Beating the sicilian 2 by Jose Manuel Vargas

Beating the Sicilian 2 JOHNNUNN

B.T.Batsford Ltd, London

First published 1990 Â©John Nunn 1990

ISBN 0 7134 6445 3 A CIP catalogue record for this book is available from the British Library All rights reserved. No part of this book may be reproduced, by any means, without the prior permission of the publisher Typeset by Latimer Trend Co Ltd, Plymouth, Devon and printed in Great Britain by Dotesios (Printers) Ltd, Trowbridge, Wilts for the publishers, B. T. Batsford Ltd,

4 Fitzhardinge Street, London WIH OAH

A BATSFORD CHESS BOOK Adviser: R. D. Keene GM, OBE Technical Editor: lan Kingston

Contents Preface 2 3 4 5 6 7 8 9 10

Najdorf Variation Scheveningen Variation Classical Variation Pelikan Variation Dragon Variation Kan Variation Maroczy Bind Taimanov System Sicilian Four Knights Lowenthal Variation

l l Pin Variation <2:lf6 12 2 13 2 a6 14 Unusual Lines Index of Variations .

19 40 62 77 94 110 125 134 140 145 149 1 55 1 59 16 8

Preface This second edition of Beating the Sicilian does not need such a lengthy introduction, as the con cept of a 'repertoire' book has become a familiar one. However I will take time to detail some changes which have been made to the original Beating the Sicilian. The earlier work was based on the premise that a book like T he N aj dorf for the T ourn ament Player provides a surfeit of information for the average White player, who only wants to know one good line to play against the Najdorf. It therefore developed a complete repertoire for White against the Sicilian, covered in sufficient de tail for most players to meet any line of the Sicilian with confi dence. To my surprise I found that the book became popular not only amongst the intended audience of club players, but also at the Grandmaster level. There were faults in Beating the Sicilian. I recommended some unusual sidelines, but this recom mendation was self-defeating, because many of these lines were suddenly played in Grandmaster games. At this level the reasons why the lines were in fact sidelines became abundantly clear. Thus the book contributed to its own

rapid dating. This time I have adopted a different policy. In every variation I have recom mended one of the current main lines against that variation. Such main lines cannot be refuted, although the whims of fashion may lead to them becoming more or less popular as the years go by. In this way the lines recom mended in the book should re main valid for years to come (un til the third edition?) and readers may be confident that the effort put into studying them will earn a long-term reward. One side-effect of this policy is that the analysis has become considerably more detailed. You cannot learn the main line of the Yugoslav Attack without studying a fair number of variations, but as we only need to examine Black alternatives the work required is not too onerous. In any case the growth of opening theory at all levels of chess has forced players to devote more ef fort to the study of opening theory in order to maintain (or increase) their level of success. Unfortunately this has led to scrapping the policy of offering alternative White choices such as 6 a4 against the Najdorf. The book would have grown too big

vi P reface (and expensive) if these had been left in. In any case, these lines have not changed greatly since the first book, so much of the old material is still valid. I have maintained the arrange ment of material from the earlier book. The biggest changes in the repertoire are the abandonment of the Classical Scheveningen (now only the Keres attack is recommended), the switch to 9 J..xf6 in the Pelikan, adopting the main line of the Yugoslav Attack against the Dragon, and moving to 5 �b5 against the Taimanov. Needless to say, all chapters have been thoroughly updated, even if variation the is chosen unchanged. All material up to October 1 989 has been consulted, including /nf ormator 47, N ew in Chess yearbook 1 2 and ChessBase magazine 13. In some cases, usually at a fairly late stage of the analysis, I have given more than one move for White where there is genuine doubt as to which is the best alter native. I have devoted special at tention to a few relatively unusual lines which in some cases have a 'cult' following. Examples of this are the Pin Variation (Chapter 1 1 ) and 2 . . . �f6 (Chapter 1 2). These lines are probably not very good for Black, but they are tricky and

are sometimes used as a surprise weapon. All too often a White player, when confronted by one of these unusual lines, will decide to play safe and not enter the main theoretical paths. In doing this he often throws away any chance to gain the advantage. A short sur vey of the critical lines is enough to enable one to counter such unusual lines, but I have given a fairly detailed coverage for the sake of completeness. One problem with the division of material into games is that while it is convenient for the reader who starts at page l it is not so suitable if the book is to be used as a reference work. Finding which game contains a particular line is not so easy, but an unusually detailed index should help with this problem. I have again included plenty of diagrams at critical moments to provide further assistance. Finally I would like to thank all those readers of Beating the Sici lian who wrote with comments, suggestions and games. Grand masters are usually only too will ing to offer their opinions, and it is useful to get balancing feedback from the club players who provide the main readership of a book such as this.

Najdorf Variation

Of all the lines in the Sicilian which Black can adopt, the Naj dorf has developed the largest body of theory. Whole books have been written on mere sub variations of the Najdorf, for ex ample the Polugayevsky variation and the infamous Poisoned Pawn. Devising a counter which is viable in tournament play, while at the same time necessitating relatively slight book knowledge, has proved especially hard. The Naj dorf starts with the moves I 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 ll. b5+ while the secondary point is to prepare queenside expansion by b5. Some of White's sys tems against the Najdorf are spe cifically aimed at preventing e5, while others allow Black to play this move in the hope of exploiting the backward d-pawn later. 6 ll.g5 and 6ll.e2 are the most common replies, but as both lead to reams of analysis I have settled on a less familiar system which offers good attacking chances while retaining an ele ment of solidity. This system, based on 6 f4, has become more popular since the first edition of this book was published and

games by players such as Belyavsky, Dolmatov and A. Sokolov have shed new light on several lines. In games 1 -4 we investigate the replies most com monly encountered in practice. The main division is between those lines involving an early e5 and those in which Black de lays this move or omits it entirely. We will postpone consideration of the former lines until games 3 and 4, and concentrate first on the e5. alternatives to Game I F. Olafsson-Sax Novi Sad 1976

1 2 3 4 5 6 1 B

e4 �f3 d4 � xd4 �c3 f4 (1)

c5 d6 cxd4 �f6 a6

2 N aj dorf Variation 6 f4 is a flexible move; White gives little away regarding his piece deployment, and waits for Black's reply before deciding where to put his bishops. 6 <E)bd7 This move is designed to re serve the option of playing e5 under more favourable circum stances if White should develop his pieces to unsuitable squares. 6 'Wfc7 is game 2, but there are other playable moves: ( l ) 6 ... <E)c6 7 <E)xc6 bxc6 8 e5 <E)d7 (8 . . . dxe5 9 'Wfxd8 + '\t;xd8 1 0 fxe5 <E)d5 1 1 jj_d2 is good for White) 9 jJ_c4!? dxe5 1 0 0-0 e6 1 1 f5 jj_c5+ 1 2 fph1 with good at tacking chances. (2) 6 .. . g6 7 jJ_d3 and after a subsequent . . . 'Wfc7 or <E)bd7 there will probably be a transposi tion into lines considered in game 2. (3) 6 e6 (after this White may transpose into various lines of the Scheveningen, but since these lines do not form part of the repertoire recommended in this book, we suggest an independent alternative which promises good chances for White) 7 'W/f3 (White's advantage over similar lines is that his bishops are not committed, so he can force through g4-g5 very quickly) and now: (3a) 7 ... �bd7 8 g4 e5 9 <E)f5 is dangerous for Black, for example 9 g6 (9 . . . exf4?! 1 0 jj_xf4 <E)e5 1 1 jJ_xe5 dxe5 1 2 g5 <E�d7 1 3 jJ_c4 gives White a tremendous attack)

1 0 g5 gxf5 1 1 exf5 and Black is under heavy pressure. (3b) 7 ... 'Wfc7 8 g4 b5 9 g5 �fd7 (9 b4 1 0 <E)cb5 axb5 1 1 gxf6 gxf6 1 2 jJ_xb5+ jJ_d7 1 3 jj_ d3 gave White a good game in Lau-Schuh, West German Ch. 1 987) 1 0 a3 jJ_b7 1 1 jj_e3 (1 1 jJ_g2 g6 12 'W/f2 <E�c6 1 3 <E)de2 h6 1 4 f5 was also promising in Smyslov Kamsky, New York Open 1 989) <E)c6 1 2 jj_ h3 b4 13 axb4 <E)xb4 1 4 0-0 <E)c5 15 madl g6 1 6 .!!d2 jJ_e7 17 K df2 with advantage to White, Timman-Hjartarson, Bel fort World Cup 1 988. (3c) 7 .. 'W/b6 8 <E)b3 'Wfc7 9 g4 b5 1 0 jj_ d3 jj_b7 1 1 g5 <E�fd7 1 2 jj_e3 <E)c6 1 3 'W/h3 b4 (1 3 <E�c5 1 4 0-0 <E)b4 1 5 l!adl <E)bxd3 1 6 cxd3 <E)xb3 1 7 axb3 'W!d7 1 8 f5 exf5 1 9 jj_ d4! was good for White in Ulybin-Labunsky, USSR 1 987) 1 4 <E)e2 g6 1 5 0-0 -0 jJ_g7 1 6 K hf l 0-0-0 17 f5 <E)ce5 1 8 'lt;b l with a slight plus for White, Uly bin-Magerramov, Uzgorod 1988. 7 jJ_e2 (2) This position frequently arises via the move order 6 jJ_e2 <E)bd7 .

Naj dorf Variation 3 (instead of the more common 6 e5) 7 f4. 7 eS This move is most common, the following lines being somewhat unpleasant for Black: ( 1) 7 g6 (an attempt to reach positions akin to the Dragon, but here White can exploit an interest ing tactical resource) 8 g4 Qlc5 (8 . . h6 9 f5 Qlc5 10 J.,f3 e5 1 1 <El b3 gxf5 12 exf5 e4 13 J.,g2 Qlxb3 14 axb3 )!!g8 15 h3 J.,xf5 16 'i$'d4 J.,e6 was unclear in Belyavsky Ljubojevic, Bugojno 1984, but this line is certainly risky for Black) 9 Q�b3! b6 (not 9 Ql fxe4? 10 Qlxe4 '2:!xe4 1 1 �d4 Qlf6 1 2 g5 winning, nor 9 Qlxb3 10 axb3 J.,g7 1 1 g5 Qld7 1 2 J.,e3 with a clear plus for White, Liberzon-Savon, Sukhumi 1972) 10 g5 Qlfd7, Arnason-Tringov, Plovdiv 1 986, and now 1 1 J.,f3 J..b7 12 �e2 supporting e4 and preparing J.,d2 and 0-0-0 gives White a promising position. (2) 7 Qlc5 8 J..f3 *b6 9 '2:! b3 (once again this move, put ting the question to the c5 knight, gives White the advantage) Qlxb3 10 axb3 g6 11 e5 dxe5 1 2 fxe5 Qld7 13 Qld5 'i'ird8 14 J.,g5, Gips Iis-Quinteros, Olot 1 973 and White stands very well since 14 . . <El xe5? loses to 15 <El f6 + exf6 1 6 tij'xd8 + f!'xd8 1 7 J.,xf6 + (3) 7 �b6 8 Qlb3 g6 9 litd3 (preparing jle3, when Black is driven back in confusion) �c7 1 0 g4 Qlc5 1 1 Qlxc5 *xc5 1 2 j_e3 �aS 1 3 b4! with advantage, . . .

. . .

Torre-Quinteros, Leningrad IZ 1973. (4) 7 bS (I suggested this move in 1982 but a few months later found a good reply; 8 Qld5! J.,b7 (9 Qlc6 was the threat and 8 Qlxd5 9 exd5 gives White a superb outpost at c6) 9 Qlxf6+ '2:!xf6 (or else White has a positio nal advantage) 1 0 e5 dxe5 l l fxe5 followed by e6, when Black has problems developing his kingside p1eces. 8 Qlf5 Inexperienced White players sometimes try 8 <£)b3 but after the reply 8 b5 White should be thinking about equalizing! Natu ral moves like 8 <£)b3 quite often turn out badly in the Najdorf, which is one reason why it is so popular with Black players. 8 Qlc5 9 Qlg3 White's knight manoeuvre solidly defends the sensitive e4 square and he is now ready to complete his development by 0-0 and J..e3. If Black does nothing to stop this then White's position will be very promising, so his next move is directed against both the developing moves mentioned above. 9 �b6 Or 9 j_d7 10 0-0 j_e7 1 1 j_e3 ( 1 1 a 4 i s also possible, when 11 g6? 12 J.,e3 J.. c6 1 3 fxe5! '2:!fxe4 1 4 '2:!cxe4 Qlxe4 1 5 e6! was very good for White in Kiprich nikov-Petkevich, USSR 1975) � c8 1 2 fxe5 dxe5 13 <£)d5 j_e6! . . .

4 N aj dorf Variation 1 4 <E:Jxf6+ (or 1 4 tt:Jxe7 �xe7 1 5 Jtg5 .:E:lcd7 1 6 c3 with an edge for White) Jtxf6 1 5 �cl Jte7 with a small plus for White, Tsesh kovsky-Tarjan, Riga 1 979. 10 Kb1! (3) It may seem to be a major con cession to play such a move merely in order to achieve j_e3, and for a long time 1 0 f5 was played in order to delay the pain ful decision to waste time with l!l b l . However, it is now recog nized that White should retain the option of opening up the position by fxe5 and so the immediate .1. b l is thought best today. The situation in this line is typical of opening variations in which Black plays very actively. If he can keep his initiative going then every thing turns out well for him, but if he falters and has to start moving backwards the result is usually disastrous. Which of these alter natives actually occurs frequently depends on obscure tactical points, as happens here. 3 B

10 10

Jtd7 JJ.. e7 sets a neat trap

which Karpov once fell into-1 1 Jte3?! exf4 1 2 j_xf4 .:E:lcxe4 1 3 .:E:lgxe4 .:E:lxe4 1 4 .:E:lxe4 �b4+ and Black wins a pawn, although White probably has sufficient compensation to draw. The best line against 1 0 j_e7 is 1 1 fxe5 dxe5 1 2 j_e3 �c6 (1 2 �c7 1 3 0-0 0-0 1 4 <ffh l b5 15 � xf6! Jtxf6 1 6 .:E:ld5 �c6 1 7 .:E:lxf6+ gxf6 1 8 �el is also dangerous for Black) 1 3 0-0 0-0 1 4 .:E:ld5 .:E:lxd5 1 5 exd5 �c7 1 6 b4 .:E:la4 1 7 .:E:le4 and White's powerful queenside majority gives him a substantial Vogt-Fernandez, advantage, Halle 1 978. 10 exf4 1 1 j_xf4 .:E:le6 is a recent idea for Black, which has had success in two games. After 1 2 itd2 .:E:ld4 13 itd3 (or 1 3 .:E:ld5 .:E:lxd5 1 4 exd5 Jte7 1 5 JJ..c3 .:E:lf5 with rough equality), Kinder mann-King, Dortmund 1 988, continued 1 3 Jte7 1 4 .:E:lh5 0-0 15 .:E:lxf6+ Jtxf6 1 6 .:E:ld5 �d8 1 7 0-0 j_e5 1 8 �h5 j_e6 1 9 j_g5 f6 20 Jte3 b5 with equality, while Vogt-H. Olafsson, Thessaloniki 01. 1 988 went 1 3 . . �c5 1 4 .:E:ld5 .:E:lxd5 1 5 exd5 �xd5 1 6 0-0 .:E:lc6 1 7 j_e4 �c5 + 1 8 <ffh l .:E:le5 19 .:E:lf5 J..xf5 20 K xf5 �c7 and this too ended in a draw. The second line looks risky for Black, but King's handling of the position affords White few chances for an advantage. However White has an interesting option on move 13, namely 13 .:E:lh5!? .:E:lxh5 1 4 j_xh5 j_e6 (or else .:E:ld5) 1 5 0-0!? To castle into a double check appears . . .

N aj dorf Variation suicidal, yet this is the only way to make j_e3 into a threat! After 15 g6 16 J,.e3 J,.g7 17 <E)a4 'fl!a7 18 c3 or 15 . . . � xc2 + 1 6 fihl IE}e3 17 'fl!a4+ White is clearly better, so Black should probably play the safe 15 <E)c6+ 16 * h 1 g6, although even here 17 jtg4! looks at least slightly better for White. 1 1 fxe5 dxe5 12 j_e3 Here, too, this is White's best. The threat of b4 forces Black to move his queen, and since he must try to keep up the momentum it is natural to attack the e4 pawn. Everything then hinges on whether White can successfully gambit this pawn. 'fl!c6 12 13 0-0 0-0-0 Black would like to play 13 <E) cxe4 14 <21 cxe4 <21 xe4 15 J. f3 jtc5 but White wins material by 1 6 �xe4 j_xe3+ 1 7 ff h l 0-0 18 <E)c3 'fl!c8 1 9 IE}d5 J..c5 20 b4 j_d6 21 <E)b6. If Black can't take the e-pawn then his whole plan looks suspicious, but he could have minimized his disadvantage by 1 3 !it, d8! As played his king is subjected to a devastating attack. 14 b4! White forces Black to take the poisoned pawn. This would also have been the reply to 13 . . . j_e7. 14 � cxe4 1 4 . . . <E)e6 1 5 <E)d5 threatening c4 and b5 is also horrible. IE}xe4 15 IE}cxe4 'flfxe4 (4) 16 <E)xe4

. . .

17 j_b6 K e8? Now Black gets mated. The best practical chance was to jetti son the exchange by 1 7 . . . J..e7, although 18 J..f3 'fl!g6 1 9 jtxd8 � xd8 20 'f!ld5 'f!lb6+ 21 fi h l f6 22 'f!lf7 attacking the vulnerable kingside pawns should win for White in the long run. 18 J.. f3 'W!g6 19 jt xb7+ '*b8 19 '\ffxb7 20 'fl!xd7+ fixb6 2 1 'flfxe8 jtc5+ is impossible because White takes the bishop with check. 20 J.. e3 The simplest. Material is level and White's attack against the naked Black king must be deci sive. 20 j_b5 21 jtf3 'W!d6 21 . . . jtxf l loses to 22 'f!ld7. f5 22 'fl!e1 22 jtxfl 23 'fl!xfl followed by b5 or H dl wins. 23 c4 jt xc4 axb5 24 b5 25 'fl!a5 e4 Resigns 26 .!il(xb5 +

6 Naj dorf Vari ation 26 jlxb5 27 tyxb5+ *c7 28 H c l + *d8 29 j_b6+ * e7 30 jlc5 wins. Game 2 Nunn-Griinfeld England-Israel Telex Match 1981 c5 1 e4 d6 2 �f3 3 d4 cxd4 4 � xd4 � f6 a6 5 �c3 tyc7 6 f4 If Black wishes to delay e5 (or even dispense with it alto gether) then this is probably the most reliable way to go about it. Black avoids the tactical problems resulting from a quick e5 by White and can continue his deg6, jlg7, velopment by �bd7 and maybe b5 and 0-0 as well. 7 �f3 (5) This is more accurate than 7 j)_d3, when 7 e5 8 �f3 b5 transposes into a relatively com fortable line for Black. 5 B

range of move orders. The main question is whether Black can exploit White's early �f3 by play ing 7 . . . e6. The analysis runs 7 e6 8 j_d3 and now: (I) 8 b5 9 e5! dxe5 (9 . . . b4 1 0 �e4 �xe4 1 1 jlxe4 d5 1 2 Ad3 is slightly better for White, while Sax-Guerra, Dubai 01. 1 986 con tinued 9 � fd7 10 0-0 �c6 !I f!h l .f/J_e7 1 2 tye l 0-0 1 3 tyg3 f5 1 4 exf6 �xf6 1 5 j_d2 with advan tage to White) 1 0 fxe5 �fd7 (1 0 �g4 1 1 tye2 j_b7 1 2 jle4 also gives White an edge) Il 0-0 �c6 1 2 J_f4 .f/J...b7 1 3 *hi �c5 1 4 �e4 �xe4 1 5 jlxe4 �b4, Reeh Schulz, West German Ch. 1 987, and now 1 6 �g5!? gives White dangerous attacking chances. (2) 8 . <2lc6 9 0-0 and now: (2a) 9 .f/J...e7 10 tyel and Black has a range of possibilities. The passive 10 � d7 1 1 �g3 0-0 1 2 f!h l �e8?! 13 e5! �b4 1 4 f5! gave White a strong attack in Hazai-Karolyi, Hungary Ch. 1 986, while 10 . . . 0-0 is met by 1 1 e5 dxe5 1 2 fxe5 �d7 1 3 J_f4 �c5 1 4 ,.g3 �xd3 1 5 cxd3 *h8 1 6 �e4 and White i s better accord ing to Sax and Hazai. Perhaps the best line is 10 �b4 1 1 e5 �xd3 1 2 cxd3 �d5 1 3 .:£)xd5 exd5 1 4 tyg3 0-0 1 5 * h i dxe5 1 6 fxe5 jlf5 1 7 �d4 j)_g6 1 8 .f/J_d2 with a level position, Sax-de Firmian, New York Open 1 987. (2b) 9 b5 1 0 'l!!te l j_b7 I l *hi jle7 1 2 e5! dxe5 1 3 fxe5 ®d7 1 4 .f/J_f4 �c5 1 5 �e4! (as in line I above, except that the . . .

. . .

7 �bd7 After this we reach a standard position which can arise by a wide

N aj dorf Variation 7 moves �e 1 and . . . .fi.e7 have been added) <£)xe4 ( 15 <£)xd3 1 6 cxd3 0-0 17 �g3 'f!h8 1 8 <£)f6! �d8 19 <£)g5 is very good for White) 1 6 .fi.xe4 h6?! ( 16 . <E:lb4 17 .fi.xb7 �xb7 1 8 �g3 g6 may be better, although 19 .!i.h6 cuts Black off from the kingside), Sax A. Sokolov, Reykjavik 1 988, and now 1 7 a4 b4 1 8 �f 2 gives White good attacking prospects. (3) 8 . . . <£)bd7 9 0-0 11.e7 (9 b5 may be met by the simple 1 0 �e1 , or even b y 1 0 e5!? dxe5 1 1 fxe5 <E:lxe5 12 <E:lxe5 �xe5 1 3 �f3 and now 13 . . . lilt b8?! 14 .!i.f4 'i'ifc5 + 15 \tih l Ji.b7 1 6 .fi.e4! <E:lxe4 1 7 .fi.xb8 f5 18 li;l\ ae1 was good for White in Nicevski-Mar kiewicz, Dembica 1 987, so Black should have played 13 ... � a7 1 4 .!i.f4 �h5 1 5 'i'lfg3 with an unclear position) 1 0 �e l 0-0 11 e5! <E:le8 12 'ltg3 <£)c5 13 .fi.e3 1l_d7?! 14 .fi.xh7+ ! \tixh7 15 <E:lg5 + \tig8 ( 15 .fi.xg5 1 6 fxg5 �h8 17 /i;il,xf7 is unpleasant) 1 6 '/!l!i'h4 .fi.xg5 17 fxg5 Ji.c6, Wedberg1 onescu, Berlin 1 988, and after 1 8 exd6 <E:lxd6 (not 1 8 . . . �xd6 19 /i;il,ad1 �e7 20 'l!!fb4) 1 9 .fi.xc5 <£)f5 Black does not have enough for the pawn. g6 8 1l_d3 8 . e5 9 a4 transposes to game 3, while 8 e6 is line 3 in the above analysis. 9 0-0 White's strategy in this line is rather crude. He intends a straightforward attacking build up on the kingside by *e 1-h4, f5,

.!i. h6, and <E:lg5. Of course Black is also playing moves while all this is going on but if he continues naively with his development without taking specific counter measures he can easily fall victim to White's attack. 9 .!i.g7 10 �e1 (6)

10 0-0 This move is probably already an inaccuracy. The alternatives are: ( l ) 10 ... <E:lc5 1 1 e5 dxe5 1 2 fxe5 <£)fd7 1 3 .fi.f4 <£)e6 1 4 .fi.g3 <E:lb6 (after 1 4 0-0 1 5 fi'h l Black has to find an answer to <E:ld5) 1 5 a4 .!i.d7 1 6 a5 �c8 1 7 <£)e4 and White's initiative proved decisive in Sax-Minic, Rovinj Zagreb 1 975. (2) 10 . . . b5 (probably the best move, aiming to complete Black's development before he gives White a target to attack by castling) 1 1 e5 (attacking moves like �h4 serve no function while Black's king is still in the centre) dxe5 1 2 fxe5 <£)g4 1 3 e6 fxe6 14 �h4 with an unclear position. For the pawn White has some

8 N aj dorf Var iation

initiative and Black has problems finding a refuge for his king. White also has the option of opening lines on the queenside by a timely a4, and in practice Black will not have an easy defensive task ahead of him. b5 1 1 .-h4 12 f5 At one time White invariably played * h1 before proceeding with his attack. This type of con solidating move is often a symp tom of chess laziness, in that White does not want to be both ered with calculating the conse quences of Black's queen check in every variation and so simply rules it out, even though it may cost him a vital tempo. jj_b7? 12 Black continues to play normal Sicilian moves without realizing how critical his position has become. The point is that after White's fxg6 Black does not want to play hxg6 when �g5 gives White a permanent mating threat on h7. However, the recapture . . . fxg6 invites White's knight to come in at e6 and Black's . . . jj_b7 removes a vital defence from this square. 1 2 . . . �c5 was essential, when 1 3 jj_h6 b4 may enable Black to defend. 13 fxg6 In a game Velikov-Valenti, Pernik 1 979, White played 13 j}_e3 (laziness again) when Black missed his second chance to play �c5 and lost after 1 3 b4? 1 4 �d5! jlxd5 1 5 exd5 �xd5 1 6

fxg6 hxg6 1 7 �g5 �5f6 1 8 gf3 with a crushing attack. fxg6 13 After 1 3 . . . hxg6 1 4 �g5 Black cannot move either knight since �h5 is met by g4. White can just build up by �f3 and � afl to eliminate the defensive knights at f6. 14 .f)g5 �c5 ( 7) Too late! 1 4 �b6+ (1 4 �h5 1 5 J..e3 is good for White) 1 5 \ffh l �h5 was best, but even then 1 6 j}_d2 threatening both �e6 and .:2)d5 gives White a pro mising attack.

15 );! xf6! �xf6 16 �xh7 + 'if1fB 17 j}_e3 White's material investment is very slight for such a strong at tack. The main threat is 1 8 .:2)d5 jj_xd5 1 9 exd5 attacking g6 and preparing b4 followed by the occupation of e6 by White's knight. 17 .:2) xd3 1 7 . . . e5 1 8 .:2\ d5 j}_xd5 1 9 exd5 e4 20 .il.. e2 � e8 21 b4 followed by .f)e6 + is also winning.

N aj dorf Variation 9 'ltd7 18 cxd3 Black cannot meet the threat of �d5 by 18 e6 since 19 �xe6+ ! li( xe6 20 � f l + <lieS (20 � f6 2 1 il! xf6 + wins the queen) 21 'lt'g8+ is decisive. j}_ xd5 19 .!2ld5 'lt'f5 20 exd5 2£le6 + � xe6 21 'lt'xe6 22 dxe6 Resigns 23 j}_h6 23 J..xh6 24 'ith8 + "'ffl 25 ii f l + wins everything.

Game 3 Nunn-Cserna Lugano 1984 1 e4 c5 2 �f3 d6 3 d4 cxd4 4 .!2lxd4 .!2lf6 5 .!2lc3 a6 e5 6 f4 Black's most popular move. After all, a6 was designed to prepare . . . e5 and the determined Najdorf player will generally play e5 unless it is absolutely im possible. 'lt'c7 7 .!2lf3 At one time it was held that Black should prevent the active development of White's bishop at c4 and so this move was almost universal. But more recently 7 �bd7 has become the most popu lar move. We consider this in game 4. 8 a4 The alternative is 8 J..d3. The continuations after 8 a4 and 8 J..d3 are rather similar, but there

are some important differences. Firstly 8 a4 expends a tempo, but this is not especially serious since White can often omit * h 1 (after 8 j_d3 b5 White usually has to play *hi since Black's check on b6 gives him extra defensive possibi lities). More significantly, a4 re serves the c4 square for White's use (by 'lt'e2 and J..c4 or .!2ld2 and j_c4) and in some lines the added pressure White can exert on f7 by these manoeuvres improves his chances considerably. 8 .!2lbd7 9 J.. d3 (8) 8 B

9 g6 Black's main decision is whether the f8 bishop should go to e7 or g7. There are two other lines, one in which Black commits himself to . . . j}_e7 at once and one in which he postpones the decision: (1 ) 9 J.. e7 10 0-0 0-0 1 1 .!2lh4 g6 (this is usually unavoidable in lines where Black plays . . . j_e7, for example 1 1 J..d8?! 1 2 .!2lf5 g6 1 3 <2\h6+ fig7 14 f5 b6 15 g4 with a dangerous attack, Sax• . .

10 N aj dorf Var iation

Bukic, Vrbas 1980) 1 2 f5 d5!? 13 exd5 e4 (13 �xd5? 1 4 �xd5 �c5+ 1 5 �e3 �xd5 1 6 fxg6 hxg6, Cramling-Gallagher, Oak ham 1 984, and now White could have won by 1 7 �g4!, e.g. 17 . . . �f6 1 8 �xg6 j_xg4 1 9 �xe7+ 'fih8 20 <2)xd5 <2lxd5 2 1 j_h6, 17 . . . j_c5 1 8 <2lxg6 J..xe3+ 19 fih l "ii\'e6 20 J..f5, 17 . . . �b6 1 8 "ii\'g3 and White has threats a t b6 and g6, or finally 17 . . . J.. xh4 1 8 "ii\'xh4 followed by j_c4 and j_h6, and in every case White has a winning position) 1 4 j_e2 j_d6 1 5 g3 b6, Sax-Andersson, London 1 980, and now 16 <2)g2 is best, followed by j_h6 and �e3 sup porting the pawns at d5 and f5, when White should have the advantage. (2) 9 . . . b6 (the problem with this delaying move is that Black may have trouble getting castled) 1 0 0-0 j_b7 11 �e1 g6 ( 11 . . . J..e7 1 2 fi h l 0-0 1 3 �h4 g6 1 4 fxe5 dxe5 1 5 j_h6 H e8 1 6 �f5! J..f8 1 7 � xf8 � xf8 1 8 �e3 is slightly better for White, Szna pik-Ostermeyer, Oslo 1 983) 1 2 fxe5 dxe5 1 3 �h4 J..e7 (1 3 J..g7 14 J..h6 0-0 transposes to the main line) 14 j_g5 h6 1 5 f1h 1 (1 5 �d2? J..c5+ and 1 6 �h5) f1f8 (or 1 5 0-0-0 1 6 j_e3 with an automatic attack against Black's weakened queenside) 1 6 �d2! f1g7 1 7 j_e3 �c5 18 j_c4 �af8 19 �f2 j_d8 20 j_xc5 "ii\'xc5 21 'i!J\,1'xc5 bxc5 22 a5 H. Ranta nen-Nunn, 1 98 1, Helsinki although White is distinctly

better. This i s an example of the advantage of having c4 available for the bishop. 10 0-0 � g7 11 �e1 0-0 12 fxe5 dxe5 13 �h4 b6 14 J.. h6 j_b7 15 � g5 (9) This position is the natural re sult of White's blunt play. Although its evaluation has fluc tuated over the years, the scales have recently tipped in White's favour. It is very rarely seen today because Black players steer well clear of it. 9 8

};i(fc8 15 The only other reasonable move is 15 "ii!'d6 (15 ... �h5? 1 6 j_xg7 *xg7 17 Hxf7+ and 15 1ii! ae8 16 g4! are to be avoided, while 15 . . . l;:l: fe8 1 6 .�xg7 fixg7 17 �xf7! 'ifi>xf7 18 �xh7+ 'lt'e6 19 �xf6 + ! fixf6 20 };i(fl + 'l!'e6 2 1 ..l\'l.c4 + *d6 22 liild l + f1c6 23 J.. d5+ *c5 24 j_xb7 Resigns was Rantanen-Morris, Gausdal 1 978), but 1 6 Hadl causes serious problems. After 1 6 . . . �h5 White • • •

N aj dorf Variation 11 plays 18 j_xg7 'f5xg7 19 J..e2 �c5 + 20 'I;> hi, when 20 . . . �hf6 loses to 2 1 � xd 7 and 20 . . . '21 df6 loses to 2 1 j_xh5 �xh5 22 � xf 7+ Other 16th moves are almost as bad. 16 '15h1 Black intends to meet moves such as 16 g4 and 16 gad! by 'ii'Jc5 + and 'ii'Jf8, when the rook on c8 prevents J..c4 and Black successfully defends. Un fortunately White has a tactical idea which cuts across Black's plan to bring his queen to f8. 16 �d6 17 j_xg7 * xg7 18 � xti! This hardly counts as a sacri fice, since White immediately gains three pawns for the piece, while Black's king is left floating around in the middle of the board. * xti 18 19 •xh7 + *e6 20 •xg6 (10) There was even a second good line in 20 �xf6+ �xf6 21 •xb7 since the attempt to liquidate by 21 •c7 allows 22 j_xa6. 10 B

20 ftte7 21 � ad1 As is so often the case, tt ts better to spend time cutting off the king's escape route than to give pointless checks which only serve to drive the king into safety. 21 �h7 22 �g3 * e7 Black cannot play 22 li;i!.g8 because of the check at c4. 23 �d5 + jt xd5 �h6? 24 exd5 This error allows White to force the king onto the back rank, cut ting off both Black rooks from the kingside. However, even the best line 24 e4 is good for White after 25 �de l H g8 26 \Wc7! 'ii!Jg6 27 g3 � ac8 28 d6 + *e6 29 �b7, and with the fall of the e4 pawn White's rooks can get to grips with the Black king. *d8 25 d6 + 26 J.. f5! � a7 It is hardly surpnsmg that Black has no reasonable move. The immediate threat is 27 j_xd7 *xd7 28 �xe5. � xd7 27 j_xd7 Resigns 28 it xf6

Game 4 Nunn-King Bundesliga 1986/7 1 2 3 4 5 6 7

e4 � f3 d4 � xd4 �c3 f4 � f3

c5 d6 cxd4 � f6 a6 e5 �bd7

12 Naj dorf Variation In my view this is Black's best response to 6 f4. Although White's bishop can now be de veloped more actively at c4, Black saves a vital tempo by missing out 'f!lc7 and this gives him good chances to equalize. In fact cur rent theory suggests that White's best plan is to ignore the option to play �c4, and to proceed with his normal development by �d3. Admittedly Black benefits from missing out -wtc7, but it is not clear that the alternatives to '$c7 fully equalize. 8 a4 White cannot do without this as 8 �c4 allows 8 . . . b5 9 � d5 )i:i( b8 1 0 �g5 (after 10 fxe5 dxe5 1 1 jtg5 �b4 Black was slightly better in Hort-Andersson, Wijk aan Zee 1 979) �xd5 1 1 wxd5 'itf/e7 1 2 0-0 h6 with equality, Korchnoi-Hort, Zurich 1 984, while 8 jtd3 allows 8 . . b5, just the line White is trying to avoid. �e7 8 If Black relents by 8 . 'l!!fc7, we reach game 3. 8 d5 is a speculative recent idea, but if White replies 9 exd5 (better than 9 fxe5, as played in Ulybin-Odeev, USSR 1 989) e4 1 0 �e5! 11.. b4 1 1 �c4 �b6 1 2 jtb3, he gains the advantage after 12 �fxd5 1 3 a5 �xc3 1 4 �xd8+ wxd8 1 5 �xf7+ 'l!/e7 1 8 bxc3 �xc3+ 1 7 fl>f2 or 1 2 �bxd5 1 3 0-0 �xc3 1 4 bxc3 0-0 15 *d4. This analysis has not been tested in practice, but it looks good. . • .

