•. ic6? 23 1:ιd6! Φf7 (Black can't bail ουΙ with 23 ... 0-0 οη account of 24 ':xe7 1Ixf4 25 1Id8+ while 23 ... h6 24 .*.h5+ g6 25 1Ixe6 .*.d7 26 .*.xg6+ ~d8 27 1Id6 is also very good for Whίte) 24 .*.g4! and the black position is caving ίη; οη 24.••eS, 25 b4! is the most efficient while 24••..*.xe4 25 .*.xe6+ Φe8 26 1Idd7! is also the end. 2c232) 22•.ixa2! 23 1:tdd7lDc6! (ηοι 23 ...lbc8 24 .*.h5+! g6 25 1Ixh7 :'xh7 26 .*.xg6+ Φf8 27 1:ιχh7 when White's passed pawns should prove decisive) and now: 2c2321) 24 .*.hS+!?? (if you are a raving lunatic, then here is your winning attempt) 24 ... g6 25 1Ixh7 1:ιχh7 26 .*.xg6+ Φd8! 27 :'xh7 lDd4 28 Φd2 :'xc2+ 29 ~d3 :'c4 30 h4 and the race is οη between the white pawns and the blackmating attack. 2c2322) 24 1:ιac7 (sensible) and now 24•••1:ιf8 loses Ιο 25 1Ixg7 :'xf4 26 :'g8+ :'f8 27 :'c8+, 24 .•..*.b3 Ιο 25 :'xg7! while 24••• lDb8 25 .*.h5+! g6 26 1Ixh7 is also good for White. However, Black has just one saving move, 24•••lDb4!, after which White

has nothing better than ιο take a draw by perpetual check. So, perhaps theory is right and the position after 13 g4 can be assessed as =, but there is certainly plenty of rope for Black Ιο hang himselfby! Returning Ιο the position after 12 .....a5 (D):

13 'i'a7! This required careful calculation as Black has a dangerous looking exchange sacήfice. Instead 13 e51ed ιο great complίcations ίη Polίhron iade-J.Polgar, Thessaloniki wom OL 1988 after 13 ... .*.c5 14 'ifd3 .*.xe3 15 "xe3 b4 16 .*.xb7 (16 exf6? bxc3 17 bxc3 "xa2) 16... bxc3 17 1Id3 (17 .*.xc8? cxd2+ 18 "xd2 "xa2 is very good for Black) 17 ...lDg4 and now White should have played 18 "g3 when Judίt and Zsuzsa Polgar give the vaήation 18 .....b5 19.*.xc8 "xb2+ 20 'itί>dl "bl+ 21 Φe2 "xc2+ 22 ΦeΙ "bl + 23 Φe2+ with a draw by perpetual check. 'i'c7 13 ..• 13•.•.*.a3 fails ιο 14 lDbl, but 13•••:.xc3 is more complίcated. ΜΥ

Taimαnov

idea was 14 bxc3 'iWxc3 (14 ... j.a3+ 15 ~dl) 15 :Ιd3! andafter 15 ...'iWal+ 16 Φd2 j.b4+ 17 Φe2 'iWxhl 18 :ΙdΙ 'iWxh2 19 'iWb8+ Φe7 20 'iWxb7! White has a very strong attack. Α sarnple varίation runs: 20... e5 21 c3! j.d6 (21 ... j.a3 22 'iWxa6) 22 :Ιχd6! ~xd6 23 j.c5+! and now the lίnes 23••. Φχc5 24 'iWc7 and 23••. Φe6 24 ί5 are checkmate. 14 e5!? Ι was a bit nervous about my queen οη a7 so Ι opted for the ending at once. However 14 :ΙeΙ deserves serious consideration: 14••• b4 can be met by 15llJd5! and 14••• j.b4 by 15 j.b6! (15 j.d4 J:ιa8) 15 ... 'iWb8 (15 .....c6 16 e5lM5 17 :Ιχd5! wins for White) 16 'iWxb8 :Ιχb8 17 j.d4 with an edge for White. 14 j.xf3 15 'iνxc7 :Xc7 16 gxf3 lbg8 Ι was quite surpήsed by this retreat and had been expecting 16••• Μ, after which Ι intended 17 lba4 lbd5 18 :Ιχd5! exd5 19 :ΙdΙ with good long-term compensation for the exchange. 17 lbe4 lbb6 18 :Jιdl Ι was ηοΙ interested ίη 18 lbd6+ j.xd6 19 :Ιχd6 lbf5 20 :Ιχa6 0-0 when BIack has good play for the pawn. 18 lbf5 19 j.b6 :Ιb7 20 j.cS :Ιc7 21 j.b6 :Ιb7

22 j.cS

Variation 179 :Ιc7 (D)

23 .txf8 overestimated my chances ίη the resulting double rook ending. Ι should have played 23 b4! and after 23 ... .te7! 24 j.xe7 lbxe7 25 lbc5 lbd5, 26 :Ιχd5 is another promising exchange sacήfice. 23 :Ιxf8 24 lbd6+ lbxd6 25 1hd6 J:ιa7 26 :b6 Φe7 (6! 27 :dd6 Now Black wil1 get a rook Ioose arnongst my weak kingside pawns. Ι didn' t like the look of 28 ha6 :Ιχa6 29 :Ιχa6 fxe5 30 fxe5 :Ιxf3 31 :Ιb6 g5 32 :Ιχb5 g4! when anything can happen, so Ι played the solίd ... 28 :Ιd3 ... and offered a draw, which was accepted. After 28 ... fxe5 29 fxe5 :Ιί5 30 :Ιe3 :Ιh5 31 :Ιe2 :Ιh3 32 :Ιf2 g5 33 :Ιd6 the game is level. Both players have one active rook and one passive rook tied down ιο their weak pawns. Ι

9 SiciIian Four Κnίghts This rather antiquated system has never real1y caught οη as the generally passive nature of Black's posiιίοη has proved unattractive Ιο most players. Black plays 1 e4 c5 2 ~f3 e6 3 d4 cxd4 4 ~xd4 ~f6 5 tα3 ~c6 (of course this can arise from other move orders, ίη parιicular via 2 ... ~c6). Ιη some ways this is akin Ιο the Kan and Taimanov Systems since Black leaves the f8-b4 diagonal open for his bishop, but instead of playing ... a6 he develops a piece. Naturally this is ίη Black's favour unless White has some direct method of exploitίng the ornίssion of ... a6, so 6 ~db5 is the οηlΥ move to cause problems. Black then very often continues 6 ... d6 and after 7 .tf4 e5 8 ..tg5 we have transposed Ιο the Pelikan, considered ίη Chapter 4. The ροίηι of this move order is that Black avoids the possibility that after 1 e4 c5 2 ~f3 ~c6 3 d4 cxd4 4 ~xd4 ~f6 5 ~c3 e5 6 ~db5 d6 White might play 7 ~d5 or 7 a4. Since we are recommending the main line with 7 ..tg5 this transposίtion is ηοΙ a wouy and therefore 6... d6 just leads ιο the earlίer chapter οη the Pelίkan. After 6 ~db5, 6 ... d5 loses Ιο 7 exd5 exd5 8 ..tf4 and 6 ... ..tc5 7 .tf4 followed by ..td6 is unpleasant for Black, so we need οηlΥ consider 6 ... ..tb4 ίη the cuuent

chapter. The tactical lίne recommended ίη the Μι edition has suffered a serious setback ίη recent years, so this time (as ίη BTS2) we οηlΥ analyse the positional continuatίon 7 a3 ..txc3+ 8 ~xc3 d5 9 exd5, which eίther gives Black an isolated pawn after 9 ... exd5 or gives White a lead ίη development after 9... ~xd5 10 ..td2. This is a safe lίne for White ίη which he is lίkely Ιο secure a small but permanent advantage. Ιη practice ίι is easy for White Ιο allow the position to slίde towards a draw, and ίη some ways ίι is an annoying lίne Ιο meet because instead of the sharp struggle typical of most Sicilίan lίnes, White is trying Ιο exploit a slight positional edge. Nevertheless ίι is even more unpleasant for Black, who can only win ifWhite takes exceptional risks, and so this line is relatively unpopular. Game24 Mokry - Β. Stein Gausdal1988 1 e4 2 ~f3 3 d4 4 ~xd4 5 ~3

cS e6 cxd4 lbf6

6 ~bS 7 a3

..tb4 ..txc3+

Siciliαn

8 tαιc3 9 exdS(D)

9 ••• exdS The knight capture is quite rare these days, but it's interesting Ιο note that Ulf Andersson seems quite happy ιο defend the unpleasant ending that arises from line '1' below. After 9••.lbxdS 10 .id2 Black has: 1) 10••.lbxc3Il.ixc3 'ii'xdl+ 12 lIxdl f6 (12 ... e5 13 .id3 .ie614 0-0 f6 15 f4 is similar, Ιvkοv-Gligοήc, Amsterdam 1964) 13 f4 (for some reason Ehlvest prefened the passive 13 f3 and after 13 ....id7 14.id3 0-0-0 15 1:d2 e5 16 .ie4 .ie6 17 Φe21:χd2+ 18 ΦΧd2lbe7 19.ib4 lbd5 20 .ixd5 .ixd5 21 Φe3 the players agreed Ιο a draw ίη Ehlvest-Andersson, Skelleftea 1989) 13 ....id7 14 .ic4 0-0-015 ο-ο 1:he8 (15 ... Φc7 16 1:del 1:he8 17 1:f3 .ic8 with a very unpleasant positίon for Black, Fischer-Addison, USA Ch 1962/3) and now instead of 16 b4lbe7 17 b5 00 18 .id2lbb6 19 .id3 lba4 20 1:bl lbc5! (the fifth successive move by this knight) which led Ιο a roughly level game

ίη

Four Knights 181

Psakhis-Andersson, Manila OL 1992, Ι would suggest 16 %:ιdeΙ, α lα Fischer, which prevents ... e5 (unless Black doesn't mind a rook ρenetrat ing Ιο f7) without making a mess of the queenside. 2) 10.....h4 11 'ii'f3 ο-ο (playing 11 ... lbe5 12 'ii'g3 and 11 ...lbd4 12 'ii'd3 just make matters worse) 12 0-0-0 lbxc3 13 .ixc3 e5 14 .id3 .ig4 (14 ... 'ii'g4 15 .ie4 'ii'xf3 16 .ixf3 with the typical favourable ending for White, Minic-Gerusel, Halle 1967) 15 'ii'e4 'ii'h6+ 16 .id2 'ii'g6 17 f3 .ie6 18 'ii'xg6 hxg6 19 .ie3 and again White has a promising ending, Tal-Matulovic, Kislovodsk 1966. 3) 10..:.b6 11 lbb5 ttJd4 12 lbxd4 'ii'xd4 13 .ib5+ .id7 14 .ixd7+ ΦΧd7 15 ο-ο left Black's king badly placed ίη Kaplan-Siaperas, Siegen OL 1970. 4) 10.....f6 11 'ii'h5 ο-ο 120-0-0 lbxc3 (12 ... 'ii'xf2 13lbxd5 exd5 14 .id3 is ηο better) 13 .ixc3 'ii'f4+ 14 1:d2 e5 15 .ib5 with advantage, Matulovic- Κοkkοήs, Athens 1969. 5) 10..•0-0 11 'ii'h5lbf6 12 "h4 'ii'd4 13 .ig5 1:d8 14 'ii'xd4 ':xd4 15 .id3 b6 16 0-0-0 is the same story as ίη all the other lines, Gufeld-Κhasin, USSR Ch 1966. 10 .id3 ο-ο Or 10..•d4 (after 1O... 'ii'e7+ 11 'ii'e2 'ii'xe2+ 12lbxe2lbe5 13 .ib5+ .id7 14 .ixd7+ ΦΧd7, LiberzonBronstein, USSR 1972, White could have played 15 .ie3 1:he8 16.id4 with a slight advantage) with the variations:

182

Siciliαn

Four Knights

Ι) 11 'jj'e2+ J.e6 12 tLΊe4 tLΊxe4 13 'jj'xe4 'jj'dS (D) and now:

la) 14 'ii'xd5 J.xd5 15 J.f4 ο-ο 16 ο-ο :fe8 17 :fe 1 f6, FernandezIllescas, Pamplona 1991, and now instead of 18 :xe8 :Ιχe8 19 J.g3 g5!, which was followed υρ with ... J.f7 -g6, White should have prevented Black's kingside expansion by 18 h4!, after which he could have claimed an edge. 1b) 14 J.f4 0-0-0 15 ο-ο g5 16 J.d2 h6 (16 ... tLΊe5 17 "xd5 J.xd5 18 J.f5+) 17 Wxd5 J.xd5 18 f4 g4 19 :ΙaeΙ h5 20 b4 and White is better, IIlescas-Sorokin, PaIma 1989. 2) 11 tLΊe2 J.f5 (11 ... 0-012 ο-ο transposes ίnΙo the main line) 12 ο-ο J.xd3 13 Wxd3 ο-ο and now: 2a) 14 b4!? (White plans to help himself Ιο the d-pawn) 14...'jj'b6 (14 ...:c8?! 15 J.b2 :e8 16 :adl tLΊe5 17 "xd4 "c7 18lbc3 was very good for White ίη Jonsson-Κrist jansson, Akureyri 1994) 15 J.b2 (15 J.g5 Ι?) IS ...:fd8 16 :fdl (perhaps 16 :adl, but then White won't be able to play the ending given below) 16...tLΊe5 (16 ... a5 17 b5 tLΊe5 18

J.xd4! is very good for White) and now: 2al) 17 J.xd4 tLΊxd3 18 J.xb6 axb6 19 cxd3 with an extra pawn for White but aIso positional problems which will make this advantage extremely hard Ιο reaIise. 2a2) 17 'jj'g3lbc6 18 c4 (18 "'h4 looks strong, but Black has the unpleasant reply 18 ... :Ιe8!) 18 ...dxc3 19 J.xc3 :xd 1+ 20 :Ιχd 1 :d8! with just a smaIl plus for White. 2b) 14 J.g5 h6 15 J.h4 :e8 16 :Ιad1 :Ιc8 17 c3! (a more dynamic move than 17 :fe1 :e618 Φf1 "'c7 19 J.g3 "b6 20 b4 with just a microscopic advantage for White, Karpov-Kuzmin, Leningrad 1977) 17 ... dxc3 18 "h3 "e7 19 tLΊxc3 "'e6 (19 .....e5 20 f4 "'e3+ 21 "'xe3 :xe3 22 J.xf6 gxf6 23 tLΊd5 is very good for White) 20 "xe6 fxe6 (20...:xe6 21 J.xf6 :xf6 22 :d7 b6 23 :e1 favours White) 21 J.xf6 gxf6 22 tLΊe4 with an endgame advantage for White, Estevez-Chaviano, Santa C1ara 1983. 11 ο-ο (D)

11 ...

d4 (D)

Siciliαn

This is the most logical because forces White ιο decide where his knight is going immediately. Ιη order ιο avoid liquidation ίι seems that 11 ... d4 should be met by the relatively passive ll)e2; after other 11th moves White can usually arrange ιο meet ... d4 by the more active ll)e4, for example: 1) 11 •.•a612 .tf4 (12 .tg5 is also promising) 12... d4 13 ~ ll)d5 (the line 13 ....tf5 14 .tc7! il1ustrates why Black should ηοΙ have delayed) 14 .td6 ':'e8 15 .tg3 f5? (suicide, but even 15 ....tf5 16ll)d6 .txd3 17 'ifxd3 :e7 18ll)f5 is very awkward) 16 ll)d6 ':'f8 17 .tc4 .te6 18 ':'el 'ifd7 19l1)xb7 "xb7 20 ':'xe6ll)a5 21 .ta2 1-0 Vukcevic-Ervin, USA 1976. 2) 11 •..h6 12 .tf4 d4 13ll)b5 (13 ll)e4 as ίη line '1' is also possible) 13 ... ll)d5 14 "f3! .te615 ':'fel (15 ':'adl "d7 16 h3 :ad817 .th2 "e7 18 "g3 also gave White some advantage ίη Ciric-Rossolimo, Vr§ac 1969) 15 .....d7 16 .te5! ll)xe5 17 ':'xe5 ll)e7 (17 ... ll)f6100ks more resilient but Black is still ίη difficulties after 18 "f4 ':'fd8 19 ':'dl) 18 'ii'e4 ':'fd8 19 'ifh7+ Φf8 20 'ifh8+ ll)g8 21 ':'ael (threatening .th7, as ... f6 is ηο longer a defence) 21 ... .td5 22 ':'xd5! _xd5 23ll)c7 'ifd7 24 .th7 1-0 Timoshenko-Chernov, Bucharest 1993. 3) 11••..tg4 12 f3 and now: 3a) 12.••.te6 13 .tg5 h6 (both 13 ... ':'e8 14 "d2 d4 15 ll)e2 a6 16 ll)g3, Planinc-Andersson, Sombor 1970 and 13 ... 'ifb6+ 14 <itthlll)d7 ίι

Four Knights 183

15 f4! f5 16 "f3, Matulovic-Benko, Vrnjacka Banja 1973 were also bad for Black) 14 .th4 g5 15 .tf2ll)h5 16ll)b5 and according ιο Taimanov White has a clear plus. 3b) 12•••.thS 13 .tg5 'ii'b6+!? 14 <itthl ll)e4 (this is imaginative play by Black but ίι fails ιο equalize) 15 ll)xe4 dxe4 16 .txe4 'ii'xb2 17 1t'b1! "xbl 18 ':'fxbl f5 19 .td3 b620 ':'b5 .tg6 21 ':'dl with a somewhat better ending for White, EhlvestRomero, Logrono 1991.