. . •

9 10

� d3 0-0 (11)

0-0

10 �c5 It is risky to take the pawn by 10 exf4 1 1 Axf4 'l!!fb6+ 1 2 'itrh 1 li(xb2 1 3 'l!!le l , for example 13 'l!l!l'b4 ( 1 3 �c5 1 4 liil b l �xd3 1 5 cxd3 'l!!fc2 1 6 d4 l;[e8 1 7 �f2! 'l!!ld3 1 8 Jiilc l Af8 1 9 Jge2 with advantage to White, Ciocal tea-Danner, Timisoara 1 982) 1 4 �b l 'l!!fc5 1 5 �d5 �xd5 1 6 exd5 A f6 ( 1 6 .ztd8 1 7 c4 !E:lf6 1 8 jte3 'f<!tc7 1 9 jtd4 gave White a dangerous attack in Ledermann Lau, Ramat-Hasharon 1 982) 1 7 c4 (or 1 7 '/!l;l'e4 g6 1 8 1J..h611..g7 1 9 Axg7 'l!/xg7 20 �g5 '2\e5 2 1 -Wth4 h6 22 �e4 'l\l'!'xd5 with a double edged position, Korolev-Lipiridi, corr. 1 984) 'l¥fc7 (a draw was agreed here in Hazai-Marin, Warsaw 1 987) 1 8 'l!!lg3! �e5 1 9 jtg5 Jtxg5 20 jtxh7 + ! \t1xh7 2 1 !2Jxg5+ wh6 (2 1 wg8 22 • h4 Ji:!! e8 23 lll, beI also gives White a winning attack) 22 �h4+ 'l!/g6 23 .i, b3 f5 24 lf,!g3 '2:)g4 25 � xg4 fxg4 26 � xf8 'l!!fe7 27 � f7 �e8 28 �h7 + *xg5 29

N aj dorf Variation 13 �xg7 + *f6 30 �h6+ fl f5 3 1 �g5 + * e4 32 �e7+ 1-0, Vogt Womacka, E. Germany 1989. 1 1 *h1 This is not the only reasonable move, but judging by recent re sults it is the most dangerous for Black. d5 11 This is the tactical justification of Black's play, but there are quieter alternatives: (I) 1 1 ... .:2\ xd3 12 cxd3 �a5 13 'i!lfe l exf4 14 l£)d5 �d8 15 l£)xf4 ..lll..d7 16 j_d2 was good for White, Mateo-Byrne, New York 1986. (2) 11 . . . Yiic7 12 �e l ( 12) (threatening 13 fxe5 dxe5 14 �g3, when Black has no natural way to defend the e5 pawn) and now: 12 B

(2a) 12 .. ..�d7 13 fxe5 dxe5 14 'ii1'g 3 .M,ae8 15 �xe5 j_d6 16 �d4, but Black's compensation looks inadequate in view of White's control of d5. (2b) 12 ... j_e6 13 f5 .lll_d7 14 g4!? ( 14 j_g5 J_c6 15 jlxf6 J_xf6 16 g4 is also slightly better for White) J_c6 ( 14 l£)xg4 loses to 15 l£)d5 �d8 16 Jggl .:2lf6 17

J.h6 l£)e8 18 Xxg7 + l£)xg7 19 litg3 J..f6 20 li!gl with a winning attack) 15 g5 .:2lh5 (retreating to d7 gives White a completely free hand on the kingside, e.g. 15 l£)fd7 16 f6 j_d8 17 lith4 l£\e6 18 .:2ld5 and wins) 16 f6 j_d8 (not 16 gxf6? 17 lith4 winning) 17 �h4 g6 18 j_e3 (White's kingside attack has come to a temporary halt, so the time has come to bring the remaining pieces into play. There is a positional threat of 19 j_xc5 dxc5 20 j_c4, followed by the occupation of d5. Thus the c5 knight must move, but 18 . . . l£\e6 19 l£)d5 J_xd5 20 exd5 l£\ef4 2 1 J..xf4 exf4 22 J..f5! followed by Jl_g4 is good for White, hence Black's next move) l£)xd3 19 cxd3 *h8 20 l£\e2 d5? (a very natural move aiming to weaken e4 and activate the c6 bishop against the White king. However, it also weakens the important e5 pawn and this turns out to balance White's weak spot at e4. 20 ... jld7, intending lilfc2, is refuted by 2 1 �fcI when 2 1 �a5 surprisingly loses to 22 Xc4! and Black has no defence to 23 b4; the best defence is 20 ... litd7! aiming for counterplay by attack ing a4, when White should con tinue 2 1 �g3 jlxa4 22 l£\xh5 gxh5 23 lilfxh5 jlb5 24 M a3 and Black still has to find a defence to the threat of .Mgl -g4-h4) 2 1 l£\g3 dxe4 (Black cannot play 2 1 litd6 because of 22 d4!, when both 22 . . . dxe4 23 l£\xe5 and 22 .. . exd4 23 e5 followed by Jtxd4 '

14 N aj dorf Var iation leave the c6-hl diagonal blocked by a Black pawn) 22 dxe4 \'td6 23 \'te6 24 _§_ c5 liil, adI! 'it b4 (23 M.g8 25 J.d6 leads to the loss of e5) 24 .:t:lxe5 Q)xg3 + 25 hxg3 J.xe4 + 26 \'txe4 + (after 25 * g l £lb6 27 J.xb6 �xb6 + 28 �f2 Black cannot meet the threats of �xe4 and \'th6) 26 �xe4 ..l\1xe4 + 27 * h2 (surpris ingly the exchange of queens does not stop the attack. White's im mediate threat is 28 J. c5 and the lines 27 . . . Ji:i:c8 28 �d7 *g8 29 .:t:lg4! heading for h6 and 27 . . . 'fi'g8 28 j}_c5 �e8 29 .:t:lxf7! ffxf7 30 1itd7+ 'fjle6 3 1 f7 M.f8 32 �d4 �xf7 33 I, d6+ both win for White) j}_c7 28 ..l\1c5 .�xe5 29 Axf8 lii\xf8 30 �fe 1 .�c2 3 1 !ltd2 Axa4 (Black cannot get two pawns for the exchange since 3 1 J..xg3 + 32 *xg3 Axa4 33 'fJ e7 threatens both �xb7 and M.xf7) 32 �xe5 h6 33 gxh6 'fi'h7 34 g4 * xh6 35 g5 + 'li' h7 36 .1i\ e4 jlc6 37 1ith4 + *g8 38 'fi'g3 Resigns, Nunn-Portisch, Brussels 1986. (2c) 12 . . . exf4 13 Axf4 M.e8 j}_e6 14 .:t:ld4 �b6 15 (after 13 .�.e3 .:t:lg4 16 j}_g 1 .:t:le5 17 .:t:lf5 �.xf5 18 .:t:ld5 �d8 19 exf5 Af6 20 £l.e2 .:t:led7 2 1 .:t:lxf6+ .:t:lxf6 22 j}_f3 White's two bishops gave him the advantage in Short-Gal lagher, British Ch. 1987) 14 .:t:ld4 J..d7 15 ..l\1g5 '$d8, Hazai Novikov, Camaguey 1987, and now White should have taken the chance to activate his bishop by 16 �c4!, pointing it at the sensi-

tive square f7. The key tactical .:t:lxa4 17 'jJ xa4 J..xa4 line 16 b5 18 Q)cxb5 axb5 19 ( 17 'fJ xa8 �xa8 20 J..xb5 is good for White) 18 .:t:lxa4 b5 19 .:t:lc6 �c7 20 .:t:lxe7 + �xe7 2 1 J..xf6 gxf6 22 .:2lb6 bxc4 23 .:t:ld5 �e5 24 .:t:lxf6 + 'fjlg7 25 �h4 h6 (25 M.h8 26 �g4 + ) 26 M.f5! turns out �xb2 well for White since 26 loses to 27 �g5+ ! (3) 1 1 . . . exf4 12 J_xf4 ..l\1d7!? 13 �e2 'jJc8 14 a5 lii(e8 15 J.. e3 :tf8 16 ..l\1d4! (intending .:t:lg5) and now: (3a) 16 . . . jle7 17 b4 .:t:le6 18 j}_b6, 16 . . . h6 17 .:t:ld2 .:t:lg4 18 Ac4 and 16 . . . .:t:le6 17 J..b 6 �e7 18 .:t:lh4 g6 19 �ae1 are clearly good for White. (3b) 16 . . . J..g4 17 'iire3 Ah5 18 j}_xf6 �xf6 19 .:t:ld5 �d8 20 .:t:ld4 j}_g6 2 1 .:t:lf5 ll\e5 22 b4! .:t:lxe4 23 Axe4 ,ll\c4 24 �ae 1 ,ll\cxe4 25 'i'lrxe4 'fJ xe4 26 .Ill xe4 was very promising for White in Ulybin Pigusov, Pavlodar 1987. (3c) 16 . . . .:t:lfxe4 (Ulybin's recommendation in Informator, but it appears to have a tactical flaw) 17 .:t:lxe4 .:t:lxe4 18 J_xe4 £1. b5 19 c4! (Ulybin only con sidered 19 'l'lre3, which leads to a J._xc4 20 draw) ,ll\xc4 (or 19 �c2 J..xfl 2 1 Axh7 + *h8 22 �f5 with a winning attack) 20 �d3 '!Jc5 2 1 Axh7+ *h8 22 �b3 .�xf l 23 j}_xc5 dxc5 24 ifi(xfl *xh7 25 �xf7 with a large ad vantage for White. 12 .:t:lxe5 ( 13) Q�fxe4 12

Naj dorf Variation 15 marginally better for White in Kindermann--de Firmian, Biel II 1986, but 19 b4 .:E�e6 20 c4 itc6 21 l£)xc6 be 22 !!xd8 !!xd8 23 c5 looks more dangerous) 14 j_e3 (White may also play 1 4 'tllte l l£)e6 by first, so as to meet 13 14 itd 1 attacking e4) b6 15 'tllte1 J..b7 16 t�tg3 and now: (2a) Belyavsky-Chandler, Vienna 1986 continued 16 Black has two important alter l£)e6?! (with the idea of exchang j_c5, but this ing bishops by natives: (I) 12 . . . l£lcxe4 13 itxe4 dxe4 plan fails tactically) 17 11 ad1 (not 1 7 f5 J.. d6!, but now f5 is a serious 14 'l{!te2 and now: ( l a) 14 . . . itf5 1 5 g4 jtc8 16 threat) Ac5? (Black's position !!:d l �e8 1 7 g5 l£)d7 18 l£)c4 e3 was uncomfortable in any case) 1 8 (18 . . . l£)c5 19 b4 l£)e6 20 l£)d5 f5! Ad6 (the point is that 18 and 18 . . . b6 1 9 J.. e3 Ab7 20 jtxe3 1 9 fxe6 fxe6 loses to 20 l!xf6! ;gxf6 2 1 �d7) 19 !!xd6 l£)d5 are also good for White) and now there are two favourable 'tlltxd6 20 fxe6 and White won. (2b) Black tried 16 ... � ad8 in lines for White, either 19 itxe3 b5 20 axb5 itb7+ 21 fl>gl axb5, Psakhis-Balashov, Irkutsk 1986, Kengis-Loginov, Pavlodar 1 987, but 1 7 !!ad1 l£)cd7 18 J.. d4 l£)xe5 and now 22 1: xa8 J..xa8 23 l£)xb5 1 9 fxe5 l£)d7 20 b3! (20 J..xa6! was �c8 24 .:E�cd6 �c6 25 fl>f2 leaves even stronger) followed by jtc4 Black with inadequate play for gave White a decisive attack. the two pawns, or the simple 19 Black should have played for !!xd 1 18 exchanges by 1 7 �xe3. )!! xd1 !!d8, but White is still ( l b) 14 . . . �d4 1 5 11d l 'tlltb4 16 a5 J.. d8 17 �a4 �e7 is Logi slightly better. dxe4 13 J.. xe4 nov's suggestion in Informator, 14 l£)d5 ( 14) but now 18 b3! appears good for Better than 14 jte3, played in White. Tilburg (2) 12 . . . dxe4 13 j_e2 (Black Belyavsky-Portisch, f6 15 J_xc5 1986, when 1 4 gains time, but the pawn on e4 obstructs Black's pieces) 'tlltc7 ( 13 jtxc5 1 6 �xe4 'tlltx d1 17 .1, axd1 'tlltx dl 1 4 .i,xd l j_e6 15 j_e3 fxe5 1 8 l£lxc5 J..g4 1 9 11 de1 exf4 Xfd8 1 6 g4 g6 1 7 g5 l£ld5 18 led to equality. The idea of 1 4 �d5 i s to eliminate the e7 bishop; �xd5 j_xd5 19 b3 l£)e6 20 .:E�c4 J_xc4 21 J..xc4 jtc5 22 j_xc5 Black's remaining bishop will be l£lxc5 23 * g2 .1, ac8 24 '11> f2 was obstructed by the e4 pawn, while

16 N aj do rf Varia tion White's can become active along the b2-g7 diagonal. 14 B

jl e6

Or:

(I) 14 j1d6 1 5 l£)c4! and Black has immediate difficulties since the only natural developing move 1 5 jL e6 loses a piece to 1 6 l£)xd6. Otherwise White can proceed with l£)db6, or b4 fol lowed by 1:1. b2. (2) 14 f5 15 b41£)d7 1 6j;j_e3! is good for White, when both 16 1£\ f6 and 16 1£\ xe5 1 7 fxe5 jle6 lose material to l£)xe7+ and jlc5. (3) 14 f6! 15 l£)xe7 + fi(xe7 16 l£)c4 .�e6 1 7 L2\e3 f5 leads to a typical position. White continues with moves such as b3, 1:1. b2 and )llte l when White's bishop is more effective than Black's. However the opposite coloured bishops will exert a drawish tendency, particu larly if Black can exchange .£Jd7-f6-d5. knights by 15 l£)xe7 + fi(xe7 f6 16 f5 16 llil,ad8 1 7 �g4 j;j_c8 ( 17 . . . f6 18 fxe6 fxe5 1 9 j;j_g5 llil, xfl + . . .

. . .

20 �xfl lif8 2 1 �xf8 + �xf8 22 h3 is very good for White) 1 8 fi(g3 (not 18 f6 �xf6!) f6 19 l£)g4 leads to a position much like the game, except that Black's bishop is on c8 instead of f7. Although Black is exerting pressure on f5 from c8, in my view the bishop will be needed for defending the kingside and therefore the game continuation is better. 17 l£\g4 17 fxe6 fxe5 1 8 J:l.e3 l£)xe6 19 'l!l{d5 may give White a minute advantage, but the move played is much more interesting. The reply 17 jL c4 is ineffective after 1 8 �f4 and the bishop will soon be driven away by b3 in any case. 17 J:l. f7 18 '1$'e1 ?! I decided that it was time to start developing my queenside pieces, but I should have spent just one more tempo improving my position by 1 8 a5! It looks strange to put a pawn on a black square when White's plan is to imprison Black's bishop using the pawns on c2, b3, a4, e4 and f5, but it is very useful to have the option of attacking the knight on c5. Not only may White push it away by b4 at a later stage, but by prevent ing .. . b6 White can also attack it by jla3. The extra possibility of winning the e-pawn by getting rid of this knight (e.g. after a5, b3, 1:1. b2, tl' h4 and � aeI) would have made Black's position even more uncomfortable . 18 aS!

Najdorf Variation 1 7 Black seizes on the mistake and secures the c5 square. lilfd8 19 b3 This is not very logical, since in a few moves Black decides that he needs a rook on f8 to support f6, so he could have saved time by 19 b6 followed by ill( ad8--d6. Ji d6 20 Ab2 <1Jh8 21 -rtg3 b6 22 J!lael � f8 23 tth4 24 a e3 ( 15) White tries a little trick; per haps Black won't notice the threat of 25 ll h3 jtg8 26 �e5.

ttd8 24 Unfortunately he does! Now White has to decide on a plan. At first sight his position looks very promising: since both f6 and h7 are weak, the knight on c5 cannot move without losing the e-pawn and Black has no real counter play. However it is hard to find a concrete way White can improve his position, since almost all his pieces are already on their best possible squares. Black's potential counterplay down the d-file limits

the action of White's rooks and although White can hope to win the e-pawn by a timely �f2xe4, this would not be enough to win the game. I could only conceive of two possible ideas, namely �f2h3-f4 coupled with ll h3 to aim at the weak g6 square, or g4--g5 to open up the long diagonal. If White wants to play g4 the knight must move, so it seems natural to play �f2, which keeps both plans open. White therefore aims for the optimum arrangement of rook on e3 and knight on f2, with the bishop on c3 to prevent counter play by �. d2. Black tries to prevent White arriving at this set up. 25 <IJgl Avoiding immediate back rank problems and future long diag onal troubles after g4. White does not need to play jtc3 straight away because 25 liil. d2 loses to 26 �xf6 jtg8 27 �h5! 25 jt g8 Now, however, 26 )(ll. d2 is a threat. 26 jtc3 If Black does nothing White will be able to play 27 .:E}f2, so he deflects the rook away from e3. 26 � dl � d6 27 � eel If correctly followed up this is a perfectly good defence. 27 gxe l 28 �xe l �d6 29 �e3 ,l4d8 30 �d l �e7 31 lii(d4! .!.!xd4 32 jtxd4 is still not completely equal, since if the Black queen leaves the defence of the e-pawn,

18 Najdorf Variation .:&.xc5 and �xe4 wins it, while otherwise White intends '/11!\'f4-b8. Thus Black's choice was probably correct. 28 .£)£2 Sometimes an oversight is the best chance to win! It seemed to me that this was the moment to set up the position with QJf2 and Jii\ e3, but it fails for tactical rea sons. However, there seems to be no other winning attempt. 28 '&l'c8! Since Black obviously cannot play �xf5 because of QJxe4, I decided to continue with my plan. There was nothing better in any case. 29 Jii\e3 After having made this move I suddenly noticed that Black could play 29 'i'i'\'xf5 30 QJxe4 �e6! with an immediate draw, but the confident way I had made the move evidently led my opponent to accept that the pawn was invul nerable. � d5? 29 30 g4!

White not only achieves the position he has been aiming for, but does so in a very favourable form since the rooks on d5 and f8 are both vulnerable to the ma noeuvre <2:1 h3-f4-g6. �c7 30 Black could perhaps have offered more resistance, but the twin possibilities of QJh3-f4 and g5 make this position very un pleasant. 31 QJh3 QJd3 After 31 J..f7 32 QJf4 gd6, the continuation 33 QJg6+ J..xg6 34 fxg6 h6 35 g5 blows up Black's position. 32 cxd3 �xc3 33 QJf4 g5 34 QJxd5 J.,xd5 35 �el Thanks to the earlier * gI Black has no real counterplay and he is soon forced to give up. 35 �d4 J..xe4 36 dxe4 37 �c3 �d5 Resigns 38 �c4

Scheveningen Variation

This line is popular with many of the world's leading players, including Kasparov, and so one would hardly expect there to be a clear way for White to obtain an advantage. The characteristic feature of the Scheveningen is Black's pawn centre at d6 and e6 covering all the central squares on Black's 4th rank. Thus Black avoids the slight weakness at d5 inherent in the Najdorf and Dra gon systems. Behind the cover of his modest but solid pawn centre Black intends to complete his de velopment in peace. The most common move order for Black to adopt if he is aiming for a Sche veningen is I e4 c5 2 �f3 e6 3 d4 cxd4 4 �xd4 �f6 5 �c3 d6, but Black can invert his second and fifth moves in this line. The amount of theory in the Scheveningen has increased dra matically since the first edition of the book, so it is no longer feas ible to cover two major systems within the confines of a relatively short book. Therefore I have de cided to concentrate on the Keres Attack, which is currently con sidered a more critical test of the Scheveningen than the Classical System. The Keres Attack starts I e4 c5 2 �f3 e6 3 d4 cxd4 4 �xd4

�f6 5 �c3 d6 6 g4. This kingside pawn push aims firstly to drive the knight from f6, thereby making it harder for Black to break open the centre by . d5, and secondly to gain space on the kingside and dissuade Black from castling there. Although White often obtains good attacking chances with this system he must not neg lect his development since Black can often break the position open by d5 and even if this loses a pawn White can find his own king stuck in the centre. In other words, a balance must be struck between furthering White's own kingside ambitions and restrain ing Black in 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 in games 5, 6, and 7, or he may continue his own plans and allow the knight to be driven back to d7. In this case Black may choose 6 .. a6, 6 JJ...e7 or 6 . . . �c6. The specific choice of sixth move may not be very important because there are many transposi tions. In game 8 we deal with lines specific to 6 �c6 (i.e. those involving an early 'Wfc7), and

20 Scheveningen Variation the other lines may be found in game 9. Apart from these two major options, there is a third possibility for Black, namely to counterattack by 6 e5. With this move Black loses a tempo, but he hopes to prove that White's g4 has only served to weaken his kingside. This move is a lot better than its reputation, and it has recently seen a surge of popularity; it is covered in game 8. Game 5 Nunrt-Bischoff Lugano 1986 1 2 3 4 5 6

e4 �f3 d4 �xd4 �c3 g4

c5 e6 cxd4 �f6 d6 h6 (16)

Black avoids having his knight driven away from f6. For a long time White almost played 7 g5 hxg5 8 j_xg5, Anatoly Karpov being one supporter of White's point of view. Although this con tinuation gives White a lead in

development it has defects, not the least being that Black's rook on h8 is activated and presses down on White's weak h-pawn. Now 7 h4 is considered to give White better chances than 7 g5 and is currently the most popular line. White wants to continue with Itg l and g5, driving away the knight after all. Black can either aim for counterplay by preparing d5 (game 6), or he can aim to nip White's kingside play in the bud with a timely . . . h5 (game 7). 7 h4 a6!? Although this is a natural move, it has only become popular recently. The main variation is 7 �c6, and this will be exa mined in games 6 and 7, but there is one other important possibility, namely 7 j_e7 (7 e5 8 �f5! jle6 9 g5 �xe4 10 �xg7+ j_xg7 l l �xe4 d5 1 2 gxh6 jlxh6 13 jlxh6 �xh6 14 'Wf"d2 �xh4 1 5 j_b 5+ � c6 1 6 0-0-0 was very good for White in Stanciu-Vegh, Ulan Bator 1986) 8 � f3 (17) and now:

(1 ) 8 . . . �fd7 9 itg3 �c6 1 0 j_e3 a6 1 1 0-0-0 'Wfc7, Ljubojevic-

Scheveningen Variation 21 Timman, Brussels SWIFT 1 986, and now Ljubojevic recommends 12 J..e2 as slightly better for White. (2) 8 . . . g6!? 9 g5 hxg5 1 0 j_xg5 a6 l l 0-0-0 e5 1 2 �de2 J,.g4 1 3 titg3 <E:�bd7 14 f3 j_e6 15 J,.h3 J..xh3 1 6 K xh3 !! c8 1 7 f4 was unclear in De Wit-0 11, Groningen 1984/5. This interesting idea does not seem to have been repeated. Perhaps 1 4 f4 is better, hoping to prove that the exposed position of the g4 bishop is a liability. (3) 8 . . . h5 9 gxh5 and now: (3a) 9 <E:�c6 1 0 j_b5 (an attempt to exploit Black's move order; 10 <E:�xc6 bxc6 l l J,.g5 may be better, when Black has to prove that he has something better than l l <El xh5 transpos ing to line 3b) j_d7 l l j_xc6 bxc6 1 2 e5 <E:�d5?! (1 2 . . . dxe5 1 3 <E:�xc6 j_xc6 1 4 titxc6+ 'ltff8 15 h6 gxh6 1 6 j_d2 �b8 17 0-0-0 titb6 1 8 titxb6 K xb6 is equal according to Ljubojevic) 1 3 exd6 j_xd6 1 4 J..g5 titb6 1 5 0-0-0 j_ e5 16 <El xd5 cxd5 17 c3 with an edge for White, Ljubojevic-Timman, Bugojno 1 986. (3b) 9 <E:� xh5 1 0 J,.g5 (1 0 j_e3!? is an interesting untested idea, offering the h-pawn in return for a quick attack with 0-0-0) <E:�c6 ll <E:�xc6!? bxc6 1 2 0-0-0 j_xg5+ ?! (accepting the sacrifice turns out to be too risky; Black should develop by 1 2 . . . K b8) 1 3 hxg5 titxg5+ 1 4 'ltfbl f1Je7 (not 14 d5? 1 5 exd5 cxd5 1 6 <E:�xd5 exd5 1 7 H xd5 nor 14 . titc5? 1 5

e5! and White wins in both cases, while 14 tite5 15 j_e2 g6 1 6 tite3 intending f4 gives White a dangerous initiative) 1 5 J..e2 g6 1 6 �xd6! f'lxd6 17 titxf7! (some what surprisingly Black has no defence) a5 1 8 �d l + 'ltfe5 1 9 j_xh5 �xh5 20 f4+ titxf4 2 1 titg7+ Resigns, Sobura-Pienia zek, Poland 1988. 8 J..g 2 White abandons his plan to play ii£ gl and g5 because after 8 Kgl d5 9 exd5 <E:�xd5 1 0 <E:�xd5 'l!ltxd5 l l J..g 2 titc4 12 c3 j_e7 1 3 g5 <E:�d7 1 4 tite2 titxe2+ 1 5 'ltfxe2 <El b6 White had no advantage in Karpov-Kindermann, Vienna 1 986. 8 <E:�c6 Or: ( I ) 8 . . . g6 9 g5 hxg5 10 J..xg5 j_e7 l l *d2 e5 1 2 <E:�de2 j_e6 13 0-0-0 .ebd7 1 4 f4 tita5 (or 1 4 titc7?! 1 5 fxe5! dxe5 1 6 <E:�d5 J..xd5 17 exd5 !I c8 1 8 K hf l ! with advantage for White, Ghinda Bonsch, Halle 1 987) 1 5 '\tfb l <E:�b6 16 b3 with some advantage for White, Gufeld-Georgadze, USSR 1 981 . (2) 8 . . . d5 9 exd5 <E:�xd5 10 <E�xd5 exd5 is given as unclear by ECO. However in distinction to 8 I;Igl d5, White's rook is still defending the h-pawn, so White might consider 9 e5 �fd7 10 f4, when 1 0 . . j_e7 l l h5 and 10 . . . h5 l l gxh5 look good for White, so the critical reply is probably 1 0 d5 9 g5 is also titb6. 8 interesting.

22 Scheveningen Variation 9 g5 hxg5 10 hxg5 �xh1 + 1 1 J,.xh1 <21d7 (18) If Black attacks the g5 pawn White continues 11 <Elxd4 1 2 'ljxd4 <Elh7 13 e5! <Elxg5 (13 dxe5 14 '!jh4 traps the knight) 14 exd6 j_xd6 15 ftlxg7 (15 j_xg5 'ljxg5 16 �xd6 �g l + is less clear) with a tremendous attack for no material loss.

12 J.. g2 This move was the result of lengthy thought, but even so it wasn't the best. White has very few constructive moves apart from g6 and 12 g6 <Elxd4 13 gxf7+ fJ'xf7 14 �xd4 '!jh4 (this represents the main advantage of 7 . . . a6 over 7 J,.e7, since in the corre sponding position with the bishop on e7, White's g6 is much more effective) 1 5 J..g2 <Ele5 is obscure. White might be able to claim a slight plus after 1 6 J,.e3 'ljg4 1 7 * f l , but both kings are unhappily placed and I wanted to find some thing clearer. 1 2 j}_e3 is bad after 12 <21c4 <Elde5 threatening (1 3 'lje2 <Elxd4 loses a pawn). Thus the only direct alternative to

12 g6 is 12 f4, but I was reluctant to create a huge empty space around my king. However after the next move White finds himself committed to f4 in any case, so it would have been better to play it at once. As a result of Nunn-Bischoff, subsequent White players took my advice and played the more accurate 1 2 f4!, with the continuation 1 2 . . . '!jb6 (or 12 . . . �xd4 1 3 �xd4 'ljb6 l 4'1jxb6 �xb6 1 5 a4!? J..d7 1 6 a5 'Zlc8 1 7 Ji..e3 Ji..c6 1 8 0-0-0! fJ'd7 1 9 J,.f3 �e7 20 ii..g4 with a clear plus for White, Ghinda-Vogt, Halle 1 987) 1 3 <Elde2 g6 14 b3 'ljc5 (1 4 �c7 1 5 J..b2 b5 1 6 'Wt'd2 J..b7 17 �dl 0-0-0 18 <Ele3 J..e7 1 9 0-0-0 �b6 20 fjbl fJ'b8 21 <2:Jcl was also a little better for White in Griinfeld-Bischoff, Munich 1 987) 15 �d2 b5 1 6 j}_b2 J..b7 17 0-0-0 0-0-0 18 fJ'bl 'ljf2!? (or 1 8 j_e7 1 9 �cl! f6 20 gxf6 'E)xf6 21 '-Z\d3 with an edge) 1 9 �e l! j}_e7 20 'E)dl 'tll!'c5 21 <2:Je3 and White has a small but permanent advan tage, Short-Kindermann, Dort mund 1986. 1 2 Ji..g 2 and 1 2 f4 lead to similar positions, so it is well worth studying Nunn-Bischoff even if you intend to play 12 f4. 12 g6! The point of 1 2 J,.g2 is that White improves the position of his bishop (particularly in the g6 line given above) while Black lacks useful moves. 12 <21de5 allows 1 3 f4 with gain of tempo 'E)c4 14 b3 ltb6 1 5 <2\ce2), (13 12 . 'tll!'c7 (or J.. e7) allows 13 g6

Scheveningen Variation 23 and 1 2 . �b6 1 3 �b3 leads to a loss of time after a subsequent j_e3. Black's move is the best, cutting out g6 by White and again posing the question as to how White can improve his position. 13 f4 Now the defect of 12 j_g2 is revealed. In the analysis of 1 2 f4 we saw that Black generally plays g6 in any case, while White's j_ g2 is usually not necessary. Therefore Black gains a tempo, although in this type of position an extra move is not especially valuable. Now that e5 is denied to Black's knights, White threatens simply .Jte3 so Black's rely is more or less forced. �b6 13 14 �de2 White can only complete his development by �d2 (or 'i'i1 d3), b3, lit b2 and 0-0-0, so before playing f4 I had to make sure that Black couldn't use the four free tempi to harass White's central ized king. �c5 14 Black settles for finishing his own development by b5 and . . .§tb7. There wasn't much choice, as if the queen quits the b(r.gl diagonal White can play j)_ e3. 15 'l!ll' d3 It would have been slightly better to play �d2 as the position of the queen gives Black a tactical opportunity in a few moves. b5 15 16 b3

16 j)_ e3 is countered by 16 �b4! jJ_ b7 16 17 j)_b2 l! c8 A provocative move. Black de cides to keep his king in the centre to help c-file counterplay. After 1 7 . 0-0-0 18 0-0-0 White has a very slight advantage as in the ex amples given after 1 2 f4. 18 0-0-0 � b4 Black goes in for tactics and given his choice last move there was no other consistent continua tion, for otherwise he has no compensation for the long-term handicap of a poor king position. � xc2 19 �d2 b4 20 lft xc2 *f2! 21 *b1 This is the point of Black's combination. The immediate 21 . . bxc3 22 �xc3 leaves Black in a very poor position as the exchanges have not created any weak spots in White's position to provide counterplay. bxc3 22 .§th1 At this stage Black offered a draw but although White must adopt the much less satisfactory recapture with the bishop, thus leaving e4 weak, I decided to play on. 23 .§t xc3 � a7? A misjudgement. The main merit of Black's combination is that his queen has become a nui sance by taking up residence in the heart of White's position, the more so as White cannot contem plate a queen exchange which

24 Scheveningen Variation would relieve Black of any wor ries about his king. Bischoff re treats it to the passive square a8 in return for inconvenient but not really serious pressure on e4. He should have played 23 �c5 24 J.. d4 t,;-h2! (24 . . �xe4 25 t,;-d3! loses a piece while 24 t,;-h4 25 �c3 e5 26 J..e3 looks good for White) when White has problems with his e4 pawn. 25 t,;-e3 e5 26 J..b2 J..g7 creates a very awkward threat of . exf4, so White would prefer 25 <2) c3 t,;-xd2 26 H xd2 with equality. 24 j_b2 t,;-a8 <2) c5 25 'ii!1 e3 26 <2) c3 26 <2) g3 with the idea of f5 is also promising. 26 J.. g7 ( 19) After 26 a5 the reply 27 �d4 ties Black up by preventing J..g7. 19 w

27 <2)d5! Before playing 26 <2)c3 I had to think very carefully about the tac tics initiated by this move, since if White had been reduced to the passive 27 H e t (27 H xd6 j_xc3 is

bad since Black wins after 28 j_xc3 j_xe4+ 29 J,.xe4 *xe4+ 30 'ii!1 xe4 <2)xe4 or 28 f(xc3 j_xe4+ 29 j_xe4 *xe4+ 30 ff a l *h i + 31 J..c l <2)xb3+ ) he would have no advantage. 27 J,. xb2 The most natural move. 27 exd5 (27 . . . J..xd5 28 exd5 J,.xb2 29 dxe6 wins) 28 J..xg7 �xe4 (28 dxe4 29 H xd6 is worse as the undefended �c5 prevents �d8, while 29 . . . �d3 allows H xd3) 29 J..xe4 dxe4 gives White a strong attack on the black squares and Black's queen is bur ied on a8. However it is not easy to find a concrete way to con tinue. 30 H xd6 li;!! d8! 31 �d4 .i, xd6 32 �xd6 t,;-d8 (or else J..f6) 33 'i!lf f8 + *d7 34 �xf7 + wins a pawn, but in view of the opposite coloured bishops and Black's passed e-pawn this might not be enough to win. White should therefore prefer the slow build-up by 30 J..f6, with many defensive problems for Black. 28 �b6 j_ xe4 + After 28 t,;-a7 29 �xc8 j_xc8 White can't take the bish op, but 30 b4 picks up the knight instead, when Black has insuf ficient compensation for the lost exchange. *a7 29 f!' xb2 The only defence. Now 30 �c4 d5 31 'i!lfd4 (31 �d6+ f!' d7 and 31 j_xe4 dxe4 also seem satisfac tory for Black) fails to 31 �a4+ ! 32 bxa4 H b8+ followed by �xd4 .and . . . J..xh l .

Scheveningen Variation 25 �d3+ 30 �xd6 Black's moves continue to be forced. 30 jj_xh 1 3 1 �d4! (at tacking the �c8 and threatening �h8+ ) g b8 (3 1 . . 'g c7 32 �h8 + and mate at d8) 32 ll!fxc5 (threat !=! d7) li;li, d8 (32 ll!fe7 33 �d7 �a8 34 �e5 wins) 33 �e5 gives White a decisive attack, e.g. 33 . . �xd6/'iit e7/�b8 34 'l¥t'h8+ or 33 .i4. b8 34 �xe6+ fxe6 35 �h8+ 31 lfj a3 3 1 �xd3 is met by 31 jj_xd3. 31 �c6! I had overlooked this ingenious defence when I played 26 �c3. Other moves lose quickly: ( 1 ) 31 . . . .i4, c5 32 !,ii xd3 �a5+ (32 . �xb6 33 j}_xe4) 33 \11 b2 jj_xd3 (33 jj_ xh l 34 �d4) 34 jj_c6+ '1id8 (or else a knight check wins the queen) 35 �d4+ 'flc7 36 �a8+ '1i b8 37 �d8 mate. (2) 31 . . . "i!fc7 32 �xd3 jj_ xd3 33 �xc8 �xc8 34 �xd3 �c l + 35 '1i b4 \Wxh l 36 �xa6 with a won ending. (3) 31 . . . Axh1 32 ll!fh3 �b8 33 �xe6+ fxe6 34 \Wh8+ win ning the queen. 32 �xc6 The best, as 32 �d5 �xe3 33 �f6 + is a draw and 32 !=! xd3 (32 �xd3 Axd3 33 Axc6+ lfj f8 is fine for Black) jj_xh I 33 �c4 is about equal. �e7 + (20) 32 33 \WC5! This move was a visible shock for my opponent. 33 J!i!, c5 J..xh1

34 b4 �xc5 35 \Wxc5 is not so good as Black can avoid the exchange of queens by 35 "itd8. 33 �xc5+ 33 �xc5 34 �c8+ \Wd8 35 K xd8+ lfj xd8 transposes except for the position of Black's king, which makes no real difference. 34 I bc5 �xc5 35 b4! White is a1mmg for a good knight v bad bishop ending, but he must avoid 35 '1i b4 �d3+ 36 '1ic3, which allows 36 �f2! Knight v knight endings are still slightly better for White, but far less promising than the position arising in the game. 35 A xh 1 36 bxc5 A d5? With very little time left to reach move 40 Black blunders. He should have tried 36 'll e7 (not 36 . . f6 37 gxf6 <1Jf7 38 �d7 and j}_c6 is impossible, while even if Black's king were on d8 as in the note to Black's 33rd, 36 'll c7 37 \11 b4 \11 c 6 38 lfjc4 followed by \11 d4 and �c4-e5 wins) 37 f1b4 e5! (37 . . . f6 38 lfja5 Ab7 39 �c4

26 Scheveningen Variation followed by fi b6 should win) 38 fxe5 fie6, but even here White wins: 39 .£) c4 J..d5 40 a3! j}_xc4 (there is not much choice as 40 j}_ moves allows fi a5 and 40 fie7 4 1 2E:J e3 and 42 fia5 wins) 41 *xc4 fixe5 42 'ltJ b4! (42 a4? f5 43 gxf6 *xf6 44 fi d5 'ltJe7! 45 fi c6 g5 46 fi b? g4 47 c6 leads to an ending of � + a- Jl, v �, which should be a draw, but Black must avoid 44 . . g5? in this line, when 45 fid6 g4 46 c6 g3 47 c7 g2 48 c8 (Q) g 1 (Q) 49 �f8 + wins his queen) fi d5 (42 f5 43 gxf6 *xf6 44 fi a5 g5 45 c6 wins) 43 a4 and Black is in zugzwang. exd5 37 12) xd5 White wins because Black's a pawn has moved, while White's can still advance either one or two squares. 38 fi b4 fl d7 39 fi c3 and Black lost on time, but 39 fic6 40 'ltJ d4 a5 41 a4 and 39 flc7 40 fi d4 'fic6 4 1 a3 a5 42 a4 lead to the same fatal zugzwang. Game 6 Karpov-8passky Tilburg 1980 1 2 3 4 5 6 7 For 7 game 5. 8

e4 Q) f3 d4 iZ) xd4 �c3 g4 h4 a6 and 7

c5 e6 cxd4 12)f6 d6 h6 12)c6 (21) !J_e7 see

:l!! g1

Another idea is 8 . . . 12) d7 (for 8 h5 see game 7) 9 g5 hxg5 and now White had a good game after both 10 J_ xg5 'l!!l' b6 1 1 12)b3 a6 1 2 h5 fi!c7 13 fi!e2 b5 1 4 0-0-0, Sochi Tseshkovsky-Zarubin, 1981 , and 10 hxg5 12) de5 11 j}_e3 j_d7 12 !J_e2 a6 13 f4, Matulovic Simic, Smederevo 1981 -my per sonal preference is for 1 0 j}_xg5, since Black's knight is badly placed at d7 9 .�b5 j}_d7 10 exd5 Q) xd5 exd5 11 � xd5 12 j}_e3! Karpov's play in this game forced a reassessment of the whole line. Hitherto White had played 12 �e2 + , with rather uninspiring results, but Karpov's idea of !J_e3, "l!lfd2 and 0-0-0 launching a direct attack (even if this means giving up the h4 pawn) strengthens White's play consi derably. 12 !J_ e7 12 �xh4 13 'l!!f e2 exd4 1 4 j}_xd4 + �e7 1 5 j}_xd7+ fixd7 16 j}_e5 H d8 is not so clear, but 13 �d2 J..e7 1 4 0-0-0 followed by

Scheveningen Variation 27 �f5 gives White a strong attack as in the game. It is curious that Kasparov recommends 1 2 . . . •xh4 both in ECO and in his book with Nikitin on the Sche veningen, but in neither case does he mention 1 3 itd2, even though it was given in Informator 30. 13 it d2 (22) 22 B

A xh4?! 13 0-0 1 4 �f5 Axf5 1 5 gxf5 13 fr'h7 16 0-0-0 is also very good for White, but Black has better survi val chances after 13 � xd4 14 Axd7+ 'l!ltxd7 15 •xd4 �f6 16 •b4 �e7, when he went on to draw in Marjanovic-Cebalo, Yugoslav Ch. 1 982, although he needed to defend accurately until move 64 to achieve this! 1t f6 14 0-0-0 �xd4 1 5 �xd7 + 'l!ltxd7 14 16 �xd4 0-0 17 f4! followed by g5 0-0 15 g5! hxg5 16 and 14 �xg5 �xg5 17 I;I xg5 �xd4 1 8 .l. dg l g6 1 9 •xd4 both give White a winning attack. � xf5 15 �f5 16 gxf5 a6 Black could not castle without . . .

. . .

losing his vital h6 pawn, but now his king is permanently pinned down in the centre. bxc6 17 �xc6 + 18 �c5! Now White only needs to rip open the d-file by c4 to finish Black off. Although Spassky launches an ingenious counterat tack his inability to bring the h8 rook into the game enables Kar pov to repulse the threats. 18 K b8 19 b4 H b5 Black's only chance is to eliminate the deadly bishop. 'f;>d7 20 .!:! gel + 21 c4 � xc5 22 bxc5 A g5 itb8 23 cxd5 J..g5 After 22 24 � e3 �xe3 25 fxe3 •e5 26 dxc6+ fixc6 27 itd7+ White should win easily enough. itf6 (23) 23 f4 With the point that 24 fxg5 it a l + 25 fr'c2 •xa2 + 26 fr'd3 •xc4 + 27 f!'e3 hxg5 gives Black four pawns and a tremendous at tack for the rook.

cxd5!

28 Scheveningen Variation Liquidating to a winning end ing. 'i'lfa1 + 24 'i'lfxa2 + 25 f1c2 'i'lfxd2 + 26 *d3 26 'it b3+ 27 'i'lfc3 also forces the queens off. jlxf4 27 X xd2 Although Black has two pawns for the exchange all White's pieces are very active and Black is unable to organize himself against the advance of the c-pawn. cxd5 28 B, a2 h5 29 �:I xa6 h4 30 <11 d4 3 1 f1 xd5 �b8 gxf6 32 f6 jl g3 33 � xf6 f1d8 34 � xti + 35 �f8 + Resigns As 35 f1 d7 36 c6+ *c7 37 K e7+ * b6 38 B, xb8+ jtxb8 39 B. b7 + ends the game. Game 7 Nunn-Sax Rotterdam 1989 c5 e4 e6 �f3 d4 cxd4 �f6 � xd4 d6 �c3 h6 g4 h4 �c6 h5 Jlii g1 This is currently the most fashionable move. White's best reply is to take on h5 and since the recapture �xh5 leaves the knight badly placed, Black nor mally returns it to f6. The net 1 2 3 4 5 6 7 8

effect of this is to reach a position similar to that after 6 g4 h6 7 g5 hxg5 8 jlxg5, but with White hav ing the two extra tempi h4 and �gl . This might seem to be a great improvement, but in fact the disadvantages almost balance the advantages. White has two prob lems: firstly the h-pawn can become weak without the defence of the rook and secondly Black's �b6 effectively pins the f pawn against the undefended � g l , so it is hard for White to play f4. 9 gxh5 <E) xh5 Black players have been known to try 9 �xh5 1 0 �g5 �h8, but this is quite pointless. Black reaches the same position as in the main line, but having forfeited the right to castle kingside. 10 jt g5 (24) 24 B

10 �f6 The alternative is the immedi ate 10 �b6, which attempts to avoid the loss of time inherent in 10 �f6. The problem is that the knight is genuinely badly placed at h5, so Black gains

Scheveningen Variation 29 nothing by keeping it there. After 10 � b6 1 1 � b3 a6 1 2 jt e2 g6 (12 � f6 transposes to the analysis of 1 1 . . . � b6 in the main line) 1 3 �d2 Black may play: ( l ) 13 . . . jte7 14 K g2! (White need not offer his f-pawn by cas tling immediately) jtd7 1 5 0-0-0 I;I c8 1 6 '�i b l "i!Jic7 1 7 a3 b5 1 8 jtxb5 axb5 1 9 � xb5 � b8 20 jtxe7 �xb5 2 1 jt xd6 with a clear plus for White, Lobron-Marjano vic, Reggio Emilia 1 985/6. (2) 13 . . . jtd7 14 K g2! �c7 1 5 0-0-0 b 5 1 6 a3 �e5 1 7 �d4! )[i( h7 1 8 f4 � c4 1 9 jt xh5 1!l( xh5 20 � e 1 jtc6 2 1 � d5 jt xd5 2 2 exd5 e 5 23 �d3 JJ.. e 7 24 �d4! and again White stands well, Motwani Roca, Dubai 01. 1 986. (3) 13 'fltc7 (this is even worse than the lines above because White need not spend time on K g2) 14 0-0-0 b5 1 5 a3 jtd7 jtxb5! axb5 1 7 � xb5 �b8 1 8 �xd6 + jtxd6 19 � xd6 �xd6 20 K xd6 with advantage to White, Govedarica-Mokry, Trnava 1 987. ll jte2 This flexible move, which pre pares a possible h5, has gained in popularity, even though White sometimes has to sacrifice his f2 pawn after �d2 and 0-0-0. In reply the immediate �b6 turns out badly because h5-h6 becomes strong, so Black normally bides his time with a6. a6 ll Or: ( 1 ) l l . jt e7 1 2 �d2 � xd4 1 3 . .

�xd4 'itb6 1 4 JJ.. b5 + \t>f8 ( 1 4 . . . jtd7 1 5 J.. xd7 + * xd7 1 6 �d2 is good for White) 1 5 �xb6 axb6 1 6 0-0-0 e5 was unclear i n Ljuboje vic-Adorjan, Linares I 985, but it is hard to believe that there is no way White can exploit the weak b pawns. Perhaps 1 6 � a4! is best. (2) l l . �b6 1 2 � b3 a6 1 3 h5 �c7 (or I 3 . . . JJ.. d 7 1 4 h6 K h7 1 5 �d2 �g8 1 6 jte3 �c7 I 7 hxg7 .1. xg7 I 8 0-0-0 with a plus for White, Alzate-Frias, Dubai 01 1 986) 1 4 h6 �d7 I 5 hxg7 jt xg7 1 6 �d2 JJ.. f8 17 0-0-0 b5 1 8 a3 Ab 7 19 � h I K xh I 20 K xh I � ce5, Hellers-Sax, New York Open I 987, and now 2 I .1, h8! � g6 22 1!1 h7 would have given White a very dangerous attack. 12 �d2 JJ.. d7 Playing for b5 is a new idea, but the critical continuation is 'lt b6 1 3 probably he older 1 2 � b3 (25) and now: . .

( l ) 13 . . . 'ltc7 (this doesn't make much sense; Black may as well keep his queen on b6 and dare White to offer his f-pawn) 1 4 h5 � xh5 ( 1 4 . . . b 5 1 5 a 3 jt b7 I 6

30 Scheveningen Variation O-O-O b4 1 7 axb4 � xb4 1 8 tr d4 d 5 1 9 h6! !;l xh6 2 0 jt xh6 e5 2 1 .zt f4! won for White in Luthar-Bonsch, East German Ch. 1 989) 1 5 � h i g6 1 6 .ztxh5 gxh5 1 7 "i!i!(e2 b5 1 8 ». xh5 K xh5 1 9 trxh5 .zt b7 20 0-0-0 b4 2 1 � e2 � c8! 22 � d2 <2Je5 23 <2Jg3! aS 24 f4 with a clear plus for White, A. Rodriguez Douven, Amsterdam 1 987. (2) 1 3 . . . _zt d7 14 h5 <2J xh5 (Black should take everything on offer; the passive 14 0-0-0 1 5 h6 J� h7 1 6 0-0-0! !fJ_ e7 1 7 !fJ_ e3 trc7 1 8 !;l xg7 � xg7 1 9 hxg7 � g8 20 J� g l <2J e5 2 1 .ztd4 <2J g6 22 lith6 !fJ_ c6 23 !fJ_ d3 was very good for White in Korolev-Agzamov, USSR 1 983) 1 5 � h 1 g6 1 6 0-0-0 trxf2 (once again the crucial move; 1 6 trc7 1 7 !J;j_ xh5 gxh5 1 8 tre2! JJ.. e 7 1 9 JJ.. xe7 <2J xe7 20 � xh5 � xh5 2 1 trxh5 <2J g6 was played in Tseshkovsky-Mokry, Trnava 1 986, and now 22 trh2! was promising for White) 1 7 e5! trf5! ( 1 7 � xe5 18 <2J e4 trf5 1 9 tre3! JJ.. c6 20 <2J bd2! gives White a crushing attack) and the evalu ation of the whole line depends critically on this position. In A. Dieren Rodriguez--Grooten, 1 987, White played 18 exd6 trxg5 19 trxg5 JJ.. h 6 20 trxh6 � xh6 2 1 <2Jc5 <2Je5! 2 2 <2J 3e4 0-0-0 23 <2J xd7! *xd7 24 � h3 .)I e8, and although Rodriguez gives 25 b4! as unclear, in my opinion this line is not convincing. Therefore in Chandler-Hellers, Thessaloniki 01. 1 988 White tried 18 !J;j_xh5, but after 1 8 . . . � xh5 19 � xh5 gxh5

20 exd6 <2J e5 2 1 <2Jd4 trg4 22 .zt e7 <2Jc4 23 trd3 <2J e5 24 tre3 � c4 25 trd3 �e5 there was only a draw by repetition. It seems to me that White's best chance is to follow the Rodriguez-Grooten game, but instead of 21 � c5 play 2 1 <2J e4! white intends a combination of <2J f6 and <2J bc5, while after 2 1 0-0-0 22 � f6 � dh8 23 � xd7 *xd7 24 � c5 + * c8 25 _zt f3 White has dangerous threats. b5 l3 0-0-0 14 � xc6! The immediate 1 4 tre3 is met by 1 4 tr b6. JJ.. xc6 14 15 fte3 (26) 26 B

White has the unpleasant threats of 1 6 e5 and 1 6 <2Jd5, while after 1 5 tra5 White can afford to take time out for 1 6 * b l because 1 6 b 4 may be met by 1 7 <2Jd5 <2J xd5 1 8 exd5 jt xd5 1 9 � xd5 trxd5 20 JJ.. f3 . Black is therefore forced into the unpalatrc7. table 1 5 15 trc7 JJ.. xd5 16 <2J d5 e5 17 exd5

Scheveningen Variation 31 After 1 7 g c8 White replies 1 8 c3, and Black has achieved nothing positive, while giving away the chance to castle queen side. 1 7 . . . e5 is better, but even so White's lead in development and Black's exposed king give him dangerous attacking chances. 18 *bl?! Chess laziness. Of course * bI is a desirable move, but by giving Black a free tempo White's attack loses much of its momentum. The immediate 1 8 f4! was correct, when White has a clear advan tage. Now by accurate defence Black survives the immediate cri sis. .:2) h7 18 Eliminating the g5 bishop makes it easier to flee with the king, should that prove necessary, and ultimately the opposite-col oured bishops might provide a drawing mechanism. .:2\ xgS 19 f4 Not 1 9 f6 20 fxe5 dxe5 2 1 J.,h5 + 'fr'd8 2 2 d 6 J.,xd6 23 g xd6 + '!ifxd6 24 g d l '!ifxd l + 25 J., xd l fxg5 26 '!ifxe5 with an excellent position for White. 20 K xg5 A difficult choice, as although White may win a pawn by 20 fxe5 (20 hxg5 g6 21 f5!? is probably also slightly better for White) dxe5 2 1 !! xg5 0-0-0! 22 li xe5 j}_ d6, Black completes his de velopment and the opposite coloured bishops become an im portant factor. K c8 20

Now 20 . . . 0-0-0 is bad because of 2 1 a4!, so Black must adopt a different defensive plan. 21 c3 '!ifc5 Of course this is only possible when White had not exchanged on e5. Black gains time to reorga nize his defences. 22 '!ifg3 exf4 K c7! (27) 23 '!ifxf4 An excellent move. Black's rook covers the vulnerable square f7 and when it arrives at e7 the attack on the bishop will gain more time for Black.

24 a4! White's only chance to make something of his waning initiative is to create a new target on the queenside. 24 '1l e7 g6 25 j}_d3 axb5 26 axb5 Not 26 J., h6? 27 '!itf6. 27 '!it d4 The ending now represents the best winning chance for Whit�. Although White's gain of a pawn is only temporary, the passed b pawn combined with an exposed

32 Scheveningen Variation Black king gives White a nagging advantage. 'lt xd4 27 ftd8 28 Jl_ xb5 + !;l xd4 29 K e1 + 30 ft a2 J.. e7 l;l h1 � gg4 31 The h-pawn is doomed, so White switches to harassing Black's king. In this the opposite coloured bishops prove a big help. .1, a4 32 !:[ 8xh4 i: xh4 Ill xh4 33 33 . . . J.. xh4 34 H a7 is worse, since White threatens the f-pawn directly and the d-pawn indirectly via .ii! d7 + � a8 + ? 34 A careless check which drives the Black king to a better square. White should have cut the king off by 34 !:[ a7 ! (threat � d7 + ) g5 35 l:l d7 + � e4 (after 34 '* e8 36 i: xd6 + ft f8 37 i: c6 the d-pawn becomes dangerous) 35 jtc6! with some winning chances because Black cannot challenge White to a pawn race (35 . . . g5 36 b4 g4 37 b5 g3 38 b6 wins because the mate threat gains a tempo). 34 '*c7 35 J.. c6 35 .l, a7 + fl b6 36 H xe7 fj xb5 37 )ll xf7 '\tr'c5 is an easy draw. 35 .l, e4! White is effectively a tempo down over the above line since after 36 l!l( a7 + '* b8 White must waste time with his rook. This tempo makes all the difference and Black can now draw comfor tably.

36 37 38 39 40 41 42 43 44

b4 H a7 + H xf7 b5 ff a3 l! b7 + b6 K c7 + !:[ b7 + Draw

jtf6 .b8 j_xc3 H b4 .ill b1 fjc8 j_d4 .b8

Game 8 Karpov-Dorfman USSR Ch. 1976 1 e4 c5 e6 2 � f3 3 d4 cxd4 4 � xd4 �f6 5 �c3 d6 6 g4 � c6 (28) I must confess to have taken some liberties with the move ord er of Karpov-Dorfman, which actually continued 6 . . . Jt e7. We transpose back in a few moves, but this move order makes it easier to explain the proposed repertoire. Of the sixth move alternatives, we only consider 6 e5 here; the others may be e5 found in game 9. After 6 . (played by Murei and Suba) 7 j_b5 + j_d7 8 j_ xd7 + '/txd7 (8 � bxd7 9 � f5 is awful for Black) 9 � f5 h5 10 f3 hxg4 1 1 fxg4 K h3!? 1 2 'lte2 (better than 1 2 �e3 'ltc6 1 3 '£) ed5 � bd7 1 4 London Howeii-Suba, �e2, � xd5 1 5 1 988, and now 1 4 � xd5 � b6 is unclear) 'ltc6 ( 1 2 � xc3?! 1 3 bxc3 'ltc6 1 4 0-0 '*xe4 1 5 �xe4 <2l xe4 1 6 &J[ e 1 • . .

Scheveningen Variation 33 L£J c5 1 7 Jl. a3 '£l ba6 1 8 I: ad I Jf!i d8 19 h4 g6 20 '£J e3 is good for White, Wahls-Mainka, Luxem bourg 1 989) 1 3 0-0 '£l bd7 14 g5 �xc3, Tisdaii-Suba, Preston 1 989, and now 1 5 '£J xg7 + Jl. xg7 16 gxf6 �xc2 ( 1 6 '£l xf6 I 7 bxc3 tyxe4 1 8 tyd3 d5 1 9 tyg3 is promising) 1 7 fxg7 *e7 ( 1 7 0-0-0 1 8 trg4) 1 8 Jl. g5 + ! f6 1 9 tyf3 i s very good for White.

'£J d7 g5 h4 a6 Or 8 '£J xd4 9 tyxd4 L£Je5 1 0 Jl.e2 '£Jc6 1 1 �d3 Jl. e7 (or 1 1 a6 1 2 Jt f4 trc7 1 3 0-0-0 '£J e5 14 '*d4 Jl. d7 15 h5 with strong pres sure for White, Lutikov-Malich, Leipzig 1 977) 1 2 Jtf4 (were it not for this move, exposing the weak ness of d6, Black's scheme would be viable-this is one of the few situations in the Sicilian where a direct attack on d6 works) 0-0 1 3 0-0-0 e5 1 4 Jl. e3 Jl. e6 1 5 <21 d5 ty a5 16 a3 Jtxd5 1 7 tyxd5 tyxd5 1 8 1:: xd5 with the type of ending Sicilian players have nightmares about, Nunn-Jansa, Dortmund 1 979. 9 Jte3 'f!l c7 7 8

The repertoire proposed in this book involves meeting lines with out . . . tyc7 by the new idea Jl. c4 (see next game), while if Black plays . . . "W!fc7 preventing Jtc4 then White replies with tre2 and 0-0-0. Unfortunately White can not guarantee to play Jl. c4 in every line, because if White plays Jl.c4 without preparing it by Jl.e3 then the reply <2! de5 is very awkward. 10 tre2 Karpov's move is very logical in that it prepares queenside cas tling as quickly as possible, while the fl bishop and h 1 rook are left at home since it is not yet clear which square is best for these pieces. At e2 the queen sets up tactical chances down the e-file and avoids attack from a !Jiack knight arriving at c4. 10 Jl.e7 We are now back in Karpov Dorfman. b5 1 1 0-0-0 Tactical ideas for White are already in the air, for example 1 2 '£J f5 exf5 1 3 '£Jd5 trd8 1 4 exf5, but although this is quite good for White Black can improve by 1 2 '£Jf5 b4! tr xc6 12 '£l xc6 13 Jl. d4 b4 (29) Black forces White to sacrifice on d5, but this move was itself virtually forced as 1 3 0-0 1 4 K g I gives White a crushing at tack, e.g. 1 4 b4 1 5 <21d5 exd5 1 6 exd5 'f!l xd5 1 7 trxe7 'f!lxa2 1 8 g6 hxg6 1 9 �xg6 and wins.

34 Scheveningen Variation

exd5 14 � d5 � g8 15 j_ xg7 'W/c7 16 exd5 17 J.. f6 The position of White's bishop at fl is shown up as a defect since 1 7 !! e l l£l e5 1 8 j_ xe5 dxe5 19 f4 exf4 achieves nothing when d6 is impossible. If the other rook could come to e l Black would be finished. '£) e5 17 dxe5 18 J.. xe5 19 f4 Now White wins a third pawn for the piece since 1 9 . . . e4 fails to 20 d6 j_ xd6 2 1 'Wfxe4 + Black's king must remain stuck in the centre so one must consider Kar pov's sacrifice correct, although in the subsequent play Black's resourceful defence almost saves the game. J.f5 19 20 J.h3 White takes time out to neutral ize Black's counterplay as 20 fxe5 at once allows the unclear 20 � c8 21 � h2 'l!lfa5. J.xh3 20

21 li:i( xh3 ![ c8 22 fxe5 After this Black activates his queen and Karpov is obliged to play with extreme accuracy to maintain his advantage. In his notes Karpov suggested 22 b3 to prevent the following manoeuvre. 22 �c4! 23 � dd3 .f4 + 23 . �xa2 24 d6 (threat d7 + ) g c4 (24 g c5 25 � f2 and 26 dxe7) 25 dxe7 � a l + 26 * d2 �xb2 27 g d8 + fij xe7 28 � d7 + fj>xd7 29 �xc4 and 23 g xg5 24 hxg5 'itxa2 25 d6 j_ xg5 + 26 lii,\ he3 !i( c4 27 'itg2 are good for White. � c4! 24 *b1 The rook follows the queen's path with the aim of causing White some problems on the back rank. � e4 25 d6 26 � he3 Ji! xe3 �xh4 (30) 27 � xe3 If Black attempts to save his A f8 bishop by 27 J.d8 (27 2 8 'l!lfxa6 i s even worse) h e is crushed after 28 Ji! f3 'it g4 29 e6

Scheveningen Variation 35 fxe6 30 d7 + winning the queen, so he quite rightly decides to grab as many pawns as he can while White is taking his bishop. 28 trf3! Naturally not 28 dxe7 at once since Black exchanges queens by 28 tr h 1 + White's advantage lies in the insecure black king, which causes trouble even when Black restores material equality. 28 'ltxg5 28 jl xg5 29 e6 fxe6 30 � xe6 + fJ' d8 (30 . . . fJ'd7 3 1 'lt f7 + ) 3 1 'ltc6! and 28 � xg5 29 'ltc6 + fJ'f8 30 dxe7 + fg xe7 31 a3 win for White. 29 K el 29 'ltc6 + fJ'f8 30 dxe7 + trxe7 3 1 'lth6 + � g7 is a little better for White and this may in fact be his best line. 'ltg2? 29 29 . . . 'lt g4 was better, when it is far from clear if White can do more than draw. K g6 30 trf5 'ltd5 31 K fl fj' xe7 32 dxe7 M aterial equality is re-estab lished but Black's king position makes his defensive task difficult. Detailed analysis of this position would take us too far afield, but Black does not seem to have any real improvements hereafter and the task of defending both his king and his pawns soon over stretches his forces. aS 33 'ltf4 fj'e8 34 '!th4 + 'ltf3 35 'ltxh7

36 'lth8 + * e7 37 �h4 + fj'e8 38 �c4 'ltb7 39 b3 One of the most impressive features of this game is the way Karpov managed to conduct all the tactical operations with a vul nerable black rank. Many players, through nervousness or laziness, would have wasted a tempo on b3 earlier, and this might well have cost the game (note that although b3 was a good idea at move 22, the point was to prevent the 'ltc4-f4 manoeuvre rather than to give the king air). � e6 (31) 39 31 w

40 K gl?! Perhaps Karpov assumed that the exposed king must succumb quickly to the combined attack of White's queen and rook, but the task is much more difficult than appears at first. I suspect that if Karpov had realized this he would not have been so hasty in giving back the pawn, because he could have waited for a more favourable moment.

36 Scheveningen Variation 40 41 42 43 44 45

J� xe5 * e7 K g8 + trh4 + *d 7 tyf6! H e7 *d6 ty f5 + trxa5 H e5 45 tye4 would have lasted longer, but the result is not in doubt. 'fle6 46 tyd8 + f6 47 * b2! trg7 � f8 48 'fld5 49 trc8 + Resigns 50 tyc4 +

Game 9 Nunn-Thorsteins Lugano 1986 1 2 3 4 5 6

e4 � f3 d4 � xd4 �c3 g4 (32)

c5 e6 cxd4 �f6 d6

32 B

6 jLe7 We also need to consider those lines which are special to 6 . . a6. After 6 . . . a6 7 g5 <E)fd7 8 h4 b5 (8 �c6 9 jLe3 will lead to game 8 or to the main line of this game) 9

a3 A b7 (9 � b6 is less accurate since after I 0 h5 White may meet 10 jLe7 with the dangerous piece sacrifice 1 1 trg4 e5 1 2 .:2! f5 g6 1 3 hxg6 fxg6 14 jLe3! gxf5 1 5 exf5; in Nunn-Walden, Notting ham 1 983, the continuation 1 0 . � 8d7 1 1 � h3 �c5 1 2 g6 f6 1 3 � g3 e5 1 4 � c6 trc7 1 5 � b4 was good for White) 10 Ae3 ( 1 0 h5 at once is not so good since 1 0 §Le7 awkwardly attacks the g5 pawn) � b6 (10 . . . �c6 1 1 tr e2 �de5 1 2 0-0-0 � c4 1 3 <E) xc6 jLxc6 14 f4 �a5?! 1 5 �d5! gave White a crushing attack in Alex ander-Lundholm, corr. 1 970-- 1 , while 10 . . . �c5 1 1 tyg4!? � bd7 12 0-0-0 � e5 1 3 tyg2 � c4 1 4 A xc4 bxc4 1 5 h 5 intending g6 is good for White according to Boudy) 1 1 h5 � d7 1 2 � h3! (the discovery of this move led to a reassessment of many lines in the Keres Attack-the point is to defend the rook in preparation for g6) Black may play: ( I ) 12 . . . � e5 1 3 g6 hxg6 1 4 hxg6 K xh3 1 5 gxf7 + � xf7 (Black avoids the displacement of his king) 1 6 A xh3 and Black's weak pawn at e6 gives White a good game, Torre-Vogt, Polanica Zdroj 1 977. (2) 12 . . . d5 1 3 g6 e5 reaches a complex position. In two games White played 14 � f5 and after 14 . . . hxg6 15 hxg6 � xh3 16 gxf7 + *xf7 1 7 A xh3 .:2l f6 1 8 A xb6 �xb6 1 9 �xd5 A xd5 20 exd5 � d8 2 1 � e3 A c5 22 A e6 + Yakovic-Espig, Leipzig 1 986 and ,

Scheveningen Variation 3 7 14 . . . d4 1 5 h6 hxg6 1 6 hxg7 JJ.. xg7 1 7 <2\ xg7 + 'ffj e7 1 8 JJ.. xd4! exd4 19 'it xd4 K xh3 20 0-0-0! , Fogarasi-Espig, Budapest Open 1 987, White had a clear advan tage. However, as Lukacs and Hazai point out, the improvement 14 . . . <2\ f6! leads to a totally unclear position. I therefore suggest 14 exd5!? exd4 ( 1 4 . . . <2\ xd5 1 5 <2\ xd5 JJ.. xd5 1 6 <2\ xb5) 1 5 JJ.. xd4 with an enormous at tack, e.g. 1 5 . . <2\ xd5 1 6 <2\ xd5 JJ.. x d5 17 )J:! e3 + JJ.. e6 18 'ltf3! winning. (3) 12 . . . JJ.. e7 1 3 g6 JJ.. f6 14 tyg4 'lte7 15 gxf7 + flxf7 16 f4 � hc8 ( 1 6 � ac8 1 7 f5 e5 1 8 <2\ e6 � hg8 1 9 h6! g6 20 � g3 is also good for White) 1 7 f5 <2\e5 1 8 fxe6 + fi f8 1 9 t)'g2 � xc3! 20 bxc3 <2\ a4 2 1 lli, g3 ! <2\ xc3 22 il_ d3 'Wfc7 23 e7 + ! turned out well for White in Fernandez Garcia-D. Cramling, Barcelona 1 986. (4) 12 . . . <2\c5 3 g6 'Wfe7 1 4 gxf7 + 'Wfxf7 1 5 'W/g4 JJ.. c 8 1 6 e5! dxe5 17 � f3 'ltb7 18 <2\ de2! i#fc7 19 0-0-0 JJ.. d 7 20 JJ.. g5 aS? 21 <2\g3! with a massive advantage for White, ldelstein-Barash, Israel 1 987. 7 g5 <2\ fd7 <2\c6 8 h4 9 JJ.. e3 0-0 If Black plays 9 a6 White may again reply 10 il_ c4, with play similar to that in the main line. Black has one more possi bility, namely 9 <2\ b6 10 f4 and now: ( I ) 10 . . . 0-0 1 1 )!lJtf3 d5 1 2

0-0-0 JJ.. d7?! 1 3 exd5 exd5 1 4 JJ.. d 3! JJ.. c5 1 5 <2\ xc6 JJ.. xe3 + 1 6 trxe3 bxc6 1 7 h 5 � e8 1 8 trd4 'Wfe7 1 9 h6 life3 + 20 ty xe3 � xe3 21 hxg7 with a decisive advantage for White, Lanka-Strautinsh, corr. 1 986. (2) 10 . . d5 1 1 JJ.. b 5 ( 1 1 e5!? is possible) JJ.. d7 12 exd5 exd5 1 3 'W! f3 JJ.. b4 1 4 0-0 JJ.. xc3 1 5 bxc3 <2\ xd4 1 6 JJ.. xd7 + 'ltxd7 1 7 JJ.. xd4 0-0 1 8 f5 <2\c4, Tseitlin-Lukin, Leningrad 1 987, and now 1 9 � fl! intending h5--h6 should be good for White. 10 JJ.. c4!? (33) .

An intriguing new idea. White plans a similar strategy to that in the Velimirovic Attack, which arises after 1 e4 c5 2 � f3 <2\c6 3 d4 cxd4 4 <2\ xd4 <2\ f6 5 Qlc3 Q�c6 6 JJ.. c4 e6 7 JJ.. e3 JJ.. e 7 8 'i/!i'e2 0-0 9 0-0-0, one typical variation being 9 . . . a6 1 0 il_ b3 fltc7 1 1 g4 <2\ xd4 1 2 Ji;!t xd4 b5 1 3 g5 <2\d7 14 h4, with an obvious similarity to Nunn-Thorsteins. One major dif ference stands out. In the Velimir ovic Attack White cannot play g4

38 Scheveningen Variation under the most favourable pos sible circumstances; he must either waste time preparing g4 by l l .1, hg I , or he must reconcile himself to the relatively unfavour able recapture with the rook on d4 (the reason being that l l g4 <2} xd4 1 2 J_ xd4 fails to 1 2 . e5). In the Keres Attack position White has already achieved g4-g5 without playing K hg 1 , and there is no reason why he should not take on d4 with his bishop. There fore if Black plays just as in the Ve1imirovic Attack, he will end up with an unfavourable version of that system, and White's kingside breakthrough by h5 and g6 will have added force. On the other hand White is two moves away from castling, so the question is whether Black can do something quickly while White's king is still in the centre. It is worth adding that 1 0 'ii!i' e2 is perfectly playable here, just as in Karpov-Dorfman above, but having played 0-0 instead of �c7 improves Black's prospects. <2} xd4 10 Perhaps the most natural at tempt to exploit White's king position is to open the centre by 10 <2} b6 1 1 J_ b3 d5, but after 1 2 �e2 it is not clear what Black' s next move is. 1 2 . . . e5 loses a pawn and 12 . . . <2} xd4 1 3 J_ xd4 dxe4 14 0-0-0 followed by <2} xe4 prepares a sacrifice on f6. Maybe 12 . . . jtb4!? is best, with an unclear position. 1 1 �xd4?!

Inconsistent. White should stick to his idea and play 1 1 J_xd4. I rejected this because of 11 d5 12 exd5 exd5 13 <2} xd5 .!!K e8, but a little more thought would have shown that after 1 4 <2}e3 ! Black has no real compensa tion for the pawn. Indeed, after 1 4 <2} e5) �c7 (intending White can go over to the attack himself by 1 5 J_xf7 + * xf7 1 6 �h5 + \fl f8 1 7 J_xg7 + * xg7 1 8 <2} f5 + with a large advantage. 11 a6?! Black misses his chance. 1 1 <2}e5 would have forced White's bishop to abandon the active b3g8 diagonal and after 1 2 J_ e2 Qlc6 1 3 �d2 a6 the position would have been roughly level. 12 0-0-0 Now White can set up the type of position he is aiming for. 12 b5 13 J_ b3 <2} c5 14 f4 Defending g5 m preparation for h5-h6. �aS 14 Black intends to meet the ad vance of the h-pawn by e5, relying on a tactical point to se cure the e5 square, but unfortu nately White has a tactical counterpoint! 15 h5 b4 e5 (34) 16 h6 White cannot take twice on e5 because of the discovered attack along the 5th rank, but by means of a queen sacrifice White can convert his lead in development

Scheveningen Variation 39

into a crushing attack against Black's king. 17 � d5! � xb3 + Black must remove one attack ing piece. The immediate capture of the queen leads to mate after 1 7 exd4 1 8 <i) xe7 + \tr h8 1 9 hxg7 + lti xg7 20 J.. xd4 + f6 2 1 gxf6 + J1l; xf6 2 2 H dg l + \ft f8 23 Jiit g8 + ltixe7 24 H xh7 + H f7 25 a xf7. 18 axb3 j}_ xg5 Black decides to decline the offered queen. After 1 8 . . . exd4 1 9 <E::J xe7 + lti h8 20 j_xd4 f6 (20 � g8 2 1 hxg7 + H xg7 22 Ii!, xh7 + mates) 21 g6! Black is strangely helpless against the threat of 22 hxg7 + lti xg7 23 Ii!, xh7 mate, for example 2 1 H g 8 (21 gxh6 2 2 K xh6, 21 hxg6 22 <E::J xg6 + lti g8 23 h7 + ltif7 24 <E::J x f8 and 21 >�tal + 22 ltid2 'tltxd l + 23 lti xd l j}_g4 + 24 ltid2 hxg6 25 <E::J xg6 + \t'h7 26 <E::J xf8 + � xf8 27 hxg7 + 'iJtjxg7 28 . . •

. . .

K g l are no better) 22 hxg7 + f;j xg7 (22 K xg7 23 :;: xh7 + K xh7 24 j}_xf6 + K g7 25 J;� h l + mates) 23 K xh7 + '1Jf8 24 X f7 + 'll e 8 25 <E::J xg8 j}_ e6 26 <E::J xf6 + 'll d8 27 .I f8 + and in return for the queen White wins almost all Black's pieces. 19 fxe5 White must not repeat the offer because after 1 9 fxg5 exd4 20 <E::J e7 + \t'h8 he cannot play 2 1 J.. xd4 due to 2 1 >�txg5 + . 19 J.. xe3 + After 1 9 dxe5 20 >�t xe5 J.. xe3 + 21 fJ b l White will at the very least win Black's queen. 20 >�txe3 g6 Or else hxg7 wins instantly. f6 21 >�tg5 Black must jettison material to meet the threats of 22 �e7 + and 22 >�t f6. 22 <E::J e7 + Even stronger than taking on f6. After 22 'll h8 23 � xg6 + hxg6 24 >�txg6 K a7 (or 24 H g8 25 >�t xf6 + \t'h7 26 >�tf7 + '1J h8 27 h7 >�t a l + 28 'll d 2 ll. g2 + 29 lti e3) 25 exf6 there is no defence to the threat of f!/g7 + . Black has only one alter native, but then it is White who can exploit a line-up on the 5th rank. 22 * ti Resigns 23 e6 +

Classical Variation

In the first edition of this book I . e6 7 tl!l'd2 a6 8 0-0-0 jtd7 is the christened the line 1 e4 c5 2 � f3 subject of game 1 1 while 6 . . . e6 7 � c6 3 d4 cxd4 4 � xd4 � f6 5 �d2 a6 0-0-0 h6 appears in game �c3 d6, which can also occur 1 2 . Finally some players have with the move order 2 . . . d6 and 5 experimented with the omission e6, not fearing the dou . . . �c6, with the name 'Classical of 6 Variation' This nomenclature bled pawns resulting from jlxf6, seems to have caught on, so I will and the most popular of these jld7, forms the basis keep the name in this edition. The ideas, 6 line I am recommending against of game 1 3 . Unusual lines involv e6 are dealt with in game the Classical is 6 jlg5, called the ing Richter-Rauzer Attack even 1 0, while the others are in game though the treatment used today 1 3 . doesn't st!em to owe anything to Richter. This line is very common Game 10 in practice, so there is a large body Anand-lnkiov of theory. In general I will keep to Calcutta 1986 the main lines in the proposed 1 e4 c5 repertoire, but where there are interesting sidelines I will give 2 �f3 d6 3 d4 them a brief mention. cxd4 4 �xd4 The idea of 6 jlg5 as it is � f6 5 �c3 �c6 played today is based on a quick e6 6 .zlg5 '{lfd2 and 0-0-0, exerting pressure Other moves are considered in down the d-file and restraining Black from active play in the game 1 3 . 7 �d2 (35) J.. e7 centre. Black's most solid reply is 7 a6 appears in games 1 1 the natural 6 e6 7 ttd2 jle7 8 0-0-0 0-0, but despite its solid and 1 2. There are two other alter appearance it can often lead to natives: sharp tactical play. This is (I) 7 h6 (7 j_ d7? 8 � db5 covered in game 1 0. Sometimes is just a mistake) 8 j_ xf6 gxf6 (8 Black players postpone 0-0 so �xf6 9 � db5 and 1 0 0-0-0 as to delay exposing the king to a wins the d6 pawn) 9 0-0-0 (playing possible pawn storm. The line 6 for 0-0 is also slightly better for . . .