12 tί)e2 The available evidence suggests that this offers the best chances for an advantage. After 12ll)e4 .tf5 13 .tg5 .txe4 14 .txe4 "d6 (the diehards of this system currently favour this move over 14 ...h6, which allows White ιο gain a small advantage after 15 .th4 g5 16 .txc6 bxc6 17 .tg3 "d5 18 f4!, as ίη Κir.Georgiev Chandler, Leningrad 1987) 15 .txf6 (nobody has tried to keep the bishop pair) 15 .....xf6 16 "d3 (16 'ifh5 g6 17 "b5 :ae8 followed by 18 ... :e7 is also ΟΚ for Black, but perhaps 16 'iff3 is a better try for the advantage)

184

Siciliαn

Four Knights

16 ... g6 17 :adl :ad8 18 'iWb5 :d7 19 :fel <j;g7 with an equal garne, Westeήηeη-Κhenkiη, Gausdall991. 12 ••. i.g4 There are several alternatives, with line '3' recently proving the most popular of Black's 12th move choices. 1) 12...:e8 is an interestίng new idea: la) The main point is to meet the standard 13 ~gS with 13 ... :e5. For exarnple, Kamsky-Lautier, Paris Immopar rapid 1992 contίnued 14 i.f4 (14 'iί'd2 ~f5!) 14 ...:d5 15 lίJg3 lίJe5 16 i.xe5 :xe5 17 'iί'd2 'iί'b6 18 :ael :xel 19 :xel i.e6 20 b3 :d8 and BIack was very close to equa1ity. lb) 13 b4, launching an immediate assault οη the d-pawn, seems to pose Black the most problems. For example, Yudasin-Sorokin, Moscow 1992 continued 13 ... i.g4 14 f3 i.h5 15 i.b2 a616 :f2! (threatening lίJxd4) 16... lίJe5 (οη 16... ~h8, 17 g4 would be a ήsky way to win a pawn, but Ι can't see an adequate reply to 17 i.c4!, as 17 ... lίJe5 18 i.b3 d3 fails to 19 .ixe5) 17 .ixd4lίJxd3 18 'iί'xd3 .ig6 19 'iί'd2 :c8 20 c3 and White was a pawn υρ for nothing. 2) 12...h6 13 b4! (if this is good against ... :e8 then ίι must be at least as strong agaίnst the slow ... h6) 13 ... a6 14 i.b2 ~h8 15 'iί'd2 (15 i.c4!?, as above, looks good to me) 15 ... b5?! (15 .....d6 is only;t according to Bangiev) 16 'iί'f4lίJd5 17 'iί'g3 (17 'ί!ff3!?) 17 ... lίJde7 18 :adl 1:Ia7 19 c3 :d7 20 cxd4lίJd5 (PetrushinBangiev, Simferopoll989) and now

21.ibl would have been very good forWhite. 3) 12..:ii'd5 (D), and now White has several ways to handle the position:

3a) 13 c4 (aίming for active piece play) 13 ... dxc3 14 lίJxc3 'ί!fa5 15 .id2!? (15 i.f4 is more common but has not been extremely successful) 15 ... :d816lίJb5'iWa6(16 ... 'iί'b617 .ie3 'ί!fa5 18 b4) 17 'iί'c2 lίJe5 18 i.e2 i.g4 19 .ic3 :ac8 20 :adl lίJd5 21 .ixg4lίJxg4 22 'iί'd3 :d7 23 'iί'f5 'iί'e6 24 'iWxe6 fxe6 25 h3 (25 lίJxa7 is οηlΥ equal after 25 ... lίJxc3! 26 bxc3 {26 :xd7?? lίJe2+ 27 ~hl lίJxf2+! mates} 26 ...:xd 1 27 lίJxc8 :d3 28 1:Icl :d2 29 h3 lίJxf2 30 1:Ιn :d8! 31lίJe7+ ~f7 and White has ηο square for his knight) 25 ... a6 26 lίJa7 1:Icd8 27 i.a5 :a8 28 hxg4 1:Ixa7 29 .ic3 :a8 30 :fel :ad8 31 .ie5 with an edge for White, Dvoirys-Maliutin, USSR 1991. 3b) 13lίJΙ4 "d6 14lίJhS lίJxhS (14 ... lίJg4 15 .if4 'iί'd5 16lίJg3 .ie6 17 :e 1 1Iad8 18 h3lίJf6 19 'ί!fd2 was also a little better for White ίη Schlemermeyer-Stein, Bundesliga 1993)

Siciliαn

15 "xh5 h6 16 :e1 j.d7 (D) and now:

3bl) 17 "Ι3 :fe8 18 j.f4 'ii'f6 19'ti'g3 ~h8 20h3 j.f5! 21 j.xh6!? gxh6 (21 ... j.xd3 22 j.g5) 22 'ii'f4 CfJe7 23 g4 'ii'g6! 24 j.xf5 CfJxf5 25 'ti'xf5 'ti'xf5 26 gxf5 ~g7 27 :adl 1/2-1f2 Smiήn-Κhenkin, Elenite 1993. White's extra pawn is rneaningless. 3b2) 17 'ii'h4 (again White plans j.f4 and this is probably rnore accurate than '3b' as the black queen is kept away frorn her ideal square, f6) 17 ... :ae8 18 j.f4 'ii'd5 19 'ii'g3 :e6 (preferable was 19...~h8 although White would still have an edge) 20 :xe6 'ii'xe6 21 j.d2 'ii'e5 (21 ...:e8 22 h3 'ii'e5 would have been better than the garne) 22 :el 'ii'xg3 23 hxg3 :e8? (this turns a difficult endgarne ίηΙο a lost one - without rooks οη the board White's king will becorne Ιοο active; the last chance for Black Ιο rnount a successful defence was 23 ... j.e6 followed by ...:d8) 24 :xe8+ j.xe8 25 f4 j.d7 26 ~f2 ~f8 27 .te4 q;e7 28 ~e2 ~d6 29 ~d3 b6 30 j.b4+ ~e6 31 j.f3 f5 32 ~c4 a5 33 j.d5+ ~f6 34

Four Knights 185

j.d6 and faced with the destruction of his queenside by .tc7 Black called ίι a day, Tiviakov-Sorokin, St. Petersburg 1993. 3c) 13lbg3 was recoιnrnended ίη BTS2 but has still ηοΙ been tested ίη a seήοus garne. The only exarnple is Cavello-Mellado, frorn the Oviedo rapid 1991. Play continued 13 ... lbe5 14 j.f4 lbxd3 15 'ii'xd3 .te6 16 :adl :fd8 17 :d2 :ac8 18 :el 'ii'c4 19 'ii'xc4 :xc4 20 c3? lbd5 21 .tg5 dxc3! with advantage Ιο Black. White should have tried 20 .te5 with the threat of 21 b3; 20••':dc8 can be rnet by 21 :ee2 and Ι (1G) suspect that after 20...lbg4 21 b3 lbxe5 22 bxc4 lbxc4 23 :d3 that Black doesn't have qώte enough for the exchange. Returning Ιο the rnain line after 12 ... j.g4 (D):

13 f3 This appears best, because the exchange οη e2 usually sirnplifιes Black's defensive task: 1) 13 j.f4 :e8 14 :el 'ii'b6 15 b4lbe5 16 .txe5 :Xe5 17 'ii'd2 j.xe2 18 :xe2 :xe2 19 j.xe2 g6 20 :dl

186

Siciliαn

Four Knights

.:td8 21 .tf3 gave White a sma11 but lasting advantage ίη Lobron-Chandler, Biel 1987. 2) After 13 .tg5 'ifd614 ':el (14 'iί'd2 .txe2 15 'ifxe2 .:tfe8 16 'iί'dl tt1e5 offers White nothing, ShortWiedenkeller, Esbjerg 1984) Black has tried: 2a) 14.•• &5 15 'iί'd2 tt1d5 16 h3 .txe2 17 .txe2 h6 18 .th4 was the continuation of R.Mainka-B.Stein, Dortmund 1987. Black now tried Ιο imprison White's bishop by 18 ... f5 19 c4 tt1f4 20 .tf3 tt1g6 21 .tg3 f4 but this rebounded badly: 22.td5+ ~h7 23 .:te6 'ifd7 24 'iί'd3 tt1ce7 25 .th2 tt1f5 26 .:tael ':ab8 27 'ife4 b6 28 g3 '6'd8 29 h4 h5 30 '6'f3 ~h6 31 'iί'xf4+ ~h7 32 'iί'e4 .:tc8 33 '6'f3 ~h6 34 .:tle5 1-0. 2b) 14•••.:te8 15 'ifd2 (15 f3.th5 16 tt1f4 h6 17 tt1xh5 hxg5 18 tt1xf6+ 'iί'xf6 19 'ifd2 ~e5 20 1:Ie4 1/2_1/2 Short-Chandler, Hastings 1987/8) 15 ....txe2 (better than 15 ....:tac8 16 tt1g3 with a clear edge for White) 16 .:txe2 ':xe2 17 'ft'xe2 .:te8 18 'iff3 tt1e5 19 'iff4 'iί'b6 20 .txf6 ~xd3 21 'iί'g3 'iί'xf6 22 '6'xd3 'ifb6 and Black drew easily ίη Kudrin-Rogers, London 1988. 13 ... .th5 14 j,g5 'ifd6 The position is the same as ίη Short-Chandler above, except that the moves .:tel and ....:tfe8 have been omitted. This difference favours White, as we shall see. 15 'i'el This is a good square for the queen. The immediate threat is 'ifh4,

but White a1so intends Ιο step υρ the pressure οη d4 by 'iff2 and ':adl. 15 'ifd2 is an interesting a1temative ifWhite plans Ιο meet 15 ... .:tad8 by 16 'iί'f4!; Fishbein-Langua, Chicago 1991 continued 16....tg6 17 .txf6 gxf6 18 'ifxd6 .:txd6 19 f4 .txd3 20 cxd3 .:te6 21 tt1g3 tt1e7! 22 l1acl tt1d5 23 .:tc5 .:td8 and with his hole οη e3 White had ηο rea1 advantage ιο speak of. Ι (JG), persona11y, would have been loathe Ιο exchange queens off after wrecking my ορρο nent's kingside; 18 'ifh6 and 18 'ifh4 both look much more testing. 15 .tg6 (D)

16 ':dl 1Ue8 Black must ηοΙ exchange οη d3 as this gives White a free tempo Ιο increase the pressure οη d4 by 'iff2 and.:tfdl. 16,..tt1d5 was tried ίη TopalovSion, Leon 1993 but after 17 'iff2 tt1b6 18 .tc ι! .:tac8 19 b3 .:tfd8 20 .tb2 the d-pawn was a1ready ήΡe for plucking. Perhaps Black's best plan is ιο ρυι a rook οη the c-file Ιο try ιο exploit the fact that White's queen is ηο longer defending the c2-pawn.

Sicilian Four Knights 187 17 'ift:z :ad8 18 :d2 :d7 Black decides Ιο meet White's plan passively, even though being forced οη the defensive is usually a sign that an isolated pawn position has gone wrong. However Mokry's suggestion of 18•. ':e5 appears ηο better after 19.th4 threatening .tg3. 19 :fdl :ed8 Ιη the line 19•••.txd3 20 :xd3 'ife5 21 .txf6 'ii'xe2 22 'ii'xe2 :xe2 23 .txd4 :xc2 24 .tc3 Black succeeds ίη exchanging his weak isolated pawn, but only at the cost of giving White a dominant bishop and chances of peneιrating Ιο the seventh rank. 20 .tb5! Removing a vital defender ίη creases the pressure οη d4 intolerably. Black's reply leads Ιο material 10ss, but even 20•••'ifeS 21 .txc6 bxc6 22 .txf6 gxf6 23 f4 'ifb5 24 lΔg3, threatening f5, is unpleasant. 20 ••. h6 (D) 21 tZ)xd4! hxg5 Black's moves are all forced, since 21••• tZ)xd410ses Ιο 22 .txd7.

22 lΔxc6 'ifxd2 23 :xd2 :xd2 24 tZ)xd8 :xt:z 25 φxt:z .txc2 26 lΔxb7 White is a pawn υρ and his active king makes the task of converting his mateήaΙ plus ίηΙο a ροίηΙ relatively simple. 26 .tb3 27 tZ)a5 .td5 28 1Δc6 .txc6 29 .txc6 Φf8 30 Φe3 Φe7 31 Φd4 Φd6 32 .tb5 ι-ο Black did ηοΙ wish Ιο see the technical phase of the game.

10 Lowenthal and Kalashnikov Variations These lines start 1 e4 c5 2 ~B ~6 3 d4 cxd4 4 ~xd4 e5 and are slightly akin Ιο the Pelikan ίη its use of an early ...e5. The 'old' Lowenthal runs 5 ~b5 a6 6 lL!d6+ -txd6 7 "xd6 Wf6 and Black hopes that his lead ίη development will compensate for his dark-square weaknesses and lack of the two bishops. Cuπent theory suggests that this is a vain hope and White should be able Ιο maintain an advantage. This is the subject of Game25. Black has an interesting and dynamic alternative, which has been pioneered by Sveshnikov and other Soviet players. This runs 5 ~b5 d6, and is generally known as the Kalashnikov Variation. Here White has the choice between 6 c4, aiming for a firm grip οη d5, οι 6 ~lc3 as ίη the Pelikan. Ιη BTS2 both lines were examined but with theory developing rapidly it was time Ιο make a choice - and the solid 6 c4 has emerged οη Ιορ. This is examined ίη Game 26. Game25 Liberzon - Franzoni Biel1980

1 e4

c5 αι

3 d4 cxd4 4 lDxd4 eS 5 ~bS a6 6 ~d6+ -txd6 7 "xd6 "r6(D) 7.....e7 8 Wdl ~f6 9 ~3 threatening -tg5 is good for White because he will secure control of d5.

8 "dl White has a wide variety of queen moves and most of them are good! There seems little doubt that 8 "c7, which has always been highly regarded theoretically, gives White a good game but Ι have ηοΙ recommended ίι here because White must always be careful that his queen is ηοΙ trapped. The simpler Wdl seems preferable. One should note that 8 "xf6 is also quite good, for example 8 •••lΔxf69 tα3 and now:

Lδwenthαl aΜ Kαlashnikov

1) 9...d51OJ.g5 d4 (10 ...lbb411 J.xf6 gxf612lbxd5lbxc2+ 13 ~d2 lbxal 14lbc7+ Φe7 15 lbxa8 J.e6 16lbb6 J.xa2 17 Φc3 and 1O...lbxe4 I1lbxd5 ο-ο 12 J.e3 are also good for White) 11 J.xf6 dxc3 12 J.xg7 ':g8 13 J.h6lbb4 140-0-0 lbxa2+ 15 ~bl J.e6 16 ':d6 ':g6 17 J.e3 lbb4 18 J.c5 with advantage for White according to GligοήC. 2) 9...lbb4 (D) and now:

Variations 189

J:td8 14 ~bl J.b7 15 g4 f6 16lbd5 lbxd5 17 1Wxd5+ is also promίsing for White, Glίgoric-Benko, Dublin 1957. 9 lbc3 (D)

2a) 10 J.d3lbxd3+ (10 ... h6 11 b3 d6 12 J.a3lbxd3+ 13 cxd3 Φe7 14 f4 ~e6 15 f5+ ~e7 16 ':dl ':e8 17 d4 exd4 18 ':xd4 ~f8 19 J.xd6+ ~g8 20 ο-ο b5 21 e5 1-0 was a drastic finish, Byme-Evans, USA Ch 1981) 11 cxd3 h6 12 b3 with an edge for Whίte. 2b) 10 ~d2 d5 11 a3 d4 12 axb4 dxc3+ 13 ~e3lbg4+ 14 ~e2 f5 15 bxc3 lbf6 16 ':a5 lbxe4 17 f3! lbd6 18 ':xe5+ and Black has very little for the lost pawn, Velίmirovic-Ris tic, Yugoslavia 1979. 8 ... Wg6 8 •••lbge7 9lbc3 ο-ο (9 .. :l'g6 will transpose to the next note) 10 J.e3 b5 11 1Wd2 Wg6 12 f3 d6 130-0-0

9 ... d5!? For a time this move caused a revival of the Lowenthal, but now White has found a way Ιο defuse the complications and liquidate Ιο a favourable ending. The older lίne runs 9 ...lbge7 10 h4! h5 (10...d5 11 h5 'ίi'd6 12 h6! g6 13 exd5 and now both 13 ... lbd4 and 13 ... lbb4 allow thereply 14lbe4, whίle 10... h6 11 h5 'ίi'f6 12 J.e3 ο-ο 13 'ίi'd2 b5 140-0-0 b4 15lba4 a5 16lbb6 ':b8 17 Wd6 gave White an excellent ending ίη Boleslavsky-Sakharov, USSR 1957) 11 J.g5 d5 (the only move that makes sense; e.g. 11 ... b5 12 'ίi'd3 J.b7 13 0-0-0 ':d8 14 'ίi'd6 'ίi'xd6 15 ':xd6 f6 16 J.e3 lbc8 17 ':d2lb6e7 18 J.d3 d6 19 :hd 1 was very good for White ίη Hazaί-Csom, Warsaw 1987) 12 exd5 (the tempting 12 J.xe7 is met by 12 ... d4!) and now we have: 1) 12...lbd413 J.d3 J.f5 14 J.xf5 lbexf5 15 'ίi'd3 ':c8!? (this move is