Classical Variation 41 White) 1 1 J.. e 2! J.. c6 ( 1 1 Y!fc5 is still bad, this time because after 1 2 Y!fxc5 dxc5 1 3 e5 �d5 1 4 � xd5 exd5 1 5 J.. f3 j_ c6 1 6 j_ a5 ! White prevents . . .1. d8) 1 2 f3 Y!fc7 1 3 g4 j_e7 1 4 g5 <E}h5 1 5 Y!fg 1 ! (not 1 5 f4 at once because of 1 5 . . . e5) and White's threat of f4 proved very hard to meet, Tal Sosonko, Wijk aan Zee 1 982. 0-0 8 0-0-0 Or 8 � xd4 9 Y!fxd4 0-0 1 0 e 5 dxe5 1 1 Y!fxe5 (36) and now: .

White, e.g. 9 .1. d 1 a6 1 0 j}_ e2 h5 1 1 0-0 J.. d 7 1 2 � b3 Y!fc7 1 3 f; h 1 0-0-0 1 4 f4, or 9 J.. e 2 h5 1 0 0-0 a6 1 1 f; h l ! J.. d 7 1 2 f4 Y!fc7 1 3 .l. f3 ! � xd4 1 4 Y!fxd4 j_e7 1 5 I;� d 1 h4 1 6 K fd3 K d8, van der Wie1-J. Piket, Leiden 1 986, and now 1 7 J.. f3 gives White an edge) a6 1 0 f4 j_d7 1 1 j}_e2 h5 ( 1 1 Y!tb6 1 2 j_h5 Y!fxd4 1 3 Y!fxd4 � xd4 1 4 .l, xd4 .H g8 1 5 g3 j_e7 1 6 .i. fl j_c6 1 7 f5 .A g5 1 8 j_ e2 with ad vantage, Bondarevsky-Botvin nik, USSR Ch. 1 95 1 ) 1 2 f; b 1 Y!fc7 (or 1 2 . . . Y!lb6 1 3 � b3 0-0-0 1 4 K hfl �a5 1 5 B. f3 � xb3 1 6 axb3 f; b8 1 7 �a4 Y!fa7 1 8 f5 with a clear plus for White, Keres Botvinnik, Moscow 1 956) 1 3 .A hfl 0-0-0 ( 1 3 J.. e 7 1 4 K f3 � xd4 1 5 Y!fxd4 Y!fc5 1 6 Y!td2 j_c6 1 7 � e3 Y!fa5 1 8 a3 H d8 1 9 j_c4 gives White an edge, Liberzon Botvinnik, USSR 1 967) 1 4 � b3 f; b8 1 5 .1. 0 J.. e 7 16 J,! h3 h4 1 7 Y!f e 1 and Black has not yet equa lized, Vasyukov-Shamkovich, Dubna 1 973. (2) 7 � xd4 8 Y!fxd4 j_d7 9 0-0-0Y!fa5 1 0 J.. d2 a6 ( 1 0 . . . Y!fc5 1 1 Y!fxc5 dxc5 12 � b5 is good for . . .

36 B

(1) 11 � d7 is bad after 1 2 J.. x e7 Y!fxe7 1 3 Y!fc7! (2) 11 Y!fb6 12 j}_e3 � g4 gives White a favourable ending after 1 3 J.. xb6 <E} xe5 1 4 j_c7 � g4 1 5 � g3 <2:l f6 1 6 j_ b5 ! a6 1 7 j}_e2 b5 1 8 J.. f3 K a7 1 9 j_d6, as in Vasyukov-Boleslavsky, USSR 1 957. (3) 11 J.. d7 12 h4 � c8 ( 1 2 Y!fe8 transposes t o line 4) 1 3 lil h 3 (White's rook can come to d3 or g3) Y!fc7 ( 1 3 . l;;!!, c5 14 �e3 �c8 1 5 M g3 \t'h8 1 6 \t' b 1 fol lowed by the push of the h-pawn gives White a dangerous attack) 1 4 �xc7 � xc7 1 5 � b5 J.. x b5 1 6 . . .

. . .

42 Classical Variation j_xb5 and White's two bishops is to sacrifice the d-pawn, for ex give him an edge in this ending. ample by 9 h6 1 0 jlxf6 j_xf6, (4) 11 iJ}te8 1 2 h4 j_d7 ( 1 2 but a number of games prove that a6?! 1 3 <2:l e4 <2:l d5 1 4 J.. xe7 White can gain an advantage. A <2:l xe7 1 5 h5 <2:1 c6 1 6 iJ}tg3 iJ}te7 1 7 second plan is to allow jt xf6 and h6 g6, Hubner-Timman, Belfort then recapture with the pawn. 1 988, and now 1 8 j_ c4 is very This leaves Black's king some good for White) 1 3 � h3 j_c6 1 4 what exposed, but White has no � g3 itb8 1 5 llte3 H d8, Tal immediate method of launching Timman, Brussels SWIFT 1 988, an attack. More serious is that and now 16 J.. d 3 ! j_ d6 1 7 J.. xf6 Black's central pawns are inflex j_xg3 1 8 <2:le2! gxf6 1 9 <2:l xg3 * f8 ible and White may have enough 20 ith6 + *e7 2 1 � e l � d5 22 time to start a kingside pawn <2:l f5 + Ji xf5 23 J.. xf5 gives White storm. The assessment of this line the advantage according to Tal. depends on the speed of the re The position after 8 0-0 is spective attacks. Black's third one of the most important in the plan is to counter White's pres whole Sicilian Defence and de sure on d6 directly by playing 9 a5 I 0 a4 d5 (moreover the spite decades of practical ex d5 is just about perience no definite assessment immediate 9 can be given. Although 9 <2:1 b3 possible). The final plan is to iJ}tb6, was often played in the late fifties counterattack f2 by 9 and early sixties, it fell into disuse gaining enough time to defend d6 � d8. Apart from the first and 9 f4, which has always been by regarded as the main line, became plan all these lines are playable. virtually universal. However 9 <2:1 b3 has been regaining popular ity and now rivals 9 f4 for the distinction of being considered the 'main line'. 9 <2:l b3 (37) Notice that 9 j_xf6? is bad since Black can play 9 jlxf6 1 0 <21 xc6 bxc6 1 1 iJ}txd6 itb6 when the threats to b2 and f2 are more than enough compensation for the pawn. The main point of 9 <2:1 b3 is that 9 a5 it unveils an attack against the d6 Apart from the two major alter pawn and so prepares j_xf6. natives of 9 a6 and 9 it b6, Black has four main methods of there are a number of less com countering White's plan. The first mon ideas: . . .

Classical Variation 43 ( 1 ) 9 . . . dS 1 0 J.. xf6 J_xf6 1 1 exd5 jlxc3 ( 1 1 � b4 1 2 a3 /2l xd5 1 3 � xd5 exd5 14 �xd5 is slightly better for White since although Black has the two bis hops it is not easy for him to avoid the exchange of queens) 1 2 �xc3 exd5 and now: ( l a) 13 g3 j_g4 1 4 � d2 �e7 1 5 A g2 lr;! fe8 1 6 'iti' b l � ac8 1 7 �c5! 121 b4? ( 17 J.. f5 is better) 1 8 �xe7 � xe7 1 9 � c l with the superior ending for White, Psak his-Aseev, Sevastopol 1 986. (1 b) 13 � d4 should give White a small but safe advantage. ( l e) 13 jibS!? �g5 + and now both 14 �d2 ytxg2 1 5 !i[ hg l and 1 4 f} b l d4 1 5 'i!lfc5 ytxg2 give White an attack in return for the pawn. (2) 9 . . � aS 10 fJ b l � xb3 1 1 cxb3 a6 1 2 f4 b5 1 3 A xf6 gxf6 1 4 j_ d 3 \9i h8 1 5 f5 b4 1 6 <E:� e2 e 5 1 7 jl c4, Anand-Mateo, Dubai 01. 1 986, is worth mentioning because it is a perfect example of what Black should avoid. His king position has been weakened without any compensating queen side attack and Black has played e5 at a moment when White can reply jl c4 to gain control of d5. (3) 9 . . h6 (it now seems estab lished that this line is good for White) 10 Axf6 J.. xf6 1 1 �xd6 jl xc3 ( 1 1 � b6 12 �c5 �c7 1 3 g 3 JJ... e 7 1 4 yte3 a6 1 5 f4 b 5 1 6 lJ... g2 � e5 1 7 yte2 was also good for White in Marjanovic-Barlov, Yugoslav Ch. 1 985) 1 2 bxc3 �h4 .

1 3 g3 �f6 (not 1 3 -�xe4? 1 4 jld3 and j_h7 + ) 1 4 ·�c5 e 5 1 5 JJ... c4 jle6 (or 1 5 J_g4 1 6 hl[ d6 �g5 + 1 7 \fi b2 J.. f3 1 8 � e l � ac8 1 9 h4 �g4 20 �e3 � fd8 2 1 � xd8 + .l! xd8 22 _!d5 JJ... g2 23 !! g 1 �f3 24 g4 with advantage to White, Lobron-Kunsztowicz, Bad Neuenahr 1 984) 1 6 JL xe6 �xe6 1 7 � d6 �h3 1 8 '/l!(e3 � fd8 19 � d5 and now both 19 . .I deS 20 � hd 1 � c7 2 1 f4 exf4 22 '!Wxf4 �e7 23 <E:�c5, Chandler Torre, London 1 984 and 19 . . . �g2 20 � hd l �xh2 2 i �c5, Klovan-Tal, Jurmala 1 983, were good for White. We now move on the major lines: (4) 9 . . . a6 10 J.. xf6 and now: (4a) 10 . . . JJ... xf6 (this is dubious) 1 1 �xd6 Ji_ xc3 (or 1 1 ·� b6 1 2 f4 J.. e 7 1 3 'i'l!{d2 a5 1 4 a4 � d8 1 5 i/..d3 with advantage to White, Shaposhnikov-Boles lavsky, USSR 1 950; as usual 1 2 j_ xc3 1 3 bxc3 �e3 + 1 4 ff b2 �xe4 loses to 1 5 Ji..d3 and 1 6 Ji.. x h7 + ) 1 2 bxc3 (the position is the same as after 9 h6, except that Black has played a6 in stead; the verdict is unchanged) � f6 ( 1 2 �h4 1 3 g3 �[6 1 4 � c 5 e5 1 5 JJ...c4 �.g4 1 6 g d6 �g5 + 1 7 f4 and White was clearly better in Ivanovic-Popo vic, Novi Sad 1 984) 1 3 i'l!i'g3 e5 1 4 �c4 j_ e6 1 5 jlxe6 �xe6 1 6 g d5 and again White had won the opening battle, Benjamin-Chris tiansen, USA 1 984. (4b) 10 . . . gxf6 1 1 �h6 . .

44 Classical Variation (although this line had not been (4b2) 19 b3!? 20 axb3 (20 played very often, in my opinion cxb3 axb3 2 1 a3 is also possible, it offers White the best chances; when 2 1 <2:l xd4 22 <2:l xd4 'lta4 the plan is a general kingside is unclear) axb3 21 <2:lxb3 ji_b7 22 pawn advance by g4, f4, h4 and <E)c3 � b4 23 � hg l 'ltc6 24 � g3 g5) * h8 1 2 � h5 (Black is now � b6 25 'lte2 d5 was played in virtually forced to lose time with Psakhis-Kotronias, Dortmund his queen because he must free the 1 989, and now the best move is 26 f8 rook to defend h7 by ii;! g8- 'lte3!, when Kotronias gives 26 g7) �e8 ( 1 2 � g8?, Ernst <2:l xd3 + 27 H xd3 'i'l!l'xe3 + 28 Chandler, London 1 988, and now � dxe3 ji_ d6 29 e5 fxe5 30 fxe5 l 3 'l!lf xf7! '!i!, g6 1 4 f4 j_d7 1 5 � d3 J.. e7 intending h6 as unclear. e5 1 6 ii;! g3 ! j_e8 1 7 'i'�!!' d 5 exf4 1 8 However after 3 1 <2:l b5! I doubt if Black has enough for the pawn, Jilt xg6 hxg6 1 9 'iii!' d 2 g 5 2 0 <2:l d4 h6 32 gxh6 � xg3 33 �xd4 2 1 �xd4 is good for White) e.g. 3 1 '14. xg3 ji_ xh4 34 K g7 intending 1 3 f4 b5 (in view of the note to White's 1 5th move it might be � d6. (5) 9 �b6 (the counterat more accurate to play 1 3 . . . ii g8) 14 jj_d3 li( g8 1 5 g4 (Tisdall points tack against f2 nullifies White's out the possibility of 1 5 �d5, threat to take on f6, so Black gets time to support his d-pawn by based on the tactical point 1 5 H d8; White's usual reaction has exd5 1 6 exd5 '!i!, g7 1 7 dxc6 J.. g4 1 8 �d5 .Rxd l 1 9 ,ii;i xd l with fan been to start a kingside pawn tastic compensation for White's storm, but he must be careful slight material deficit; of course because too many pawn moves Black should play 1 5 J.. b 7 1 6 might encourage Black to open up d5) 10 f3 (38) <2:l xe7 �xe7 and i t i s far from the centre by certain that the exchange of minor and now: pieces favours White, especially as he has spent two tempi achieving it) '14. g7 ( 1 5 b4 1 6 e5 clears e4 for the knight with gain of tempo) 1 6 h4 b4 1 7 <2:le2 a5 1 8 g5 a4 1 9 <2:l bd4 and now: (4b l ) 19 <2:lxd4 20 <2:l xd4 j_d7 2 1 gxf6 ji_xf6 22 e5 dxe5 23 � hg l �g8, Arnason-Inkiov, Plovdiv 1 986, and now Inkiov gives 24 <2:l f3 ! ji_ e8 (24 . . . exf4 25 � xg7 J.. xg7 26 <2:lg5 wins) 25 (5a) 10 a6 1 1 h4! (this is the j_ e4 (25 fxe5 looks even better to move which has put 1 0 a6 out me) as good for White. of favour) .f;l d8 1 2 h5 'ltc7 (when . . .

. . .

Classical Variation 45 the pawn reaches hS Black faces a so Anand-Benjamin, Wijk aan tricky problem -is he willing to Zee 1 989 continued 1 4 exf3 let his black squares be weakened and now 15 �c4 fxg2 1 6 J.. xg2 by allowing h6, or should be play ftfc7 1 7 §td6 �d8 1 8 jtg3 -wte7 19 � heI is crushing for White) 1 3 h6 himself, even though this aS 1 4 � a4 makes the subsequent advances of exdS j_xc3 ( 1 3 "*a7 1 5 d6 b6 1 6 �e3! lil!. b8 1 7 White's g-pawn very strong? .Z1 b 5 was very good for White in After 1 2 h6 1 3 J.. e3 'fjc7 1 4 �f2 �d7 1 S g4 � ceS 1 6 � g l bS Mokry-Conquest, Gausdal 1 989) 14 �xc3 exd5 ( 1 4 . . . � b4 is 1 7 gS White had a very dangerous attack in Hellers--J. Piket, worse as 15 d6! � xd6 16 A c4 Amsterdam 11 1 98S) 1 3 g4 (or 1 3 makes it hard for Black to com h6 g6 1 4 �f4 � e8 l S J.. xe7 -wtxe7 plete his development) 1 5 'i!!f c S 1 6 -wte3 bS 1 7 11 e2 � b8, Marti followed by j_ b5, with advantage to White. novic-Popovic, Yugoslav Ch. (Sb2) ll . . . a6 1 2 j_e3 'f#c7 1 3 1 986, and now 18 � d2 followed by � hd l is slightly better for 'i!i!i'f2 (thanks to the threat of j_ b6 White according to Martinovic) Black has no time for . . . dS) � d7 1 4 h4 bS with a position similar to bS 1 4 j_e3 �d7 1 S gS � ceS 1 6 g6! b4 ( 1 6 fxg6 1 7 f4! � c4 1 8 line Sa above. Admittedly White J.. xc4 bxc4 1 9 Z!dS! exdS 20 has spent a move on '\fi b ! , but this �xdS + '\fi h8 2 1 hxg6 � f6 22 is certainly not a waste of tempo, � xh7 + � xh7 23 Ji h l wins) 1 7 and while White's chances are not gxf7 + 'l¥ xf7 1 8 �dS! exdS 1 9 quite as good as in line Sa he has '*' xdS + '\fif8 ( 1 9 'itre8 20 h6 g6 fair attacking chances. In Sax 2 1 f4 � g4 22 j_c4 is also good for Wilder, Lugano 1 989, White White) 20 'l!lfxa8! R_b7 2 1 'l!lfa7 adopted the rather strange plan with advantage to White since 2 1 l S h5 J.. b7 16 �g3 J.. f8 1 7 J.. gS � c6 i s met b y 22 �d4, � e8 and now Black is at least Serper-Brodsky, USSR 1 986. equal, but the simple l S g4 is (Sb) 10 . . . Zii, d8 1 1 * b l (a better. 10 a4 useful semi-waiting move; the re This appears to be the best res ply 1 1 dS is bad for tactical reasons, so Black normally plays ponse to the advance of the a pawn. 11 a6, when White can switch d5 10 to his kingside attack plan withll jl b5 (39) out allowing dS) and now: This move was introduced by (Sb 1 ) ll . . . d5?! 12 �xf6 j_xf6 (Black should not play 1 2 . . . Tal in his game against Sisniega in the 1 98S Taxco Jnterzonal, dxe4? because of 1 3 j_xe7! Zil xd2 although Vitolinsh was appar 14 �xd2! when 14 � xe7 1 5 � c4 Y¥tc7 1 6 � b5 wins the queen, ently the originator.

46 Classical Variation 39 B

Q) b4 1l Or: (1) 11 Q)xe4 1 2 Q) xe4 dxe4 1 3 � xd8 ..§.. x d8 1 4 _§_xd8 Q) xd8 1 5 Q) c5 f5 (or 1 5 b6 16 Q) xe4 Jtb7 1 7 � he l jj_d5 1 8 0 � c8 1 9 Q) c3 jj_ a8 20 jj_ d 7 � c7 2 1 �b5 � c5 22 � d6 !2Jc6 23 Q)c3 with a slight plus for White, Rohde Joshi, USA 1 986) 16 l:!\ d6 '!fl> f7 1 7 � hd l f!/e7 1 8 i.:..d7 (Tal-Sis niega, Taxco 1 985) and now 18 jj_xd7 was just slightly better for White according to Tal. In the game 18 � ti 1 9 Q) xe6 Jtxd7 20 JZJ c7 .R.xa4 21 Ji) xa8 led to a quick White win . (2) 1 l dxe4 1 2 'ltxd8 Jtxd8 h6 1 4 1 3 lii( he 1 Q)a7 (or 13 J.. xf6 A xf6 1 5 Q) xe4 with a n edge for White) 14 _§_c4 h6 1 5 Jtxf6 gxf6 1 6 Q)xe4 f5 1 7 Q)d6 Jtc7 1 8 g 3 b6? ( 1 8 H d8 1 9 Q) b5 Q)xb5 20 A xb5 would have been slightly better for White according to Tal) 1 9 Q) xf5! and White won in the famous game Tal-Korchnoi, Montpellier 1 985. (3) 11 Q) a7 and now there are two tempting lines for White: (3a) 12 Jte2 Ad7 ( 1 2 . . . b5 1 3 . . .

. . .

exd5 bxa4 1 4 d 6 axb3 1 5 dxe7 'Wfxe7 1 6 cxb3 was unclear in Oll Ryskin, USSR 1 987, but 1 3 J.. xf6 J.. xf6 1 4 <EJ xb5 is better according to Oil) 1 3 J.. xf6 j}_xf6 14 exd5 _t xc3 1 5 'f!txc3 _txa4 1 6 dxe6 'f!te7 1 7 exf7 + * h8 (Black has some initiative for the two pawns, but not nearly enough) 1 8 _tc4 � ac8 19 � he l 'f!tg5 + 20 'f!td2 'f!tg6 21 _te6 � c7 22 'f!td3 'i!lt'xg2 23 'f!td6 and Black's counterat tack has collapsed, Gelfand-Rys kin, Minsk 1 986. (3b) 12 Jtxf6 Jtxf6 13 exd5 _t xc3 14 11f'xc3 Q) xb5 1 5 axb5 a4?! (15 exd5 16 Q) d4 ttb6 is rela tively best, but still good for White after 1 7 H he l ) 1 6 dxe6! 'l!l!fg5 + (the point is that 1 6 . . axb3 1 7 � xd8 � a 1 + 1 8 'i91 d2 )j xd8 + 1 9 \tie2 !! xh 1 fails to 20 llrc7 )j f8 2 1 e7 � e8 22 'f!td8) 1 7 'i!it'd2 'f!t f6 1 8 Q) d4 and Black has very little for his minus pawn, Hoffman-Timoshchenko, Buda pest Open 1 989. (4) 1 1 J.. b4 (an untested suggestion by Tal) 1 2 exd5 exd5 1 3 'f!tf4 with an edge for White. 12 � bel (40) Or 12 e5 Q)d7 1 3 J.. xe7 ttxe7 14 f4 (in this French Defence type of position, the exchange of black-squared bishops theoreti cally favours White, but with the kings castled on opposite sides of the board the game is more likely to be decided by the speed of the respective attacks rather than by the endgame advantage of the better bishop) Q) c5 (after 1 4 . . . . . .

Classical Variation 47 b6 1 5 � he l � c5 1 6 � d4 A d7 1 7 * b l � ac8? 1 8 g4 � fd8 1 9 f5 � e4 20 � xe4 dxe4 2 1 c3 �d3 22 A xd3 exd3 23 * f4! A xa4 24 � xd3 White had an excellent position in Balashov-Khalifman, Minsk 1 986, but 1 7 � fc8 intending . A xb5 was better) 1 5 <E'J xc5 * xc5 1 6 h4 (the idea is to bring the rook to g3, not only helping the kingside attack, but also providing useful defence along the third rank) b6 and now 1 7 h5 was quite unclear in Kinder mann-Felsberger, Vienna 1 986. Perhaps 1 7 � h3 was more accur ate because in some lines the h pawn plays no important role on h5, but in any case the position is very double-edged. 40 B

dxe4 12 Or: (I) 12 . . h6 and now: ( l a) 13 Axf6 A xf6 14 exd5 exd5 1 5 <E'Jxd5 j't g4 1 6 f3 lt g5 1 7 <E'Je7 + ! ( 1 7 <E'Je3 Jt f5 1 8 <E'J d4 j't g6 19 g3 j'txe3 20 �xe3 � c8 gave Black enough for the pawn in 011-Temirbaev, Kuibyshev 1 986) f/!xe7 ( 1 7 . . . j't xe7 1 8 .

)ll!J xd8 � fxd8 19 � xd8 + iii( xd8 20 il, xe7 jie6 2 1 <E'J d2 gives Black nothing for the pawn) 1 8 iii( xe7 Jtxe7 19 i'ite2 j'tg5 + 20 fl> b 1 A f5 2 1 <E'Jd4 jtg6 2 2 '\t1 a l ! with excellent winning chances for White, Oll-Kha1ifman, USSR 1 987. (I b) 13 exd5!? exd5 ( 1 3 hxg5 14 d6) 14 '!ioi'e3 Ae6 1 5 JJ.. f4 iJ, c8 (Winsnes-Khalifman, Groningen 1 985/6) and now 1 6 * b l (intend ing -2]d4) -2]h5 1 7 ..iLe5 -2] c6 1 8 Jtxc6 bxc6 1 9 A.d4 i s good for White according to Donaldson. ( l e) 13 Ji.e3 �c7 14 1i. f4 e5 1 5 exd5 exf4 1 6 d6 gives White an edge, Andrijevic-Kapetanovic, Yugoslavia 1 988. (2) 12 . . . Jtd7 (perhaps the most solid move) 1 3 e5 ( 1 3 exd5 .iL xb5 1 4 d6 'ilfxd6 1 5 � xd6 jJ_ xd6 1 6 £ xd6 j'tc6 gives White an edge, as does 13 Ji.xd7 'li\' xd7 14 e5 � e8 15 j'txe7 ljj xe7 16 f4 tz:Jc7 1 7 -2J d4 .:t:Jc6 1 8 �db5 .:t:J xb5 1 9 -2] xb5 � ac8 even though a draw was agreed here in Rachels0. Gurevich, Boston 1 988) � e8 14 h4 <E'Jc7 1 5 -2]d4 -2] c6 1 6 ..iLxe7 �xe7 1 7 �g5 and again White has a small advantage, Wang Zili D. Gurevich, Belgrade, 1 988. 13 )litxd8 White has a speculative alterna tive in 1 3 � xe4 -2] xe4 14 �xd8 Ji.xg5 + 1 5 "Wfxg5 <E'J xg5 1 6 h4 e5 ( 1 6 . . . h6 1 7 hxg5 hxg5 1 8 £ d6 offers White reasonable play for the pawn as Black's development is very difficult) 1 7 hxg5 1i. f5 1 8 � d 2 !! fc8 ( 1 8 . . f6 was more

48 Classical Variation cautious) 1 9 c3 A e6 20 <E) a l B, c5 2 1 J.. d 7! with some advantage to White, Vitolinsh-Inkiov, Junnala 1 985. � xd8 13 1 3 . . . A xd8 1 4 <E) xe4 is good for White after 1 4 . . . <E) xe4 1 5 A xd8 <E) xf2 1 6 B, d2 or 1 4 . . . A e7 1 5 <2\xf6 + J.. xf6 1 6 j}_ xf6 gxf6 1 7 )!lt d6. <E)bd5 14 <E) xe4 15 c4! <2\ c7 Or 1 5 <E) b4 1 6 B. xd8 + � xd8 1 7 l! d 1 A e7 1 8 <E) d6 with a clear plus. 16 ll xd8 + � xd8 jte7 17 M d1 18 <E) xf6 + gxf6 19 Ae3 <E) xb5 20 axb5 (41) 41 B

The liquidation has left White with a clear advantage. Black still has problems developing his pieces and White's queenside majority is ominously near to creating a passed pawn. True, Black has the two bishops, but White can always force an exchange (e.g. by � c5) if they show signs of becoming active.

20 f5 Black would like to play 20 . . . e5 to free his white-squared bishop, but after 2 1 �c5 A xc5 22 <2\ xc5 j}_ g4 23 l! d5 l, c8 24 b3 Black's queenside pawns are in big trouble (24 . K c7 25 b6 K e7 26 <E) e4). Black therefore prepares to duck the exchange of bishops after � c5 by �g5 + . e5 21 <E) c5 In lnformator Anand gives 2 1 f4 22 jtd4 f6 (22 e5? 23 jtxe5) as only slightly better for White. I find this assessment unduly modest, since after 23 <E) a4 e5 24 <E) b6 g b8 25 <E) d5! J.. d 8 (25 *f7 26 <E)xe7 ff/ xe7 27 jtc5 + *e8 28 B, d6 is also unpleasant) 26 � a7 l;li a8 27 jtb6 Black is in serious trouble. f4 22 <E)d7 f6 23 �b6 Practically the only legal move! * f7-e6. Black intends 24 �c7 The threat of <21 b6 forces liqui dation into a winning rook and pawn ending. ll._ xd7 24 25 � xd7 ll._ c5 11._ xd6 26 11._ d6 After 26 � xf2 27 lii xb7 White's two connected passed pawns roll forwards. 27 l;1 xd6 � c8 Black goes for counterplay, but he cannot repair the fundamental defects of his position. After 27 � f8 28 *c2 e4 29 M e6 f5 30 * c3 White's active king will de cide the issue. .

Classical Variation 49 a4 28 b3 axb3 29 rff b2 \t f7 30 \txb3 Black offers a pawn to get his own majority moving, but the two connected passed pawns outweigh anything Black can do with his e pawn. \te6 31 H d7 + e4 32 H xb7 e3 33 H a7 fxe3 34 fxe3 H d8 35 H a2 The immediate 35 \te5 is also met by 36 fg c3, and if 36 fge4 then 37 b6. * e5 36 fgc3 Black could have quite reasonably resigned instead. *e4 37 b6 38 b7 H d3 + .i, d2 39 rff b4 .l, xa2 40 b8 (�) *d3 41 �b7 + 42 �d5 + *e2 43 c5 Resigns

Game 1 1 Ernst-Popovic Subotica 1987 1 e4 c5 <2l c6 2 <2l f3 3 d4 cxd4 4 <2l xd4 <2l f6 d6 5 �c3 e6 6 jj_g5 7 'i(1d2 a6 8 0-0-0 (42) Jl.d7 8 . . . h6 is examined in game 1 2. Other ideas are dubious, e.g. 8 . . .

!fJ.. e 7 (8 �b6 9 <2l b3 !fJ_d7 l O !fJ_e2 'W/c7 l l f4 h6 1 2 !fJ_xf6 gxf6 1 3 !fJ.. h5 is good for White, Pan chenko--Csom, Las Palmas 1 978) 9 f4 file? (9 � xd4 10 '*'xd4 'lit a5 l l e5 dxe5 12 fxe5 <2l d5 1 3 JJ.. xe7 � xe7 l 4 JJ.. d 3 .:t:Jc6 1 5 'W/h4 � xe5 16 <2le4 f6 17 H hfl with a very dangerous attack for White, Adler-Bannik, USSR 1 978) and now: ( I ) 10 fg b1 JJ.. d 7 ( 1 0 0-0 l l !fJ_e2 H d8 1 2 jj_ f3 h6 l 3 h4 <2l xd4 14 �xd4 b5 1 5 iYf2 !fJ.. b7 1 6 g4 was good for White in Kavalek Larsen, Montreal 1 979) 1 1 <2l f3 H d8 1 2 JJ.. d3 b5 1 3 � he 1 b4 1 4 <2l e2 a 5 1 5 <2l g3 0-0 1 6 e 5 <2ld5 1 7 \\lfe2 was unclear i n Przewoznik Bielczyk, Katowice 1 986. (2) 10 JJ.. e2 <2l xd4 1 1 �xd4 b5 12 e5 dxe5 1 3 fxe5 � d5 14 .JJ.. xe7 <2l xc3 1 5 J.. f3 ! <2l xd 1 1 6 J.. d6! and White went on to win quickly in Tal-Larsen, Montreal 1 979 after 16 �c4 1 7 iYb6! <2l f2 1 8 jj_c6 + J.. d 7 1 9 .JJ.. xd7 + * xd7 20 � b7 + * d 8 2 1 '/!i!1"xa8 + �c8 22 �a7 Resigns. (3) 10 1L xf6 gxf6 l l g3 .il_d7 12 f5 � xd4 13 � xd4 £ c8?! ( 1 3 . . .

50 Classical Variation 0-0 and 1 3 b5 are possible improvements) 14 * b l b5 1 5 �d2 �c5 1 6 j_d3 h5? 1 7 fxe(l fxe6 1 8 <i:Je2 intending <i:l f4, Short-Larsen, London 1 986, and White has the advantage. 9 f4 b5 Black has two major alternatives, 9 h6 and 9 !P.. e 7: (l) 9 h6 (9 Jl c8 10 <i:l f3 �a5 1 1 * b l b 5 1 2 e 5 b4 1 3 exf6 bxc3 1 4 fxg7 !P.. xg7 1 5 �xd6 '!!A c7 16 <i:Je5 A xeS 1 7 fxe5 1:! g8 1 8 h4 is good for White, Yanofsky Olafsson, Dallas 1 957) 1 0 !P.. h4 (43) and now: . . .

43 B

g5 1 2 <i:J xe4 gxh4 1 3 "Wt"c3) (l l 1 2 <E:� f5 "Wt" a5 (12 "Wt"b8 1 3 !P.. xf6 gxf6 1 4 <i:Je4, 12 . !P.. e7 1 3 <E) xd6 + \fi' f8 1 4 <i:J xb7 "Wt"c7 1 5 �d2 and 1 2 "Wt"c7 1 3 !P.. xf6 gxf6 1 4 <i:Jd5 "Wt"d8 1 5 "Wt"e3 ! are all very pleasant for White) 1 3 <E) xd6 + J. xd6 1 4 .M xd6 0-0-0 (14 �c7 1 5 lilt d2 is unsatisfactory after 15 0-0-0 1 6 "Wt"f2 <E:�e7 1 7 .!id3 !P.. c6 1 8 f5 e 5 1 9 � hd l or 15 "Wfxf4 1 6 J_e2 <i:Je4 1 7 <i:J xe4 "Wfxe4 1 8 "Wtf2, Gligoric-Barden, Bognor Regis 1 957, while 14 . . . <i:Je7 1 5 K d 1 .:E� g6 1 6 <i:Je4! "Wtxe1 17 <i:Jd6 + 'ltle7 18 J. xe l <i:Jd5 1 9 <i:Jx7 <i:Jxf4 20 g3 <i:J g6 2 1 !P.. g2 and 14 <i:J b4 1 5 a3 <i:J bd5 1 6 �e5 !P.. c6 17 !P.. c4! are both very good good for White) 1 5 K d l ! �c7 (15 . . . <i:Je7? 16 <i:Jd5 wins, 15 . . . g5 1 6 fxg5 hxg5 1 7 !P.. g3 gives White very strong pressure on the dark squares and 15 e5 1 6 fxe5 K he8 1 7 jj_ g3 <i:J xe5 18 <i:J b5 ll_ g4 1 9 K xd8 + K xd8 20 <i:Ja7 + * b8 2 1 <i:Jc6 + bxc6 22 �xe5 + is a very good ending for White) 1 6 "Wt f2 <i:Je7 1 7 jj_ d3 jj_ c6 1 8 f5 e5 1 9 M he l <i:Jed5 20 <i:J xd5 K xd5 (20 _ztxd5 2 1 "Wta7) 2 1 �g3 e4 22 "Wfxc7 + \fi' xc7 23 ll_xf6 exd3 24 !P.. xg7 K hd8 25 .zte5 + with good winning chances for White, Spassky-Rabar, Goteborg 1 955. ( l e) 10 g5 1 1 fxg5 <i:Jg4 and now: ( l c l ) 12 <i:lf3 hxg5 ( 1 2 jj_ e7 1 3 g6! J. xh4 1 4 gxf7 + \fi' xfi 1 5 "Wtf4 + is good for White, or 14 . . . * f8 1 5 "Wtxd6 + jj_ e7 1 6 t�t xd7 <2:Jf2 1 7 "Wtxb7 <i:J xd 1 1 8 "Wfxc6 and • . •

• •

. • .

. • •

. . •

• . •

. • .

• • .

( l a) 10 !P.. e7 (10 . . . <i:J xd4 1 1 �xd4 !P.. c6 1 2 !P.. c4 is clearly good for White, while 10 . . . K c8 1 1 <i:l f3 � a5 is similar to the note to Black's 9th move) 1 1 .:E� f3 ! b5 1 2 e5 ( 1 2 !P.. xf6 !P.. xf6 1 3 �xd6 )i,i( a7 14 e5 !P.. e7 1 5 �d3! -a5 1 6 * b 1 I:t c7 1 7 �e3! 0-0 1 8 �e4 was also good for White in Geor Nikolaev gadze-Makarychev, 1 983) b4 1 3 exf6 bxc3 1 4 �xc3 gxf6 1 5 f5 and White stands well, Thiemann-Reynolds, corr. 1 966. ( l b) 10 <i:J xe4 1 1 �e1 .:E� f6 • • •

. • .