190

Lδwenthαl αΜ Kαlαshnikov Vαrίαtions

an interesting new υΥ; 15 ... f6 16 .te3 is known to be good for White) 160-0-0 ο-ο 17 ΦbΙ b5 18ltJe2?! and now 18•••e4?! 19 'ii'xe4 :ce8 20 'ii'xe8 :xe8 21 ltJxd4 f6 22 .tf4 ltJxd4 23 :xd4 'ii'xg2 24 :hdl, despite his queen minus, gave White the advantage ίη Gazik-Gross, Stary Smokovec 1991. However, Ι can't see anything wrong with 18•••ltJxc2, as 19 :c 1 runs into 19 ...ltJb4 20 'ii'e4 ltJxd5!. Ι think White should have played the ca1m 18 :cl, e.g. 18•••b4 19ltJe2ltJxe2 20 'ii'xe2ltJd4 21 'ii'd3 :xc2 22 'ii'xg6 :xcl + 23 :xcl fxg6 24 d6 with an excellent ending; or 18•••f6 19 .te3 'ii'xg2 20 ltJe4, with the idea of c3, should be favourable for White. It would be interesting Ιο see further tests of this line. 2) 12•••ltJb4 13 .txe7 Φχe7 14 .td3! (much better than the oftenrecommended 14 d6+, since White reaches the same type of ending, but with his d-pawn securely defended) 14•••ltJxd3+ 15 'ii'xd3 'ii'xd316 cxd3 (D) andnow:

17.••.tf5 18 Φd2! :h619 :hel 'iPd6 20 J:r.acl, Sveshnikov-Panchenko, USSR 1977 or 17••..tb7 18 0-0-0 Φd6 (suggested by Sveshnikov) 19 d4, or fma11y 17•••:b8 (suggested by Baumbach) 18 0-0-0 b4 19 axb4 1:txb4 20 :hel :'xh4 (or else :e4) 21 d4!, and ίη all cases White has a good ending. 2b) 16...:h6 17 0-0-0 :g6 18 ':he 1 ! :xg2 (the lines 18 ... f6 19 d4 and 18 ...Φd6 19 d4 are a1so good for White) 19 :xe5+ (19 d4!? is interesting) 19 ... Φd6 (19 ... Φd8 20:del .td7 21 d6 threatens ':xh5) 20 d4 followed by ltJe4+ and again White has the advantage. 10 ltJxd5 'ii'xe4+ ltJd4 11 .te3 This move, which is the only reasonable reply ιο the threat of ltJc7 +, is the idea behind 9 ...d5!? 12 ltJc7+ Φe7 12...'iPd8? allows White to take the a8-rook, while after 12...Φf'8? White can either play 13 1:tcl or take the exchange by means of 13 'ii'd3 ltJxc2+ 14 Φd2 'ii'xd3+ 15 .txd3 ltJxe3 16 ltJxa8 ltJd5 17 1:tacl - a pleasant choice! 13 :cl!

2a) 16•••b5 17 a3 and Black canrecover his pawn, for example

ηοΙ

Until this move was discovered Black had been doing rather well against 13ltJxa8?! and 13 'ii'd3. 13 ... .tg4 13...:b814 c3ltJf5 15 'ii'd5! ltJf6 16 'ii'xe4 ltJxe4 17 .ta7 +- Ga1lagher-Verdier, Nimes 1992. 14 'ii'd3 'ii'xd3 15 .txd3 ':d8 16 h3 (D)

Lδwenthal αnd Kαlαshnikov Vαriαtions

16 ••• -tc8 The alternative is 16••. -th5 17 f4 f6 18 Φf2 Φd6 19 c3 and now: 1) 19..•Φxc7 is bad due ΙΟ 20 cxd4+ Φb8 21 fxe5 fxe5 221Σc5. 2) 19••. .!ί)b3 20 axb3 Φχc7 21 -te4 .!ί)e7 22 1Σhe1 .!ί)d5 23 fxe5 .!ί)χe3 24 Φχe3 fxe5 25 1Σfl 1Σde8 26 1:[f5 -tg6 27 1Σg5 -txe4 28 1Σχg7+ Φc6 29 Φχe4 with a winning posiιίοη for White, Marjanovic-Simic, Yugoslavia 1983. 3) 19•.. .!ί)c6 20 -tb6 exf4 21 c4 .!ί)ge7 22 -te4 .!ί)c8 23 c5+ Φd7 24 .!ί)d5 .!ί)Χb6 25 cxb6 Φd6 26 .!ί)Χf4 -tf7 27 1ΣhdΙ+ Φe5 28 Φe3 with a clear plus for White, Winsnes-Hillarp, Rilton Cup 1988. 17 f4 exf4 17•••Φd6 is met by 18 .!ί)a8!. 18 -txf4 .!ί)e6 Otherwise White castles and Black is unable Ιο develop his king's rook while e8 is covered. 19.!ί)xe6 -txe6 20 ο-ο The outcome of the opening is very favourable for White. He has two active bishops supporting a queenside pawn majority and while so

191

many pieces remain οη the board Black's king is ηοΙ well placed οη e7. 20 .!ί)f6 21 a3 lί:Κι5 22 -td2 1:[d7 23 1ΣοοΙ 1Σc8 24 1Σf3 b5 25 b3 h6 26 c4 bxc4 27 -txc4?! It was more important ιο dήve away Black's centralised knight than Ιο keep the queenside pawns intact. After 27 bxc4 .!ί)f6 28 -tb4+ White has a passed pawn and an attack against Black's king. 27 ••. :a8?! (D) Α passive and nervous moνe . 27•.. 1Σc6 is better.

28 a4 Preparing both -tc l-a3+ and b4b5. 28 ... Φd6 Αη attempt Ιο bring the king over to help ίη the fight against White's queenside majority, but two rooks and two bishops are a formidable attacking force and the king soon runs ίηΙο trouble.

192

Lδwenthαl αΜ Kαlαshnikov Vαriαtions

q;c7 29 b4 30 b5 axb5 1Id6 31 .itxb5 32 J:ιdl With the sneaky threat 33 ~f4 lbxf4 34 :tc3+ winning the exchange. 32 q;b7 33 .tb4 1Idd8 34 J:ιrd3 Q;c7 Trying Ιο υηρίη the knight. 35 :Ιcι+ Q;b6 35••• ~b8 36 1Ic5 1Ia7 37 ~c6 wins mateήal. 36 .tc5+ Q;a5 37 :cdl 1Iac8 38 ~e7 1:[e8 39 ~xe8 1Ixe8 40 :XιlS+ 1-0 Game26 Arakhamia Tbilisi 1991

Chiburdanίdze -

1 e4 2

3 d4 4 lbxd4

c5 lbc6

cxd4 eS

5 lbb5 d6 6 c4 ~e7 6•••~OO generally leads Ιο a transposition. 7 lblc3 a6 8 lba3(D)

~oo

Altematives are: 1) 8•••h6 (Black plans ... ~g5 but wants Ιο recapture οη g5 with a pawn; this turns ουΙ ιο be a rather dubious idea) 9 ~e2 (or 9lbc2 ~g5 1Ο ~xg5 hxg5 11 'iνd211h6 120-0-0

wsa good for White ίη Wedberg0stenstad, Espoo Ζ 1989) 9 ... ~g5 10 lbc2 lbf6 (10 ... ~xcl 11 :xcl lbf6 12 ο-ο ο-ο 13 'iί'd2;t) 11 ~xg5 hxg5 12 'iVd2 1Ih613 'iί'xg5 :g614 'ife3 :xg2 15 0-0-0 with a clear plus for White, Hjartarson-Spraggett. Manila ΙΖ 1990. 2) 8••• lbr6 (abandoning the idea of exchanging dark-squared bishops is ηοΙ really ίη the spirit of this line) 9 ~e2 (9 ~d3 also deserves attention) 9 ... 0-0 10 ~e3 .te6 11 ο-ο :c8 12 'iί'd2lba5 (12 ... h6 13 1Iacl was slightly better for White ίη Dolmatov-Guseinov, ΚIaipeda 1988) 13 lbd5 ~xd5 14 exd5 b615 1Iacllbb7 16 f4lbd7 17lbc2 a5 18lba3 with advantage Ιο White, Dolmatov-Minasian, USSR 1988. 9 .td3 (Ο) Αι the time of BTS2 this line was very much ίη its infancy. Ιι has now developed ίηΙο one of the most solid lines ίη the Sicilian and plays a role ίη the repertoire of a number of strong grandmasters. Even so, ηο clear consensus has been reached οη White's best line. Apart from the text. 9 ~e2 and 9lbc2 are seen quite

Lδwenthal αΜ Kalαshnikov Variαtions

frequently and 9 .te3 and 9 lbd5 appear from time Ιο time. Ι have opted for 9 .td3, as opposed ΙΟ .te2, firstly because ίι is safe from capture by a black knight οπ d4 and, secondly, because ίι feels more harmonious. The queen still has an open diagonal Ιο the kingside and after the inevitable lbd5, ... somethingxd5, exd5 will be a more tempting οριίοη. The οηlΥ drawback of the move is that d6 is under less pressure, but ίι is quite rare for White Ιο mount a seή ous attack against this pawn.

193

:f2 lbg6 17 :d2 'Wh4! was υnclear ίπ Conquest-Fossan, Gausdall991, but 13 lbd5!? looks like a possible improvement Ιο me, whίlst the aggressive 12 f4 is ηοΙ the οηlΥ way Ιο treat the position, e.g. 12 .tbl!? or even 12lbc2. 2) 11••• lbf6 12 lbc2 ο-ο 13 "'d2 (D) (13 b4 :c8 14 00 is widely quoted as ;1;) with a further bifurcation:

:c8

9 ... is more common for Black Ιο play an immediate 9••..tg5. After 10 0-0 (1 Ο lbc2 can be played ίί White wishes Ιο deny Black the opportuηίΙΥ of playing ... lbd4) 10•••.txcl (10 ... h6 has been played bυt makes little sense as Black will soon have Ιο capture οπ cl anyway; e.g. Wang Zίli-Arakhamίa, Sydney 1991 was good for White after ll00lbf6 12 lbc2.txcl 13 :xcl .txd5 14 cxd5 tΔe7 15lbe3) 11 :Xcl we have: 1) 11•••lbd4 12 Ι4!? exf413 .tbl lbc6 14 'ili>hllbge7 15 :xf4 ο-ο 16 Ιι

2a) 13......b6 14 'ili>hl :ac8 15 b3 lbd4? (15 ...:td8 was more solίd) 16 lbe3"'a5 17 ί4! exf4, FedorowiczSalov, WijkaanZee 1991, 18lbed5! and Black is ίη seήουs trouble after both 18_ixd5 19 exd5 and 18._"'d8 19 'iνxί4, mainly because οί the knight οη d4 which is hanging ίη the mίddle οί nowhere. 2b) 13•••:c8 14 .l:.fdl 'Wb6 15 .tf1 :fd8 16 b3 'Wa7 17 h3 'iνb8 (Black has reached his optimυm Ρο sition, often πο! a good sign as it leaves you with nothing to do; White now builds υρ slowly, first tying Black down to the defence οί d6 and then switching operations ιο the kingside) 18 'We3 'Wa7 19 'Wd2 'Wb8

194

Lδwenthαl αΜ Kαlαshnikov Vαriαtions

20 :e1 'ila7 21 :cd1 tΔe7 22 :e3 :d7 23 :d3 ~e8 24 Φh2 :dc7 25 g3 'ilc5 26 a4 (Black was actually threatening to play ... b5) 26 ... 'ilb6 27 :b1 :d7 28 ~g2 :dd8 29 f4 f6 30 f5 ~f7 31 h4 with a clear advantage Ιο White, Timoshenko-Kiselev, Bucharest 1993. White can anticipate a successful pawn-storm as Black lacks both counterplay and space. 10 ο-ο ~ι5

11 1Dc2 12 :Xcl

J.xcl ~?

This move is Ιοο ambitίous. Black would do well Ιο concentrate οη her development, after which the game would resemble the lίnes ίη the note ΙΟ Black's 9th move. 13 b3 b5 This move cannot really be condemned, even if swallowing one's pride and returning the knight Ιο c6 is the lesser evi1.

14 lDe3 15 ~xc4!

bxc4

Ιι is the knights that are going cause the damage.

15 16 bxc4 (D)

Ιο

So Black has achieved her strategic aim of wrecking the white pawn structure, but at a great cost. Her own pawns οη a6 and d6 are very weak, and more significantly, her kίng is still stuck ίη the middle and the knight οη g8 and rook οη h8 are ηοΙ exactly playing a full part ίη the struggle. The positίon can already be assessed as '±'. 16 ... aS Faced with the threat of 17 "'a4+, Black relίnquishes control of b5. As 16••.'ild7 would have been well met by 17 ~cd5 'ifc6 (17 ...:b818 c5) 18 c5! dxc5 19 ~c4, Ι think ίι was tίme for Black Ιο cut her losses and jettison a pawn. After 16•••~e7 17 "'a4+ ~c6 18 'ifxa6 ο-ο there is still some hope. 17 ~b5 Φe7 18 (D) 18 "'a4 looks lίke a tempting alternative, but the maίn purpose behind moving the queen is Ιο connect the rooks and threaten 19 :fd 1 :c6 20 c5!.

"'f3

~c4 Β

18 19 :tfdl

'iνb6

:c6

Lδwenthαl αΜ Kαlashnikov Vαrίαtions

20 :bl 'fi'cS 21 ω! After playing an important role ίη luring the black major pieces Ιο unfavourable squares, the knight now clears the way for the rook and once agaίn looks towards d5. 21 ... ltJf6 And this knight finaHy develops itself, at least nine moves ιοο late. 22 :bS 'fi'a7 The queen must retreat as otherwise :b7+ would prove decisive. 23 lbcd5+ .i.xd5 24 cxd5 :b6? (D) Α blunder, but 24••• AcS 25 lbf5+ 'iii>d7 26 Axc5 'ίi'xc5 27 lbxg7 is also pretty hopeless.

195

Φf8 25 lbfS+ The h8-rook never managed move. 26 'fi'e3 :xa7 27 'fi'xa7 28 :b8+ lbe8 29 lbxd6 1-0

:a6

ιο

Ρίn

Variation

There is ηο generally accepted name for this variation, which runs 1 e4 c5 2 lDf3 e6 3 d4 cxd4 4 ι!ί)Χd4 lDf6 5 lDc3 .i.b4. There is certainly a ρίη involved, so 'Ρίη Variation' is areasonable name. Up until about 15 years ago this was thought Ιο be a very poor line for Black, but round about 1979 ίι suddenly reappeared with Black's play being based οη a new idea involving an exchange sacήfice. After a few years, during which ίι was used ίη occasional grandmaster games, ίι entered a decline and is now very rarely seen. However it is worth studying because there are a 10t of tricky tactics ίη the Ρίη Variation, and White players who do ηοι know the conect antidote may well find themselves ίη trouble. Game27 Wagman - Barle

Biel1981 1 e4 2 00 3 d4 4 ι!ί)χd4

cS e6 cxd4

till6

5 ω .i.b4 6 eS (D) The only move Ιο cause Black any difficulties.

6 ...