Classical Variation White should win) 1 3 jj_g3 jle7 1 4 jle2 � ge5 and now: ( l c l l ) 15 'fl'bl b5 ( 1 5 f6 1 6 h4! gxh4 1 7 � xh4 �a5 1 8 � f5! � xh l 19 � xh l exf5 20 .M h8 + jlf8 2 1 jlxe5! � xe5 22 �xd6 is good for White, Mokry-Banas, Trnava 1 986) 1 6 a3 M b8 1 7 �a2 a5 1 8 � c l � xf3 1 9 gxf3 e5? (19 . �e5! is unclear) 20 .il.. f2 jj_e6 2 1 jj_ e3 with an edge for White, Riemersma-A . Rodriguez, Dieren 1 987. ( l c l 2) 15 � hfl f6 ( 1 5 '*'a5 16 'fl' b l f6 17 �e l b5 18 jj_d3 0-0-0 19 � d2 \fJ b7 20 h3 * a8 2 1 jl f2 was a little better for White, Vogt-Barczay, Zalakaros 1 987) 1 6 � xe5 �xe5 1 7 jl xe5 dxe5 1 8 � f3 �c7 1 9 � d3 .M d8 20 h3 jlc8 2 1 l! xd8 + �xd8 22 'i!ire3 ·ii!i' a5 (22 'ilfc7 preventing jj_c4 ap pears better) 23 jj_c4 � c5 24 'il!fe2 jj_d8, Vogt-Wirius, Zalakaros 1 987, and now 25 \fJ b l gives White an edge. ( l c2) 12 � xc6 (this seems the better choice) jlxc6 1 3 jj_e2 � e5 1 4 g3 �g6 1 5 \fi b ! h5 ( 1 5 jj_e7?! 1 6 gxh6 jj_xh4 1 7 gxh4 �xh4 1 8 �xd6 was good for White in Marjanovic-Popovic, Belgrade 1 987) with a further branch: ( l c2 1 ) 16 �e3 jle7 (better than 16 . . . jj_ g7 1 7 �d5! exd5 1 8 exd 5 + '*e7 1 9 �xe7 + *xe7 20 dxc6 bxc6 2 1 c3 with an edge for White, Chandler-Bellin, Com monwealth Ch. 1 985, or 16 . . . '*e7 1 7 � hfl jj_g7 1 8 � f2 * f8 1 9 � dfl , Mainka-Popovic, Dort-

mund 1 988 and White is better) 1 7 � hfl �c7 1 8 �f2 � e5 1 9 h3 0-0-0 20 g4 hxg4 21 jj_xg4 � xg4 22 hxg4 � h7 is unclear, Jansa Banas, CSSR Ch. 1 986. ( l c22) 16 .M hfl ( 1 6 lil, dfl !? intending 17 �d I attacking h5 is an interesting idea) �c7 ( 1 6 jj_g7!? is better) 1 7 e5! 0-0-0 1 8 J., d3 � xh4 1 9 gxh4 J.. g7 20 *f2! and White stands well, Tsesh kovsky-Fahnenschmidt, Baden Baden 1 988. ( I d) 10 . . . b5 1 1 jj_xf6 � xf6!? 12 e5 ( 1 2 j}_xb5 axb5 1 3 � dxb5 �d8 14 �xd6 + jlxd6 1 5 trxd6 �e7 16 e5 �xd6 1 7 ii:i!. xd6 * e7 is unconvincing) dxe5 1 3 � dxb5 �d8 14 �d6 + jlxd6 1 5 �xd6 exf4 16 �e4! ft(e7 1 7 ft(c7 ii:il. a7 1 8 .£Jd6 + \fi f8 1 9 �b6 and White I S at least slightly better, Sobura Berebora, Poland 1 988. (2) 9 . . . jl e7 1 0 � f3 b5 (other moves are inconsistent, e.g. I 0 tllf c 7 1 1 e5 dxe5 1 2 fxe5 � d5 1 3 � xd5 exd5 1 4 jlxe7 � xe7 1 5 jj_ d3 0-0 1 6 �g5 � c6 1 7 trh5 or 10 h6 1 1 jj_xf6 gxf6 1 2 f5 trc7 1 3 \fi b ) 0-0-0 14 jl c4 with a clear plus for White in both cases) 1 1 e5 b4 (not 1 1 dxe5 1 2 fxe5 b4 1 3 exf6 bxc3 1 4 ft(xd7 + and White wins) 12 exf6 bxc3 1 3 Yllf xc3 gxf6 1 4 jj_ h4 (44) and now: (2a) 14 . . . aS ( 1 4 tllf a 5 1 5 jl xf6 � b4 1 6 jj_c4 H c8 1 7 a3 is good for White) 1 5 \fi b ! � b4 ( 1 5 !! b 8 1 6 g4 � b4 I 7 a 3 .M c8 1 8 tll! b3 � d5 1 9 K xd5 with an excel lent position for White, Gligoric Conrady, Dublin 1 957) 1 6 a3 K c8

52 Classical Variation 44 B

1 7 '\t b3 �d5 1 8 l;l xd5 exd5 and now either 1 9 � d4 or 1 9 '\txd5, with very good compensation for the exchange. (2b) 14 . . d5 1 5 * b l with a further branch: (2b I ) 15 . . . a5 1 6 J.. b5 M c8 1 7 � d4 � xd4 1 8 J_ xd7 + '\txd7 1 9 '\txd4 K g8 20 g 3 .- b s 2 1 M be l is good for White. (2b2) 15 . . . �b4 1 6 � d4 'll!!f a 5 ( 1 6 . . . K c8 1 7 'itb3 '\ta5 1 8 J.. e l ! J.. a4 1 9 '\ta3 with a clear plus for White, Matanovic-Jansa, Lugano 1 968) 1 7 a3 � c6 1 8 '\tg3 �xd4 1 9 �t xd4 K b8 20 K d3 * f8 2 1 J.. e2 with a small plus for White according to Lukin. (2b3) 15 . . . � a5 1 6 f5 l, c8 1 7 'it d2 '\tc7 1 8 fxe6 fxe6 1 9 j_d3! � c4 20 J.. xc4 '\txc4 2 1 K he l K g8 (2 1 0-0 is most simply met by 22 g4!) 22 h3 K b8 23 *a I h5 24 g3 with a small advantage for White, Tseshkovshy-Lukin, USSR 1 982. gxf6 10 J_xf6 J.. xc6 l l � xc6 White's chances lie in the fact that Black's king has no really safe spot, and in an attack against .

the weak e6 square by f5, j_d3 and �e2-f4. Black must be care ful about playing . . . b4; he may gain time by driving the knight away, but he may also allow the bishop to become active at c4. In general the exchange of queens favours Black, so White must make sure he avoids this. 12 'ite1 (45) Probably the most accurate move. Black's immediate am bition is to bring his queen to an active square, so White often plays 1 2 *e3 to prevent . . . '\tb6, but after 'ite7-a7 White has to waste time avoiding the queen swap. 12 lill' e I also prevents 1 2 . . '!'lfb6 because of 1 3 �d5, but it also tucks the queen away from the unwelcome attentions of her opposite number.

12 J.. e7 Or: ( I ) 12 b4 1 3 �d5 a5, A. Ivanov-Ermolinsky, U SSR 1 98 1 , and now 1 4 M d4! intending .!11 c4 is good for White. (2) 12 '!'lfa5 1 3 * b l 0-0-0 1 4 j_ d 3 \t¥ b8 1 5 '!'lfh4 '!'lfc7 1 6 f5 was • . .

. . .

Classical Variation 53 promising for White in Kuzmin Tukmakov, USSR eh. 1 977. (3) 12 jte7 1 3 j_d3 J.. g7 ( 1 3 j_ b7 1 4 \t;' b l 0-0-0 1 5 a4 b4 1 6 <2\a2 a 5 1 7 c3 d5 1 8 cxb4 axb4 1 9 exd5 K xd5 20 jte3 jt d7 2 1 li;l c l + \t;' b8 2 2 J_ e4 with some advantage for White, Spraggett Mednis, Lugano 1 985) 14 jtg3 (I prefer 14 fi' b l intending f5) llii g8 1 5 � he I * ffi 1 6 'Wf h4 f5 l 7 jtxe7 + fi' xe7 1 8 exf5 J.. xc3 1 9 bxc3 K xg2 with equality, Kud rin-Christiansen, USA Ch. 1 986. jtb6 13 j_d3 14 \t;'b1 Safer than 1 4 'Wf g3 b4 1 5 <21e2 \t;'d7 1 6 f5 e5 1 7 jtf3 jtc5, Menc inger-Damljanovic, Bled 1 984, with an unclear position. b4 14 The alternative is 1 4 h5 1 5 f5 b4 (after 1 5 jtc5 White i s a little better after the solid 1 6 K fl !?, but 1 6 �g3 fl'd7 1 7 �h3 K ag8 18 K he1 � g5, Hellers-van der Wiel, Amsterdam 1 986, is also possible, when 1 9 g3 gives White an edge) 1 6 <2\e2 e5 1 7 jtg3 ( 1 7 <2\g3 a5 1 8 �e2 h4 1 9 <2\fl was unclear in Watson-Yrjola, Kecs kemet 1 988) J.. ffi? 1 8 'W/f3! and White stands well, A. lvanov Anikaev, USSR 1 979. ECO suggests 1 7 . . . d5, but 1 8 'W/ f3 looks good in this case too. aS 15 <2\e2 16 f5! More active than 1 6 'W/g3, as played hitherto. White threatens 1 7 fxe6 fxe6 1 8 <2'! f4, so Black's reply is virtually forced. . • .

e5 16 17 <2\g3 (46) With the simple plan of J.. c4d5. 46 B

17 *c5 a4 18 ,.e2 There is nothing more Black can do to prevent White's bishop manoeuvre, so his only chance is to aim for queenside counterplay 19 J.. c4 0-0 Black needs his other rook if this queenside play is to genuinely worry White, but there is an obvious danger to his king after 0-0. * h8 20 <2\h5 21 Jii hfl a3? Overlooking that White has an immediate mating threat, but even the best line 2 1 K fc8 22 b3 K a7 23 � f3 axb3 24 J.. xb3 is very unpleasant for Black. 22 � f3 White can afford to abandon the queenside since he has a forced win on the other flank. 22 axb2 23 H h3

54 Classical Variation Intending 24 <E:� xf6 Jl xf6 25 'i4J'h5. 23 � g8 Resigns 24 � xf6 After 24 ilxf6 25 !l xh7 + fj' xh7 26 'i4J'h5 + fl g7 27 'i4f xf7 + flh6 (27 flh8 28 'i4J'h5 + flg7 29 'i4J'g6 + and mate next move) 28 'i4fxf6 + flh7 29 'i4fh4 + flg7 30 'i4J'g5 + the rook is the first of many Black pieces to disappear.

(I) 9 'i4fc7 (9 .:2:J g4 1 0 .£) xc6 bxc6 1 1 il c5 i s good for White) 10 f4 ,ile7 1 1 Jle2 Jl d7 (or 1 1 .:2:Ja5 1 2 e5! dxe5 1 3 fxe5 �xe5 14 _R f4 �c5 1 5 .:2:Ja4 'i4fd5 16 '!ti b 1 ! with a very strong attack for the sacrificed pawn) 1 2 .:2:1 b3 � a5 13 .:2:J xa5 'i4fxa5 14 fl b 1 ilc6 1 5 il f3 'i4fc7 16 � he 1 (the im mediate 1 6 g4 may be even better) lii( c8 1 7 g4 with advantage to White, Chiburdanidze-Lanka, USSR 1 980. (2) 9 ;Le7 10 f4 .:2:J xd4 1 1 ;L xd4 b5 1 2 il e2 transposes to line 3 . (3) 9 .:2:J xd4 1 0 ;L xd4 b5 1 1 f4 .ile7 ( 1 1 b4 1 2 A xf6 'i4fxf6 1 3 .:2:J e2 � b8 1 4 .:2:l d4 !;K b6 1 5 ilc4, Ta1-Radu1ov, Malta 01. 1 980 and 11 _R b7 12 Jl xf6 gxf6 1 3 ild3 'i4J'b6 1 4 fJ b 1 ;L e7 1 5 f5 e5 1 6 ;Le2 followed by ilf3, Jansa-Spassov, Sochi 1 980, were both slightly better for White) 1 2 .ile2 (a promising pawn sacrifice) b4 ( 1 2 ilb7 1 3 A f3 b4 1 4 .il xf6 .il xf6 1 5 .:2:1 e2 'i4r a 5 1 6 a 3 is good for White) 13 .:2:J a4 .:2:J xe4 1 4 'i4fe3 � f6 1 5 _R f3 and now: (3a) 15 . . . d5 1 6 fl b 1 (this has • • .

Game 12 Short-Ljubojevic Amsterdam (Euwe Memorial) 1988 c5 1 e4 d6 2 �f3 cxd4 3 d4 4 �xd4 �f6 5 �c3 <E:�c6 6 il g5 e6 a6 7 'i4J'd2 h6 8 0-0-0 Black once again aims to further his queenside ambitions by . . . ild7 and b5 before committing his king, but in this line he first of all forces White to decide where to put his bishop. 9 il e3 (47) This move is currently thought best. White intends to play f4 and develop his king's bishop to e2 or d3. If Black castles kingside then White can launch a direct attack by h3 and g4-g5. 9 il d7 Black has a number of possible moves, but the basic rule is 'f4 against anything' The alterna tives are:

. . .

• . .

. . .

Classical Variation 55 been played in practice, but in fact 1 6 g4 0-0 1 7 � b6 !![ b8 1 8 * b I may be more accurate, transpos ing into Short-A. Rodriguez be low but without allowing Black so much choice) 0-0 ( 1 6 A d7 1 7 � b6 1i b8 1 8 g4 A b5 1 9 h4 * ffl 20 g5 �e8 2 1 f5 was also danger ous for Black in Chandler-Kos ten, Hastings 1 988/9) 1 7 � b6 1i b8 1 8 g4 Ad6?! ( 1 8 �d7 1 9 � xd7 jt xd7 20 h4 jt f6 2 1 g5 Axd4 22 Jiil xd4 was better, even though White still has a danger ous attack) 19 g5 hxg5 20 fxg5 � d 7 21 � xc8 li( xc8 22 g6 with a clear plus for White, Short-A. Rodriguez, Subotica 1 987. (3b) 15 . . . � b8 1 6 A a7 Ad7! (not 16 . . . � b5 1 7 jt b6 �d7 1 8 � c5 and White wins, Balashov Tukmakov, Sverdlovsk 1 987) 1 7 � b6 jtb5 1 8 \fi b ! !? (after 1 8 g4 0-0 1 9 Axb8 �xb8 20 g5 hxg5 2 1 fxg5 jtd8! 22 gxf6 �xb6 2 3 '*t xb6 jt xb6 Black had sufficient com pensation in Khalifman-Ionov, USSR 1 988) 0-0 1 9 f5 f/fc7 (the only move) 20 jtxb8 li xb8 2 1 � a8! �d8! (not 2 1 . . . �c8? 22 'fr!! a 7 � ffl 23 fxe6 fxe6 24 <Elc7 and the knight excapes, nor 2 1 'fr!/d7 2 2 fxe6 fxe6 2 3 Rhel d 5 24 � b6, followed by taking on e6) 22 'fr!! a 7! (after 22 fxe6 ;ii xa8 23 exf7 + \fixf7 24 � xa8 'i#xa8 25 Jii he ! �b7 Black's active minor pieces are at least as valuable as White's rooks) d5! (22 . . . exf5 23 � he ! � e4 24 � xe4 fxe4 25 ;a xe4 J.. f6 26 :t::J c 7 and 22 e5 23 � he I! 1L d7 24 � c7 JL xf5 25

� xa6 are good for White) 23 � c7 JLc6 (threatening 24 . . . � b7, and if 24 � xa6 then 24 ;a a8 wins) 24 fxe6 � b7 25 <Elxd5! � xa7 (not 25 � xd5 26 exf7 + fifE 27 �xa6 � c7 28 '*d3 :!i d7 29 �h7 � f6 30 f/fh8 + *xf7 3 1 �xd8 and wins) 26 � xf6 + JL xf6 27 � xd8 + JLxd8 28 exf7 + \fi xf7 29 JLxc6 (normally an extra pawn offers some winning chances in a rook and opposite bishop ending, but White must overcome the problem of his inactive king) A f6 30 � d l ;ii c7 3 1 JL f3 JL e7, Nunn van der Wiel, Lucerne 1 989, and now 32 g d5! � c5 33 1i d3 leads to an ending in which Nigel Short believes White has significant win ning chances. The plan is a3 or c3, followed by advancing the king to attack Black's remaining queen side pawn. Certainly Black will be tortured for a long time. 10 f4 b5 Against other moves White adopts the same general plan of JL d3, * b l , and then a kingside pawn advance, but he has to be careful against 1 0 JLe7, because 1 1 jtd3 allows the awk ward 1 1 <El g4! Therefore the best answer to 1 0 jt e7 is 1 1 h3 b5 1 2 JLd3, transposing into the note to White's 1 2th move. 11 JLd3 JLe7 ( 48) Two alternatives are 1 1 . . . � c8 1 2 fi b ) � a5?! 1 3 e5! b4! 14 � ce2 dxe5 1 5 fxe5 �d5 1 6 liii, hfl � c4 1 7 J.. xc4 � xc4 1 8 <El f4 � xe3 1 9 \'¥(xe3 Jl.c5 2 0 � g6!, Hazai Szabo, Hungary 1 983, and 1 1 . . .

56 Classical Variation .:£) xd4 1 2 J.. xd4 b4 1 3 �e2 lta5 14 J.. xf6 gxf6 1 5 * b 1 , Psakhis lvanovic, Sochi 1 979, with a clear plus for White in both cases. Against other moves, such as 1 1 lta5 o r 1 1 ltc7, White proceeds with his plan by 1 2 * b l . . . .

tives are better, for example 12 flf c7 1 3 h3 <2) a5 ( 1 3 . . . <2) xd4 1 4 J.. xd4 J.. c 6 i s probably the most sensible, but even here White has an edge) 1 4 g b4 1 5 <2) ce2 <2) c4 1 6 j_ xc4 fifxc4 1 7 <2) g3 a5 1 8 � hfl lta6 1 9 g5 with advantage, Hodg son-Csom, Tel Aviv 1 988, or 12 <2) xd4 1 3 j_ xd4 b4 1 4 <2) e2 ftfb8 1 5 <2) g3 ( 1 5 h3 is more con sistent) 0-0?! ( 1 5 e5 is better) 1 6 e5! <2)d5 1 7 <2) e4 � d8 1 8 ftf f2 dxe5 1 9 fxe5 J.. e 8 2 0 � hfl and White is better, Hazai-Lobron, Rotterdam 1 988. 0-0 13 <2) ce2 14 h3 Black's problem is that White has an automatic attack by h3, g4, <2) g3 and g5 while Black has to struggle to create any counter chances at all. It is curious that Black's troubles stem from the apparently innocuous 8 . . . h6, which in this type of position creates a fatal kingside weakness. 14 flf c7 ltb7 15 g4 <2) xd4 16 <2) g3 17 J.. xd4 j_c6 All Black has achieved is to create a threat to e4, which delays White's attack by precisely one move. 18 � he t lit fe8? This doesn't help the belea guered kingside. The last chance was to play 1 8 <2) d7 intending e5 to block the deadly long diagonal, but in this case Short gives the line 1 9 g5 hxg5 20 lit g 1 (threat <2) h5) e5 2 1 <2) f5 lit fe8 22 . . .

. • .

12 *b1 12 * b I has been played a number of times recently, but there is a strong argument for the immediate 1 2 h3, e.g. 1 2 . . . <2) xd4 1 3 Jl.. xd4 J.. c6 ( 1 3 . . . b4 1 4 � e2 a5 1 5 g4 j_c6 1 6 <2) g3 d5 1 7 "i!ite2 "tl!l' b8 18 � hfl b3 19 cxb3 a4 20 * b l axb3 21 a3 gives White a modest advantage, Murei Lobron, Lyons 1 988) 1 4 "tl!l'e3 ( 1 4 M del 0-0 1 5 * b l .:£) d 7 1 6 g4 was slightly better for White in Timoshchenko-Tukmakov, USSR Ch. 1 978) b4 1 5 <2)e2 fife? 1 6 e5 dxe5 1 7 j_ xe5 ftfb7 1 8 f5 <2)d5 1 9 ftfg3 with dangerous threats for White, Timosh chenko-Sirov, Moscow GMA 1 989. 12 b4 This move has been criticised, but it isn't clear that the alterna-

Classical Variation 57 fxe5 <2!xe5 23 .,;}_ xe5 dxe5 24 k! xg5! j_xg5 25 �xg5 f6 26 J.. c4 + fi'ffi 27 �h5 and mate at h8. hxg5 19 g5 20 fxg5 <2! d7 (49) After this we are treated to a king-hunt in the style of the 1 9th century. 20 � h7 was objec tively better (anything is better than being mated), but 2 1 h4 loc king the knight out of play is very good for White. 49 w

21 j_ xg7! * xg7 22 <2! h5 + \tlg6 There is no choice as 22 * g8 23 g6 fxg6 24 "�th6, 22 . . . \tlf8 23 g6 J.. f6 24 <21 xf6 <2! xf6 25 � f1 \tl e7 26 �g5 and 22 \tlh8 23 g6 J.. ffi 24 H g l fxg6 25 � xg6 <2!e5 26 � h6 + ! all lead to disaster. \tl xh5 23 e5 + 24 t!'f4 j}_ xg5 All Black's moves are forced. \l;'h4 25 "�txf7 + fi>·g3 26 �h7 + The main problem when play ing such positions with White is trying to keep a broad grin off . • .

your face. The main problem when playing such positions with Black is to avoid looking at the broad grin on your opponent's face. \tlh2 27 �h5 Or 27 H g8 28 �g4 + \tlh2 29 'lt g l + flxh3 30 A fl + flh4 (30 . . . j_g2 3 1 K d3 + and 32 Axg2) 3 1 �h2 + \tl g4 32 .i;i d4 + \tlf5 3 3 �h7 + H g6 34 J.h3 mate is a nice line given by Short. 28 'flfxg5 In fact White could have forced mate by 28 *e2 + \tl xh3 (28 j_ g2 29 ;i;i h l + 1trg3 30 �g4 + lff f2 3 1 li:!( hfl + Axfl 32 lli!. xfl + lff e 3 33 )!!; e l + 1f!>f2 34 �g l + w f3 35 i1H1 mate) 29 �h5 + J. h4 30 H e3 + l;l, f3 (30 . . . ltrg2 3 1 "�t xh4 or 30 . . . J. f3 3 1 H h 1 + 1trg2 32 �xh4) 3 1 � h 1 + wg2 32 *xh4 etc., but there is nothing wrong with winning Black's queen (and having his king on h2). 28 H g8 29 � d2 + jtg2 30 jtf4 + Ji;i g3 31 J.e4 �xe4 Resigns 32 'i!lj'xe4

Game 13 Kupreichik-Kuzmin Minsk 1982 1 e4 c5 �c6 2 <2!f3 cxd4 3 d4 <2! f6 4 <2! xd4 d6 5 <2! c3 6 Ag5 (50)

58 Classical Variation 50 B

6 jtd7 This is the most popular alternative to 6 e6, but there are other moves: (I) 6 . g6 (this appears to be an underrated move) 7 j't xf6 exf6 8 j_c4 (if Stoica's recommenda tion below is effective, White should prefer 8 j_b5 jtd7 9 0-0 J.. g7 10 � de2 with a slight plus) J.. g7 (after this Black must sacri fice a pawn, but 8 A. e7 9 �d2 followed by 0-0-0 is depressing for Black) 9 <2) db5 0-0 1 0 �xd6 f5 l l 0-0-0 �a5 ( l l �g5 + 1 2 f4 �xg2 1 3 e5 is given as good for White by theory, but Stoica suggests 1 3 �g4!, threatening both . . . 't'fxf4 + and . . <2) xe5; at the moment I can't see a good reply for White) 1 2 flfc7 a6 (the lines 12 . . j_ xc3 1 3 bxc3 'lit'a4 1 4 <2)d6, 12 . . �b4 1 3 <2) d 6 J..xc3 14 bxc3 fif xc3 15 J.. xf7 + lii h8 1 6 'fi b l and 1 2 fxe4 1 3 �xa5 <2) xa5 14 j_d5 j_h6 + 1 5 lii b l are all good for White) 1 3 'iif x a5 <2) xa5 14 � c7 � a7 1 5 j_b3 j_ xc3 1 6 bxc3 fxe4 (Kholmov-Cher nikov, USSR 1 982) and now 1 7 . .

. . .

<2)d5! is good for White according to Kholmov. (2) 6 . . . �aS j't xf6 gxf6 8 jtb5 A.d7 9 <2) b3 �c7 10 <2)d5 �d8 1 1 �h5 e6 1 2 <2) e3 a6 1 3 J.. e 2 �c7 14 0-0-0 J,.e7 1 5 'fi b 1 0-0-0 16 f4 was good for White in S. Nikolic Gufeld, Kislovodsk 1 968. (3) 6 . . . a6 7 �d2 <2) xd4 (other moves transpose to the main var iations) 8 �xd4 e5 9 �a4 + (White has no trouble keeping a slight advantage by 9 �d3 J.. e 6 lO 0-0-0 !;! c8 1 1 <2)d5 J.. x d5 1 2 A. xf6, but with � a4 + he is play ing for more) jtd7 10 �b3 b5 1 1 j'txf6 gxf6 1 2 jte2 j_e6 1 3 <2)d5 j't h6 14 a4 with some advantage for White, Marjanovic-Stoica, Istanbul 1 988. (4) 6 . . . �b6 (this has become popular recently) 7 <2) b3 (7 j_ e3!? is interesting, e.g. 7 �xb2 8 'i:Jdb5 �b4 9 jtd2 �c5! l O J.. e 2!? �b6 1 1 .!!! b1 <2) e5? 1 2 j_e3 � a5 1 3 !;! b3 g6 1 4 g a3 i!!t d 8 1 5 J.. x a7 <2) ed7 1 6 f4! and White stands well, Balashov-Petrienko, Voro nezh 1 987, but 1 1 . . . �d8 was the critical test) e6 and now: (4a) 8 j_f4 <2)e5 9 j_e3 �c7 l O f4 <2) g6 1 1 �f3 J.. d 7 1 2 J.. d 3 j_e7 13 0-0-0 j't c6 14 <2) d4 was good for White in Greenfield-Schren zel, Israel 1 983, but 8 e5 may be better. (4b) 8 �d2 a6 9 0-0-0 jte7 (9 . . . jtd7 l O f4 'fll c 7 1 1 j_e2 b5 1 2 J.. xf6 gxf6 1 3 J.. h 5 J.. h6 1 4 g hfl was a little better for White, Kho1mov-Petrienko, USSR 1 980, while 9 . �c7 l O '* b l b5 1 1 . .

Classical Variation 59 �f4! '2!d7 1 2 .§t xb5 axb5 1 3 '2l xb5 itb8 1 4 l£l x.d6 + A xd6 1 5 'ltxd6 �xd6 1 6 l!l. xd6 l£l ce5 was unclear in van der Wiel-Sprag gett, Wijk aan Zee 1 985) 10 * b 1 ( 1 0 h4 is a promising alternative which anticipates Black's 0-0) 'ltc7 1 1 f4 Ad7?! ( Black allows a central breakthrough) 1 2 e5 dxe5 1 3 fxe5 '2!d5 ( 1 3 l£l xe5 1 4 A f4 A c6 1 5 tt e 1 is very good for White) 1 4 '2lxd5 exd5 1 5 A xe7 � xe7 16 'ltg5 A f5 1 7 '2! d4 Ag6 18 Ad3 and White has a dangerous lead in development, lvano vic-Popovic, Vinkovci 1 982. (4c) 8 Ad3 (aiming for 0-0 has been the most popular reply, and it has also scored well) Ae7 9 0-0 a6 1 0 a4 ( 1 0 lfi h 1 'ltc7 is also possible, e.g. 11 a4 b6?! 1 2 f4 A b7 1 3 f5 � e5?! 1 4 � d4, Kinder mann-Liberzon, Beersheva 1 984 or 1 1 f4 Ad7 1 2 "�te2! h6 1 3 Ah4 g5?! 1 4 jl g3, Byrne-J. Benjamin, USA Ch. 1 984 with advantage to White in both cases although Black' s play was doubtful in these examples) � a5 1 1 �xa5 ( 1 1 A e3 �c7 1 2 � xa5 'ltxa5 1 3 h3 Ad7 1 4 f4 Jl;!i, c8 15 itd2 A c6 1 6 f5 e5 1 7 b4 'f/!tc7 1 8 b5 gave White some Grunfeld-lvanov, advantage, Toronto 1 984) *xa5 1 2 Ad2 ith5 1 3 � e l � g4 1 4 h3 �e5 1 5 Ae2 �h4 1 6 'l!ltd 1 g5 1 7 f4 gxf4 1 8 � xf4 �g3 1 9 Ag4 � g6 20 � e2 � h4 2 1 g3 and White wins, Iva novic-Piachetka, Stara Pazova 1 988. 7 Ae2 (51) White usually plays 7 itd2

when Black continues 7 � xd4 8 'l!ltxd4 *'a5. The move Ae2 is well motivated since if Black exchanges on d4 he has lost a tempo over the usual line, so he has to change his plan. 51 B

7 a6 A flexible reply, but 7 . . . �a5 may be better. This and other options: (I) 7 �b6 8 � db5 (threa tening 9 Axf6 and 1 0 �d5, while at the same time preventing e6) 1i c8 9 0-0 a6 10 il, xf6 gxf6 1 1 �d5 !i1d8 1 2 � bc3 e6 1 3 � e3 and White's knights proved well placed in Vogt-Mascarinas, Pola nica Zdroj 1 977. (2) 7 e6 8 '2! db5 � b8 9 a4 Ae7 1 0 �d2 a6 1 1 � a3 fijc7 1 2 � d 1 � d8 1 3 � c4 jlc8 1 4 Ae3 'l!lt b8 1 5 Q) b6 with an edge for White, Spassky-Hort, Moscow 1 97 1 . (3) 7 'iii' aS 8 .�xf6 gxf6 9 0-0!? and now: 3a) 9 'l!ltg5 10 '2! f5! � g8 ( 1 0 Axf5 1 1 f4 'l!ltg6 1 2 A h 5 is good for White) 1 1 '2! g3 with an edge for White. . . .

. . .

60 Classical Variation (3b) 9 . . �e5 (too ambitious) 10 <2\f3 �c5 ( 1 0 � a5 is better even though this is an admission that Black's last move was a mis take) 1 1 <E\d5 � c8 1 2 c3 a6 1 3 � d4 .fJ.. g 7 1 4 b4 >WJa7 1 5 JJ.. h5 with advantage to White, Stoica Kotronias, Istanbul 1 988. (3c) 9 . . � xd4 10 >Wfxd4 >WJc5 (this is Black's best; 10 � c8 1 1 <2\d5 �c5 1 2 � d2 'Wfxc2 1 3 >WJe3 �c4 14 'l!ll\' f4 gives White a strong initiative in return for the pawn) and now Stoica gives 1 1 '/l!\'xc5 dxc5 12 ..§.. c4 as slightly better for White. I cannot see any White advantage after 1 2 e6, so I prefer 1 2 <2\d5 � c8 ( 1 2 0-0-0 1 3 il_h5 and now 13 . il_e6 14 <2\ f4 and 13 . . . Qe8 1 4 Ji;!, ad l are good for White) 1 3 � ad l and White probably does have an edge. White can also consider 1 1 'lt"d3 intending *h I and f4. 8 .fJ.. xf6 gxf6 9 0-0 There is a second possibility which, like the main line, is based on the move � f5, namely 9 l£! [5 'lt"a5 1 0 0-0 l:l: c8 1 1 � d5 'Wfd8 1 2 � de3! � e5 1 3 f4 � g6 1 4 >WJd3! h5 15 l:! ad l b5 16 a4 and White is better, Sznapik-Hawelko, Poland 1 984. 9 'lt"b6 9 e6 10 'lfl h l JJ.. e 7 1 1 f4 �xd4 1 2 >WJxd4 'lt"a5 1 3 M ad l is clearly good for White, Geller Hort, Palma de Mallorca 1 970. 10 �f5! (52) Many players would have auto matically retreated the knight to .

b3 but Kupreichik realizes that he can afford to give up his b-pawn. Although Black can organize e6 to expel the knight we have already seen in Vogt-Mascarinas above that knights on c3 and e3 can be well placed.

52 B

10 0-0-0?! e6 but it Black prepares turns out that this is too slow. He should have tried 10 . . . e6 ( 1 0 �xb2? at once fails to 1 1 � d5 and 1 2 U, b l ) l l l£J xd6 + il_xd6 1 2 'Wfxd6 >WJxb2 1 3 � fd l and now: ( 1 ) 13 . . � e5 14 �d5! exd5 1 5 >WJxf6 <E)c4 ( 1 5 M g8 1 6 B, ab l >WJc3 1 7 K xd5 loses at once) 1 6 c3 K f8 ( 1 6 � g8 1 7 JJ.. h 5) 1 7 � ab l 'Wfxe2 1 8 � e l followed by exd5 + forcing Black to play JJ.. e6, when White continues dxe6 with a winning attack. (2) 13 . . . 0-0-0 14 �d5! exd5 1 5 exd5 'lt"c3 ( 1 5 '*'e5 1 6 dxc6 '*'xd6 1 7 � xd6 _§_ xc6 1 8 l:! xf6 wins a pawn) 1 6 dxc6 'i:i;i'xc6 1 7 >WJf4 followed by J.. f3 when Black has problems with his exposed king and his weak f-pawns. (3) 13 . . . � d8! 14 t�rg3 (the .

Classical Variation 61 sacrifice 1 4 l;l ab l '*xc3 1 5 l;l xb7 flops after 1 5 jt c8 16 g e7 + * f8 ) *a3 1 5 .zl_h5 and White still has some pressure although far less than in the game. l1 a4 1 1 <E)d5 *a7 only leaves White with the problem of meeting e6. l1 In the Sicilian, Black can nor mally only contemplate castling queenside when White has also played 0-0-0, since in a race between attacks on opposite wings the missing Black c-pawn gives White a large head start. 12 a5 'Wfc7 13 <E) a4 Black's . . . <E) b4 has stopped <E)d5 by White but there are other ways to reach b6. 13 e6 13 'Wfxa5 14 c3 <E) c6 1 5 b4 'Wic7 1 6 'W/b3 gives White two free tempi and an open a-file for his attack. jt xa4 14 <E) d4 White could not prevent this exchange by playing 14 <£) b6 + * b8 1 5 <E)d4 as then 1 5 e5 followed by 'Wfxc2 would con fuse the issue.

15 M xa4 d5 Black's only chance is to find some counterplay quickly, or else he will be crushed by c3 followed by b4-b5. 16 c3 <E) c6 Naturally not 16 dxe4 1 7 cxb4 e 5 when 1 8 � b 1 threatening M c 1 wins. 17 exd5 M xd5 18 .ztf3 )ii( d6 18 liii xa5 19 <E) xc6 � xa4 20 �xa4 bxc6 21 .zt xc6 is one of those positions in which the op posite-coloured bishops increase the strength of an attack to alarm ing proportions. lli!, g8? 19 lill c4 Black overlooks the threat. 1 9 . e 5 was necessary but even then 20 �xc6 )ii( xd 1 21 � xd 1 bxc6 22 £:I xc6 gives White an ending with an extra pawn and the better posi tion, while some players might prefer 20 .� xc6 bxc6 21 'Wf g4 + and 22 LE:\ f5 . 20 'li'J a4 So simple; c6 collapses and with it Black's position. 20 � g5 21 _zt xc6 bxc6 22 LE:\ xc6 Resigns

Pelikan Variation

This line arises after 1 e4 c5 2 lE}f3 � c6 3 d4 cxd4 4 � xd4 � f6 5 <E}c3 e5. Black is willing to accept a backward d-pawn in return for active piece play and, in some variations, the two bishops. The historical background to this line is rather obscure since many players have adopted it over the years with different ideas in mind. The names of Lasker and Pelikan are associated with it, but the modern handling probably owes most to the Soviet Grandmaster Sveshnikov. We have given Pelik an's name to the whole system with 6 e5, reserving that of Sveshnikov for the 8 b5 varia tion, today considered the main line. It has gained many other adherents in recent years and is regarded as an excellent way to play for a win with Black, since unbalanced positions arise in almost every line. Very recently it has suffered a decline, probably more because of changing fashions than for any clear objec tive reason. There is a second move order by which the Pelikan can arise, namely 1 e4 c5 2 � f3 � c6 3 d4 cxd4 4 ® xd4 � f6 5 � c3 e6 (or 2 . e6 and 5 ® c6) 6 � db5 d6 7 jt f4 e5 8 jtg5, reaching the same .

position as after 1 e4 c5 2 lE}f3 � c6 3 d4 cxd4 4 � xd4 lE} f6 5 <2l c3 e5 6 � db5 d6 7 jtg5, but in one extra move. To avoid the con fusion of having two different move numbers in each position I will take the 5 e5 order as standard, although in practice the two move orders are equally com mon. Since Black is incurring stra tegic weaknesses White's most logical (and best) lines are those in which he limits his immediate am bitions to nullifying Black's piece play and only later turns his mind to the exploitation of his long term advantage. Our recommen dation for White runs 6 � db5 d6 7 ji_g5 a6 8 � a3 b5 (the less common line 8 jL e6 is con sidered in game 1 4) 9 jtxf6 gxf6 1 0 �d5 f5 ( 1 0 jtg7 is an important alternative) 1 1 jtd3. In this unbalanced position Black pits his two bishops and central pawn majority against White's control of d5 and superior pawn structure. The offside knight on a3 can be an important factor, and White usually aims to bring the knight back into the game by playing c4. All the lines of the Sveshnikov are covered in game 15.

Pelikan Variation 63 Game 1 4 Karpov-Nunn London 1 982 1 2 3 4 5 6

c5 e4 l£l c6 <2\ f3 cxd4 d4 l2) f6 l£l xd4 e5 12l c3 <2\ db5 (53)

53 B

d6 6 All Black's 6th move alterna tives give White a clear plus: (I) 6 . a6 7 <2\ d6 + jj_ xd6 8 'j!jxd6 'l!je7 9 'l!jxe7 + <2\ xe7 (9 * xe7 1 0 jt g5 <2\ b4 1 1 0-0-0 is similar) 10 jj_ g5 and White has undisputed control of d5. (2) 6 . . . jj_c5 7 jj_ e3!? (7 12l d6 + is also good) jtxe3 8 <2\ d6 + * f8 9 fxe3 'it'b 6 1 0 12l c4 'l!jc5 1 1 'j!jd6 + 'l!jxd6 1 2 l£lxd6 and f7 is about to come under heavy attack by j_c4. (3) 6 . . . j_b4 7 a3 jj_ xc3 + 8 l£l xc3 d6 9 jt g5 h6 (9 a6 1 0 l£l d 5 is also very pleasant for White) 1 0 j_ xf6 'l!jxf6 1 1 <2\ b5 and Black will lose his d-pawn for insufficient compensation. . .

(4) 6 . . . h6 (by preventing j)_g5 Black avoids the loss of control of d5 as in line I, but the move is really just too slow) 7 <2\d6 + jtxd6 8 'l!jxd6 'l!je7 9 l2l b5 (Spassky introduced this pawn sacrifice-the older lines 9 'l!jxe7 + fjxe7 1 0 b3 and 1 0 j_e3 also give White a favourable end ing) 'l!jxd6 (9 0-0 10 'l!jxe7 <2\ xe7 1 1 iE:l d6 or 1 0 b3 l£l xe4 1 1 'l!jxe7 l£lxe7 1 2 jl a3) 1 0 l2) xd6 + * e7 1 1 �f5 + frf8 1 2 b3 d5 ( 1 2 l£lxe4 1 3 jj_ a3 + *g8 and now 14 f3 or 14 iE:l d6) 1 3 jj_ a3 + * g8 1 4 exd5 �xd5 1 5 <2\ d6 I! b8 1 6 A c4 jJ_ e6 1 7 0-0-0 and White has a very pleasant position, Spassky Gheorghiu, Bath 1 973. 7 J.. g5 a6 Black must meet the threat of <2\ d5 so the only other move is 7 jte6, but then White does not need to retreat his b5 knight to the bad square a3 and can gain the advantage by 8 l2)d5 � c8 (8 jt xd5 9 exd5 <2\e7 is good for White after 1 0 � f3 or 1 0 c3 a6 1 1 'i!ita4) 9 c3 a6 1 0 l£l a3 j_ xd5 1 1 J_ xf6 gxf6 1 2 'l!jxd5 'l!ja5 1 3 jt c4, Jansa-Danek, CSSR Ch. 1 982, with a firm hold on d5 in both cases. 8 l2l a3 jJ_e6 8 b5 is the line popularized by Sveshnikov and is examined in game 1 5. Other moves are defini tely inferior: ( 1 ) 8 . . jte7 (Black commits the bishop too soon) 9 <2\c4 jt e6 (9 . . . 12l d4 1 0 j)_ xf6 J_ xf6 1 1 iE:ld5 b5 1 2 l2) cb6 � b8 13 l2) xc8 � xc8 .

64 Pelikan Variation 1 4 c3 �e6 l S a4! is very good for White, Averbakh-Korchnoi, semi-final USSR. Ch. 1 9SO, 9 . . . 0-0 1 0 � xf6 J.. xf6 1 1 � xd6 doesn't give Black enough for the pawn and 9 bS 10 �xf6 gxf6 1 1 �e3 gives White a crushing bind) 1 0 J.. xf6 gxf6 1 1 � e3 (thanks to Black's �e7 he cannot now dislodge the knight by � h6) '*l' d7 1 2 L£)cdS fol lowed by �d3 and '*l'hS, once again with a total white-squared bind. (2) 8 . . . d5 (not correct) 9 � xdS � xa3 1 0 bxa3 �a5 + 1 1 �d2 'l'!!f x d2 + 1 2 � xd2 -2J xdS 1 3 exdS � d4 1 4 il_d3 followed by 0-0 and f4, when White has two bishops in an open position and a moderately relevant extra pawn. 9 �c4 i;i, c8 (54) 9 iJ.. e7 transposes to line 1 of the last note, while the alternative 9 . -E:l d4 (9 b5 10 Jt xf6 �xf6 1 1 �e3 and 1 2 � cdS is very good for White) often leads to the knight being driven back with loss of time, e.g. 9 �d4 10 J.. xf6 gxf6 ( 1 0 *xf6 1 1 � b6 .M b8 1 2 . . .

54 w

� cdS �d8 1 3 c3 leaves Black a tempo down on Karpov-Nunn) 1 1 � e3 l! c8 1 2 itd3 ith6 ( 1 2 . hS 1 3 0-0 h4 1 4 � cdS il_g7 l S c3 � c6 1 6 •f3 is excellent for White, Bronstein-Pilnik, Moscow l 9S6) 1 3 0-0 0-0 14 � cd5 and again Black's knight is expelled by c3, Dely-Flesch, Hungary Ch. 1 96S. �xf6 10 il_xf6 With this move Black accepts a loss of time to prevent his pawn structure being damaged but as a result he is driven into a passive position. The important alterna tive 10 gxf6 is met by 1 1 � d3 'E:l e7 (11 .£l d4 1 2 � e3 trans poses to the last note while l l . l! g8?! 1 2 0-0 J.. h6 1 3 � dS f5 1 4 �hS! J.. f8 1 S � cb6 f4 1 6 � xc8 1l. g4 1 7 �xh7 l! g7 1 8 )llf x g7! 1l. xg7 19 � cb6 gave White too much for the queen in Matulovic Arnason, Zeman 1 983) 1 2 � e3 il. h6 ( 1 2 . . . �b6 1 3 0-0! �xb2 1 4 � cdS il_ xdS 1 S � xdS � xdS 1 6 exd5 �d4 1 7 � f3 'ftj e7 1 8 a4! M c7 1 9 M fd 1 .-c3 20 I:� ab l .-as 2 1 .-e4 gave White more than enough for the pawn in Mednis Lombardy, USA Ch. 1 978 - this type of pawn sacrifice in return for white-squared pressure and attacking chances occurs fre quently in the 1 0 gxf6 line) 1 3 0-0 ii_ xe3 1 4 fxe3 � b6 I S .-o hS (IS .-xb2? 16 �dS j_xdS 1 7 exdS i s bad for Black) 1 6 �dS Jt xdS 17 exdS � h6 18 � ab l .-as!? ( 1 8 . 'i! c7 1 9 c4 was good for White in Tseshkovsky. . .

. .

Pelikan Variation 65 Chandler, Minsk 1 982) 1 9 e4 f5 (19 'ltxa2? 20 'lte3 � h8 2 1 �b6 wins) and there i s n o doubt that White is better, although it is not clear by how much. In de Vrnjacka Firmian-Matulovic, Banja 1 983 White played 20 a3 'fJf8 and now 2 1 c4 or 2 1 exf5 is good for White, but in Kolosovs kaia-Suspanova, corr. 1 986 Black muddied the waters by 20 a3 f4!? 2 1 c4 �g6 22 � fc 1 'ltd8 23 b4 � h4. Therefore the simple 20 exf5 may be stronger, when White's effective bishop and Black's mis placed rook on h6 guarantee White at least a slight advantage. 1 1 �b6 It is very dangerous for White to take the pawn by 1 1 � xd6 + g d8! 1l._xd6 1 2 �xd6 when 1 2 13 'lllf c 5 � d4 1 4 1l._d3 �g5 1 5 'fJ fl � c8 1 6 '!tt b4 b5 gives Black ad equate compensation. � b8 11 �d8 12 � cd5 'i!i(rg6 1 3 ttd3 1l._e7 14 12 ttJc7 + \t1d8 15 � cd5 followed by 0-0-0 gives White a dangerous at tack. J... e7 13 c3 Attempting to develop the bish op more actively runs into trouble g6 14 �a4! 1l._ xd5 1 5 after 1 3 � xd5 1t g7 1 6 _t xa6 .ii a8 1 7 jj_ xb7 � xa4 1 8 _txc6 + with a clear plus for White. 14 1l._c4 Karpov correctly steers clear of the complications resulting from the win of a pawn by 14 � c4 0-0 1 5 � xe7 + �xe7 16 'ltxd6 �h4,

and quietly consolidates his grip on d5. White's knights are rather clumsily placed but Black's possi bilities for active play are very f5 is his limited and aiming for only constructive plan. 14 0-0 15 0-0 J.. g5 16 a4 *h8 Chekhov suggests the imagina)'!tl(e8 intending 1 7 tive 1 6 jj_d8, but after 1 7 a5 (with the queen on e8 Black can no longer �xa5) j_d8 meet this move by 18 b4 \tr h8 19 ,!,;\ a2 f5 20 exf5 J... xf5 2 1 � e3 White has the better chances. 17 �e2 (55) This move and the next are evidence of Karpov's understand ing of the position. White would like to play 1 7 a5 defending the b6 knight and thereby freeing the tangle of minor pieces, but at the moment it just allows 1 7 � xa5! 1 8 )g( xa5 .§, xd5. I t seems natural, therefore, to prepare a5 with 1 7 b4 but in Sznapik-Simic, Smederevo 1 98 1 , Black obtained active play by 1 7 b4 f5 1 8 b5 � a5 1 9 J.. d 3 g6 and equalized comfor tably. Karpov's first concern is to take the sting out of . . . f5 by preparing to answer it with exf5 and f4. For this purpose 'Wt'e2, which pins the e-pawn against the loose bishop on e6, and his next move 'fJ h l removing the king from the vul nerable diagonal are excellent preparation. Only when Black's counterplay is completely neutra-

66 Pelikan Variation lized does White return to the exploitation of his queenside space advantage and d5 control.

17 g6 I made use of Karpov's chess lesson five years later in the game Nunn-Manor, London ( Lloyds Bank) 1 987, which continued 1 7 a5!? 1 8 J,i ad l J\l.h6 (after 1 8 g6 1 9 jla2! Black should avoid 19 J;Je7 20 'i'Jii b5 and 19 jlh6 20 Qlc4 f5 2 1 exf5 gxf5 22 J;J db6 with a clear plus for White, but even his best line 1 9 f5 20 exf5 gxf5 2 1 Ql c4 f4 22 f3 is slightly better for White) 1 9 \\> h l .:2:J e7?! ( 1 9 g6 was more solid, as in Karpov-Nunn) 20 e, xe7 �xb6 (20 �xe7 21 jl xe6 fxe6 22 trb5 is very bad) 2 1 .:2l f5 and Black has no compensation for his serious weaknesses. White went on to win. 18 \\>hi ji_h6 19 b4?! Afterwards Karpov thought that this was still too soon and that 1 9 � ad I or � ae l would have been better. f5 19 gxf5 20 exf5 . . •

. . .

jl xd5 21 f4 Black hopes for salvation in the opposite-coloured bishops, but White's bishop has a fine outpost at d5 whereas Black's is rather useless. e4?! 22 .:2:J xd5 22 .:2:Je7 at once is more logical, based on the fact that White cannot win a pawn by 23 fxe5 .:2:J xd5 24 jl xd5 dxe5 25 *xe5 + because of 25 ji_g7 26 'lte6 � f6. 23 a5 ji_ g7 24 !i acl .:2:J e7 .:2:J xd5 25 � fdl 26 Axd5 'ltc7 27 � c2 'lte7 .i, bc8 28 'l!il'e3 29 c4 14. c7 White has two ways of making further progress. He can either play b5 to leave Black with an isolated pawn on the queenside, which will be hard to defend when his bishop is operating only on the kingside, or he can prepare g4 to attack Black's king. For the mo ment White is not sure which plan offers the best chances. 30 g3 � e8 31 14. g2 trf6 (56) 32 g4 This doesn't have the desired effect and it would probably have been better to try the other plan. If Black moved his forces to the queenside White could then have contemplated g4 later. fxg4 32 33 M. xg4 'ltc3 34 [4. g3 �xb4

Pe/ikan Variation 67 42 l! gh4 h6 White cannot now play 43 .!! xh6 + J.. xh6 44 !! xh6 + due to 44 'ftl'xh6 + 45 Y:llf x h6 + H h7. 43 l! g4 ,!! e5 Forced in order to meet 44 !it g6 by 44 *'xg6. 44 � gg5 H c8 (57)

If Black swaps queens Karpov gives the line 34 ytxe3 35 B, xe3 .!, ce7 36 b5! J.. h6 37 H fl I;!; f8 38 ,g xe4 � xe4 39 J.. xe4 .!. xf4 40 l! xf4 J.. xf4 4 1 j_ xb7 and wins. 35 H dg l ttb2 Stopping White's threat of 'ii!' d 4. There isn't much Black can do ytf6 36 ,!! g5 to meet the threat of ltr g4 fol �at + 37 H lg4 lowed by !it g6. �b2 + 38 ff g2 * h7? 45 ltr g4! ,!! ce7 39 ffh3 Going under without a fight. I �f6? 40 f5 !:il, xd5 46 This was the sealed move (move should have tried 45 4 1 in the game, which started with cxd5 !it c2 although White is still the 2 . . . e6 move order) and, as so winning even after this. often happens, after a long period �f8 46 H g6 of difficult defence a player's relief 47 �g5 �xf5 + at reaching the time control re Or else 48 H gxh6 + .fl_xh6 49 sults in a casual sealed move. �g6 + is the end. White obviously has considerable H xf5 48 �xf5 pressure for the pawn but after 40 49 H xg7 + ff xg7 50 H xf5 Resigns H f8 (Karpov also suggests 40 � a l threatening . . . 'ftl'fl + ) it Game 1 7 is likely that Black can draw. Short-sax Black's passive queen move gives Candidates' Match White the freedom of action he Saint John 1988 needs to mount the decisive as sault. 1 e4 c5 H ffi 41 H h5 12\ c6 2 <El f3

68 Pelikan Variation 3 d4 cxd4 4 i£) xd4 i£) f6 5 i£)c3 e5 d6 6 i£)db5 7 jlg5 a6 b5 8 i£) a3 gxf6 9 jl xf6 9 �xf6 lO �d5 �d8 1 1 c4 i£)e7 ( 1 1 b4 1 2 �a4 j_d7 1 3 i£)b5! axb5 1 4 �xa8 �xa8 1 5 J}Jc7 + wins material) 1 2 cxb5 €:l xd5 13 'l!lfxd5 ( 1 3 exd5 is also safe and good) j_e6 14 �d2 d5 1 5 bxa6 J.. xa3 1 6 J.. b5 + ! gives White a clear advantage. 10 i£)d5 (58)

f5 10 This is an already an important moment for Black. At one time 1 0 f5 was the only move played in this position, but recently the alternative lO j_ g7 has become popular. The idea behind this move is to delay Black's the matic break f5, giving first priority to removing the powerful knight from d5 by t;J e7. Theory gives no clear-cut recom mendation against this new plan, so we give a more detailed cover-

age than usual: l O J.. g7 ( 1 0 J.. e6 1 1 c3 J.. g7 1 2 i£) c2 f5 1 3 exf5 J.. xf5 1 4 i£\ e3 leaves Black a tempo down over a standard line and after 14 . . . J.. e6 1 5 �f3 0-0 1 6 j_d3 or 14 . . . J.. g6 1 5 a4 White has good chances) and now: ( 1 ) 1 1 -tt h5 ( 1 1 i£)e3 i£) e7 1 2 � h5 is the same, while 1 1 c4 0-0 1 2 cxb5 <E! d4 is very risky) i£) e7 (better than 1 1 f5 1 2 exf5 �a5 + 1 3 c3 b4 1 4 �g5! � xd5 1 5 �xg7 bxa3 1 6 �xh8 + fr d7 1 7 b3 with advantage to White, Sir ias-Ochoa de Echaguen, Havana 1 984) 1 2 �e3 f5! 1 3 exf5 e4 1 4 0-0-0 0-0 and now: ( l a) 15 f3?! �b6 16 �g5 d5 1 7 i£1 xd 5? ( l 7 f6 'l¥t' xf6 1 8 � xf6 j_ xf6 1 9 i£)xd5 /2l xd5 20 ll( xd5 J.. e6 was an improvement, but Black is at least equal) i£) xd5 1 8 !;! xd5 frh8! and Black is better, Klinger-Vaiser, Szirak 1 985. (l b) 15 h4 ( 1 5 g4!? is untested) d5!? 16 l! h3 �d6 1 7 � g3 �e5 1 8 c3 b4 1 9 i£) ac2 bxc3 20 bxc3 l! b8! (not 20 . '1!!4' xc3? 21 K d4!, nor 20 . . 12l xf5? 2 1 K xg7 + !) 2 1 12l d4 j_d7 2 2 K d2, Vakhnov Shipkov, USSR 1 987 and now 22 frh8 23 12l g4 � f4 24 f6! J.. xf6 25 i£) xf6 �xf6 26 �g5 \'4fxg5 27 hxg5 M b6 is roughly equal. This murky variation must be a good place to look for a White im provement! (2) 1 1 j_ d3 i£) e7 1 2 12l xe7 (not 1 2 '2! e3 f5! 1 3 'fllf h 5 d5!) 'fllf xe7 (59) with a branch: (2a) 13 c3 f5 14 i£) c2 0-0 ( 14 .

. . .

Pelikan Variation 69

d5?! I 5 exf5 e4 1 6 j_ e2 and 14 j_b7 I 5 121e3 fxe4 I 6 121 f5 �f6 I 7 ll.xe4 d5 I 8 J,.xd5 H d8 I 9 �g4 j_ xd5 20 121xg7 + '11v f8 2 I 121h5 are good for White, but 14 �g5 I 5 0-0 J.. b7 I 6 f3 0-0 I 7 exf5 d5 was unclear in Feigeison Malyshev, USSR I 988) I 5 0-0 d5! (not 15 . . . fxe4?! I6 J.. xe4 lit b8 1 7 121 b4, while 15 � b8 I 6 exf5! e4 I 7 Ji e I 1l. xf5 I 8 121 e3 1l. g6 I 9 121 d5 �e5 20 j_c2 f5 2 I f4 'jff e 8 22 121c7 �c6 23 121 e6 led to a slight plus for White, Dolmatov-Simic, Belgrade I 988) 1 6 exf5 e4 I 7 1l. e2 � d8 I 8 121d4 j_xd4 I 9 cxd4 j_xf5 20 �d2 j_ e6 2 I f4 f5 Draw, Psak his-Dolmatov, Klaipeda I 988. (2b) 13 c4 f5 I4 0-0 (not I 4 •d2?! j_ b 7 I 5 exf5 e4 I 6 ll.e2, Hardicsay-Vaiser, Berlin I 988 and now I 6 e3! I 7 fxe3 j_xg2 I 8 H g i �h4 + 1 9 '11v d l �xh2 20 '\1v c2 j_ e4 + 2I 'fj' b3 bxc4 + 22 121 xc4 � b8 + 2 3 '\1v a3 ll. f8 wins for Black) 0-0 (I4 j_b7!? is interesting, e.g. 15 cxb5 fxe4 I 6 bxa6 exd3 1 7 axb7 �xb7 I 8 �xd3 d 5 1 9 � ad l and now 1 9 0-0 20 �xd5 *xb2 2 I 121 c4 �xa2 22 � xe5 �xd5 23 Ji xd5 j_ xe5 . • .

• • .

. . .

. . •

should be a draw, while after 19 . . . �xb2 20 M bl •xa3 2 I � b8 + *e7 22 !;!; b7 + '\1ve6 23 Ji,i: b6 + 'fj' e7 White unwisely dec lined the draw by repetition and lost in Solozhenkin-Skvortsov, Berdiansk 1 987; 1 5 t�tf3! is proba bly the best reply) with a wide choice for White: (2b I ) 15 cxb5?! d5 and I 5 'ti!i'e2 j_ b7! I6 )1:( ad l � fd8 1 7 cxb5 ( 1 7 exf5?! bxc4 I 8 .£Jxc4 d5 1 9 121e3 e4 20 ll. b 1 � f6 is also good for Black) fxe4 1 8 j_xe4 d5 I 9 bxa6 j_xa6 20 1l.d3 1l_ b7! and Black has very good compensation for the pawn. (2b2) 15 �h5!? );;t, b8 (15 . . . bxc4 1 6 .£J xc4 d5 1 7 exd5 e4 1 8 J.. e2 and 15 . . . 'Wt'b7 1 6 l!( fe i d5 17 exd5! e4 I8 ll.fl j_xb2 I 9 E( ab i , Vogt-Chekhov, Berlin I 988, are good for White) I 6 exf5 e4 I 7 � ae l j_b7 and now: (2b2 1 ) 18 � e3 bxc4 19 1l_ xc4 ( 1 9 � h3 h6 20 ll.xc4 d5 2 I J.. xd5 j_ xd5 22 f6 � xf6 23 �xd5 li xb2 24 •xe4 !i( xa2 25 12\ c4 *e6 26 '*d3 'Wfe2 27 �d5 �e6 28 'Wtd3 *e2 29 •d5 �e6 Draw, Smagin Vaiser, Sochi 1 988) d5 20 1l_xd5 .ffi.xd5 2 I f6 �xf6 22 � xd5 l!( xb2 23 � xe4 lit d8 24 �c4 �c3 25 �xa6 •c5 26 lili, c4 •f5 was unclear in Hjartarson-Yusupov, Munich 1 988. (2b22) 18 �g4 liii fe8 1 9 cxb5 ( 1 9 f3 bxc4 20 j_xe4 � f6 was also good for Black in Kuijf-Schmitt diel, Luxembourg I 988) d5 20 bxa6 il_c6 2 I a7 !i( xb2 22 l!( c i j_d7! 2 3 j_ b5 h5! 24 f6 hxg4 25

70 Pelikan Variation fxe7 ll_ xb5 and Black is better, Bosboom-van der Wiel, Nether lands Ch. 1 988. (2b23) 18 cxb5 d5! 19 bxa6 ll.. c6 20 )ii e3 � xb2 2 1 ll_ b l l!il fb8 with a clear plus for Black, Kolotilin Shipkov, USSR 1 984. (2b3) 15 � et fxe4 ( 1 5 . . ll_ b7 1 6 exf5 e4 1 7 �g4 � fe8 1 8 cxb5 d5 19 bxa6 j_xa6 20 j_ xa6 � xa6 2 1 �b5 was good for White in Ivanovic-Ivanov, Belgrade 1 988) 1 6 j_xe4 � b8 1 7 cxb5 axb5 and now 18 �b3 j_e6 1 9 j_d5 J,i fc8 20 !!! ad l J.. f5! 2 1 j_e4?! (2 1 � c l would have been equal) j_xe4 22 � xe4 �e6 gave Black some ad vantage in Klovans-Ivanchuk, Frunze 1 988. However 18 iitd3!? is interesting ( 1 8 �c2 f5 1 9 j_d 5 + J.. e 6 20 � b4 i s also pos sible), when Nunn-Kosten, Bri tish !-hour Ch. , continued 1 8 f5 1 9 ll_d5 + ft h8 20 � xb5 e4 2 1 '(!ltb3 �e5 22 a4 j_d7 2 3 � e2 !( fc8 24 I;i d l � c5 25 ll_c4 f4? 26 � xd6! and wins. Slim evidence, but there are unexplored ideas in this line. At the moment Black's chances in the I 0 ll.. g7 line appear quite good, but a number of lines above, such as 1 b and 2b3, de serve further investigation by White players. J.. e6 1 1 ll.. d3 f4 1 2 g3! is 11 �g5 ( 1 1 also good for White) 1 2 g4! ft d8 1 3 gxf5 ll_ xf5 1 4 �e3 J.. e 6 1 5 �d2 gives White a positional ad vantage. 12 0-0 (60)

j_ g7 12 Or 1 2 j_ xd5 ( 1 2 f4?! 1 3 c4! is very good for White after 13 bxc4 1 4 j_c2! J.. g7 15 Jj_ a4 I;;i, c8 1 6 � xc4 or 13 . . . b4 1 4 l:!Ya4! j_ d 7 1 5 � b5) 1 3 exd5 � e7 1 4 �xb5 j_g7 1 5 � c3 e4 1 6 j_c4 �a5 ( 1 6 �c7 1 7 �e2! puts pressure on the a6 pawn) and now: ( 1 ) 17 jj<d2 � c8 1 8 j_b3 (not 18 .£) xe4? �b6 19 'j#g5 � g8 20 .£) d2 j_c3! 2 1 'j#xg8 + .£) xg8 22 bxc3 'j#a5 with a clear advantage for Black) j_xc3 1 9 bxc3 � xc3 20 ii(( h6 \frd7 21 g ae l ;g, cg8 22 'ii!' e 3 (22 g3 � g6 23 � c l � hg8 is unclear) 'j#xe3 23 g xe3 .£) g6 24 f3 .£) h4 25 ll_ a4 + 'lfd8 with a sharp but balanced ending, Malishauskas-Krasenkov, Vilnius 1 988. (2) 17 .£)e2! ll_ xb2 ( 1 7 �c5 is well met by the exchange sacri fice 1 8 b3!) 1 8 � b l J.. e 5 was played in Short-van der Wiel, SWIFT Brussels 1 987. Here Short played 19 � c l and although White was still slightly better Black managed to escape from his difficulties. However White could

Pelikan Variation 71 have tried other moves, for ex ample 1 9 l! b7 l! c8 20 jt b3 iJ. c7 2 1 .!;i b8 + X c8 22 .!;i xc8 + .:21 xc8 23 12:Jg3 tz:Je7 24 �h5 appears good for White, as does the sim ple 1 9 12:J g3, and if 1 9 !:! c8 then 20 �e2. f4 13 �h5 Not 1 3 0-0? 1 4 exf5 Jt xd5 1 5 f6 e4 16 fxg7 !:! e8 17 �xd5, nor 13 h6?! 14 f4! opening the position while Black's king is still in the centre. bxc4 14 c4 After 1 4 . . . b4 ( 1 4 j'J_ xd5 1 5 exd5 12:Je7 1 6 M ad 1 b4 1 7 tz:J b 1 12:J g6 1 8 g 3 with advantage to White, Vogt-Georgadze, Halle 1 978) 1 5 12:J c2 X b8 White has a variety of promising ideas, since with d5 secured he can play on the queenside with b3 and a3, or on the kingside with g3. Finally Ql e 1 -f3-h4-f5 can be awkward. The speculative 14 0-0 1 5 cxb5 12:J d4 1 6 Ql c2 Ql xb5 ( 1 6 .:2) xc2 1 7 Jt xc2 axb5 1 8 Jt b3 is a safe positional plus for White) is dubious after 17 a4 ( 1 7 Ql cb4 12:J d4 1 8 � c I is also promising) .:2)a7 1 8 a5 .:2)c6 1 9 b4 f5 20 .:2) b6 � a7 2 1 exf5 J.. f7 22 �h3 �f6 23 j_e4 12:J e7 24 � ad ! and White keeps control of d5. 0-0 15 Jtxc4 1 6 l! acl (6 1 ) tz:J e1 The position after 1 6 .!;[ ac l is the key to the whole line and is critical for the assessment of Sveshnikov's variation. Black has a wide range of options: ( I ) 16 . . . 12:Jd4?! 1 7 12:Jc2! 12:J xc2

(17 f3 18 Qlxd4 fxg2 19 12:J f5! gxfl ( ilt) + 20 fi' xfl Jt xf5 2 1 exf5 h6 22 lli!, c3! *g5 23 �xg5 hxg5 24 12:J e7 + 'lrh7 25 M h3 + Jt h6 26 f6 e4 27 Jt d5 wins for White) 1 8 � xc2 frh8 1 9 .!;i fc 1 M a7 2 0 b4 M g8 2 1 a4 with an excellent posi tion for White, Mokry-Vodichka, Decin 1 979. (2) 16 . . . H a7!? and now: (2a) 17 12:J xf4!? exf4 ( 1 7 Jt xc4? 1 8 l! xc4 12:J a5 1 9 � a4 exf4 20 !:! xa5 Jt xb2 2 1 .:2) c4 is good for White) 1 8 J.. xe6 .:2) e7! (not 1 8 fxe6 1 9 )i:l xc6) 1 9 Jt c4! (after 1 9 Jtd5 or 1 9 J.. f5 Black can take the bishop and then play 20 j_xb2, when the sacrifice 21 M b 1 Jt xa3 2 2 � b 3 Jt c 5 2 3 � h3 doesn't work because Black can advance his f-pawn and defend the second rank) 12:J g6 (not now 19 Jt xb2? 20 1! b l j_g7 2 1 1! b3 intending M h3) 20 Ji:I c2 � e7 21 ltd5 ll[ e5 22 � d 1 �h4 23 12:J c4 l! h5 24 h3 l! g5 25 fii h 1 Jte5 and Black has enough for the pawn, Dvoris-Basagic, Sibenik 1 988. (2b) 17 K fd 1 � b8!? ( 1 7 <�i h8? 1 8 .:2) xf4 exf4 1 9 Jt xe6

72 Pelikan Variation really does work because f7 hangs after l£) e7) 1 8 b3 ( 1 8 <2} x f4 exf4 1 9 jt xe6 l£) e7 is still unclear) <2} d4 1 9 <2}c2 1£\ b5 and now 20 l£) cb4! appears good for White. (3) 16 'ltfh8 1 7 � fd l l£) d4 1 8 L2! c2 is slightly better for White. (4) 16 �b8 (the main line) 1 7 b3 (62) and now: . • .

. . .

gua 1 988, and now 24 �d2! f3 (24 l£)xf5 25 �xd5 �xd5 26 ]txd5 e3 27 l£)e4 is very good for White) reaches the critical posi tion. Black has some compensa tion for the pawn after 25 l£) xe4 �xd2 26 l£\xd2 fxg2 27 'fpxg2 l£\xf5, but 25 g4! appears quite unpleasant for Black (25 ]th6 26 �d4). (4c) 17 . �d7 18 � fd l <�ih8 (not 18 ]t g4? 19 �g5, while 18 . . . -2l d4 19 1£\ c2 1£\ xc2 20 .liii xc2 *h8 2 1 � d3 ]tg4 22 �h4 f5 23 f3 fxe4 24 fxe4 is a little better for White) 1 9 �h4 and now: (4c 1 ) 19 f5!? 20 l£) xf4 exf4 2 1 j_ xe6 ·Wf xe6 22 !! xc6 fxe4 23 � cxd6 �e8 reaching a very sharp position. It is perhaps surprising that nobody has repeated 1 9 f5!?, because i n the only game played, Stanciu-Brkovic, Pernik 1 983, the continuation 24 l£\c4 e3 25 fxe3 (not 25 � e l jtc3) fxe3 26 � e l �fl! 27 l£) xe3 iiiK be8 28 l£\c4! � e4! 29 �g3 was satisfac tory for Black since he could now have forced a draw by 29 . ]td4 + 30 .!;it xd4 .14 xd4 3 1 'l!l!fe5 + �g7 32 lltxg7 + * xg7. It seems to me that it was not necessary to play l£\ c4 at once, and White should first try to defuse the ad vance of the e-pawn. Therefore 24 �g4 is logical, and after 24 . . . e3 simply 25 �f3 (25 �e2!?), with l£\c4 coming next move, and Black has to prove that he has enough for the pawn. In any case no other players have been con vinced by Black's sacrifice. . .

. . .

(4a) 17 <�i h8 1 8 l£)xf4!? exf4 1 9 Jlxe6 l£) d4 20 ]tf5! l£) xf5 2 1 �xf5 ]t b 2 2 2 l£\c4 ]txc l 23 M xc 1 is a very promtsmg exchange sacrifice, while 17 . Jl xd5 1 8 ]t xd5 o2l b4 1 9 .!! fd l l£\ xa2 20 �c6, with l£\c4 to come, leaves Black very badly placed. (4b) 17 �a5 (a recent idea) 1 8 1£\ b l !? ( 1 8 l£) c2 �xa2 should lead to a draw after 19 �h4 h6! 20 l£) f6 + ]txf6 2 1 '*'xf6 ]t xc4 22 �xh6 ]txfl 23 �g5 + ) 'fp h8 (not 18 � xa2? 1 9 l£) bc3 �a5 20 ll;i a l lltc5 2 1 :l;ii, xa6) 19 l£) bc3 f5 20 l£\ e7 <2} xe7 2 1 ]t xe6 :�;!( f6! (not 21 fxe4? 22 l£) xe4 d5 23 ll! c5 lit b4 24 1£\ g5 h6 25 K c7) 22 exf5 e4! 23 llte2! (not 23 l£)xe4? .!! xe6 24 1£) g5 � xf5 and Black wins) d5, Estevez-Timoshchenko, Mana• . .

• . .

Pelikan Variation 73 (4c2) 19 Jt xd5 20 J.xd5! (stronger than the old continua tion 20 g xd5) � b4 (after 20 . . . €)d4 2 1 l!l c4! both 21 . . f3 22 �c2 �e2 + 23 fl fl � f4 24 gxf3 and 21 . . . f5 2 2 � c2 fxe4 23 � xd4 exd4 24 Jt xe4, Svesh nikov-Vyzmanavin, Moscow 1 1 1 987, are good for White) and now: (4c2 1 ) 21 l!l d2 f5 22 � c4 (22 � h3 should be met by 22 . . . � xd5 23 li;ii xd5 � b7 24 'lifd3 fxe4 25 'itt xe4 f3 rather than 22 . . . 'life7 23 exf5 e4 24 � c4 g f6 25 li;ii e l e3 26 fxe3 M h6 27 'li!'g4 J. c3 28 litxf4 Zil f6 29 �ed I it xd2 30 !:ll xd2 with advantage for White, Klovans-Vyzmanavin, USSR 1 987) � xd5?! (22 .1i b5 23 !:ll cd I �xd5 is better) 23 �xd5 fxe4 24 !:ll xd6 �f5 25 h3! h5 26 �e7 e3 27 fxe3 fxe3 28 li d7! and White stands well, Renet-Korch noi, Lugano 1 988. (4c22) 21 Ji c3!? � xa2 (after 2 1 . . f5 22 jj, h3 1t f6 23 ii_ e6! both 23 . . . 'Wfe7 24 �h5 fxe4 25 � c4 d5 26 �d6 and 23 . . . 1t xh4 24 ii_ xd7 are good for White) 22 l;i!, h3 h6 23 � c4 � bd8 24 g4! and White has dangerous attacking chances since if Black exchanges on g3 the route is open for White's knight to move to f5 . 17 � fdl There is an interesting alterna tive in 1 7 � c7!? Jt xc4 ( 1 7 �xc7 1 8 1t xe6 'ii!' b 7 1 9 1tb3! is good for White since 19 . . . �xe4 loses to 20 Jt c2) 1 8 � xc4 � c8 1 9 .

I! fc I ; perhaps 19 . . . 'lifd7 is the best reply. 17 K c8 18 �xe7 + 'lifxe7 (63) Vp to this point the game has followed Matanovic-Sax, Buenos Aires 1 978, which continued 1 9 �e2 *h8 20 1txa6 f5! with dan gerous counterplay. Black went on to win and later it was sug gested that 20 .I c2 followed by doubling rooks on the c-file would have been good for White. Short prefers to double rooks on the d file with his queen still actively posted on h5.

19 .li c3! Black has no immediate threats, so White can afford to take time out to prepare b3 fol lowed by }i! cd3. Black must aim for f5, since this provides the only possible counterplay to off set the backward d-pawn. The immediate 1 9 f5 fails to 20 exf5 . 19 'fih8 Short's innovation caused Sax to use a lot of time over this and his next few moves, so that before long White was an hour ahead on

74 Pelikan Variation the clock. 1 9 !K c6? 20 A xe6 � xc3 2 1 A f5 wins. f5 20 b3 20 d5 2 1 exd5 A xd5 22 !K xd5 �xa3 fails after 23 Ad3 e4 (23 . . . h6 24 � f5 and 23 . . . f5 24 A xf5 are even worse) 24 A xe4 h6 25 �xc8 jj( xc8 26 h3. Black could have tried 20 A d7, but then White can keep the advantage by 2 1 12J b I ! (intending ll cd3 and 12Jc3) f5 22 I! cd3 fxe4 23 !! xd6 followed by 12Jc3. h6 21 l!l( h3 �xe6 22 A xe6 23 I! hd3 .!t cd8 Although this looks passive it is the best way to defend the d pawn, for example 23 � c6 24 �f3 or 23 . . . fxe4 24 liil xd6 'Wte7 25 12J c4 with a clear plus for White in both cases. There is little point leaving a rook on the c-file when it will soon be blocked by 12Jc4. 24 �e2! It is easy to win a pawn, but if this involves allowing Black to advance his central pawns Black's powerful bishop may well provide enough compensation. One such line is 24 12Jc4 (24 �f3? �g6! 25 exf5 �xf5 followed by e4 is also bad) fxe4 25 liil xd6 I! xd6 26 liil xd6 �e7 27 a xa6 e3 28 fxe3 fxe3 and the threats to White's king practically force him to play 29 � xh6 + (29 �xe3 �c5 30 'Wt'e2 e4 threatening A d4 is awkward) A xh6 30 �xh6 + \fig8 3 1 'ltxe3, but 3 1 •f6 32 h3 �fl + 33 * h2 � f4 + is a likely

draw because the knight and queenside pawns cannot assume a stable defensive configuration. White correctly prefers to keep the bind and cash in later. fxe4 24 24 . �g6 25 f3 d5 (25 . . . fxe4 26 '*xe4 is good for White) 26 li xd5 a xd5 27 exd5 e4 28 12! c4 e3 may seem unclear, but White's passed pawn is just as far ad vanced as Black's and he has an extra pawn. 25 �xe4 f3 26 �c4! (64) Much better than 26 a xf3 d5 27 'i!i!'h4 (27 liil xf8 + A xf8 attacks a3) /4 fe8 followed by the advance of the central pawns, with good compensation for the pawn.

26 li!l f4 26 fxg2 27 � xd6 � xd6 28 liil xd6 '/!tff7 29 liil d2 and 30 �e3 is also very good for White. 27 �d5 'itt g4 Black cannot avoid a miserable ending. Although he now suc ceeds in playing d5 his bishop remains blocked by the e5 pawn. 28 lii[ xf3 lii[ xf3

Pelikan Variation 75 29 �xf3 �xf3 30 gxf3 d5 J.. f6 31 'll fl Despite White's doubled pawn the position should be a win since Black's d-pawn is easily block aded by White's king and the bishop will be impeded by the central pawns. Moreover White has a powerful queenside pawn majority. 32 i£jb6 d4 33 '81e2 33 .!;l c l ! was more accurate, activating the rook before Black can play j_g5. Then 34 I! c8 is a threat, and after 33 "'/ g7 34 � c7 + <1Jg6 35 � d7 White gets his rook behind Black's passed pawn and follows up with *e2d3. However the move played should also be sufficient to win. 33 j}_ g5 J.. f4 34 i£lc4 � g8 35 h3 Black will have to play this sooner or later, for otherwise White plays *d3 and � g l , seiz ing the open file for his own use. 36 b4 X g2 The rook must move up the g file to allow the Black king to cross, and at g2 it delays White's ltrd3. * g7 37 a4 axb5 38 b5 *f6 39 axb5 40 b6 <1Je6 41 b7 H g8 El 118 (65) 42 � b1 The win still requires some work. White first secures the ad-

vanced pawn and moves his king up. Black is paralysed and can only adopt a waiting strategy. 43 )!i b5 '1fl> d7 44 l£l a5 wc7 White's only worry is that Black might try to liquidate to a rook and pawn ending by j_ g5--d8xa5. This might prove awkward to win as White would be left with only f- and h-pawns. The immediate 44 j_ g5 45 � xe5 J.. d 8 fails to 46 lii( d5 + w e6 47 lii( xd8 . '1id6 45 '1id3 46 h4 Definitely ruling out . . . j_ g5. 46 if!J c7 47 l!i b2 j_ h2 48 we4 White's plan is to play §l c2 + at a moment when Black must reply \fld7 to prevent � c8. Then White will seize the g-file by � c l -g l . Finally the penetration of the rook combined with the advance of the White king to c4 and b5 will decide the game. The immediate 48 El c2 + fails to 48 . . . \fl b6! 49 )ii c8 * xa5, so White

76 Pelikan Variation must blockade the e-pawn before starting his plan. 48 J.. f4 Black can only delay the end. This move covers cl to prevent � c l -g l . 49 .!i\ c2 + *d7 Or 49 * b6 50 � c8 Jii,( xb7 5 1 '-£)xb7 * xb7 5 2 � f8 (threat )jj[ xf4) j_ h2 53 � f6 and wins. 50 .d3! Black is in zugzwang. The rook cannot move because of '-£)c6, the king cannot move because of � c8 and . . h5 doesn't help.