Black's two alternatives are close to 10sing by force: 1) 6.....aS 7 exf6 .i.xc3+ 8 bxc3 'iWxc3+ 9 'iWd2 'iWxal 1Ο c3 (threatening 11lDb3 'iWbl 12 .i.d3) 1O... 'iWbl 11 .i.d3 'ίi'b6 12 fxg7 :g8 13 "h6 and White wins. 2) 6...lDe4 7 "g4 and now: 2a) 7..•lDxc3 8 "xg7 :f8 9 a3 and now the lines 9..iaS 10 .i.h6 'iWe7 11lDb3, 9.....aS 10 lDb3 "d5 11 .i.d3 and 9...'iνb6 10 axb4 'iWxd4 11 .i.h6 'iWe4+ 12 Φd2 'iWd4+ 13 .i.d3 lDe4+ 14 Φc 1 are winning for White so Black must try 9...lDbS+ 1Ο axb4lDxd4 11 .i.g5 'iWb6 12.i.h6 'iWxb4+ 13 c3lDf5 14 cxb4lDxg7 15 .i.xg7 :g8 16 .i.f6 but White's darksquare pressure gives him a very favoυrable ending. 2b) 7.....aS 8 'ίi'xe4 .i.xc3+ 9 bxc3 'ίi'xc3+ 10 ΦdΙ 'ίi'xal 11lDb5 d5 12 'ίi'b4 lDa6 (12 ... 'ίi'xe5 13 f4 lDc6 14 fxe5lDxb4 15lDc7+ Φd8 16

Pin Vαrίαtion 197 lLJxa8 b6 17 .i.a3 is winning for White) 13lLJd6+ Φd7 14 .i.xa6 bxa6 15 lLJxf7 :g8 16 Φd2 d4 (or else .i.a3) 17 .i.b2 'ifxa2 18:al 'ifd5 19 :a5 and White has a decisive attack (analysis by Euwe). 7 .i.d2 ΟήgίηaΙΙΥ theory gave 7 'ifg4 as best, but after 7 ... 0-0 (the new idea mentioned above) 8 .i.h6 g6 9 .i.xf8 'ifxf8 Black has reasonable compensation for the exchange, with play against c3 and e5. 7 ... lLJxc3 Or 7....i.xc3 8 bxc3 ο-ο 9 .i.d3 d6 10 'ifh5 (10 exd6 'ifxd6 11 ο-ο is at least slightly better for White, Geller-Tseitlin, Moscow 1982, and 10 f4 dxe5 11 fxe5lLJd7 12 'ifh5 g6 13 'ife2'ifc714c4 'ifb61500lLJe716 .i.c3 lLJc5 17 'ifd2 proved favourable for White ίη Epishin-Ulybin, Tbilisi 1989) 10... g6 11 'ife2 dxe5 12 'ifxe5 lLJd7 13 'ifd6 'iff6 140-0 b6 15lLJc6 lLJc5 16 c4 e5 17 'ifxe5 with a clear plus for White, Vogt-Ermenkov, Berlίη 1982. 8 bxc3(D)

Α major decision point for Black. 8 ....ta5 keeps the pressure against c3 but leaves the kingside dangerously bare. Black should adopt the alternative possibility 8... .i.e7 9 'ifg4 with four vaήations: 1) 9... μ 10 h4 h5 11 'ifg3 lLJc6 12lLJb5 Was good for White ίη Rabar-Fuster, Munich 1942. 2) The remarkable idea 9...gS 10 h4 h5 11 hxg5!? hxg4 12 :xh8+ .i.f8 was played ίη Grosar-De Waal, Sas van Gent 1986, and now 13 g6 fxg6 14 .i.h6 Φf7 15 :xf8+ 'ifxf8 16 .i.xf8 ΦΧf8 17 .i.e2 is good for White. 3) 9...Φf8 also has to be taken into account, a recent example being Z.Varga-S.Horvath, Hungaήan Cht 1991: 10 .i.d3 d6 11 f4lLJd7 12 ο-ο lLJc5 13 :aellLJxd3 14 cxd3 h5 15 'iff3 g6 16 :bl d5 17 .te3 b618 c4 .i.b7 19 f5! gxf5 20 lLJxf5 exf5 21 'ifxf5 'ife8 22 'ifg6! dxc4 23 e6 f6 24 :xf6+ .i.xf6 25 'ifxf6+ Φg8 26 .td4 1-0. 4) 9 ... 0-0 10 .i.h6 g6, but even here the main line favours White after 11 h4! (D):

8 ...

.taS?!

198 Pin

Vαriation

4a) 11...d6 12 h5 'ifa5 (12 ... dxe5 13 J..d3! exd4 14 hxg6 fxg6 15 J..xg6 wins) 13 Ο-Ο-ο!? dxe5 (13 ...'ifxc3 14 :h3 'ifal+ 15 Φd2 'ifxa2 16 J..d3 'i'a5+ 17 c3 1i'xe5 18 hxg6 fxg6 19 J..xg6 :xf2+ 20 Φcι and 13 .....xe5 14 hxg6 fxg6 15 J..d3 :f6 16 :del "d5 17 J..g5 :17 18 J..xe7 :xe7 19 J..xg6 are winning for White) 14 lΔb5 a6 15 hxg6 fxg6 16 J..xf8 J..xf8 17 :xh7! Φχh7 18 J..d3 wins for Whίte.

4b) 11•..'if&5 12 'ifg3 :d8 (after 12 ... d6 13 exd6 Black may continue 13 ... J..xd6 14 'ifxd6 :d8 15 'i'b4 'i'xb4 16 cxb4 :xd4 17 c3 with the better ending for White, or 13 ...J..f6 140-0-0 'i'xa2 15 h5lΔc6 16lΔxc6 'ifa3+ 17 Φd2 bxc6 18 J..xf8 ΦΧf8 19 hxg6 hxg6 20 J..c4 and again White has the advantage) 13 h5 d6 14 hxg6 fxg6 15 J..f4 dxe5 16 J..xe5 :d5 17 f4lΔd7 18 J..c4lΔxe5 (after 18 ...:xe5+ 19 fxe5 'i'xe5+ 20 'i'xe5 lΔxe5 21 J..b3 the e6-pawn is too weak) 19 J..xd5 'i'xd5 20 fxe5 'fi'e4+ 21 Φd2 J..d7 22 :ael "d5, as ίη Wedberg-Pokojowczyk, Copenhagen Open 1984, and now 23 :xh7! ΦΧh7 24 J:[hl + Φg7 25 'fi'h2 Φ17 26 'i'h7+ Φe8 27 'i'xg6+ Φd8 28 :h8+ should win for White.

9 "g4 White can also use the move order 9 J..d3 when after 9.•. d6 10 'fi'g4 Black has nothing better than to play 10 ... 0-0 transposing into the maίn game. Ιη Ga1lagher-Escott, London 1984 Black tried 9•••'ifc7 but suffered a horrίble fate: 10 Ο-ο!? 'ifxe5 11 :el 'i'f6 12 'i'h5 (threatening 13

J..g5) 12... J..d8 13 lΔf5! J..c7 14 J..g5 'i'xc3 15 :adl1Δc6 16 :e3 g6 17 'i'h6 gxf5 18 J..f6lΔe5 19 'i'g7!? :f8 20 :xe5! J..xe5 21 J..xe5 'ifc5 22 J..b5! (threatening J..d6; the b5square was chosen as it pins the dpawn) 22 ... 'i'xc2 23 :d4! 'i'cl+ 24 J..n f6 25 J..d6! 1-0. 9 •.• ο-ο 10 J..d3 d6 10•••lίΊc6 is not much of an improvement, as can be seen from the game PyhaIίi-Seppanen, Helsinki 1992: l1lΔxc6 bxc6 12 'i'e4 g6+ 13 h4 ί5 14 'fi'f4 :17 15 g4 fxg4 16 'i'xg4 'i'b6 17 h5! 'fi'xf2+ 18 ΦdΙ 'fi'f3+ 19 'i'xf3 :xf3 20 hxg6 hxg6 21 Φe2! :f5 22 J..xf5 exf5 23 :agl Φ17 24 :h7+ Φe6 25 :xg6+ Φχe5 26 :e7+ 1-0 (26 ... Φd5 27 c4+).

11m Byme and Mednis suggest 11 J..gS with the variations 11...J..xc3+ 12 Φη f5 13 exf6 :xf6 14 :dl and 11.....c7 12 J..f6 'i'xc3+ 13 Φe2 'i'd2+ 14 Φη g6 15 :dl. This has never been tried ίη practice but looks good to me. 11 ..• g6 Or else J..xh7+ is crushing, for example 11•••dxe5 12 J..xh7+ ΦΧh7 13 'i'h5+ Φg8 14lΔg5 :e8 15 'i'x17+ Φh8 16 1i'h5+ followed by 'i'h7+, 'i'h8+, 'i'xg7+ and lΔ17+ picking up the queen. 12 h4! The strongest line - White just plays for mate. 12 lΔgS?! h5! (but not 12...dxe5? 13 lΔxh7! ΦΧh7 14 J..g5 J..xc3+ 15 Φe2 'i'd416 'ifh3+ Φg8 17 J..f6) 13 'i'g3 dxe5 14lΔe4

Pin lZXl7 15 ο-ο Wg7 enabled Black to defend ίη Peters-Arnason, New York 1980, while 12 ο-ο dxe5 (Sigurjonsson mentions the attractive line 12 ...ltX613 .ιg5 "fic7 14 .ιf6 .ιχc3 15"fih4 ttJxe5 16 ttJg5 h5 17 "fixh5) 13 ttJxe5 fol1owed by f4 only gave White an edge ίη Sigurjonsson-Arnason, Iceland 1980. 12 ... dxe5 12•.• hS is met by 13 'Wg3 or 13 'Wf4. 13 hS fS (D)

Vαriαtion

199

If the queen defends the rook from any other square White wins by 18 ttJxf7 'Wxf7 19 :h8+. 18 'Wh4 Φf8 19 ttJxt7? Having conducted the attack so wel1 υρ to here it is surpήsίng that White should miss 19 'Wh8+ We7 20 ttJxf7 when he is material υρ with a mating attack. Fortunately White is still winning even after 19 ttJxf7? 19 .•• Φxt7 20 'Wh7+ Φοο The line 20...Wf6 21 .ιg5+ We6 22 'Wxg6+ φd5 23 0-0-0+ Wc5 24 .ιe3+ Φb5 25 :d5+ is even worse. 21 'Wxg6+ ~dS 22 1:ιh6 ttJc6 23 'Wg8+ Φό

24 :bl!

14 .ιxfS! exfS 14••• :XΙS 15 hxg6 h5 16 :xh5 'Wf6 (16 ...:xh5 17 'Wxh5 'Wc7 18 0-0-0 wins, while otherwise Black can hardly meet the threat of :h8+) 17 0-0-0 fol1owed by :dhl with a decisive attack. IS 'Wc4+ Or IS••. Wg7 16 hxg6 Φχg6 17 ttJxe5+ winning for White. 16 hxg6 hxg6 17 ttJgS 'Wc7

:17

White cοπectlΥ adheres to the rule applying to king-hunts that it is more important to cut off the king's escape route than to give check. 24 bS 2S.te3+ ttJd4 26 .txd4+ exd4 27 'Wf8+ White misses it the first time round ... 27 Wc4 28 'Wg8+ Φό 29 :xbS+! ... but spots it the second! WxbS 29 30 'WdS+ 'fic5 31 a4+ 1-0

12 Nimzowitsch Variation: 2...lbf6 Although this move surfaces from time to time, its appearances at the grandmaster level are very rare. White has a range of possible lines against 2... lDf6, which have all proved successful ίη practice. Ιη this edition, as ίη BTS2, we are recommending 3 eSlDd5 4lDc3 e6 5 ~xd5 exd5 6 d4, which is, incidentally, the most popular line as well. Play can become very sharp so ίι is certainly a good idea to familiarise yourself with the variations, as facing ίι over the board with ηο Ρήοr knowledge is a far from easy task. However, the well prepared player should be assured of gaining the advantage against 2... ~f6. Game28 Rhine - Sprenkle USA 1981 1 e4

2 ~f3 3 e5

~6 tαι5

3 ••• ~g4 4 h3 ~h6 may be met by 5 d4 or 5 c3, with advantage Ιο White. 4 ltX3 e6 Or 4...~xc3 (4 ...~b4 5 ~c4 and 4 ...~7 5 d4 cxd4 6 Wxd4 ~6 7 "e4 g6 8 .tc4 .tg7 9 ο-ο ο-ο 10 :el, Κindermann-Ostl, Bundesliga 1987, are good for White) 5 dxc3 (D) andnow:

1) 5...dS 6 exd6 Wxd6 (6...exd6 7 .tc4 .te7 8 .tf4 ο-ο 9 Wd2 followed by 0-0-0 puts severe pressure οη the backward pawn) 7 Wxd6 exd6 8 .tf4 .tg4 (8 ... d5 9 0-0-0 .te6 10 ~g5 followed by g3 and either .tg2 or ~xe6 and .th3, when Black's central pawns wilI be fortunate Ιο survive) 9 0-0-0 ~7 10 .tc4 (even better than ~xd6, since ... 0-0-0 is prevented) folIowed by ~xd6 winningapawn. 2) 5...~ 6 ~Ι4 with the following examples: 2a) 6...e6 7 We2 Wa5?! 8 ~d2! 'fIc7 9 ~c4 f6 10 ~6+ ~xd6 11 exd6 Wa5 12 h4 with advantage Ιο White, Van der Wiel-Bjelajac, Νονί Sad 1982. 2b) 6...'iVb6 7 :bl (7 b3!? would keep the ήght Ιο castle long) 7 ...h6 8 .td3 g5?! 9 ~e3 ~g7 10 "e2 d6 11 exd6 ~xc3+ 12 lDd2 ~xd2+ 13 Wxd2 e5 14 ~c4! ~d7 15 h4 and the

Nimzowitsch Variation: 2... lC/6 201

black position is a wreck, TotskyLosev, 8t Petersburg 1994. 2c) 6 •••h67 'i'e2 'fic7 80-0-0 b6 9 'ife3 e6 10 lί:κι2lbe7 l1lΔc4lbf5 (but ηοΙ 11. .. lbd5? 12 1:xd5 exd5 13 lbd6+ ~d8 14 e6 wins) 12 "h3 .i.b7 13 :gl, intending g4, and White has some advantage, Van der WielMurey, Moscow ΙΖ 1982. 3) 5••• g6 6 .i.c4 J..g7 7 J..f4 ο-ο 8 "d2 followed by 0-0-0 and h4 gives White a strong attack. 4) 5••• b6? 6 e6! dxe6 (6 ... fxe6 and 6 ... f6 are both met by 7lbe5!) 7 "xd8+ ~xd8 8lbe5 ~e8 9.i.b5+ .i.d7 1Ο lbxd7 lbxd7 11 .i.f4 and White is close to winning a1ready. 5 tαι:d5 exd5 6 d4 lbc6 If Black doesn't lίke to sacrifice a pawn he can try 6•..d6, but after 7 .i.b5+ (D) he cannot equalise, as the following analysis shows:

1) 7•••.i.d7 8 .i.xd7+ "xd7 (not 8 ...lbxd7 9 dxc5) 9 0-0 lΔc6 10 exd6 J..xd6 (10 .....xd6 11 dxc5 "xc5 12 .i.e3 is also good for White) 11 :el + lbe7 12 dxc5 .i.xc5 13 .i.g5 0-0 14 'ΙΜ3 {6 (or 14 ... h6 15 .i.xe7 J..xe7

16 :ad 1 .1:ad8 17 c4 .i.f6 18 cxd5 .i.xb2 19 d6 and White's passed pawn is very dangerous) 15 J..e3 and Black has a poor isolated pawn ρο sition ίη which he has ηο active pieces Ιο compensate (οι the static weakness. 2) 7•••lΔc6 8 ο-ο J..e7 (8 ....i.e6 is also met by 9 c4 when Black has nothing better than Ιο transpose by 9 ... .i.e7) 9 c4 J..e6 (9 ... dxc4 10 exd6 "xd6 11 d5 a6 12 .i.xc4 and 9 ... a6 10 .i.xc6+ bxc6 11 cxd5 cxd5 12 exd6 "xd6 13 dxc5 "xc5 14 J..e3 are very good for White) 10 .i.e3 "ii'b6 (White threatened exd6) 11 a4 a6 12 a5 "c7 13 exd6 "xd614 dxc5 "d8 15 .i.xc6+ bxc6 16lbe5 "c7 17 "a4 ± Unzicker-Pomar, Bad Aiblίng 1968. 7 dxc5 J..xc5 8 "ii'xd5 "ii'b6 Here there is an important alternative: 8 ..•d6 9 exd6 "ii'b6 (D) (giving υρ another pawn Ιο allow Black's pieces Ιο come ίηΙο play more rapidly) and now:

1) 10 .i.e3!? (for brave players only) 10.....xb2 (1O ... .i.xe3 11 fxe3

202 Nimzowitsch Vαriαtion: 2... EΔj6 "xe3+ 12 ~e2 .te6 13 "g5 is promising for White) 11 .tb5!? and now 11......xal+ 12 ~e2 'i'c3 13 .txc6+ bxc6 (ηοΙ 13...~f8? 14 'i'xc5 nor 13 ... ~d8? 14 'i'xf7) 14 'i'xc6+ Φf8 15 'i'xa8 'iνxc2+ 16 EΔd2 .txd6 17 %ΣbΙ leads Ιο an advantage for White, 11.. :iνxbS 12 'i'xc5 is depressing for Black, while 11••.0-0 12 ο-ο .txe3 13 fxe3 .te6 14 'i'c5 was good for White ίη BolI-Lanz, Cou. 1982, so the best line is 11•••.tb4+ 12 Φe2 'i'xc2+ 13 EΔd2 .te6 with a total mess. 2) 10 "'e4+ .te6 (10 ... Φd8 11 .tg5+ f6 12 Ο-Ο-Ο! %Σe8 13 'i'h4 is good for White) 11 "'h4! (D) (11 .tc4!? 'i'b4+ 12 EΔd2 ο-ο 13 ο-ο 1tae8 14 c3 "b6 15 EΔf3 h6 16 b4 .txd6 17 .te3 "'c7 18 .txe6 1txe6 19 "c4 and Black had insufficient compensation for the pawn, ShortVan der Wiel, Wijk aan Zee 1990) with the variations:

2a) 11•••f612.td3 0-0-0 (the line 12 ... EΔb4 13 d7+! .txd7 14 _h5+ followed by ο-ο is good for White) 13 ο-ο and now there is a further branch:

2al) 13••..txd6 14 .te3 'i'xb215 1tabl 'i'a3 16 .tc4 gives White good attacking chances for ηο sacήfice. 2a2) 13•• .%Σxd6 14 .te3 (Ι am ηοΙ sure that ίι was necessary Ιο retum the pawn since Black has ηο immediate threats; 14 a3 intending b4 appears promising) 14 ....txe3 15 fxe3 'i'xe3+ 16 ΦhΙ 'i'c5? (16 ......b6 is just slightly better for White) 17 1tael .td7 18 "g3 g6 19 EΔd2! f520 EΔb3 'i'b4 21 a3 winning material, Chandler-Arnold, Bundesliga 1987. 2a3) 13•••hS 14 .te3 .txe3 15 fxe3 'i'xe3+ 16 Φhι.tg4 (16 ...1txd6 17 1tae 1 'i'b6 18 1txe6 1txe6 19 .tf5) 17 'i'g3 (17 1tael 'i'c5 18 .te4 is good for White according Ιο Gutman) 17 ... 'i'c5181tadl ~b819.te4 EΔe5 20 EΔxe5 .txdl, Hansson-Fernandes, London 1984, and although this game has appeared a number of times ίη ρήηι, nobody seems Ιο have noticed that after 21 "xg7 Black can quite reasonably resign. 3) 11 .• ..txd6 12 .te2 (even 12 .td3 EΔb4 13 ο-ο EΔxd3 14 cxd3 ο-ο 15 'i'e4 h6 16 .te3 'i'xb2 17 1tfbl gave White a slight advantage ίη Hellers-Ivarsson, Sweden 1985) 12 ....te7 (12 ... EΔb4 13 ο-ο is good for White after 13 ... 0-0 14 c3 or 13 ... .te7 14 'i'e4 f5 15 'i'e5 EΔxc2 16 .tg5) 13 "e4 0-0-0 14 ο-ο .td5 (after 14 ...EΔd4 15 EΔxd4 1txd4 16 \i'e3 .tc5 17 _c3 White stands very well) 15 "g4+ Φb8 (15 ....te6 16 "a4), Chandler-Bartsch, Bundesliga 1985, and now 16 c4! is good for White. 9 .tc4 .txf2+

Nimzowitsch Vαrίation: 2... liJj6 203

10 ~e2 ο-ο 11 :η j.c5 Black has regained the sacήfιced pawn, but now f7 is exposed Ιο attack. 12 liJg5 (D)

12 liJd4+ The best lίne; Black transfers the knight Ιο e6 ίη order to shield f7. The alternatίve 12••• liJxe5? (12 ... d6? 13 :xf7! liJd4+ 14 ~dl j.g4+ 15 :f3+ ~h8 16 'i!i'g8+ and mate) 13 'ii'xe5 d5 14 'ii'xd5 is ίnfeήοr, e.g.: 1) 14••• j.g4+ 15 :f3 j.gl 16 ~f1! :ad8 17 'i!i'e4 :dl+ 18 ~e2 j.xf3+ 19 gxf3 :fd8 20 j.xr7+ ~f8 21 liJxh7+ ~xf7 22 liJg5+ ~f8 23 'i!i'f5+ ~e7 24 'i!i'r7+ ~d6 25 j.f4+ ~c6 26 'i!i'c4+ 'i!i'c5 27 'i!i'xc5+ j.xc5 28 :xdl :e8+ 29liJe41-0 Prokopchuk-Kuznetsov. USSR 1972. 2) 14••.:e8 15 ~f3 'i!i'f6+ 16 ~g3 j.d6+ 17 :f4! j.e6 18 liJxe6 :xe6 19 'i!i'xd6 'i!i'g6+ 20 :g4 :e3+ 21 j.xe3 'ifxd6+ 22 ~f2 :e8 23 :f4 :e7 24 j.b3 'i!i'e5 25 :el g5 26 :f3 ~g7 27 :dl f6 28 ~gl g4 29 j.d4 1-0 Spassky-Cmc, Maήanske Lazne 1962.

13 ~dl liJe6 14 liJe4 Here White has various possibilί tίes, but this move, attackίng c5 and restrainίng ...d6, looks best. 14 ••• d6 14.•• j.e7 is too passive and ίη Savkίn-Tseitlίn. Cοπ. 1972 White obtained a strong attack after 15 c3 (15 liJd6 is also good) 15 ... d6 16 exd6 :d8 17 ~c2 j.xd6 18 :xf7! ~xf7 19 liJg5+ ~e8 (19 ... ~g8 20 'i!i'e4 h6 21 j.e3 'ii'a5 22 j.xe6+ j.xe6 23 'ii'h7+ and :f1+ wins) 20 liJxe6 'i!i'f2+ 21 ~b3 'ifb6+ 22 j.b5+ j.d7 23 liJc7+! and Black resigned without waitίng to see one of the lίnes 23 ... 'i!i'xc7 24 'i!i'g8+ j.f8 25 j.f4! or 23 ... j.xc7 24 'i!i'g8+ ~e7 25 j.g5+ ~d6 26 :dl +. 15 exd6 :d8 15••• j.xd6? is a miscombination which rebounds after 16liJxd6 :d8 17 j.f4 liJxf4 (Black saw the danger ίη Zaretdinov-Pugachevsky, USSR 1977 but still 10st after 17 ...h6 18 j.e5 liJg5 19 :xf7, etc.) 18 'ifxf7+ ~h8 19 'i!i'g8+ 1-0 Unzicker-Sarapu. Siegen 1970. 16 j.d3 j.xd6 17 'ii'hS fS 18 ltJxd6 'ii'xd6 After 18••.:xd619'i!i'xf5, White's threats Ιο r7 and h7 force Black Ιο play 19 ...:xd3+. but he does ηο! gaίn enough compensatίon. 19 'ii'xr5 (D) 19 .•• 'ii'xh2 Or 19•••00 (19 ... g6 20 'iff7+ ~h8 21 'ii'f6+ ~g8 22 j.f4lίquida tes to an endίng ίη whίch White has a

204 Nimzowitsch Vαriαtion: 2... ll:f6

clear extra pawn) 20 "Ι7+ ~h8 21 "f4andnow: 1) 21 .....CS 22 .ιe3 "h5+ (after 22 ....tg4+ 23 Φd2 "a5+ 24 b4 'ifh5 25 "g5 White exchanges queens) 23 ~d2 .ιe6 24 ..g5 (24 .ιd4! looks very good to me since 24 ...~g6 ώ lows 25 .ιχg7+ ~xg7 26 "f6+ and 24 .....d5 25 "e5! "xg2+ 26 ΑΩ "g4 27 h3! is a disaster) 24 .....e8. Α.RodήgueΖ-Dίaz. Cuban Ch 1983. and now 25 Af4 intending .td4 gives White a clear advantage. 2) 21 .....e7 22 "g5 "e8. ShortMinic. Banja Luka 1985. and now Minic gives the line 23 .td2 .ιe6 (23 .....a4 24 .tc3! Axd3+ 25 ~cl ~g6 26 Af8+ mates) 24 .tc3 ~g6 25 ~d2 Ad5 26 "g3 "d7 27 Aael. assessing the final position as being slightly better for White. Ι suspect that White's advantage is considerably greater than this; he is a pawn uρ with the two bishops. and if he consolidates with ~c 1 he must be winning. Therefore Black should pIay 27 ....tf5, but after 28 Ae3 foIIowed by ~c 1 White is a pawn uρ for nothing. 20 "f7+ ~h8

21 .ιgS ':g8 22 .ιe3! t'Δd8 After this move White can gain a cIear endgame advantage with ηο ήsk. The cήtical move is 22.....xg2! 23 'iνhS! (D) (Ι gave 23 ΑΩ ίη the fιrst edition, but 23 .....g4+ 24 ~d2 "b4+ 25 ~dl ~g5! is good for Black) and now:

1) 23...~B?! 24 Af4! (intending ΑΜ and mate οη h7) 24 .....h3 (seemingly forced) 25 "xh3 .ιΧh3, Odeev-Varlamov, Cοπ. 1987, and now 26 ~d2 .tg2 (26 ... ':d8 27 Ah 1 .tg2 28 Ah2 .ιc6 29 Afh4 h6 30 .ιd4 wins the exchange) 27 Agl .ιc6 28 ΑΜ (intending Ag3-h3) 28 ... Ad8 29 Ag3 .td7 30 .td4 and White has a Iarge advantage. 2) 23...g6 24 "h4! (24 .ιd4+ ~g7 25 .ιχg7+ ~xg7 26 _e5+ ~h6 is a draw) 24 ... ~g7 (24 ... Ad8 25 "f6+ ~g8 26 "f7+ ~h8 27 .ιd2! and 24 ....td7 25 Af7 Ag7 26 .ιe4! are very good for White) 25 Agl "f3+ (25 .....d5 26 Axg6 threatening 27 "xh7+ ~xh7 28 Ah6# is crushing) 26 ~d2 .tf5 reaching a position ίη which White has a very

Nimzowitsch Vαriαtion: 2... ~f6 205

strong attack for the pawn. Ι (ΙΝ) have ηοΙ been able Ιο find a forced win for White, but Black bas a difficult defence ίη prospect, e.g. 27 1Ig3 "d5 28 1:[hl h5 29 .t.d4 (29 "g5? 1Igd8!) 29 ....t.xd3 (the threat was 30 "xh5+) 30 cxd3 Φh7 (to defend against both ':xg6 and ':g5 followed by ':xh5+) 31 ':g5 "e6 (31 .....f7? 32 "e4 ~f5 33 ':hxh5+ gxh5 34 ':xf5 is winning) 32 ':e5 "c6 33 ':cl "g2+ 34 ':e2 'iWd5 (34 .. :iVg4 35 "xg4 bxg4 36 1Ie7 wίίΙS) 35 ':c5 "d6 36 ':e7 ':ad8 37 Φc3 and Black is ίη big trouble since 38 ':cc7 is threatened and 37 ... ':c8 fails Ιο 38 ':xg7+ while 37 ... ':d710ses after 38 ':xh5+. 23 "f2?! As Rhine cοπectlΥ points ουΙ, 23 "Ι4! 'iWxf4 (23 .. :iVxg2 24 ':gl 'iWd5 25 Φd2 and ':bl) 24 1Ixf4 would have given Wbite a very favourable ending at ηο ήsk.

23 ... 24 Φd2 The bishop οη d3 is the main enemy and Black should have tήed Ιο eliminate ίι by 24••• ~5, when 25 J.xh7 (25 ':hl ~xd3) 25 ... ':d8+! (25 ...lbc4+ 26 Φc3 "e5+ 27 J.d4 "a5+ 28 Φχc4 J.e6+ 29 Φd3 and White evades the checks) 26 J.d3 tDxd3 27 cxd3 'iWd6 28 'l'h4+ Φg8 29 'l'e4 leaves White with some advantage, but ίη view οί the oppositecoloured bishops ίι isn't clear how many winning chances he has. h6 25 :tιl

26 J.c5! (D)

Wbite's ambition is Ισ gain ί4 for his queen, when a sacήfice οη h6 will be inevitable. 26 .•. "d5 26..."c7 is spectacularly refuted by 27 'iWf6! and 26.....e5 27 ':ael 'l'g5+ 28 J.e3 followed by J.xh6 also 10ses quickly. 27 "Ι4 1Id8 27.....xg2+ 28 Φc3 doesn't belp Ιο prevent ':xb6+.

28:xb6+ 29 ':h8+! 30 'ίih4+ 31 'ίih7+ 32 "g6+ 33 'ίih7+

Φι8 Φxh8 Φg8 ΦΓ/ Φι8

36 "ΧΙ5+

Φι8

ςj;f7

34 :n+ J.f5 After 34•••Φe6 or 34•••Φe8, 35 "g6+ Φd7 36 ':f7+ mates. 35 hf5+ "ΧΙ5 37 Φcι 1-0 Quite apart from his mateήal disadvantage there is ηο defence Ιο the threat of J.c4+.

13 O'Kelly Variation: 2...a6 This is generally called the O'Κelly Variation after the late Belgian grandmaster who played ίι with some regularity. Ι have Ιο confess, though, that when Ι (JG) was about 12 years old Ι knew ίι as the 'Gallagher Variation'. Basically Ι used Ιο play the Pelikan, with reasonable success except when my opponent replied Ιο 5 ... e5 with 6lί)b5, after which Ι was crushed. Therefore, Ι racked my brains and came υρ with the idea 2... a6 3 d4 cxd4 4 lί)xd4 lί)f6 5 lί)c3 e5 when the knight does ηοΙ have b5 at his disposaI. Ι was extremely proud ofmy brand-new idea, which brought me many points, υηΙίΙ one day Ι won a copy of ECO Β, and much Ιο my amazement found several pages all devoted ιο the 'Gallagher Variation'. Ι gave ίι υρ soon afterwards. Ιη fact, 3 d4 is an eποr against the O'Kelly as ίι allows Black a favourable version ofthe Najdorfwhere he can develop his king's bishop actively at c5 or b4. 3 c3 is a sensible reply, which tends Ιο lead Ιο 2 c3 Sicilian positions in which Black has played the unusual move ... a6, which is perhaps ηοΙ the best way Ιο spend a tempo. However the strongest reply of all is 3 c4, which either leads Ιο Maroczy Bind positions or Ιο a sort of hedgehog.

Ιη view of the rare οccuπence of 2... a6 ίη practice, ίι perhaps does ηοΙ rate a chapter of its own, but this did give me the excuse Ιο include another of my own (JΝ) games ίη the book!

Game29 Νοnn

- Surtees

Bαsingstoke

Open 1977

1 e4 2 00 3 c4(D)

c5 a6

3 ...

lί)c6

Or: Ι) 3••• d6 (this may lead Ιο a type of hedgehog) 4 d4 cxd4 (4 ... ,ιg4 is an interesting move, since 5 dxc5 ,ιχf3 6 "xf3 dxc5 leaves Black with a grip οη d4 Ιο compensate for the two bishops; 6 gxf3! dxc5 7 'iνxd8+ Φχd8 8lί)c3 is possible. but

O'Kelly Variation: 2... a6 207 the simplest line is 5 d5 with a positional edge for White) 5 lΔxd4lΔΙ6 (Black has also tried 5 ... g6, with similar play Ιο Chapter 7. However, the early ... a6 can prove Ιο be premature) 6lΔε3 b6 (6 ... e6 7 .i.d3 leads Ιο positions from Chapter 6, so we concentrate here οη Black's attempt ιο develop early pressure against e4, which is unique Ιο 2 ... a6) 7 .i.d3 .i.b7 8 ο-ο (ίι is more accurate Ιο play 8 'iWe2lΔbd7 9 b3 - see the following note) 8 •••lΔbd7 9 'ii'e2 e6 (D) (9 ... lΔe5 is interesting since 10 .i.c2 :c8 11 .i.a4+ lΔfd7 is unclear, so White would have Ιο allow Black Ιο take οη d3) has occuned ίη three games between co-author Νυηη and Ο' Kelly specialist Michael Franklin:

la) 10 Ι4 'iWc7 11 'ίPhl .i.e7 12 .i.d2 is good, when Nunn-Franklin, London 1985 continued 12 ... h5 13 :ael h4 14 f5lΔf8 15 fxe6 fxe6 16 lΔd5! 'iWd8 17 e5 dxe5 18 'it'xe5 lΔxd5 19lΔxe6 'it'd6 20 'iWxd6 .i.xd6 21 cxd5 1-0. lb) 10 b3 "c7 11 .i.b2 'ii'c5?! (Black is playing Ιοο ambitiously with his king stuck ίη the centre -

11 ... .i.e7 followed by ... 0-0 is a better choice) 12 :ae1! (exploiting the tactical Ροίηι 12.. :ifxd4 13lΔa4, White prepares a breakthrough by lΔd5) and now: lbl) 12.....h5 13 'iWd2 g6 14 f4 .i.h6 15 'iWf2 g5 16 .i.e2 g4 17 'ifg3 :g8 18 .i.dl! also favoured White ίη Nunn-Franklin, London 1977 since e5 is imminent, while Franklin's later suggestion of 13 ... g5 allows 14 lΔd5! with added effect as the f5square is now available. lb2) 12...bS 13 cxb5! 'iWxd4 14 bxa6 .i.c6 15 lΔb5 'ifb6 16 .i.d4 "b8 17 :cllΔc5 18 a4! (this ηυΙΙί fies the threat of ... lΔxd3 and prepares Ιο break open the c-file by b4) 18 ... e5 19 .i.e3 .i.e7 20 b4lΔxd3 21 :xc6 lΔxb4 22 a7 'iWb7 23 :b6 (heading for b8) 23 ...'iWxe4 241Δc7+ 'ίPd7 25 'iWb5+! 'ίPxc7 26 :cl+ 1Δc2 27 :xc2+ 1-0 Nunn-Franklin, Nottingham 1979, since 27 ... 'iWxc2 28 :b7+ 'ίPd8 29 :b8+ 'ίPc7 30 .i.b6 is mate. 2) 3... 00 4lΔc3 and now: 2a) 4...lΔc6 5 d4 cxd4 6 lΔxd4 .i.b4 (6 ... lΔf6 7 lΔc2! is good for White) reaching an unusual posiιίοη which does ηοΙ seem Ιο be considered by theory. 7 lΔε2 .i.xc3+ 8 bxc3 is one possibility, but Ι (JN) like 7 lΔxε6. Then 7...bxc6 8 'iWd4 looks very awkward since 8...lΔf6 and 8 ...Ί'f6 are both met by e5, so 7...dxc6 8 'iWxd8+ 'ίPxd8 is best. White will continue 9 .i.f4, intending 0-0-0+ and lΔa4 with good play against the weak dark squares at c5 and b6. If Black exchanges οη c3

208 O'Kelly Vαriation: 2... α6 White has the dream square d6 for his bishop. 2b) 4•••ttJe7!? 5 d4 cxd4 6 ~xd4 ~ec6 7 .te2 ltJxd4 8 'ifxd4 ~c6 9 1i'e3 ~d6!? 100-00-011 b3 b5!? 12 cxb5 axb5 and now ίη Akopian-Rivas, Barcelona 1992 White accepted the pawn, but after 13 .txb5 .tb7 14 i.b2 1i'c7! 15 'it'h3 .te5! Black had adequate compensation. Akopian suggests 13 .tb2 b4 14 ~b5 as an improvement which should give White an edge. cxd4 4 d4 5 lΔxd4 ~Ι6 Οι 5.••e5 (5 ... e6 6 ~c2 is still good for White, and 5 ... g6 allows 6 lbxc6 and 7 'ifd4) 6 ~f5 d5 (6 ...lbf6 7lbc3 transposes Ιο Nunn-Surtees) 7 cxd5 i.xf5 8 exf5 ~d4 9lbc3 ~7 (once again 9 .. .lbf6 transposes and 9 ...i.b4 10 i.d3 'ifxd5 11 ο-ο 'ifd7 12 :el 0-0-0 13 :xe5 was very good for White ίη Sherzer-Polovodin, Philadelphia 1991) 1Ο i.d3 (10 f6 is also promising) 1O... lbexf5 11 ο-ο i.d6 12 f4 with a dangerous inίtiative for White. 6 lbc3 e5 7 lbf5 d5 After 7•••d6 8 i.g5 (8 lbe3 controlling d5 is also good) 8... i.xf5 9 exf5lbd4 10 .td3 White's control of e4 and d5 gives him a clear plus. i.xf5 8 cxdS 9 exf5 lbd4 10 i.d3 lΔxd5 11 ο-ο (D) 11 i.b4 Black has tried a variety of moves ίη this position, but without coming

close Ιο equality, e.g. 11 •••~xc3 12 bxc3 ~c6 (once the knight has Ιο move from d4 the only asset of Black's position vanishes) 13 :bl 1:[b814 1i'f3 'iic7 15 i.e4, RavinskyΚliashchiιsky, USSR 1966,11 ••• lbf6 12 :el lbc6 13 1i'b3 i.b4 14 :dl 1i'e7 15 i.g5, Rogatsovsky-Konovalov, Cou. 1972 οι 11•••i.e7 12 i.e4 ~xc3 13 bxc3 lbc6 14 :bl 'ifc8 15 1i'g4, Matanovic-Perez, Belgrade 1961 with a clear advantage for White ίη every case. 12 i.e4! lΔxc3 13 bxc3 i.xc3 14 :bl ο-ο Ιη the game Altshuler-Fink. Cou. 1960 Black tried Ιο hold οη to the pawn but after 14•• ':b8 15 1i'g4 g6 16 i.g5 gxf5 17 i.xf5 f6 18 1i'h5+ White had a winning position. 15 :ίxb7 This simple move was suggested by Glίgoric and Sokolov as an improvement over the unclear continuation 15 'iVg4 'ifd6 16 ]1dl :ac8 (but ηοΙ 16 ... :ad8? 17 :d3 i.b4 18 f6!) 17 :d3 :c4 ίη which White lacks a knock-out blow. 'iVd6(D) 15 •.•

O'Kelly

Vαriαtion:

2... α6 209

White's main threat was 16 i-a3, the rook away from the defence of f7, followed by "h5 and if Black manages Ιο defend f7 White still has the crushing blow f6 ίη reserνe. Black's .....d6 is of course designed Ιο prevent i-a3, but alas the move 10ses by force. He had Ιο try IS•••ltb8 although 16 Itxb8 "xb8 17 f6 gives White a strong attack with ηο mateήal investment. dήνίng

Α piece of rather unnecessary flashiness since 19 'iVhS "f6 (or 19 ... ltc7 20 f6) 20 i-d5 wins quite easί1y.

16 Itb3! White utilises the undefended queen Ιο threaten i-xa8, Itxc3 and i-a3. Black's reply is forced. 16 ••• :Sc8 17 i-a3 'iVd8 If the queen moves anywhere else, 18 i-xf8 ~xb3 19 i-xg7 wins a pawn and demolishes Black's kingside. 18 Itb7 lte8(D) 19 1hf7

19 ••• "gS 19...ιhr7 20 'ii'h5+ ~g8 (20...~6 21 "xh7 q;f7 22 'iVh5+ forces the king Ιο g8 ίη any case) 21 f6 g6 22 i-xg6 "d7 23 f7+ wins, but Black can hardly hope to surνive 10ng after 10sing the νίtaι f7-pawn. 20 i-dS ~h8 21 Ι6 g6 21••• μΙ6 22 i-e7 costs material. 22 'iVa4 e4 Losing at once, but Black's king would have succumbed soon ίη any case. 23 Itxh7+ 1·0 After23 ...~xh7 24 "d7+~h625 'ii'g7+ ~h5 26 'ii'h7+ ~g4 (26 .. :ΊWh6 27 g4+ ~g5 28 f4+ exf3 29 i-cl+) 27 'ii'h3+ ~f4 28 i-d6+ lte5 29 'ii'xc8 the posίtίon speaks for itself.

14 Unusual Lines The material ίο the first thirteen chapters will be sufficient Ιο prepare the reader for the vast majοήtΥ of the games he will have as White against the Sicilian, but there remain a substantial number of unusual lines which Black might adopt. Only a few have any pretensions Ιο respectability and we concentrate οη these few ίη this chapter. Wilder eccentήcities are usually best dealt with by an applicatίon of common sense and straightforward development. The following breakdown of lines considered ίη this chapter will aid the reader ίη locating the variation he is looking for. Α) The pseudo-Dragon 1 e4 c5 2 lbf3 lbc6 3 d4 cxd4 4 lbxd4 lbf6 5 lbc3 g6. Β) Unusual lines with 2... lbc6 apart from the pseudo-Dragon: 1 e4 c5 2 lbf3 lbc6 3 d4 cxd4 (3 ... d5) 4 lbxd4 lbf6 (4 ... a6, 4 ... d5, 4 .....c7, 4 ... "b6) 5 lbc3 'ίi'b6. C) Unusuallines with 2 ... d6: 1 e4 c5 2 lbf3 d6 3 d4 cxd4 (3 ... lbf6) 4 lbxd4 lbf6 5 lDc3 e5 (S ... lbbd7, 5 ... .i.d7) 6.i.bS+. D) Unusuallines with 2 ... e6: 1 e4 c5 2 lbf3 e6 3 d4 cxd4 (3 ... d5) 4 llJxd4 .i.c5 (4 ... .1b4+, 4 .. :ii'b6). Ε) Unusual Black second moves: 1 e4 c5 2lbf3 g6 (2 ... 'ίi'c7, 2 ... b6) 3 d4.i.g7.

Α:

1 2 3 4

1lJC3 d4 lbxd4

5 ω

cS llJc6 cxd4 lbf6 g6(D)

This is an attempt by Black Ιο reach Accelerated Dragon positions without allowing White the optίoη of playing the Maroczy Bind. It has been played a few times ίη grandmaster chess, but White can obtaίn a clear advantage with accurate play. 6 lbxc6 bxc6 Or 6... dxc6 7 'ii'xd8+ ΦΧd8 8 .i.c4 Φe8 (ηοι 8 ... b5? 9 .i.xf7 e6 10 .i.g5 .i.e7 11 0-0-0+ winning, while 8 ....i.g7 9 .i.f4 Φe8 10 0-0-0 llJd7 11 .i.c7! is good for Whίte since Black is ηοΙ allowed to castle) 9 eS and now: 1) 9.••lbg4 10 f4 h5 (if Black plays 1O....i.f5 11 h311Jh6 12 g4, the

Unusuαl

bishop has Ιο go back as 12 ....txc2 13 :h2 .ta4 14 lΔxa4 b5 15 .tb3 bxa4 16 .txa4 :c8 17 :c2 Φd7 18 .te3 is winning for White) 11 .td2 h4 (or 11 ....i.f5 12 h3 tbh613 0-0-0 .i.g7 14 :hel with advantage Ιο White) 12 lΔe4lΔh6 13 .i.c3 h3 14 e6 :g8 15 exf7+ lΔxf7 16 ο-ο-ο! with a clear plus for White, MausSchlick, Bundesliga 1987. 2) 9...lΔd7 10 f4 (10 e6 fxe6 11 .txe6 .tg7 12 .te3 b6 13 0-0-0 .i.xc3 14 bxc3 1Δc5 15 .txc5 bxc5 16 .i.xc8 :xc8 17 :hel :f8 18 f3 :f4 19 :e5 only gave White a slightly soperior rook endgame ίη Shirov-Ljobojevic, Buenos Aires 1994) 1O... lΔc5 11.te3 (11.i.e2!?) 11 ....i.e6 12 .i.xe6 lΔxe6 13 0-0-0 .i.g7 14 g3 b6 15 :d3 :d8 16 :hdl :xd3 17 :xd3 f6 18 exf6 exf6 19 tbe4 f5 20 lΔd2 .tf8 (20... Φe7 21 is good for White) 211Δc4 :g8 22lΔe5 with a clearly advantageoos ending for White, Hellers-Karlsson, Ostersond Ζ 1992. 7 e5 lΔι8 After 7••.lΔdS 8 lΔxd5 cxd5 9 'ii'xd5 :b8 10 e6! (with 'ii'e5 if the pawn is taken) 10...f6 (1O ....i.g7 11 exf7+ Φf8 12 .i.h6! .txh6 131t'e5, as ίη A.Sokolov-Lotskan, Latvian Ch 1994, is certainly ηοΙ an improvement) 11 .tf4:b4 12 .i.d2 (12 ο-ο-ο!? :xf4 13 .i.b5 was also very good for White ίη Varga-Barletta, Budapest 1991) 12 ...:b6 13 .tb5 :d6 14 "c4 White has a winning position. 8 .i.c4 (D) 8 .i.g7

:a3

Lines 211

Other moves are ηο better for Black: 1) 8..•dS 9 exd6 'ii'xd6 (9 ... exd6 Ι Ο 1t'f3 d5 11 lΔxd5 cxd5 12 .txd5 'ii'e7+ 13 .te3 :b8 14 ο-ο .i.g7 15 .i.f4 with a decisive attack) 10 ο-ο "xdl 11 :xdl .th6 12 .i.xh6lΔxh6 13 :d2 tbf5 14lΔe4 is very pleasant for White, Geller-Stein, USSR Ch 1966Π.

2) 8.....a5 9 ο-ο and now: 2a) 9.....xe5 10 :el 1t'a5 (or 1O.....b8 11 "d4 f6 12lΔe4 .tg7 13 .tf4 1t'b6 14 lΔd6+ Φf8 15 "d3 .tb7 16 .txg8 :xg8 17 'ii'c4 1-0 Tiviakov-Mugerman, Pinsk 1989) 11 b4 'ii'd8 Ieaves BIack ίη a dreadfoI mess. Ιη KarakIajic-Ivanovic, Yugoslavia 1974 White won BIack's qoeen by 12lΔe4 e6 13 .i.b2 f6 14 .i.xe6 dxe6 15 lΔxf6+ 'ii'xf6 16 .i.xf6 lΔxf6, which proved sofficient ίη the end, bot Ι wouId ηο! be surΡήsed ifWhite had an even stronger continuatίon.

2b) 9•• ..i.g7 10 'ii'f3 f5 (10 ... e6 11 .tf4 and 10... f6 11 :el are also good for White) 11 .tf4 leads ίηΙο the note Ιο Black's 10th move οη the following page.

212

Unusuαl

Lines

'iνB

Relatively best, e.g. 9 ...e6 10 i.f4 'iWa5 11 Ο-Ο! Ji.xe5 12 b4 'iWc7 13 lLJb5 'itb8 14 Ji.xe5 'itxe5 15 .l:adl d5 16 .1:tfel 'itb8 17 i.xd5 cxd5 18 'itxd5 ~f8 19 'iWc5+ (even stronger than 19 'iWd8+) 19 .. 5~g7 20 1td8 'itb7 (20 ... lLJf6 21 .l:xh8 ~xh8 22 'iWf8+ lLJg8 23 lLJd6 h6 24 lLJe8) 21 'ίWf8+ ~f6 22 lLJd6 'ίWe7 23 lLJe4+ ~f5 24 'iWxe71-0Geenen-Miranda, Νονί Sad OL 1990. 10 i.f4 e6 (D) 10•••1tb8 (1O ... 'iIi'a5 11 ο-ο i.xe5 12 b4 'fic7 13 lLJb5 'itb8 14 i.xe5 'ίWxe5 15 .l:fel 'itb8 16 'ίWc3 is now immediately decisive) 11 ο-ο e6 12 :adl followed by 1tfel andjustas ίη the main lίne White has an υη pleasant bind, Andersson-Bίlek, Teesside 1972. Black has ηο way Ιο solve the problem ofhis backward dpawn and the g7-bishop is inactίve. White players must be οη the lookουΙ for the exchange sacήfice ... 1tb4xf4, which can be good for Black if he can get the e5-pawn, but provided White keeps his bind οη Black's position he can be optimίstίc about the future.

11 ο-ο The couect choice; ίη some other games White played 0-0-0, but this gives Black counterplay down the bfile. 11 ... lLJh6 We give the rest of the game Short-Sosonko, Wijk aan Zee 1986, which is a model example ofhow Ιο play such positions. Black is never a110wed Ιο free himself and is fina11y overcome by the problems resulting from the backward d-pawn: 12 1tadl 'iWc7 13 1tfel lLJf7 14 'itg3 0-0 15 h4 Φh8 16lLJa4 a517 b3 1te8 18 'ίWe3 h6 19 g4! 1tg8 20 .i.g3 .i.f8 21 'ίWb6 1ta7 22 f3 'iWxb6+ 23 lLJxb6 .tc5+ 24 .tf2 i.xf2+ 25 ~xΏ fxg4 26 fxg4 ~g7 27 lba4 g5 28 h5 1tf8 29 ~g3 lbd8 30 lbc5 1tf4 31 a4 ~f8 32 .td3 ~e7 33 .tg6 1ta8 34 1te3 :b8 35 1ted3 1tbb4 36 1txd7+ .txd7 37 1txd7+ 1-0. Β:

1 e4 cS 2 lLJC3 lbc6 3 d4 cxd4 3...d5 (highly dubious) 4 exd5 'iWxd5 5lbc3 'ίWe6+ (or 5 ...'iVh5 6 d5) 6 .te3 cxd4 7 lbxd4 'iWd7 8 lbdb5 1tb8 9 "e2 and White is probably winning already, BoJesJavsky-Gurgenidze, USSR 1960. 4 lLJxd4 lLJf6 4 ...86 5 c4 transposes Ιο Chapter 13, 4.....c7 5 lbc3 e6 transposes Ιο Chapter 8, 4 ...'iνb6 5 lLJb3 lbf6 6 lLJc3 transposes back ίηΙο the main lίne and 4 ...d5 5 lbc3 dxe4 6 lbxc6 'ίWxdl+ 7 ~xdl bxc6 8lbxe4 i.f5 9

Unusual Lines 213 .td3 0-0-0 10 Φe2 e6 11 .tf4 Φb7 12 :adl is just good for Whίte. 5 1Dc3 'iib6 6 lΔb3 e6 7 .te3 'fIc7 8 .td3 (D)

8 ... .te7 Or: 1) 8•••.tb4 (at one time White players regularly spent a tempo οη a3 ιο prevent this, but now it's ηοΙ thought Ιο be good for Black Ιο give υρ his dark-squared bishop) 9 f4 (9 ο-ο ο-ο 10 lΔb5 'iί'b8 11 f4 was good for White ίη Gheοrghίu-Fοήntοs, Monte Carlo 1968, but Black can also play 9 ... .txc3) 9 ....txc3+ 10 bxc3 d6 11 ο-ο e5 12 ΦhΙ h6 13 'iί'el b6 14 fxe5 dxe5 15 'iί'g3 Φf8 16 lΔd4! lΔh5? (16.. /iJe7 was more solid, although after 17lΔf3lΔg6 18lΔh4! lΔxh4 19 'iί'xh4 White's chances are still Ιο be Ρrefeπed) 17 'iί'g6! exd4 18 cxd4lΔf6 19 ':xf6! gxf6 20 'iί'xf6 ':g8 (or 20 ...Φg8 21 ':f1 'iί'e7 22 'iνxc6 and Whίte is clearly better) 21 .tf4 'iνd7 22 .td6+ Φe8 23 .tb5 .tb7 24 d5 ':c8 (24 ...'iί'd810ses Ιο 25 dxc6) 25 ':dl a6 26 dxc6 .txc6 27

:d5! 1-0 Vouldis-Grivas, Greek Ch 1993. 2) 8•••a6 9 f4 d6 10 '6'f3 (White may also start his kingside pawn advance immediately, for example 1Ο g4 b5 11 g5lΔd7 12 'iνd2 {12 Wf3} 12....tb7 13 0-0-OlΔc5 14 Wn! with some advantage ιο White, Belίav sky-Gufeld, Sukhumi 1972) 1O... b5 (if Black plays ....te7 we transpose Ιο the main line below) 11 0-0-0 (11 g4 is again possible) 11 ... .tb7 12 ΦbΙ lΔa5 13 lΔxa5 Wxa5 14 g4 0-0-015 g5lΔd7 16 a3 Φb8 17 Wn J.e7 18 J.d4 e5 19 fxe5 dxe5 20 J.a7+ Φa8 21lΔd5 with a clear plus for White, King-Wirthensohn, Bern 1988.