J.. h2 50 J.. f4 51 la cl The same again! Black has to free g l . *d6 52 � gl Resigns 53 *c4 The finish might be 53 � f8 54 '-£) b3 (if White gets the chance to play <El b3--c5 the winning pro cess is simplified) la b8 55 IE\c5 l;ii, h8 56 � g6 + *c7 57 *b5 d3 58 � c6 + wb8 59 . b6 mating, or 53 *c7 54 � g6 la h8 55 * c5 d3 56 � g7 + * b8 57 *b6 with a similar conclusion. . . .

. . .

Dragon Variation

The Dragon i s one o f the most controversial lines in the Sicilian. At various times over the years it has appeared to be in its death throes, only to be suddenly revived by the discovery of new ideas for Black. Practitioners re gard the variation as their private property, and defend it with an almost religious fervour against the many White players who wish to commit the heresy of mating Black down the h-file. The amount of opening analysis spawned by the Dragon is notori ous, but much of this theory is concerned with the many White alternatives, so although I am recommending one of the main lines for White the quantity of analysis is not excessive. The material is divided amongst three games. I call the continuation I e4 c5 2 � f3 d6 3 d4 cxd4 4 .:£J xd4 � f6 5 f:lc3 g6 6 .ii_ e 3 .ii_ g7 7 f3 0-0 8 �d2 ..£J c6 9 ,.Rc4 ,.Rd7 10 0-0-0 � c8 1 1 ,.R b3 � e5 1 2 h4 h5 1 3 jtg5 14 c5 the 'main line' Game 1 8 deals with all deviations from the main line before Black's 1 2th move, while game 1 9 covers Black's 1 2th and 1 3th move alter natives. The 'main line' itself is deferred until game 20. The Dragon is characterized by

the initial moves 1 e4 c5 2 � f3 d6 3 d4 cxd4 4 .£l xd4 � f6 5 f:l c3 g6, intending to develop Black's bish op actively at g7. If White plays quietly and castles kingside the bishop will still be useful in sup porting Black's minority attack on the queenside, while if White aims to attack the enemy king and plays 0-0-0 the Dragon bishop comes into its own. Hosts of White players have seen their queensides disintegrate under the laser-like power of the g7 bishop, supported perhaps by �a5 and g c8 to step up the pres sure on c3. The true Dragon player will analyse six exchange sacrifices on c3 before breakfast, and White players need to be con stantly on the alert for combi nations based on blowing open the long diagonal. The critical lines are those in which White tries to deliver mate by advancing his h-pawn, when Black has two possible defences. He may ignore the advance and concentrate on developing his own attack, or he may block White by playing h5. This lat ter approach allows a possible sacrificial breakthrough by g4, but this has to be very well-timed or it just loses material. General

78 /)ragon Variation

pn nci ples aren't much help in the I >ragon, since success or failure is determined by tactical considera tions.

Game 18 Kroncke-Schroder Correspondence, 1987 c5 1 e4 d6 2 � f3 3 d4 cxd4 <E)f6 4 � xd4 5 � c3 g6 6 J._e3 If White intends castling queen side this move is the most usual. 6 J._e2 only fits in with 0-0, since the bishop is usually better placed at fl or c4 in the more aggressive lines resulting from castling on opposite wings. J._g7 6 The Dragon differs from many other Sicilian systems in that Black often omits the typical move a6. Time is of particular importance in the Dragon and Black simply cannot afford the tempo spent on preparing b5, which can often be played without . . . a6 in case White castles queen side. The idea of playing a6 and b5 before castling has been tried, but after 6 a6 (6 . . . Q) g4? loses material after 7 J. b5 + ) 7 f3 ..f} bd7 8 'i'itd2 b5 9 a4! bxa4 (9 b4 1 0 ..f}d5 is also very good for White) 10 � xa4 J.g7 l l J._e2 0-0 1 2 0-0 Qlc5 1 3 � a3 J._b7 1 4 � fa ! �c8 15 Ql b3 White had strong queenside pressure in Kavalek-Bilek, Sousse IZ 1 967.

7 f3 (66) This is more or less forced as 7 J._c4 and 7 'if(d2 can both be met by 7 ..f} g4. 66 B

7 0-0 Or: (I) 7 . . . a6 (7 Q� bd7 exerts no pressure on d4 so simply 8 'if(d2 followed by 0-0-0, J. c4 and J._h6 gives White a strong attack) 8 �d2 Ql bd7 (8 b5 9 a4! is similar to Kavalek-Bilek above) 9 J.h6 (9 0-0-0 is also good) J._ xh6 (9 0-0 is suicidal since White's h4-h5 attack is much stronger than usual) I0 'if(xh6 b5 1 1 0-0-0 J. b7 1 2 <fi b l , Bastrikov Hasin, USSR 1 96 1 , and Black's king is stuck in the centre. (2) 7 . . . Q) c6 (this can trans pose to the main line if Black castles quickly, so we only explore lines in which Black plays .�. d7 and . . . )ji( c8 before . . . 0-0) 8 �d2 J._d7 9 0-0-0 lii( c8 10 g4 and now: (2a) 10 . . . Q) e5 1 1 h4 h5 1 2 g5 Q�h7 1 3 f4 <E) g4 and now P. Lit tlewood-Mestel, London 1 978, continued 14 ..§.. g 1 0-0 1 5 'fi b ] e5

Dragon Variation 79 1 6 Q)f3 �e6 1 7 fxe5 2£) xe5 1 8 Ql xe5 �xe5 1 9 .t) b5 � e8 and Black had no problems. However White can improve by playing f5 at some point, for example 1 5 f5 or, more provocatively, 14 f5!? since although White's dark squared bishop disappears after 14 Qlxe3 1 5 'lifxe3 0-0, it isn't easy to see a constructive plan for Black after 1 6 �h3. (2b) 10 . . . 0-0 1 1 h4 and now: (2b l ) 11 b5 1 2 <2)cxb5 ..'2l e5 1 3 h5 2l xf3 (once Black has started sacrificing he must con tinue) 1 4 Qlxf3 � xg4 1 5 '/Wg2 ( 1 5 � e2 Ql xe4 1 6 t4r e 1 ! would be a safe answer to Black's aggression) '!ifa5 1 6 a3 � xc2 + ! 1 7 lf!' xc2 � a4 + 1 8 ltfd2 'l\il' b3 1 9 Qlc3 'lf1xb2 + 20 lf!' d3 �xa3 with hor rendous complications, Mestei Christiansen, Hastings 1 978-9. This remarkable game concluded 2 1 � c l (2 1 � c l was probably better) �b4 22 �d2 I:!l c8 23 hxg6 hxg6 (23 h5!? reserving h6 for the bishop was possible) 24 � h4 � xc3 + 25 � xc3 � xe4 + 26 ltr d2 .£) d5? (26 � xf3 27 � xe4 <2l xe4 + 28 lf!' e3 � xg2 29 � xg7 'l!; xg7 30 � xg2 Q)c5 was best and the strange ending resulting looks to be better for White) 27 � xg7 'ltt e3 + 28 'ltr c2 �xf3 29 � b2! e5 30 �d2 'ltt c 5 + 3 1 \tr b l � e3 32 � c l �b6 33 '*'h2 Ql f5 34 � c8 + ltf g7 3 5 � b4 Resigns. (2b2) 11 . . �a5 1 2 'l!l b l ( 1 2 h5 � xc3 allows 1 2 <tJ b4 with to follow) � fd8? ( 1 2 . �xd4 1 3 j_ xd4 � fd8 was better according . . .

to Hiibner, who assesses the resulting position as slightly better for White) 13 .£) b3 l'lfc7 1 4 h 5 jl e6 1 5 hxg6 fxg6, Hiibner Hort, match 1 979 and now 1 6 Ql d5! would have given White a clear plus according to H iibner. (2b3) 11 . . h5 1 2 � e2 (it is by no means certain that this is White's best. Ligterink-Sosonko, Dutch Ch. 1 978 went 12 gxh5 Ql xh5 13 I:!l g l lf!'h7 14 � e2 Ql xd4 1 5 jj_ xd4 �h6 1 6 �e3 �g7 1 7 � d4 JLh6 Draw, but 1 4 ltf b l avoiding the draw may be good for White) ..'2le5 1 3 gxh5 �xh5 1 4 � d g I ( 1 4 B\ hg I i s a n alternative, but whether better or worse is not easy to decide) �c4 1 5 � xc4 (a defect of 1 2 � e2 is that this cap ture will involve a loss of a tempo) � xc4 1 6 l!l g5 (threatening 1 7 � xh5 gxh5 1 8 � g l ) � c5 1 7 ..'2ld5 e6 18 I:!l xh5 exd5 19 � xd5 � xd5 20 exd5 and White was a little better in Speelman-Liu Wenzhe, China 1 98 1 . � c6 8 ·�W d2 Or 8 d5 (other moves meet with common-sense replies, e.g. 8 . . . jj_ e6 9 <-2l xe6 fxe6 1 0 e5 or 8 . . . a6 9 0-0-0 b5 1 0 h4) 9 e5 Qle8 (9 .£) fd7 1 0 f4 Ql b6 1 1 � e2 Qlc6, Popovic-Sax, Vrsac 1 98 1 , and now 1 2 O-O-O jj_d7 1 3 .£) b3 looks unpleasant for Black since 1 3 e6 1 4 h4 gives White a very dan gerous attack) 10 f4 f6 1 1 0-0-0 fxe5 ( 1 1 Ql c6 1 2 � f3) 1 2 fxe5 <2)c6 ( 1 2 � xe5 1 3 lt:l f3 either regains the pawn with advantage or, after 1 3 . . � xc3 1 4 '*xc3 e6 .

80 Dragon Variation 1 5 h4, gives White an enormous attack) 1 3 �0 jtg4 ( 1 3 e6 1 4 jth6 leaves Black with a very bad bishop) 1 4 � xd5 !;'!. xO 1 5 gxO jt xO 1 6 jt g2 jt xd l 1 7 � xd l jt xe5 1 8 A c5 e6 1 9 � e7 + f!Jg7 20 � xc6 �xd2 + 2 1 !;'!. xd2 bxc6 22 !;'!. d7 + with a good ending for White. 9 Ac4 (67) 67 B

better, Hardicsay-Honfi, Buda pest 1 977. (3) 9 . . . � xd4 (bringing White's bishop to d4 reduces the power of the Dragon bishop on g7) 10 jtxd4 jt e6 1 1 jt b3 �a5 1 2 0-0-0 It fc8 ( 1 2 jtxb3 1 3 cxb3 � fd8 14 f!J b l s d7 1 5 g4 � ad8 b5 16 �e2 jth8 1 7 h4 and 12 1 3 f!J b l b4 14 €ld5 J..xd5 1 5 exd5 �b5 1 6 �d3 �b7 1 7 Ji:K he l a5 1 8 J.. a4 give White some advantage) 1 3 ltr b l E:i, c6 ( 1 3 b5 1 4 � he t _1xb3 1 5 cxb3 b4 1 6 1i_ xf6 bxc3 1 7 J.. xc3 _R xc3 1 8 bxc3 � xc3 1 9 ii;; e3 Jii( ac8 20 � xc3 �xc3 2 1 �xc3 Ji;i\ xc3 22 � c 1 is a fairly typical line in this variation; White's outside passed pawn gives him a big advantage) 1 4 h4 jt xb3 1 5 cxb3 b5 and now both 1 6 a3 and 16 h5 are good for White. (4) 9 '\'ita5 may be met by 1 0 J..b3, and after 1 0 1i_d7 1 1 0-0-0 we reach the main line of this game. Note that the tempting 10 � b3 �b4 1 1 J..e2 may not be good, e.g. 1 1 . . . �a5 1 2 a3!? �xb3 1 3 cxb3 '®'xb3 14 E e l jte6 1 5 ii.d l �c4 1 6 <2ld5 1i_ xd5 1 7 � xc4 jtxc4 1 8 1i_ xa7! <2l xe4! with a satisfactory position for Black. (5) 9 <E� d7 (the most import ant 9th move alternative) 1 0 h4 (68) and now: (5a) 10 . h5 (a new idea of Larsen's) 1 1 J.. b3 (the flexible 1 1 0-0-0 may be more accurate with this move order) <E:lc5 ( 1 1 � b6 12 0-0-0 <2l a5 1 3 �d3 J.. d 7 14 g4 hxg4 1 5 h5 � c8 1 6 hxg6 <2l xb3 + 1 7 axb3 fxg6 1 8 e5! dxe5 1 9 �xg6 • • •

9 jtd7 This is by far the most common in practice, but a number of other moves have been tried: ( I ) 9 . a6 (too slow) 1 0 jtb3 � a5 1 1 h4 b5 12 h5 �e5 ( 1 0 � xb3 13 axb3 j_d7 1 4 J.h6 A xh6 1 5 �xh6 e6 1 6 0-0-0 b4 1 7 � a2 � a5 1 8 * b l is good for White) 1 1 h4 b5 1 2 h5 � xh5 1 3 Ah6 e6 1 4 jl xg7 * xg7 1 5 g4 � f6 1 6 0-0-0 with a dangerous attack, Yurtaev-Shur, USSR 1 976. (2) 9 . a5 (this allows White to block the queenside) 1 0 J..b3 jtd7 1 1 a4 � xd4 1 2 1i_ xd4 1i_ e6 1 3 0-0-0 �d7 1 4 * b l !;'!. fc8 1 5 � he l a a6 1 6 �e2 � ac6 1 7 �d5 �e8 18 �b5 and White is clearly . .

. . .

Dragon Variation 81 68 B

still has some advantage) 2 1 .£ld5 J,. b5 22 f4! 'l!o/d8 23 ttf3 J,. c6 24 f5 f6 25 fxg6 j}_xd5 26 � h7 + 'ii' g8 27 'l!o/h5 Resigns, Fleck Kozul, Cesme 1 983. (5b2) 16 !JJ.. g5 .£lxb3 + 17 cxb3 l#a5 1 8 jlxe7 �e5 1 9 'i'!\"c2! and Black has a depressing choice: (5b2 1 ) 19 � xc3 20 bxc3 l;!i c8 21 'ii' b2 �b6 22 J,.g5 � c4 + 23 'ii' a l �a3 24 �b2 � b5 25 .£j xb5 'l!o/xb5 26 .§te3! with a clear plus for White, Wedberg Sosonko, Haninge 1 988. (5b22) 19 b5? 20 1ti' b 1 � fe8 2 1 <2!d5! )4 xc2 22 � f6 + J.. xf6 23 .§txf6 wins. (5b23) 19 <21 c6 20 .£l xc6 J;g xc6 2 1 '\ff b l � fc8 22 ii, d5! �b6 23 'i!Wd2 � xc3 24 bxc3 g xc3 25 � xd6 � xb3 + 26 axb3 ttxb3 + 27 lfi' c l •a3 + 28 'll' d l � a l + 29 life ) "fio\'a4 + 30 fj>e2 J,.b5 + 3 1 '\\' f2 .§td4 + 3 2 '1fl g3 J.. e5 + 3 3 f4 and wins. (5b24) 19 � fe8 20 J,.xd6 b5 21 b4 tta6 22 � b3 ..te6 23 .£l c5 � xc5 24 §t xc5 J,. xa2? 25 l;l: d6 Resigns, Kruppa-Savchenko, USSR 1 984. 10 0-0-0 (69) There are arguments in favour of the alternative move order with 1 0 h4, but we shall adopt the most usual sequence in which White castles first. 10 � a5 When the Dragon was revived in the 1 960s, this was the most popular line for Black. It frees the f8 rook to come over to c8 so that both Black rooks may participate . . .

exd4 20 l;!l h7 l;!l t7 2 1 � dh l wins) 1 2 0-0-0 .£lxb3 + 1 3 axb3 �a5 14 'iti>bl jld7 15 g4 hxg4 16 h5 gxh5?? (a blunder; Sherzer suggests 1 6 .:2) xd4 1 7 jl xd4 jl xd4 1 8 �xd4 g5, but the finesse 1 7 .£ld5! '14j'd8 1 8 J,. xd4 appears good for White) 1 7 .£l xc6 bxc6 1 8 .£l d 5 �d8 1 9 J,.h6 f6 20 � xh5 cxd5 2 1 §t xg7 *f7 22 l\1 xf8 yj,fxf8 23 �xd5 + Resigns, Sherzer-Lar sen, New York Open 1 988. (5b) 10 � b6 1 1 J,.b3 .£l a5 (or 1 1 .£je5 1 2 �e2) 12 ·� d3 J,.d7 1 3 h5 � c8 14 hxg6 hxg6 1 5 0-0-0 Ql bc4 and White has two promising lines: (5b l ) 16 J,.h6 e5 ( 1 6 J,.xh6? 1 7 l;!l xh6 e5 18 ii:ll dh l '1Wg5 + 1 9 \fr b l exd4 20 .£ld5 f5 2 1 �xd4 � xb3 22 axb3 � d2 + 23 �xd2 ytxd2 24 l;ll, h8 + 'iti>f7 25 g l h7 + *e8 26 l;!t e7 + mates) 1 7 !JJ.. xg7 '1J xg7 1 8 .£l de2 �g5 + 19 'iti b l b5 (19 )'!lt xg2? 20 J,.xc4 �xc4 2 1 b3 i s very good for White) 20 g3 b4? (20 � h8 21 � xh8 � xh8 22 a4! a6 23 J,. xc4 �xc4 24 b3 � a3 + 25 'ltl b2 b4 26 .£ld5 is Black's best chance, but White • . •

. . .

H2 DraKon Variation (2a) 11 . . b5 ( 1 1 a5? 1 2 A. h6) 1 2 A. d5! H c8 1 3 \fr b 1 ( 1 3 A.xc6 _txc6 1 4 h5 b4 1 5 � ce2 � xh5 1 6 !ll xh5 gxh5 17 � f5 '!J, c7 1 8 A.h6 was also very good for White in Wahls-Ristic, Dort mund 11 1 989) b4 14 �ce2 � xd5 <E)c3 + 15 <E) xc6 g xc6 ( 1 5 loses to 1 6 � xc3 bxc3 1 7 � xb8 )4 axb8 1 8 'lf(d5 � xb2 + 1 9 \fr a 1 � e6 20 �a5) 1 6 exd5 )!! a6 1 7 A.d4 with a n excellent position for White, Mikhalchishin-Sirov, USSR 1 988. (2b) 11 . Jiii c8 12 � b3 a5 ( 1 2 b5 1 3 h5) and White has two tempting lines: (2b l ) 13 a4 � xd4 14 A. xd4 b5 1 5 <E)d5 e5 16 jt xe5 bxa4 (after 1 6 dxe5 1 7 �xf6 + A. xf6 1 8 �xd7 l;ll a7 1 9 �d6 the only ques tion is whether the extra pawn is enough to win) 1 7 A.xf6! axb3 ( 1 7 '!ilfxb3 1 8 �e7 + \fr f8 1 9 � xc8 � xc8 20 A. xg7 + <1Jxg7 2 1 �d3 wins) 1 8 c3! g c4 (there is nothing better as White already has a deci sive attack) 1 9 � xg7 <1J xg7 20 h5 .l!l( a4 2 1 \frbl � b5 22 �g5! D' Adamo-Teipelke, Resigns, corr. 1 986. (2b2) 13 h5 a4 ( 1 3 � xd4 1 4 A.xd4 a4 1 5 A. d 5 e 5 1 6 hxg6! is crushing) 14 A.d5 .:2:J xh5 (or 14 . . e6 1 5 hxg6! exd5 1 6 A. h6! fxg6 1 7 A. xg7 f:jxg7 1 8 �h6 + <l! f7 1 9 <E) xd5 with a very strong attack) 1 5 g4 � f6 1 6 � f5 A. xf5 1 7 gxf5 � b4 1 8 fxg6! � fxd5 1 9 A_ d4! (a very nice move) A. xd4 (the lines 19 . fxg6 20 exd5, 19 . <E) f6 20 gxh7 + fJ h8 2 1 .l!lt dg 1 and 19 . . . .

in the attack. It is unpopular at the moment both because of very bad practical results for Black and because of the depressing theoretical situation detailed be low. It becomes clear that the time Black spends moving his queen to a5 only serves to present White with an important tempo by � b3 in the main line. 10 . . . l;ll c8 1 1 Ji_ b3 �e5 is covered in games 1 9 and 20, and we deal with the remaining lines now: ( I ) 10 'liif c7 1 1 �b3 M fc8 (1 1 �e5 1 2 A. h6) 1 2 h4 �e5 13 h5 � c4 1 4 A.xc4 't¥1xc4 1 5 A. h6 .:&h8 1 6 A_g5 b5 1 7 hxg6 fxg6 1 8 j).xf6 A. xf6 1 9 -t)d5 .�xd4 ( 1 9 �xa2 20 'liif h6! gives White a crushing attack) 20 � xe7 + <l!f7 (20 '\ti' f8 2 1 � xc8 .l!l( xc8 22 c3 is very good for White) 2 1 � xh7 + A_g7 2 2 c 3 � h 8 (White plays �f4 + after most moves) 23 'i!irf4 + A. f5 24 �xf5 .l!l( xh7 25 � xd6 + \fr g8 26 � xc4 A.h6 27 �xh6 � xh6 28 � e3 and with three solid pawns for the exchange White has good winning chances in the ending. (2) 10 . . . 'jjl( b8 1 1 h4 and now: . . .

. .

Dragon Variation 83 e5 20 gxh7 + *h8 2 1 Iil dg l f6 22 exd5! exd4 23 � xg7 'l!/ xg7 24 �h6 + 'I!Jf7 25 � g l all win for White) 20 �xd4 e5 2 1 gxf7 + *fll (21 *xf7 22 liii, xh7 + <8i g8 23 �gl + *xh7 24 �g2 wins) 22 � g l -EJ f6 23 � h6! -EJ xa2 + 24 -EJxa2 �c7 25 -E:l b4 wr xf7 26 � xd6 -EJe8 27 .i;;!. d3 and although material is equal, Black 's king is horribly exposed, Agdestein Karlsson, Gausdal 1 987. Iil fc8 1 l j}_b3 12 *b1 -EJe5 12 � ab8 (a bit slow, so White can afford to spend time preparing h4 with g4, which cuts out the reply h5) 1 3 g4 b5 14 h4 b4 1 5 -EJd5 -EJ xd5 16 exd5 cZ) xd4 1 7 j}_xd4 j}_ xd4 1 8 "tllf xd4 �c5 1 9 �d2 a5 20 h5 a4 2 1 hxg6 axb3 22 gxf7 + * ffi 23 "i!l.f h6 + *xf7 24 �h5 + * f6 25 g5 + Resigns, Schmidt-Bobotsov, Varna 1 964. 13 h4 .f) c4 After l 3 b5 1 4 -E:lcxb5 �xd2 1 5 A xd2 it is hard to believe that Black can have enough for the pawn. Kuzmin-Garcia Martinez, Hastings 1 973/4 continued 1 5 f)c4 1 6 Q)c3 Q)xe4 1 7 .f)xe4 j}_xd4 1 8 Ag5 Ae5 1 9 A xe7 -EJ xb2 20 � xd6! j}_ e8 2 1 � d5 with a clear plus for White. 14 A xc4 Iil xc4 15 QJ b3 ( 70) "*c7 This appears to be the best square: ( 1 ) 15 . . . �e5 (a recent idea, but there is an obvious danger of the queen being trapped) 16 Ad4

'i!iit e6 1 7 g4 ( 1 7 h5 a5 1 8 h6 Ah8 1 9 Q} d 5 a4 20 j}_xf6 j}_ xf6 2 1 cZ) xf6 + 'il!i'xf6 22 QJd4 !;!t ac8 23 � he t a3 was unclear in Hebden-Jonsson, London 1 988) � ac8 1 8 lit he 1 Ac6 1 9 Q)d5 Aa4 20 c3 b5 2 1 'lif'h2 and White has the advan tage, Koch-Ristic, Dortmund 1 1 1 989. (2) 15 . . . �a6 (dubious) 1 6 e5 Q)e8 ( 1 6 dxe5 17 Q)c5 "*'d6 1 8 "*'e2!) 1 7 h 5 ( 1 7 �d5 is also strong) � ac8 ( 1 7 J.. xe5 1 8 hxg6 followed by A d4 i s good for White) 1 8 hxg6 fxg6 (or 1 8 hxg6 1 9 A h6 A xeS 20 .!21d5 with a strong attack) 1 9 A d4 with a distinct plus for White. (3) 15 . . . "*'d8 (also poor) 1 6 Ah6 J.. h 8 (perhaps 1 6 � ffi is relatively best, but after 1 7 j}_ xg7 "*'xg7 1 8 g4 Black has no real counterplay) 1 7 h5 .i, ac8 1 8 hxg6 fxg6 1 9 e5 .!£1e8 20 exd6 e6 2 1 Ag5 and Black is in a mess, Ciocaltea Drimer, Romania 1 968. 16 Ad4 One should also take note of a promising alternative for White here, namely 1 6 g4 and now 1 6 . . .

84 Dragon Variation � e6 1 7 h5 a5 ( 1 7 Or 20 !i, c8 (20 �xb3 2 1 �c8 1 8 <21 b5 �d8 19 <2! 5d4 � d7 20 hxg6 fxg6 cxb3 e5 22 J.. e3 H xb3 23 H c l 21 e5 <2! e8 22 * h2 was very good *d7 24 �h6 � h8 25 ll. g5 ll.g7 for White in Bangiev-Ovhinikov, 26 l:![ ch l l:![ xc3 27 �xc3 was ex USSR 1 976) 1 8 hxg6 fxg6 1 9 � d4 cellent for White, Hartston-Kol a4 20 � xf6 exf6 (20 . . . �xf6 2 1 bek, Dresden 1 969) 2 1 g dh I .'£) d5 � xd5 22 �xd5 + <l/ g7 23 � xb3 22 cxb3 e5 23 J.. e 3 .i. xb3 !i, xh7 + wins) 21 <2! d5 �d8 22 24 � b5 �d8 25 J.. h6 �f8 26 <2! d4 J.. f7 23 � h2 h6 24 <2! e3 left J.. xf8 <l/xf8 27 � h8 + <2!g8 28 Black with a miserable position in i'i\ xg8 + Resigns, Savic-Simic, Gonzales-Schroder, corr. 1 987. corr. 1 979. J.. e 6 16 21 <2! xb5 � b7 The theoretical main line con Or 1 6 . . . .&_ c6 1 7 h5 a5 1 8 hxg6 �c4 22 <2!c3 � ab8 hxg6 1 9 a4 e6 (Black wants to tinues 21 take on a4 without allowing J.. xf6 and although this is good for followed by .'£) d5, but this is too White, it turns out that 22 M dh I ! slow) 20 M h4 � xa4 2 1 !i dh l is even better. Black cannot reply M xa4 because of 23 � xd6 Jl xd4 22 <2! xd4 J.. e8 23 �db5 22 with a large advantage for White, exd6 24 '*'g5 (the main threat is Wang Zili-Velimirovic, Thessalo 25 le! h8 + �xh8 26 �xh8 + , but both 25 � xf6 and 25 �xf6 are niki 01. 1 988. lesser threats) � h 5 25 � xg7 a5 17 h5 frixg7 (25 )g( c8 26 �c3) 26 hxg6 18 hxg6 ii;\ xh5 with a crushing attack. 19 a4 g b4 After 1 9 �c8 20 g4 )j, xd4 Therefore Black has nothing 2 1 <2!xd4 'llf b6 Black had some better than 22 �c8, but then counterplay in Hartston-Hollis, White has a pleasant choice. He London 1 967, but 20 g h4! looks may play 23 <2! c3, effectively gain very unpleasant for Black. ing a tempo over an already pro 20 H h4 ( 71) b5 mising line, since Black needs to double rooks on the b-file to create any threats. Alternatively 71 he may continue 23 g4 followed B by 'l!i\'h2 with an enormous attack. 22 � dh 1 The position of the queen on b7 is no better; White's attack on the h-file is just too strong. 22 j( xa4 23 <2! c3 Not 23 <21 xd6 exd6 24 '*g5 because of 24 . . li;'!, xd4.

Dragon Variation 85 23 J_xb3 Or 23 .!;1 b4 24 g4 J.. xb3 25 'f!fh2 J_xc2 + (25 J_ a2 + 26 .£) xa2) 26 * a l .£)h5 27 jJ_ xg7 'f;; xg7 (27 lll: xb2 28 lll: xh5!) 28 M. xh5! gxh5 29 'f!fxh5 f6 30 'f!fh7 + \t'f8 31 "Wtg6 mating. 24 cxb3 lll: b4 After 24 � xd4 25 'f!fxd4 'f!fxb3 26 .£)d5 e5 27 'f!fc3 White has a clear material and positional advantage. e5 25 .£) d5 26 lll: h8 + ! Resigns Because of 26 . . . J_ xh8 27 l!!f. xh8 + f'i xh8 28 'f!fh6 + .£J h7 29 !El f6 mating.

Game l9 Nunn-Mestel London (GLC) 1986 1 2 3 4 5 6 7 8 9 10 11 12

e4 .£) f3 d4 .£) xd4 .£) c3 J.. e3 f3 'f!fd2 J.. c4 0-0-0 J.. b3 h4 ( 72)

c5 d6 cxd4 L2)f6 g6 J.. g7 0--0 L2)c6 j_d7 l!!f, c8 .£) e5 h5

Or: ( I ) 12 . . . 'f!fa5 1 3 'f;; b 1 � c4 1 4 J.. xc4 l!!f. xc4 (the only difference compared to game 1 8 is that Black's rook is on f8 instead of a8, but this difference is crucial) 1 5 h5! ( 1 5 LEJ b3, as i n game 1 8, is still promising, but this may be even better) � xc3 (when the rook is on

72 B

a8 this move is good) 1 6 h6! (White saves his h-pawn because 1 6 . . . � c5 1 7 hxg7 hits the rook, so Black must play 1 7 lii: fc8 transposing to the game) l!!f, fc8 ( 1 6 . J.. h8 1 7 "ilf xc3 is very good for White) 1 7 hxg7 .!! 3c5 1 8 c3 (it seems to me that 1 8 b4 �b6 1 9 -E:J b3 i(4 xc2 20 jj_ xb6 .!! xd2 2 1 .£) xd2 axb6 22 � c l i s very good for White) 'f!fa4 1 9 J.. g5 l! xg5 20 "i!i!fxg5 with advantage to White, Moscow Shabalov-Yurtaev, GMA 1 989. (2) 12 . . . a5 13 h5 ( 1 3 a4 is also good) a4 ( 1 3 .£) xh5 14 g4 � xc3 1 5 bxc3 .£)f6 1 6 J.. h 6 a4 1 7 jJ_xg7 f'i xg7 1 8 J.. xa4! J.. xa4 1 9 � f5 + Resigns, Draskovic-Lazic, corr. 1 974) 1 4 � xa4 J_xa4 1 5 jJ_xa4 .£)c4 1 6 tf"d3 (Black suffers from the lack of a white-squared bishop) 'f!fa5 1 7 J.. b3 d5 1 8 hxg6 hxg6 1 9 g4! dxe4 20 fxe4 .£) xg4 2 1 � g 1 'f!fg5 + 2 2 'f;; b 1 � ge5 23 �h3 � fd8 24 J.. f2 � d 2 + 25 � xd2 'f!fxd2 26 J.. e3 'f!fa5 27 !El e6 fxe6 28 'f!fxe6 + Resigns, Fatali bekova-Akhsharumova, M oscow 1 975.

DraKon Variation

(3) 1 2 Qj c4 (the most im portant 1 2th move alternative) 1 3 11.. xc4 !!! xc4 1 4 h5 and now: (3a) 14 b5? 1 5 hxg6 fxg6 1 6 e5 dxe5 1 7 Qje6 i s bad and 1 4 )il(a5 1 5 * b l ( 1 5 Qjb3 is also good) transposes to line I above. (3b) 14 'fi!/c7 1 5 hxg6 fxg6 1 6 Qjde2 !!! c8 1 7 * b 1 11_ e6 1 8 Qj f4 'fi!/d7 1 9 Qj xe6 t¥xe6 20 !!! c l Qjd7 2 1 'fi!/d5 'fi!/xd5 22 Qjxd5 � e8 23 b3 !!! c6 24 c4 and White's superior pawn structure gives him a permanent advantage in �he ending. (3c) 14 .:Z) xh5 (the critical move) 1 5 g4 Qj f6 1 6 Qj b3!? ( 73) (rarely played, but promising) and now: . . .

. . .

73 B

(3c l ) 16 a5 1 7 11_ h6 J.. xh6 1 8 � xh6 (with the threat o f � dhI followed by !!l xh7 and '!Wh6) ffl g7 (intending .1i h8) 1 9 e5! dxe5 (19 Qj xg4 20 fxg4 11_ xg4 2 1 !!! dh l i s very good for White) 20 g5 .:Z) g8 2 1 'fi!/ h2 !!! e8 22 � xh7 + * f8 23 'fi!/xe5 f6 24 gxf6 and White wins. (3c2) 16 !!! xc3 1 7 bxc3 "Wtc7 1 8 e5 Qje8 19 � h2 h5 20 gxh5 . . .

. . .

11_ xe5 2 1 f4! 'l!ifxc3 (2 1 j_ xc3 22 hxg6) 22 11_ d4 J_ xd4 23 � xd4 with a large advantage for White. (3c3) 16 � e8 1 7 e5 Qj xg4 1 8 fxg4 J.. xg4 1 9 � dg l and now: (3c3 1 ) 19 h5 20 e6! 11_ xe6 (20 f5 21 Qjd5 'l!ltc8 22 Qj d4 and now 22 * h7? 23 !!! xg4! fxg4 24 'l!ltd3 won for White in Ye Jiangchuan-Velimirovic, AsiaYugoslavia 1 984) 21 11_ d4 (Gufeld's suggestion of 2 1 !!l xh5! is a good alternative, the point being that 21 gxh5 loses to 22 11_h6 li! g4 23 li!t xg4 followed by 'l!ltg5) j_g4 22 !!! xg4 hxg4 23 J.. xg7 ffj xg7 24 '*' h6 + *f6 25 Qjd5 + ff e6 26 li! e 1 + ff d7 (26 * xd5 27 'i!ll' h l + is attractive) 27 'fi!/g7 !!! f8 (27 . . . * c8 28 'fi!/xf7 * b8, Caturian-Poletaev, corr. 1 983, and now 29 Qj f6 !! h8 30 'fi!/xc4 exf6 3 1 'itxg4 gives White a winning position) 28 .:Z)d2 � c5 29 ll!( xe7 + '*'xe7 30 Qj xe7 * xe7 3 1 Qj e4 � f5 32 Qj xd6 with advan tage to White because of Black's exposed king. (3c32) 19 dxe5 and now there are two possible moves. My suggestion was 20 '*'xd8 � xd8 21 .:Z:Jd2 (or 2 1 .:Z) a5 !?, aiming to keep the knight in a more active posi tion) �xc3 (21 � b4 22 a3) 22 bxc3, when it seems to me that Black's four pawns are too far back to provide enough compen sation for the rook. However, in the 1 988 correspondence game Kauranen-Nesis White preferred to play for the attack by 20 '*'g2!?, threatening 2 1 Qjd2 and 2 1 '*'xb7. . . .

. . .

Dragon Variation 87 This policy was a success after 20 . . . "i!rc8?! (this must be bad; Black should prevent j_ h6 and secure the g4 bishop by 20 h5, even if this means sacrificing the queen side pawns) 2 1 j_h6 jj_ h8 22 �d2 � xc3 23 bxc3 jj_d7 24 �g3 (White has an extra rook and an attack) jj_c6 25 � h2 'ii!f f5 26 � h4 '(il(f6 27 jj_ g5 1i/g7 (very thematic for the Dragon; pity about the pawn on e5) 28 j_ xe7 h5 (28 . . . e4 29 jj_ f6) 29 � h3 a5 30 J.. c 5 b5 3 1 � e4 b4 32 -2) d6 � b8 3 3 � f5 Resigns. (3c4) 16 'ific7 and now there are two promising lines: (3c4 1 ) 17 e5 dxe5 ( 17 � xg4 1 8 fxg4 ]t xg4 19 exd6 exd6 20 ,2) d5 wins) 1 8 g5 ]t f5 1 9 gxf6 exf6 20 j_ h6 g5 2 1 li._xg7 rk; xg7 is recommended by Sapi and Schneider, but 22 Wre3! followed by :t)d2 and 2J de4 appears good for White. Combined with a later � d2 and :t)d5 Black will be hard pressed to defend his seriously weak pawn at f6. If Black exchanges on e4 he will be mated down the h-file. (3c42) 17 g5 (normally a bad move, but this position is a special case against White gains a lot of time) ..£) h5 1 8 :t)d5 'ii!f d 8 19 J... d4 jj_ xd4 ( 1 9 e5 20 jj_xa7) 20 -E} xd4 jj_c6 2 1 b3 j_ xd 5 (2 1 l;i, c5 22 b4 )ij( c4 23 €le3 is also good for White) 22 bxc4 .;l xc4 23 -2) [5 f6 24 li xh5! gxh5 25 )ij( g ] e6 26 gxf6 + 'ltrh8 27 ..£)e7 1-0, Thes ing-Ballmann, Zug 1 989. 13 J.. g5 ( 74) . . .