Ι4

10 'fIf3 This is the most popular choice as White keeps open the ορΙίοη of castling οη either side. However, if he is certain he wants Ιο castle kingside there is an argument for doing so at once as he may be able Ιο dispense with Wf3. For example 10 ο-ο ο-ο?! (it's better Ιο play 1O... a6, transposing Ιο the αιώη line) 11 g4!? (D) with a couple of examples:

214 Unusual

Lίnes

1) 11•••lbd712g5:e813:f3a6 14 :h3 gave White a very strong attack ίη Wedberg-Benko, New York Open 1989; the finish was 14 ... lLJf8 15 'ifh5 lLJb4 16 :n .td8 17 a3 lLJxd3 18 cxd3 b5 19 f5 exf5 20 lbd5 'ifd7 21 .td4 fxe4 22 :h4 e3 23 lLJxe3 :e6 24 :hf4 :g6 25 :xf7 'ifxf7 26 :xf7 :xg5+ 27 'ifxg5 .txg5 28 :xg7+ 1-0. 2) 11•••d5 12 e5lLJd7 13 "f3 (13 :f3 !?) 13 ...:d8 14lLJb5 'ifb8 15 c3 b6 16lLJ5d4 ~b7 17lLJxc6 .txc6 18 'ifh3 g6 19 lLJd4 .tb7, Ga1lagherEfimov, Parίs 1992, and now 20 f5! is extremely strong because after 20...lLJxe5 21 fxe6 fxe6 (Blackhas ιο play the ugly 21 ... f6) 22 lLJxe6lLJxd3 23 'ifh6, 23.•..tf8 loses Ιο 24 :xf8+ and 23.....eS ιο 24 .td4 .tg5 25 lLJxg5. 10 .•• a6(D) Black also delays castlίng since 10•••0-0 11 g4 :e8 12 g5lLJd7 13 h4 lLJb4 14 h5 .tf8 15 0-0-0 a6 16 g6 gave White an immense attack ίη Jansa-Martinovic, Vrnjat!ka Banja 1982. Another try is 10•••lLJb4; LekoAnastasian, Moscow OL 1994 continued 11 ο-ο-ο!? .td7 12 ΦbΙ a5 13 lLJd4 ο-ο 14 g4 a4 15 a3! lLJxd3 16 cxd3 b5 (16 ....tc6 17 f5 is good for White) 17 lLJa2! b4 (Black invests some material as he wants ιο steer clear of lίnes such as 17...:fc8 18 g5 lLJe8 19 f5) 18lLJxb4 d5 19 g5 dxe4 20 dxe4 .txb4 21 axb4 a3 (21 ... lLJe8 22 e5 is a1so good for White) 22 gxf6 axb2 23 "g2! g6 24 'ifxb2 "b7 25 :hel! and Black had nowhere near

sufficient compensation for the sacήficed piece.

11 ο-ο White has a choice of good lίnes. 11 g4 and now: 1) 11•••b6 (this rather odd move is designed Ιο supρort c5 ίη anticipaιίοη ofthe manoeuvre ... lbd7-c5) 12 g5lLJd7 13 0-0-OlLJc5 14 ~bl (ιο answer ... lLJxd3 by cxd3) 14 ....td7 15 h4 "i'b7 16 .te2! (now the c5-knight isn't doing much) 16... lLJa7 17 f5 lLJb5 18 .td4 .tc6 19 fxe6lLJxc3+ 20 ~xc3lLJxe6 21 :hf1 0-022 .td3 b5 23 a3 lLJc5 24 lLJxc5! dxc5 25 "i'f5 :ae8 (or 25 ... c4 26 'ife5 f6 27 gxf6 .txf6 28 "i'e6+ :f7 29 :xf6!) 26 .txg7! .td7 (26 ... ~xg7 27 e5) 27 "i'e5 ~d8 28 'ifd6 ~xg7 29 'ifh6+ ~h8 30 e5 f5 31 exf6 .te6 32 :del c4 33 ~g6:f7 34 .txn 'ifxf7 35 g6 'ifxg6 37 f7 1-0 ΚavaΙek-Ηϋbηer, Buenos Aires OL 1978. 2) 11 •••h6 12 0-0-0 b5 13 :hgl lLJd7 14 'iff2! (this possibility explains Black's preference for ... b6 ίη line '1', since once he has played ... b5 the knight οη d7 is hard to redeploy) 14....tb7 15 ~bl .tf6? (Black's

Unusuαl

· .. h6 renders the kingside Ιοο dangerous for ... 0-0 - he should haνe played 15 ... lDb4) 16 e5! dxe5 17 .txb5 ο-ο 18 g5 hxg5 19 fxg5 .te7, Estrin-Kopyloν, USSR 1973, and now 20 'iνh4! axb5 21 1:g3 would haνe giνen White a decisiνe attack according Ιο Estrin. 3) 11 ... bS 12 gS lDd7 13 0-0-0 and now: 3a) 13•..lDcs (13 ....tb7 is well met by 14 'iνh3!) 14lDxc5 dxc5 15 e5.tb7 16 .te4 1:d8 17 h4 with an edge for White, Tiνiakoν-Anasta sian, USSR 1989. 3b) 13...lDb4 14 ct>bl .tb7 15 'iνh3 ο-ο 16 1:hfl 1:fe8 17 f5 exf5 18 exf5 lDxd3 19 1:xd3 lDf8 20 f6 .td8 21 lDd4 with a great position for White, Sax-Urday, Manila OL 1992. The game finished 21 ....tc8 22 'iνf3 1:b8 23 fxg7 ct>xg7 24lDd5 'iνb7 25 lDc7 1-0.

11 ... ο-ο (D) l1...bS 12 e5, but perhaps Black can consider 11. .. .td7. Νοι

12 1:ael White can also start the pawnstorm, with 12 g4:

Lines 215

1) 12.. ':e8?! 13 g5lDd7 14 'iνh5 g6 15 'iνh6.tf8 16 'ii1ι4 .tg7 171:f3 b5 18 1:h3 lDf8 19 f5 .txc3 20 f6 .txf6 21 gxf6 h5 22 'iνf4 e5 23 "h6lDe6 24 ct>h Ι! (now nothing can stop 1:xh5) 24 .. :i'd8 25 1:n! 1-0 Akopian-Prakash, Mamaia Wch jr 1991. Α model game; eνery one of White's moνes was geared towards his kingside attack. 2) 12...bS 13 g5 lDd7 14 'iνh5 lDb4!? (14 ... g6 15 'iνh6 f6 16lDd4! lDxd4 17 .txd4 1:f7 18 f5 .tf8 19 'i'h3 was good for White ίη UlybinAkopian, USSR 1988) 15 1:f3 (15 f5 deserνes consideration) 15 ... g6 16 'iνh6 f5! 17 gxf6 (17 exf5 exf5 18 1:h3 1:f7 19 .td4 lDf8 is unclear according Ιο Griνas) 17 ...1:xf6 18 1:g3 (18lDd4 g5!) 18 ...lDf8 19lDd4 lDxd3 20 cxd3 J.d7 with a tense struggle ahead, ΒeΙΙίa-Gήνas, Vinkoνci 1989. Unless 15 f5 is νery strong this 100ks like Black's best defensiνe set-up. The knight goes Ιο b4, coνering d5 ίη some lines and ίη others allowing him Ιο remoνe the potentially dangerous bishop οη d3, while leaνing the rook οη f8 so that a timely ... f5 and ...1:f7 can be played Ιο defend h7. 12 ... bS?! (D) Played οη countless occasions but ίι now appears suspect. Perhaps Black should play 12...lDd7 and if White plays 13 g4 then 13 ... b5 14 g5 .tb7 transposes ίηΙο Νuηη-Gήνas, Athens (1) 1991, which continued 15 'iνh5lDb4 16 1:f3lDxd3!? (with the bishop οη b7 the defence with ... g6 and ... f5 doesn't work, e.g.

216

Unusuαl

Lines

16 ... g6 17 'Wh6 f5 18 exf5 exf5 19 1:th3 1:tf7 20 .td4lbf8 21.ixf5!) 17 cxd3 (17 1:th3 is aπ interesting alternative) 17 ...1:tfe8 aπd now White should have played 18 .td4! when 18... g6?? allows 19 'Wxh7+ and 18••• 19 f5! exf5 201:txf5 g621 'Wh6 lbe6 22 1:xf7! is completely crushing. 18...:ac8 is best, even though White's game is still Ιο be

Ρrefeπed.

16 .tf4 which loses Ιο 16... 'Wb6+ 17 Φhι.tb7.

16 ... ..tb7 17 'ii'a7 'ii'c6 Perhaps everyone had thought this was the end of the road for White. But... 18 .te4! lbxe4 19 tbas! ... and ίη fact it's the end of the road for Black. Lϋcke-Gήvas, Dortmund 1992 finished 19 ...'Wc5 20 ..txc5 ..txc5+ 21 'ii'xc5lbxc5 22 b4 1-0.

13 eS! dxe5 Maybe Black should settle for 13...lbd7, but White does have a number of dangerous-looking continuations, such as 14 exd6 'Wxd6 (14 .....txd6 15 ..txb5!) 15 ..te4 (15 lbe4, 15 'ίWh3 and 15 1:tdl also come ίηιο consideration) 15 ... .tb7 16 f5. 14 fxeS lbxeS? Black had ΙΟ play 14...lbd7 after which White should play 15 ..tf4! before getting οη with the attack. 15 'ii'xa8 ltieg4 15....tb7 16 'ii'a7 doesn't help. 16 g3! This position was analysed ίη /nJormator 33 by Maιulovit, who attached a '?' to 13 e5, buthe only gave

1 e4 c5 2 lbf3 d6 3 d4 cxd4 Black quite often plays 3...lbf6 ίη order Ιο avoid the line 3 ... cxd4 4 'Wxd4. White should reply 4 lbc3 when Black is obliged Ιο play 4 ... cxd4 5lbxd4 transposing Ιο norma1lines. lbf6 4 lbxd4 eS(D) 5 Μ Or: 1) S...lbbd7 6 ..tc4lbb6 (Black's development is ηοι easy because .. .e6 and .....te7 will allow a .txe6 sacrifice, while 6 ... g6 7 f3 .tg7 8 ..te3 0-09 'Wd2 is a Dragon ίη whίch Black has developed his knight ιο the ίnfeήοr square d7) 7 .tb3 e5 8 lbde2 .te6 9 .tg5 .te7 10 ..txf6 .txf6 I1lbd5 lbxd5 12 ..txd5 'Wb6 13 ..tb3 ο-ο 14 lbc3 and White's control of d5 gives him a clear advantage, R.Byrne-Cuellar, Siegen OL 1970.

Unusual Lines 217 2) S•••J.d7 6 J.gS e6 and now: 2a) 7 lL!dbS J.c6 8 J.xf6 gxf6 9 "h5 looks tempting, but Kupreichik has done well with Black from thίs position. Ροι example 9 ...a6 10 lL!d4 J.d7 11 0-0-0 lL!c6 12 lL!xc6 (12 J.c4 "a5!) 12 ... bxc6 13 J.c4 "b6 14 :hf1 :b8 15 J.b3 c5 16 "e2 J.b5 with good attacking chances for Black, Spasov-Kupreichik, MoscowOL 1994. 2b) 7 'iVd2 may be best, with a probable transposition to Chapter 3 after 7 ... a6 8 0-0-0 lL!c6.

6 J.bS+ One of the points behind 5 ... a6 is Ιο prepare ... e5 by preventing this move, so it is the οηlΥ logical reply to 5 ...e5. 6 ... lL!bd7 After 6...J.d7 7 J.xd7+ 'ifxd7 8 lL!f3 (8 lL!f5 allows Black to complicate the issue by 8 ... lL!xe4) the exchange of light-squared bishops enhances the weakness of d5.

7 lL!fS

8 J.xd7+ "xd7 (D) This is the critical moment. White has a number of possible plans, but

it is absolutely unclear which one is best: 1) 9 J.gS lL!xe410 lL!xg7+ J.xg7 l1lL!xe4 and now: la) l1 ...dS?! 12 lL!f6+ J.xf6 13 J.xf6 ο-ο (13 ...:g8 14 0-0 is also very bad) 14 "d3 e4 15 "d4 :e8 16 g4! "d6 170-0-0 "f4+ 18 ~bl :e6 19 g5 :e8 20 J.h8 f6 21 gxf6 "f5 22 :hgl+ ~f7 23 :g7+ ~e6 24 "b6+ 1-0 Camacho-Cruz Lima, Cuba 1986. lb) 11...0-0 (this pawn sacrifice is the point of the variation, but ίι may not be cοπect) 12 "xd6 f6 (12 .....f5 13 lL!f6+ J.xf6 14 "xf6 'ife4+ 15 ~f1 J.h3 16 f3 "c4+ 17 ~el "b4+ 18 J.d2 favours White) and Ι doubt if Black has enough compensation for the pawn. After 13 J.e3 (13 "xd7 J.xd7 14 J.d2 f5 15 lL!c5 J.c6 did allow Black fair compensation ίη Am.RodriguezNijboer, Amsterdam 1989) 13 .....g4 14 lL!c3! J.f5 15 "d5+:f7 16 h3 ..g6 17 ο-ο-ο! :c8 (17 ... J.xc2 18 :d2 J.f5 19 g4 is good for White) 18 :d2 h5 19 g4! White had a clear advantage ίη Κlovan-Mochalov, USSR 1981.

218

Unusuαl

Lines

2) 9lbe3 'ifc6 and now: 2a) 10 'ifd3 (this gives White a small but safe advantage) 1O....te6 11 ο-ο 1%c8 12 .td2 (12 a4.te7 13 lΔcd5 .txd5 14 exd5 "iic7 15 a5 g6 16 b3 ο-ο 17lΔc4 1%fe8 18 .te3.tf8 with equality ίη Popovit-Rajkovit, Vr§ac 1987) 12 ....te7 (12 ... g6 is a possibility, but White retains a slight plus) l31Δcd5 ~d8 (now 13....txd5 14 exd5 "iic7 15 lΔf5 is good for White; .td2 is much more useful than a4) 14 c4 ο-ο 15 1%acl 1%e8 16 b3 lΔd7 17 lΔb4 with advantage for White, Ehlvest-Kupreichik, Moscow τν 1987. 2b) 10 lΔed5lΔxd5 11lΔxd5.te6 12 ο-ο 1%c8 13 c4!? "iixc4 14 lΔb6 'ii'xe4 15 lΔxc8 .txc8 16 1%el 'ii'g4 17 f3 with a small advantage for White, Κrnit-Jovanovit, Yugoslavia 1982. 2c) 10 0·0 lΔxe4 (10 ....te6 is probably better, when White may have nothing better than 11 'ii'd3 transposing ιο line 2a) 11 lΔxe4 'ii'xe4 12lΔd5 1%b8 13 b3 .tf5 14 c4 f6 15 .ta3 Φf7 16 'ii'd2 1%d8 17 1%fel "iig418lΔe3 'ifg5 19 'ii'd5+ with advantage Ιο White, L.Schneider-Bator, Sweden Ch 1986. 2d) 10 'iff3!? b5 11 ο-ο .tb7 12 lΔcd5lΔxd5 13lΔxd5 'ii'xc2 14.tg5 .txd5 15 exd5 f6 16 1%acl 'ii'g6 17 .td2 .te7 18 :c7 with good compensation for the pawn, Kudrin-Conquest, London 1986. Ο:

1 e4 2 lΔf3

c5 e6

3 d4 cxd4 3•••dS 4 exd5 exd5 5 .tb5+ 1Δc6 gives Black aπ uncomfortable isolated pawn position after 6 1Δc3 or 6 ο-ο.

4 lΔxd4 .tcS The idea behind this move is Ιο reach a position similar Ιο that after 1 e4 c5 2lΔf3 e6 3 d4 cxd4 4lΔxd4 a6 5 .td3 .tc5 (see Game 18), but withουΙ wasting a tempo οη ... a6. Naturally this exposes Black to the possibility of lΔb5 at some point, but attempts Ιο exploit this directly don't work. White has to be a bit more careful to gain the advantage against 4 ....tc5. Other moves are: 1) 4•••.tb4+, and now White caπ transpose Ιο Chapter 11 by 5 lΔc3 lΔf6, but ίι is also possible Ιο play 5 c3 .te7 6 c4, with a Maroczy Bind position. 2) 4 •••'ifb6 5 lΔb3 'ifc7 (D) and now:

2a) 6lΔc3 a6 7 .td3 b5 8.tg5!? .tb7 9 ο-ο lΔf6 10 f4 b4 11 e5 bxc3 12 exf6 cxb2 13 1%bl g6 14Axb2 with advaπtage ιο White, Hellers-

Unusuαl

Kveinys, Os10 1992, but Black's play can probably be improved. 2b) 6 i.d3lDf6 7 ~3 (7 ο-ο d5 8 lDc3 dxe4 9 lDxe4 is completely harmless; both 9 ... lDxe4 10 i.xe4 lDd7 11 'ii'd4 i.d6, Torok-Kveinys, BudapestSpring 1992 and 9 ... lDbd7 10 lDxf6+ lDxf6 11 i.b5+ i.d7 12 i.xd7+ lDxd7 13 lDd4 a6, LutherΚ veinys, Βοηη 1993, were satisfactory for Black) a6 8 f4 d6 9 ο-ο lDbd7 10 a4 b6 11 'ife2 i.b7 12 i.d2 i.e7, Umanaliev-Kveinys, Manila OL 1992, shows Black's plan ίη acιίοη; he has reached a normal posiιίοη, but with the knight οη b3 instead of d4. 2c) 6 c41Df6 7 ~3 and now: 2cl) Black could continue with normal development by 7...a6, ίη tending Ιο adopt a hedgehog set-up; however, the position of the knight οη b3 introduces some slight differences, for example White might try Ιο meet ... d6 by i.f4, 'ii'd2 and .l:d 1 aiming for a quick attack οη d6. 2c2) 7...i.b4 was Nunn-J.Polgar, Hastings 1992/3. The game continued 8 i.d2, but 8 i.d3 is probably better since 8 ... d5 9 exd5 exd5 10 'ife2+! is good for White after both 1O ... i.e6 11 lDd4 and 10 ... 'ii'e7 11 'ifxe7+ q;xe7 12 ο-ο. Ιη view ofthis, Black would do better Ιο meet 8 i.d3 with the quiet 8 ...~6. 5 lDb3 i.b6 (D)

6 ~3 Simple development guarantees at least a slight advantage. The ambitious 6 c4 is also promising, for example 6...lDe7 71Dc3 (White must

Lines 219

prevent ... d5) 7 ... 0-0 (or 7 ... lDbc6 8 i.f4 e5 9 i.g5 f6 10 i.d2 d6 11 'ifh5+ q;f8 12 i.d3 i.e6 13 ο-ο and White is better, Howell-S.Arkell, London Lloyds Bank 1986) 8 i.f4 (ίι is essential Ιο reach d6 with the bishop before Black prevents ίι with ... f5, e.g. 8 i.e2?! f5 and i.f4 is impossible) 8 ... f5 9 i.d61Dbc6 10 i.e2 and now if Black plays 10••.fxe4 11 c5 i.c7 121Dxe4 White's hold οη d6 cannot be broken, while after 10...f4 aiming Ιο play ... lDg6-e5 (after a rook move, of course) White might even consider 11 i.h5!? 6 ... tΔe7 7 i.d3 Or 7 i.g5!? f6 8 i.h4 ο-ο 9 'ifh5 lDbc6 100-0-0 lDe5 11 i.g31D7g6 12 q;bl f5 13 f41Dc6 14 i.c4 'iff6 15 e5 'ii'e7 161Db5 a6 171Dd6 q;h8 18 h4 1Da5 19 lDxa5 1-0 ~ibarevic G.Welling, Lugano 1989. Weakplay by Black, but this direct plan could be dangerous. 7 ... 0·0 8 ο-ο lDbc6 9 i.f4 f5 (9 ... d5 10 exd5 lDxd5 11 lDxd5 'ii'xd5 12 c4 'ifd8 13 i.d6!) 10 i.d6 f4 11 'ii'h5 f3 12 g3 i.c7 13 e5 g614 'ii'g5 i.xd6

220

Unusuαl

Lines

15 exd6 ~f5 16 'fIxd8 ~xd8 was played ίη Wedberg-Nunn, Helsinki 1983, and now the continuation 17 j.xf5 :xf5 181α14:f8 19 a4! a5 20 :a3 would have favoured White. Ε:

1 e4 c5 2 Μ g6 One of the most important lines ίη this chapter, since ίι has occuued many times ίη grandmaster chess and White can probably only secure an edge against ίι. Other second moves are very unusual and can be met by normal development, e.g. 2•••b6 3 d4 cxd4 4 ~xd4 ~b7 5 ~3 lbc6 (or 5 ...a6 6 j.d3 g67 f4 j.g7 8 ~f3 d6 9 ο-ο followed by 'fIel-h4 with attacking chances) 6 ~f4 :c8 7 ~xc6 dxc6 8 'fIf3 'fId4 9 :dl 'fIc5 10 e5 :d8 11 :xd8+ ΦΧd8 12 j.e2 Φe8 13 ο-ο f5 14 e6 ~f6 15 :dl ~d5 16 j.e5 1-0 Beliavsky-Quinteros, Vienna 1986, or 2•••"c7 3 c3 (ίι is probably ηοΙ a good idea Ιο play 3 d4 since 3 ... cxd4 4 ~xd4 ~f6 5 lbc3 a6 followed by ...e5 gives Black a type of Najdorf position ίη which his king's bishop can still be developed actively at c5 or b4) 3 ... ~f6 4 e5 ~5 5 d4 cxd4 6 cxd4 d6 7 ~a3 a6 8 j.d3 e6 9lbc4 dxe5 1Ο dxe5 b5 11 ~e3 j.b7 12 ο-ο followed by a4 when Black's queenside pawn structure will be weakened. 3 d4 The attempt Ιο reach a Maroczy Bind position by 3 c4 j.g7 4 d4 can be met by 4 ...1Wa5+, when ίι is ηοι at all easy for White Ιο maintain any

advantage. j.g7 3 ••• 3•.•cxd4 4 ~xd4 transposes Ιο lines examined earlier, for example 4 •••lDc6 5 c4 and 4 ••.j.g7 5 c4 end υρ ίη Chapter 7, 4••• ιtJf6 5 ~c3 d6 is Chapter 5, and 4 •••~f6 5 lbc3 lbc6 leads Ιο line Α ίη this chapter. 4 dxc5 'iν&5+ (D) 4 •••tba6 5 ~xa6 'fIa5+ 6 c3 1Wxa6 7 'fIe2 'iνc6 8 ~e3 'fIxe4 9 ~bd2 'fIc6 10 0-0 ~611 ~ 'fIc7 12 ~M 'fId8 13 j.f4 was good for White ίη Rajna-Nagy, Hungary 1960.

lΔc3 (D)

Natural, but 5 c3 may be stronger, for example 5 •••'iνxc5 6 j.e3 'iνc7 7 ~d4andnow:

1) 7•••e5? 8 j.e3 ~f6 9 ~a3! 0-0 10 ~b5 'fIc6 11 ~xe5 'fIxe4 12 ~xΠ! with a large plus for White, Marit-Τήngοv, Bar 1977. 2) 7••• ιtJf6 8 e5 ~g8 (8 ... ~g4 9 ω! is also better for White) 9 e6 f6 10exd7+ bxd711 ω ~612 j.c4 ~h6 13 ο-ο ~π 14 :el 0-015 j.c5 j.g4 16 ~b5 'fIc8 17 j.e2 with a clear plus for White, Nunn-E.Sakhatova, Ροιι Εήη 1994.

Unusuαl

3) 7•••f6 8lL!a3lL!h6 9 j,c4 a6 10 e6, Nunn-Kotsur, Moscow OL 1994, and now 11 j,b3 (threat lL!c4) 11 ... b5 12 "d2 (threatening j,xf6) 12...lL!n 13 c4 with some adνantage forWhite. ο-ο

5 ••• lL!f6 Or: 1) 5.....xc5 6lL!d5 e6 7 b4 'ii'f8 8 lL!c7+ Φd8 9lL!xa8 j,xal 10 j,g5+ ~f6 11 j,xf6+ lL!xf6 12 'ii'd4 'ίi'e7?! 13 ~b5 b6 14lL!xb6! axb6 15 'Wxb6+ Φe8 160-0 was νery good for White ίη the game Mohrlok-Breazu, Con. 1987. 2) 5••. j,xc3+ 6 bxc3 'ii'xc3+ 7 ~d2 'Wxc5 8 j,d3 (8 j,e2lL!f6 9 e5

Lines 221

lL!g4 100-0 lL!xe5 11 j,e3lL!xf3+ 12 j,xf3 'Wc7 13 'ίi'd4 was also promising ίη the game Petroν-Limonikoν, Con. 1974) 8 ... lL!f6 (8 ... d6 9 ο-ο j,g4 10 :bl 'Wc7 11 1:tb3 lL!d7 12 :c3 lL!c5 13 h3, Ambroz-Petran, Czechosloνakίa 1979 gaνe White more than enough for the pawn) 9 ο-ο ο-ο 10 e5lL!g4 11 I:ιbllL!xe5 12 :b5 lL!xf3+ 13 'Wxf3 'ii'c7 14 j,h6 :e8 15 :f5:f8 (New in Chess gaνe 15 ... f6 as unclear, but 16 :el! appears crushing after 16 ...'ii'c3 17 j,c4+! 'ii'xc4 18 :xf6) 16 j,xf8 gxf5 17 j,xe7 d5 18 :el ~e6 19 j,f6 lL!d7 20 j,d4 'ii'd6 21 j,xf5 j,xf5 22 'ίi'xf5 lL!f8 23 "g5+ lL!g6 24 h4 :f8 25 h5 1-0 Friνaldszky Monostori, Con. 1986. 6 j,d3 "xc5 7 j,e3 7 •. :6'115 is possible, but Ι cannot find any practical examples of ίι

"&5

8 "d2

lL!c6

9 ο-ο ο-ο 10 h3 d6 11 a3 j,e6 12 lL!g5 d5 (12 ... j,d7 13 f4) 13 exd5 j,xd514 b4 'Wd815 :adl and White has a slight adνantage, Sνeshni koν-Romanishin, USSR Ch 1977.

Index of Variations 1 e4 2 00

Now: Α: 2 ... ~c6

B:2 ...d6 C:2 ...e6 Ο: 2...0thers Α

2 ...

lίJc6

3 d4 cxd4 3... d5212 4 lbxd4 lbr6 4.. :iνb6 212 4 ... d5212 4 ...'iVc7 5 ~c3 e6 172 4 ... a65 c4 208 4 ... e5 5 ~b5 d6 (5 ... ~f6 6 ~lc3 96; 5 ... a6 6 ~d6+ .i.xd6 7 'iVxd6 188; 5 ....i.c5 6 ~lc3 ~f6 96) 6 c4 i.e7 7 ~lc3 a6 8 ~a3 i.e6 (8 ... h6 192; 8 ... ~f6 192) 9 .i.d3 192 4 ... g6 5 c4 ~f6 (5 ... .i.g7 6 .i.e3 153) 6~c3 d6 (6 ...~xd4145) 7 i.e2 .i.g7 (7 ... ~xd4 145) 8 i.e3 157 (8 ~2145)

4 ...e6 5 ~3 'iVc7 (5 ... d6 6 g4165; 5 ... a6 6 ~c6166) 6 f4 ~xd4 (6 ...d6 7 .i.e3 173; 6 ... a6 7 ~xc6 173) 7 1i'xd4 a6 8 i.e2 175 5 ~c3 d6 5 ... e6 6 ~db5 i.b4 (6 ... d6 7 .i.f4 e5 8 i.g5 - see 5... e5) 7 a3 .i.xc3+ 8 ~xc3181

5 ... g6210

5 ... e5 6 ~db5 d6 (6 ... 0thers 96) 7 i.g5 a6 (7 ....i.e6 96) 8 ~a3 b5 (8 ... .i.e796; 8 ... d5 97; 8 ... .i.e6 96) 9 .i.xf6 gxf6 (9 ...'iVxf6100) 10 lDd5 f5 (10 ....i.g7 100; 10....i.e6 100) 11 .i.d3 .i.e6 12 ο-ο .i.g7 (12 ....i.xd5 102) 13 'iVh5 f4 14 c4 bxc4 (l4 ... b4 103; 14 ...0-0104) 15 .i.xc4 104 5 ...'iVb6213 6 .i.g5 e6 6....i.d7 7 .i.e2 90 6... g689 6 ... 'iVa5 89 6...'iVb6 89 6 ... a689 7 'iί'd2

7 ...h660 7 ... ~xd4 61 7 ....i.e7 8 0-0-0 ο-ο (8 ... ~xd4 62; 8 ... a6) 9 ~b3 a6 (9 ... d5 63; 9 ... ~a5 63; 9 ... h6 64; 9... a5 10 a4 d5 11 .i.b5 64; 9...'iVb6 67) 10 .i.xf6 70 8 0-0-0 h6 8....i.e773 8 ... .i.d7 9 f4 h6 (9 ... b5 10 i.xf6 gxf6 11 ΦbΙ 73; 9....i.e7 75) 10 i.h4 g5 (10....i.e7 77; 10... ~xe4 77) 11 fxg5 ~g4 12 ~xc6 77 9 .i.e3 .i.d7 9 ...'iVc782 9 ....i.e783 9...~xd484 10 f485

Index ofVariations 223 Β

1 e4 2 3 d4 3...lt:If6216 4 tαιd4 5 ω

d6 cxd4

10...

1O... e5 118

11 h41l9 Β3

5... a6 For 5 ...l2k6 see Α and for 5 ...e5 and 5 ... lt:Ibd7 see ρ. 216. ΒΙ

5 •.• e6 6 g4 h6 6 ... e543 6 ...l2k67 g5 lt:Id7 8 h4 47 6 ... a6 7 g5 lt:Ifd7 8 h4 44 6... i.e7 7 g5 lt:Ifd7 8 h4 lt:Ic6 32 7 h4 lt:Ic6 7 ...i.e729 7 ... a6 8 i.g2 30 8 .:lgl h5 8...lt:Id737 8 ... d536 9 gxh5 tαιh537

...

101i'el

lbf6

Β3:

5 6 .t.e3 6 ...a6111 7 f3 7 ... a61l1 7 ...lt:Ic6 111 8 1i'd2 8...d51l2 9 0-0-0 9....t.e6112 9 ... .t.d7113

114 lΟΦb1117

Now: ΒΙ: 5 ...e6 Β2: 5 ... g6

Β2

9... lt:Ie5 114 9 ... lt:Ixd4 10 i.xd4 i.e6 11 ΦbΙ

g6 i.g7 ο-ο

lt:Ic6 d5

5 •.• a6 6 Ι4 eS 6 ...lt:Ic68 6 ... g68 6 ...e67"f3 8 6...lt:Ibd7 7 i.e210 6..."c7 7 lbf3 lt:Ibd7 (7 ...e6 8 i.d3 14) 8 i.d315 7 m lt:Ibd7 7 .....c7 8 a417 8 a4 i.e7 8 .....c7/7 8 ...d520 9 i.d3 ο-ο 10 ο-ο 21

c 2 ... e6 3 d4 cxd4 3...d5218 4 tαιd4 a6 4 ... i.c5218 4 ....t.b4+ 218 4 ...l2k6 - see Α 4 ... lt:If65 lt:Ic3 i.b4 (5 ... d6 - see ΒΙ; 5 ... lt:Ic6180) 6 e5 196 5 i.d3 lbf6 5 ... lt:Ic6123 5 ... g6125 5 ...lt:Ie7 125 5 .. :ifb6126

224 Index ofVαriαtions

5 .. :ilc7125 S...bS 126 S ... ,ιcS 6 ~b3 126 5 ... d6 6 0-0 ~f6 134 5 ...b6 126 6

ο-ο

6 .. :ilc7135 6 ... eS 134

7 c4137 D 2.. :flc7220 2...~f6200

2...a6206 2... g6220 2...b6220

Anyone who opens with the king's ροννπ hos to hoνe ο woy of deoling with the Sicilion Defence, Block's most populor reply. Ιπ the successful first ond second editions of this book, John Νuππ recommended repertoires which were token uρ with consideroble success by players both ot club ond ot internotionol leνel . Ιπ this new edition, John Νuππ teoms uρ with onother Sicilion expert to proνide new ideos to oνercome Block's lotest defensiνe ideos ίπ the eνer-chonging Sicilion.

• Proνides ο complete repertoire for White ofter 1 e4 c5 2 Nf3 • Strotegic explonotions occompony rozor-shorp onolysis • Α coreful selection of sound ond oggressiνe lίnes

John Νυππ is one of the world's most highly occloimed chess outhors, renowned for his work οπ 011 ospects of the gome . As ο ployer, his greotest indiνiduol success wos his third ploce ίπ the World Cup ίπ 1989. His recent books include the best-selling

John Nunn's Best Games. Joe Gallagher hos played the Sicilion Defence os Black ond os White for most of his chess cαreer. His preνious book for Botsford was Beating the Anti-SiciIians, ίπ which he showed how Black should meetWhite's minor options ogainst the Sicilion. Here he shows why White should ploy the main lίne!

Other openings books from Batsford incfude: New Ideas ίπ the Pirc Defence John Νuππ Τhe Complete Pirc John Νυππ

Beating the Caro-Kann Vassilios Kotronios

Τhe Queen's Gambit for the Attacking Player

Grohom Burgess ond Steffen Pedersen Τhe Complete Slaν

Peter Wells

Beating the French Gory Lane f14.99

Semi-