74 B

13 . � c4 13 !1 c5 is covered in game 20. Other 1 3th moves: ( I ) 13 . . a6 (ignoring the threat) 14 g4 hxg4 1 5 h5 � xh5 1 6 � d5 ,g e8 1 7 ,g xh5 gxh5 1 8 'l!!t h 2 (this gives White a crushing attack just as in game 20 below) gxf3 1 9 'i!il'xh5 jj_g4 20 '/!!t h4 ,g c7 2 1 Ji h l � g6 22 'l!!t h7 + '1t1 f8 23 jj_ h6 Resigns, Hechler-Masur, corr. 1 980. (2) 13 �h7 14 J.. h6 jj_xh6 1 5 'i'l!Jxh6 � xc3 1 6 bxc3 and now: (2a) 16 . "�t a5 1 7 'fi b 1 ( 1 7 � e2 J.. b5 1 8 � he 1 <2J f6 1 9 H d4 .i, c8 20 K b4! e6 2 1 � d4 jj_d7 22 'l!!t e 3 b5 23 '1ti b2 � c5 24 � e2 � c6 25 )'tf4 � e8 26 a3 was also good for White in Tal-Mista, Dubna 1 973) )'txc3 ( 1 7 K c8 1 8 f4 <2J c4 1 9 f5 gives White the advan tage) 1 8 � e2 'l!!t c5 1 9 g4 <2)xf3 ( 1 9 . . . hxg4 20 f4 'l!!t e 3 2 1 h5 wins) 20 K d5 ! � f2 2 1 gxh5 g5 ( 2 1 '/!!t x e2 2 2 hxg6 <21 f6 23 g 7 wins) 22 hxg5 'l!!t e 3 23 l!l hd l ! and White won easily, Geller-Kuzmin, USSR Ch. 1 978. (2b) 1 6 . . 'l!!t c7 1 7 * b 1 ! �c4 (17 '/!!t xc3 leads to line 1) 18 g4 . .

. .

88 Dragon Variation hxg4 ( 1 8 � f6 1 9 � f5!) 1 9 f4 ( 1 9 h5? g5) I!l c8 20 � d3 'i!ta5 2 1 h5 g5 and now 22 e5!? was good for White in Lobron-Miles, Biel 1 986 after 22 dxe5 23 jl_ xc4 � xc4 24 � b3 'i!tc7 25 fxg5 � f5 26 g6, but according to Lobron it was simpler to play 22 �xc4! li xc4 23 � b3 'i!td8 24 fxg5 "i!ffB 25 e5! "i!fxh6 (25 � f5 26 exd6 �xd3 27 �xf8 + � xf8 28 dxe7 wins) 26 gxh6 � f5 27 exd6 exd6 28 M xd6 when White's material advantage and threats against the Black king give him a winning position. 14 'i!te2 After 1 4 ttd3 Black can reply 1 4 . . . � e5, since 1 5 'i/lj"e2 would be an invitation for Black to sacri fice at c3. 14 � a5 15 '1i'b1 For the moment Black has no threats because M xc3 isn't dangerous when the knight pre vents the queen moving to a5, so White takes the opportunity to make a useful consolidating move. 15 a6 This move represents the point of the manoeuvre �c4-a5. Black threatens to trap the knight on d4 by . . . e5 and if White were now forced to spend a tempo moving the queen Black would play b5 with good counter play. Everything depends on whether White can successfully ig, e5. nore the threat of 16 g4 ( 75)

16 e5 Or 1 6 � xb3 1 7 axb3 hxg4 (for 1 7 . . . e5, see the note to Black's 1 8th move) 1 8 h5! gxf3 ( 1 8 � xh5 1 9 fxg4 12J g3 20 *h2 � x h l 21 � xh l leads to mate) 1 9 � xf3 � g4 ( 1 9 . . . 12! xh5 20 � dg l threatening M xh5 i s very strong) 20 ll dg l (20 h6 � h8 2 1 e5 is also good, but not 2 1 h7 + �xh7 when Black can block the h-file by � h5) gxh5 (20 . . . � xh5 2 1 � xf6 exf6 22 � xh5 gxh5 23 "i!fg2 wins) 21 � xf6 exf6 and now 22 K xg4 hxg4 23 � d4 is one danger ous attacking possibility. 17 gxh5! Geller-Miles, Linares 1 983, continued 1 7 � f5 gxf5 1 8 gxf5 � xb3 1 9 axb3 jl_c6 20 � hg l *h7 2 1 1i! g2 � g8 and White went on to win, but Mestel was evidently not convinced about White's compensation, and in this I agree with him. exd4 17 18 �d5! ( 76) Geller's notes to his game gave 1 8 h6 � h8 (18 . . . dxc3 1 9 hxg7 fl xg7 20 )!;g xd6 wins) 1 9 h7 + fJxh7 20 h5 and wins, but 18 . . .

Dragon Variation 89 � xh6! 19 � xh6 � xb3 20 axb3 dxc3 is a big improvement, when the position is totally unclear. Later I discovered that Timman had suggested 18 � d5! in 1 983, but during the game I was un aware of this!

18 � xb3 The only real defence to the threat of 19 h6 � h8 20 h7 + *xh7 2 1 h5 is to play gxh5 . Unfortunately 1 8 gxh5 (trying to deflect White's rook by 1 8 d 3 1 9 H xd3 gxh5 fails after 20 � g 1 fi>h7 2 1 �xf6 + � xf6 22 � xd6 � xg5 23 � xg5 winning easily) loses to 19 � hg 1 fi>h8 20 e5 dxe5 2 1 �xe5 � xd 5 (or 2 1 � e8 2 2 � xf6 �xf6 23 �xf6 � xf6 24 � xf6 JJ. e7 25 � xd4 with a won ending) 22 �xg7 + f!Jxg7 23 j_ xd8 + f!J h7 24 jl xa5 with a strong attack and a bonus pawn to come. Thus Black tries to exchange on b3 before taking on h5, but in this position White doesn't need to recapture. If Black wanted to exchange on b3 he should have done so at move 1 6, with the co ntinuation 16 . . .

�xb3 1 7 axb3 e5 1 8 gxh5 exd4 1 9 �d5. I think White can still win even in this case, but it is much more complex: 19 gxh5 20 � dg l lt1h7 (20 f!Jh8 2 1 f4 � g4 22 � xg4 hxg4 23 h5! lt1 h7 24 h6 j_ h8 25 �d3 f!Jg8 26 h7 + wins) 2 1 f4 � g4 22 �d3! f!Jg8 (22 f!J h8 23 �xg4 hxg4 24 h5 is worse) 23 �xg4 hxg4 24 h5 (this seems to be good even when White has a tempo less) � xd5 (24 K e8 25 h6 �h8 26 h7 + lt1f8 27 � xf6 wins) 25 � xd8 Q� b4 (25 � xf4 26 '(iffl ) 26 '*' d2 d3 27 cxd3 � c2 28 W!xb4 � xb2 + 29 f!Jc I � xd8 30 d4 and White should win. 19 h6! Stronger than transposing to the above analysis by 1 9 axb3. 19 j}_ b5 Trying to generate counterplay, but there is no defence. The main line runs 1 9 �h8 ( 1 9 . . . � xd5 20 hxg7) 20 h7 + f!J xh7 21 h5 f!J g8 (2 1 �g7 22 hxg6 + lt1 g8 23 � xf6 + � xf6 24 '/j'h2) 22 � h2! (not 22 hxg6 fxg6 23 �h2 f!Jf7) gxh5 (22 � g4 23 fxg4 ytxg5 24 hxg6 mates) 23 LE) xf6 + J.. x f6 24 �xh5 li( e8 25 � dg l f!Jf8 2 6 � h6 + j_ g7 2 7 �xg7 + f!Jxg7 28 jl e7 + and mate next move. d3 20 �h2 21 cxb3 Not 2 1 c3? d2! with dangerous counterthreats. 21 LE) xd5 After 2 1 � c2 22 hxg7 � xh2 23 gxf8( �) + f!J xf8 24 g xh2

90 Dragon Variation Black loses the knight on f6. 22 hxg7 Not 22 jj_ xd8? jj_e5. .!! c2 22 Black cannot avoid a fatal loss of material. 23 jJ_ xd8 K xd8 24 � d2 Resigns 24 .!El e3 is met by 25 �f2. Game 20 Mestel-Kudrin Hastings 1986/7 I e4 c5 d6 2 .!El f3 cxd4 3 d4 4 'El xd4 'El f6 5 .!Elc3 g6 jJ_ g7 6 jJ_ e3 0-0 7 f3 8 �d2 .!Elc6 jJ_ d7 9 jJ_ c4 )iil. c8 10 0-0-0 1 1 jJ_ b3 'El e5 12 h4 h5 13 J.. g5 � c5 14 \fl bl This is not only a useful move, it leaves Black with little choice but to play b5, when curiously enough White's kingside break stronger through becomes because in one variation Black does not have the move �b6 attacking b2. b5 14 15 g4 (77) aS ( 78) 15 Declining the sacrifice, but acceptance is a major alternative for Black: 1 5 hxg4 1 6 h5 and now:

( 1 ) 16 . . . � xc3 with a further branch: ( l a) 17 'l\llt xc3 l2\ xh5 1 8 fxg4 jt xg4 1 9 � dg l 'i'!fd7 is unclear, while 17 jj_ xf6 � xb3 1 8 JJ_ xg7 )iil. xb2 + wins for Black. ( 1 b) 17 bxc3 and Black has one bad move and one good move: (1 b l ) 17 . . . .!El xh5 and now the simple 18 jJ_h6! (threat 1 9 .!:il, xh5 gxh5 20 '!lit g5) e6 19 � dg 1 gives White a strong attack for no material investment. In Ulybin Tiviakov, USSR 1 987 White chose a more complex line by 18 .!:i!, xh5 gxh5 1 9 � h2 .!Elc4 (other wise White plays �xh5 and .!:i!, h l ) 20 'l\llt x h5 f6 2 1 l2\ xb5! ll_ xb5 (White also has the advantage after 21 fxg5 22 JJ_xc4 + e6 23 l2\ xd6) 22 j_ h6 'i'!fd7 (22 �c8 23 � g l f5 24 ll_xg7 \fj xg7 25 exf5 gives White more than enough for the piece according to Ulybin) 23 jj_ xg7 \fl xg7 24 liil. gl with a crush ing attack. ( l b2) 17 l2\ xf3! 1 8 l2\ xf1 l2\ xe4 1 9 � h2 l2\ xc3 + 20 'fl> c l � a 5 2 1 hxg6 (2 1 � d4 g 3 ! IS unclear) l2\ xa2 + 22 J_ xa2 �a3 + 23 'lt> d2 �c3 + 24 * c l (24 'fl> e2 . . .

Dragon Variation 91 ttxf3 + 25 *e l ? tye4 + and 26 ttxg6 gives Black 5 pawns and attacking chances for a rook) tya3 + Draw, Lanka-Smirin, USSR 1 989. ( l e) 17 h6! (although never tried in practice I think this is a good move) and now: ( l c l ) 17 .l, xb3 1 8 hxg7 X xb2 + 1 9 '*a l ! (threat 20 j}_ xf6 exf6 2 1 ii h8 + ) • xg7 ( 1 9 . . . <2:l h5 20 J.. f6!) 20 j_ xf6 + f1 xf6 2 1 ttf4 + .g7 22 ty h6 + . f6 23 fxg4! and White has a massive attack. ( l c2) 17 . . <2:l xf3 I 8 � xf3 X xf3 1 9 hxg7 *xg7 20 jth6 + (20 e5!? is also possible, with the idea that after 20 . . . dxe5 2 1 jt h6 + . g8 22 jt g7 il,l h3 23 K xh3 gxh3 24 i!lth6 <2:l h7 25 jt xe5 Black is mated) f1g8 2I jt g7! <2:1 h5 22 �h6) 22 K h3 (2I l\1 xh3 gxh3 23 'Wf h6 <2:1 h7 24 j_ c3 e5 25 �xg6 + '* h8 26 i!lfxd6 with advantage to White. (2) 16 . . . <2:l xh5 I 7 <2:ld5 X e8 (17 l\1 xd5 1 8 jt xd5 is bad since 1 8 . . . gxf3 lets White break through by I 9 <2:l f5!) I 8 ,I xh5 gxh5 I9 tth2 g xd5 (or I 9 K c4 20 j_xc4 bxc4 2 I ttxh5 f6 22 f4! and after 22 fxg5 23 fxe5 dxe5 24 .1. hi exd4 25 tth7 + '*f7 26 K fl + '* e6 27 tyxg7 White's attack is too strong, so Karpov Sznapik, Dubai 01. I 986 finished 22 . . . <2:l f7 23 jth4! tt b8 24 � h i c 3 2 5 b 3 ttb7 26 f5 !Ele5 2 7 <2:1 e6 Resigns) 20 jt xd5 �b6 (now it becomes clear that interpolating fi b i and . . . b5 favours White . . .

since b2 is not attacked) 2 I tt xh5 e6 22 A b3 and White has a win ning attack, for example: (2a) 22 . . a5 (22 . . . <2:lg6 23 j_e3 tyc7 24 X h i and 22 . . . <2:l xf3 23 <E� f5 <E) xg5 24 <2:l xg7 * xg7 25 tyxg5 + *f8 26 e5! dxe5 27 X hi both win for White) 23 f4 <2:lc4 (23 <E)g6 24 f5) 24 j_h6 tyxd4 (24 . . . *f8 25 j_ xg7 + '* xg7 26 ttg5 + '*f8 27 ttf6 or 24 . . . ttd8 25 j_xc4 bxc4 26 e5! f5 27 J.. g5 tyc7 28 A f6) 25 X xd4 j_ xd4 26 j_xc4 bxc4 27 e5 Resigns, Short-Mandl, Bundes liga I 986. (2b) 22 ttc5 23 j_ h6 J.. f6 24 f4 <E) g6 25 <E) f5 j_c6 26 ttxg4 d5 27 <2:l g3 tye3 28 e5 jt xe5 29 fxe5 'ijxh6 30 <E) h5 K c8 3I c3 tte3 32 j_c2 '*f8 33 J.. xg6 fxg6 34 K fl + *e7 35 i!�t b4 + '*d8 36 tyd6 + Resigns, Chandler-Mestel, Brit ish Ch. Play-Off 1 986. 16 J.. xf6 The alternative I 6 gxh5 has not been played much but it could also be good, e.g. I6 gxh5 a4 ( 1 6 <2:l xh5 I 7 � d 5 I� e8 I 8 a3 was good for White in Sznapik Sehner, Slupsk I 987) and now: .

. . .

92 Dragon Variation ( I ) 17 J.. d5 b4 1 8 �ce2 e6? ( 1 8 �xd5 I 9 exd5 � xd5 20 h6 is better but still good for White according to Sznapik) 1 9 h6! J.. h8 20 h7 + ! * xh7 21 h5 exd5 22 hxg6 + *g8 23 � xh8 + *xh8 24 Resigns, Sznapik J.. xf6 + Komeljenovic, Biel Open 1 987. (2) 17 h6 J.. h8 18 J.. d 5 (after 1 8 h7 + not 18 . . . *xh7? 1 9 h5 � xh5 20 � xh5 + gxh5 21 � h2 and wins but 18 . . . � xh7! with unclear complications) *h7! (White's attack has been blocked) 1 9 a3 � b6 20 J.. a2 b4 2 1 axb4 �xb4 22 J.. xf6 �xf3! with a very dangerous attack for Black, Uly bin-Savchenko, Simferopol I 988. (3) 17 J.. xti + !? *xf7! (not I 7 !! xf7? 1 8 hxg6 � xg6 1 9 h5 � f8 20 � dg l *h7 21 e5! with a decisive attack, lvanovic-Feick, Berlin Open I 988) I 8 hxg6 + ( 1 8 f4 b4!) � xg6 I 9 h 5 �e5 20 h6 �c4 (20 J.. h 8?! 21 f4 � c4 22 �e2! gives White more chances) 21 �g2 J.. h8 22 h7 �a5, lvano vic-Kosanovic, Stara Pazova I 988, and now 23 J.. xf6! J.. xf6 (23 � b4 24 � b3 axb3 25 cxb3 .txf6 26 �d5! is good for White) is unclear after 24 �d5!? or even 24 h8(�) j_ xh8 25 �d5. 16 J.. xf6 17 gxh5 a4 18 J.. d5 Better than I 8 hxg6 axb3 I 9 �xb3 � xc3 20 gxf7 + * xf7 2 1 �xc3 (with g + 3 Jl, v J.. + � White might appear to have good winning chances, but in fact Black can just hold the balance) � c4 22

� d4 trb6 23 "tt d 3 tyxd4 24 tyxd4 J.. xd4 25 � xd4 *g6 and Black drew this ending in Chandler-Petursson, Hastings 1 986/7. 18 e6 exd5 19 hxg6 20 h5 A:l( xc3 In Wiech-Nizynski, Poland I 987, Black was successful with 20 J.. g5 2 1 •g2 � xc3 22 bxc3 �c4 23 l!;!! dg l "tt a 5 24 �xg5 �xc3 25 gxf7 + * xf7 26 �xd5 + *e8 27 '!!!!f a8 + *f7 28 �d5 + * e8 and White had nothing better than perpetual check. How ever 2 1 f4 �c4 (2 1 lli, xc3 22 bxc3 transposes to the main line, but this could be an important finesse of move order for Black if the suggestion of 21 '!!!!f h6 in the next note turns out to be strong) 22 � h2 J.. h6 23 � xd5 appears much more dangerous. 21 bxc3 Not 2 1 'j!i!rxc3? since the tempo saved by keeping the queen active is more important than the damage done to White's pawn structure. However 2 1 �h6 fxg6 22 hxg6 �e7 23 bxc3 is an inter esting alternative since White transfers his queen to a better position with gain of time. 21 J.. g5 J.. h6 ( 79) 22 f4 Not 22 �c4 23 �g2 J.. xf4 24 gxf7 + *h8 25 h6 � xf7 26 !i dg l J.. g 5 27 �xg5 � fl + 28 K xfl �xg5 29 :l;! f8 + *h7 30 l!I f7 + * h8 when White plays 3 I h7! �d8 (3 I J.. h3 32 � f8 +

Dragon Variation 93 fi' xh7 33 �t xh3 + wins the queen) 32 H g l <E:Je5 (relatively the best) 33 H g8 + -.xg8 34 hxg8( '11' ) + \ti xg8 35 K f6 with excellent win ning chances. 79 w

23 l!l dgl A very committal move. White plays for mate, but if he cannot break through the move is proba bly a mistake. Instead White should prefer 23 exd5, not only improving his material situation but also securing c6 and e6 as potential knight outposts. In this case I believe White has some advantage, although of course the position is still very complicated.

23 -.e7? Any error will have fatal conse quences in such a double-edged position. Here Black misses a good chance with 23 . . . dxe4! 24 gxf7 + (24 �g2 f5! 25 fxe5 dxe5 is unclear) fi'h8 25 �g2 (25 K g6 .:E) xg6 26 hxg6 fi'g7 27 -.h2 ,K h8 28 f5 -.f6 and Black survives) <E:J xf7 26 -.xe4 �e8! (26 . . . ,K e8 27 -.g6 is good for White) 27 -.d5 -.e3! and Black develops counterplay. In this case Black should be at least equal. j)_ xf4 24 •g2 <E:l g4 (if the Forced as 24 knight goes anywhere else Black is mated by gxf7 + ) 25 f5 <E:J f6 26 gxf7 + fi' h8 27 ttg6 wins outright. 25 g7 l!:!( b8 26 h6 <E:J g6 27 <E:JfS J.. xfS 28 exfS liteS 29 fxg6 and Black lost on time while play ing 29 f5. White wins by 30 � e I j)_ e3 3 1 lii xe3 followed by 'i!\j' xd5 + and mate.

Kan Variation

This line, which starts 1 e4 c5 2 � f3 e6 3 d4 cxd4 4 � xd4 a6, is notable for the flexibility afforded to Black, since by delaying his piece development he keeps the maximum possible range of op tions open. Because there are many reasonable choices at each move it is pointless to give precise lines against all possible move or ders, so in this chapter there will be a greater emphasis on general principles. The continuation recommended in this chapter, 5 Ad3, is the most common line in practice. At the moment Black's most popular reply is to set up a 'hedgehog' position by 5 � f6 6 0-0 d6 (or 6 �c7 7 � e2 d6). After 7 c4 Black may choose to develop his bishop on e7 immedi ately, but he sometimes brings the queenside out first in order to keep open the option of . . g6 and jl g7. The 'hedgehog' name is derived from the way Black curls up on the first two ranks, moves like Ab7, � bd7, b6, l:ii, e8, f!Jc7, J_ f8, a ac8 and � b8 being typical. Black's slow development invites White to attack, but experience has shown that the unwary attacker can easily impale himself on Black's spines, and such at-

tacks have to be well-organized if they are to stand much chance of success. Moreover White has to watch out for Black's b5 and . d5 breaks. Game 22 deals with the lines arising after 5 � f6, including the 'hedgehog'. All Black's other 5th moves, such as 5 jlc5, 5 . . �c6, 5 . . . �e7 and 5 g6, are in game 2 1 . Game 21 Kengis-Nevednici USSR 1979 1 2 3 4 5

c5 �f3 e6 d4 cxd4 Qlxd4 a6 Ad3 (80) e4

5 jlc 5 Black's are alternatives arranged in descending order of importance: (1) 5 � c6 (it is surprising • . .

Kan Variation 95 that this solid line is not p layed more frequently; although. the symmetrical position gives Black few winning chances, it is hard for White to prove any advantage) 6 � xc6 and now: ( l a) 6 . . . dxc6 7 �d2 e5 8 'l!!t h 5 j_d6 9 � c4 j_c7 (9 .Z:. f6 1 0 <E} xd6 + 'l!!f xd6 1 1 'l!!f e2 j_e6 1 2 00 gives White the chance to make his black-squared bishop a potent force, as in Jansa-Cebalo, Sme derevska Palanka 1 978 after 1 2 �d7 1 3 .l, d l 'l!!f e7 1 4 b3 0-0 1 5 a4 a5 1 6 J.. a3 � c5 1 7 j_c4!) 1 0 j_g5 � f6 1 1 'l!!f e2 h6 1 2 j_ h4 'l!!f e7 1 3 0-0-0 ( 1 3 0-0?! allowed Black to stir up trouble by 1 3 g5 1 4 j_g3 h5 1 5 f3 h4 in Ligterink Miles, Lone Pine 1 979) il_e6 14 f4 J.. xc4 ( 1 4 j_g4 1 5 il_ xf6 il_ xe2 1 6 j_ xe7 j_ xd l 1 7 j_d6 J,. xd6 1 8 � xd6 + <ff d 7 1 9 �xf7 wins material for White) 1 5 j_ xc4 b5 1 6 J.. b 3 0-0 1 7 J.. xf6 'l!!f xf6 1 8 1, d7 and White's pressure against f7 gives him some advantage, Tseshkovsky-Miles, Bled-Por toroz 1 979. ( 1 b) 6 . . . bxc6 (out of favour ever since the famous Fischer Petrosian game mentioned below) 7 0-0 d5 (7 e5 8 f4 j_ c5 + 9 <ff h l � e7 1 0 'l!!f h 5 � g6 1 1 f5 � f4 1 2 j_ xf4 exf4 1 3 f6 J.. d4 14 fxg7 j_xg7 1 5 I;! xf4 'l!!f e7 1 6 <2Jc3 is very good for White, Ravinsky Vorotnikov, USSR 1 963, while 7 g6?! 8 e5 j_ g7 9 f4 d6, van der Wiel-Anand, Thessaloniki 01. 1 984, should have been met by 1 0 exd6 'l'!txd6 1 1 �d2 'fjd4 + 1 2 . . .

. . .

<ff h l � f6 1 3 � c4 �d5 1 4 � d6 + <ff e7 1 5 c3 with a fine position for White) 8 c4 � f6 9 cxd5 cxd5 1 0 exd5 � xd5 (10 exd5 1 1 � c3 J.. e7 1 2 'l'!ta4 + l/!l'd7 1 3 K e l ! 'l!!f xa4 1 4 .z:. xa4 j_e6 1 5 j_e3 0-0 1 6 J.. c5 is Fischer-Petrosian, match. 1 97 1 , which was won by Fischer, while 10 . . 'l'!txd5 1 1 <E}c3 i>!td7 1 2 J.. g5 J.. e7 1 3 ll!l'e2 J.. b7 14 !l ac 1 0-0 1 5 .l. fd l , although keeping Black's pawns intact, gave White a dangerous initiative in Mikhalchishin-Gor chakov, USSR 1 972) 1 1 j_ e4 K a7 ( 1 1 . . . � b8 1 2 'l'!tf3 f5 1 3 j_xd5 'l'!txd5 14 �xd5 exd5 1 5 � d 1 J.. e6 1 6 � c3 � d8 1 7 j_g5 � d7 1 8 .Z:.e2 gives White the better ending and 1 1 . il_ e7 1 2 � c3 J.. b7 1 3 'l!l!fa4 + �d7 1 4 ll!l'xd7 + <ff xd7 1 5 M d 1 is also promising after 15 . . . Ji::( ad8 1 6 � xd5 j_xd5 1 7 j_xd5 exd5 18 � xd5 + <ff e 6 19 � xd8 � xd8 20 j_e3, Matanovic-Roos, Le Havre 1 966, or 15 . . . j_f6 1 6 � xd5 J.. xd5 1 7 j_xd5 exd5 1 8 � xd5 + <ff e6 1 9 Jii d2 � hd8 20 ii! e2 + , Averbakh-Taimanov, USSR Ch. 1 960) 1 2 'Wt'd4 � d7 1 3 � c3 � xc3 ( 1 3 . . . ,Zl b7 1 4 � xd5 il_xd5 15 j_ xd5 l1li, xd5 16 ll!l'a4 + wins the a-pawn) 14 'l'!txc3 and Black's uncastled king gives him plenty of problems, Belyavsky Kurajica, Sarajevo 1 982. (2) 5 . . . g6 6 c4 j_g7 (this is an attempt to reach a kind of hedge hog position, but with the bishop more actively deployed at g7) 7 � b3 and now: (2a) 7 � e7 8 � c3 d5 (8 . . • • .

. .

• . .

96 Kan Variation 0-0 9 0-0 Ql bc6 1 0 jte2 b6 1 1 J.. f4 f5 1 2 exf5 ,l:l xf5 1 3 \¥f'd2 with advantage to White, Ljubojevic Panno, Buenos Aires 1 980) 9 cxd5 exd5 1 0 J.. g5! h6 1 1 j_ xe7 jtxc3 + 1 2 bxc3 'i¥f'xe7 1 3 0-0 dxe4 1 4 J.. xe4 0-0 1 5 M e 1 \¥f'c7 1 6 \¥f' f3 Qld7 1 7 Ql d4 *g7 1 8 � ab 1 and White has a big lead i n de velopment, Donchev-Prie, Tou lon 1 988. (2b) 7 . d6 8 Qlc3 Qlf6 9 il_ f4 (9 0-0 followed by J.. f4 transposes to lines given in game 22, but White can do better here because Black has no time for Ql c6 and Qle8) 0-0 1 0 jte2 e5 ( 1 0 � e8 1 1 c 5 is particularly un pleasant when Black's queen is undefended) 1 1 il_e3 j_e6 1 2 0-0 � c6 1 3 f3 � c8 14 J!i! c l , Ljuboje vic-Rajkovic, Yugoslavia 1 980, with a good position for White. The plan of directly attacking the d-pawn by Af4 and j_e2 is a logical way to exploit Black's jtg7, and in this case it gives White the advantage. (3) 5 . . Qle7 (Black aims to play Qlc6, but only when he can recapture with a piece) 6 Ql c3 (6 0-0 <2\ ec6 7 c3 is also possible, when 7 jte7 8 j_e3 0-0 9 f4 d6 10 'E} f3 � d7 1 1 Qlbd2 gave White a small but enduring plus in Geor giev-Peev, Bulgaria Ch. 1 980- 1 ) �ec6 7 � b3 jte7 8 �h5! d 6 9 jte3 Ql d7 1 0 f4 b5 1 1 0-0-0 b4?! 1 2 � a4! e5 1 3 f5 0-0 1 4 g4! with an automatic attack for White, Mikhalchishin-Dorfman, Lvov 1 983. . .

(4) 5 * b6 6 c3! d6 (6 � c6 7 0-0 � xd4 8 cxd4 "l!!t xd4 9 � c3 is dangerous for Black) 7 0-0 <£� f6 8 a4 jte7 9 Ql d2 "l!!t c7 1 0 a5 0-0 1 1 'E} c4 'E} bd7 1 2 \¥f'e2 .l, e8 l 3 j_ g5! J.. f8?! ( 1 3 h6 1 4 jt h4 b5 is just slightly better for White) 1 4 J.. h4 with a good game for White, Lazic-Martinovic, Yugoslavia 1 987. (5) 5 . . "l!!t c7 6 0-0 <EJc6?! (Black can transpose to game 22 by 6 Ql f6) 7 Ql xc6 "l!!t xc6 (or else Black has an inferior version of variation I above) 8 c4 (8 Ql d2 is probably also good) g6 9 <EJc3 J,.g7 10 � e l � e7 1 1 jtg5 d6 1 2 "l!!t d2 with a very good position for White. (6) 5 . . . b5 6 0-0 il.. b7 7 'ilfe2 Qle7 8 a4 b4 9 Ql d2 Qlbc6 1 0 Ql4b3 Ql g6 1 1 f4 with advantage to White, Matanovic-Taimanov, Yugoslavia 1 965. One of the main advantages of 5 J.. d3 as opposed to 5 � c3 is that an early b5 by Black is hardly ever a worry, since White may undermine Black's queenside pawns by a4 without fearing a loss of time after b4. jt a7 6 Ql b3 It makes little difference whether Black plays 6 jt a7 or 6 . . A b6, since he must exchange on e3 within a few moves in any case. <2\c6 7 "l!!t e2 8 Ae3 jt xe3 9 "l!!t xe3 (81) 9 d6 Or: ( 1 ) 9 . . e5 1 0 � c3 �ge7 1 1 . . •

Kan Variation 97 81 B

� f6 10 �c3 1 1 0-0-0 White has an interesting alter native in I I g4!?, when 1 1 b5 I 2 0-0-0 0-0 1 3 g5 �e8 I 4 f4 left White with a favourable version of the main line in Anand-Ninov, Baguio City 1 987. Anand finished off efficiently: 14 . . . b4 1 5 � e2 a5 I6 <E:� bd4 <E)xd4 I7 � xd4 �b6?! 1 8 e5! J.. b7 1 9 )!;!!, hfl ! dxe5 20 fxe5 J!!!, d8? 21 j_xh7 + ! \tixh7 22 g6 + \tig8 23 ith3 � f6 24 exf6 fxg6 25 fxg7 Resigns. The critical res ponse to l l g4 is probably 1 1 �xg4 1 2 'ltg3 � f6 1 3 'ltxg7 � g8 1 4 � h6 j_d7 intending 'lte7 and 0-0-0, and it is not clear how much advantage White can claim. 11 0-0 There are quite a few alterna tives: (I) 1 1 e5 1 2 � d2 ( 1 2 j_ e2 00 1 3 f4 exf4 1 4 'ltxf4 �e8 I 5 � d4 J.. e6 I 6 � f5 was a little better for White in Psakhis-Vyzmanavin, Moscow Ch. 1 98 1 ) J.. e6 1 3 .i_ hd i 'ltc7 I 4 J.. e2 .i. d8 1 5 g4! 0-0 ( 1 5 h6?! 1 6 f4 exf4 1 7 'lt xf4 �e5 1 8 � d4 0-0 1 9 g5 d5 20 * b l was good for White, Varjomaa-Tor nefjell, corr. 1 979) 1 6 'ltg3 J.. xb3 1 7 cxb3! '2J d4 1 8 * b l b5 19 j_d3 �c6 20 f3 � fe8 2I �g2 b4 22 g5! and the outpost on c4 for White's bishop aided an already danger ous kingside attack in Lekander Schoneberg, corr. I 980. (2) 11 b5 12 � d2 ( 1 2 g4!? looks good) 0-0 1 3 .i_ hd l -wtc7 1 4 f4 b4 1 5 '2J e2 e 5 1 6 fxe5 dxe5 1 7 . . .

0-0-0 d6 I 2 J.. c4 b5 1 3 J.. d5 � xd5 I4 � xd5 j_e6 I 5 f4 0-0 with a roughly equal position, Wahls Farago, Altensteig I 987, but I 2 g d2 followed by lt hd I and J.. e2 creates a more dynamic impres sion. (2) 9 �f6 (this move order is sometimes played, but it may be inaccurate) IO e5!? (after IO � c3 Black should play I O d6; trying his luck with 1 0 0-0 is bad after I I e5 <E) g4 I 2 j_ xh7 + !) � g4 I I 'lt g3 h5 (I I � cxe5? I 2 J.. e2 h5 I 3 h3 h4 I4 itc3 wins for White) I 2 h3 � gxe5 I 3 'ltxg7 and White has some advantage. (3) 9 �ge7 IO <E) c3 0-0 I I 00-0 'ltc7 (or I I b6 I 2 f4 j_b7 1 3 � he i 'ltc7 I 4 * b i d6 I 5 ith3 <E:� b4 I 6 a3 � xd3 + I7 X xd3 � c6 I 8 ith4! with dangerous threats, Vilela-Lebredo, Cuba I 983) I 2 f4 d6 I 3 ith3 � b4 I 4 .!ll hfl ! e5 I 5 f5 d5 I 6 exd5 � exd5 I 7 � xd5 �xd5 I8 ith4 � f6 I 9 g4 e4 20 g5 exd3 2I !;K xd3 'l!lte5 22 gxf6 'ltxf6 23 �xf6 gxf6 24 )!;!!, d6 with a very good ending for White, Korlov Batakov, corr. 1 984. . . .

. . .

98 Kan Variation l:! fl �d7 1 8 � g3 � b6 1 9 .f) f5 with strong pressure for White, Byrne-Larsen, Biel IZ 1 976. (3) 1 1 . . . "f!tc7 may transpose into the main line after 1 2 f4 0-0, but it allows White the alternative of 1 2 g4!, which looks unpleasant for Black. 12 f4 "f!tc7 (82)

13 g4 There seems to be no general agreement as to whether White needs to prepare this with K hg I . The alternative attacking ideas are: ( 1 ) 13 "f!th3 � b4 1 4 g4 b5 1 5 g5 � xd3 + ( 1 5 � e8 1 6 "f!th4 f6 1 7 a3 fxg5 1 8 fxg5 � xd3 + 1 9 � xd3 K b8 was unclear in Vogt Velikov, E. Germany-Bulgaria 1 987) 1 6 K xd3 � e8 1 7 f5 b4 led to a sharp finish in Bronstein Suetin, Moscow Ch. 1 982 after 1 8 "f!th4 bxc3 1 9 K h3 cxb2 + 20 * b 1 f6 2 1 "f!txh7 + *f7 22 K g 1 J g8 23 K h6 exf5 24 Ji!t xf6 + � xf6 25 gxf6 *xf6 26 �xg8 j_ b7 27 'l!rh7 j_ xe4 28 "f!tg6 + *e5 29 "f!tg3 + Draw.

(2) 13 I, hg1 (this is probably good enough for a slight advan tage) b5 1 4 g4 b4 1 5 g5 � e8 1 6 �e2 ( 1 6 � b l a5 1 7 .!;l g4!? a4 1 8 � 3d2 j_ a6 1 9 j_ xa6 � xa6 20 X h4 g6 2 1 "f!l h3 f5 22 gxf6 I, xf6 23 �c4 may be slightly better for White, Wedberg-Spraggett, New York Open 1 987) "f!ta7 ( 1 6 a5 1 7 �bd4 � xd4 1 8 �xd4 j_ a6 1 9 * b l was a little better for White in Arnason-Suetin, Sochi 1 980) 1 7 "f!th3! g6 1 8 f5 exf5 1 9 exf5 � e7 20 �g3! "f!te3 + ?! (20 a5 is better, but still favours White) 2 1 * b l j_ xf5 22 j_ xf5 � xf5 23 � de l "f!tf4 24 � gfl tth4 25 ttxh4 � xh4 26 g e4 with an ex cellent ending for White, Arna son-Kirov, Plovdiv 1 986. 13 b5 Accepting the offer must be a critical test of White's willingness to play g4 without the prepara tory !!I hg l . In Short-Velikov, European Club Ch. 1 987 the con tinuation was 1 3 � xg4 1 4 �g3 <E) f6 1 5 � hg l � e8 1 6 * b 1 ( 1 6 f5!? i s natural) � e7?! ( 1 6 b 5 and 1 6 f6 have been sug gested as possible improvements) 1 7 � d4 ttc5 1 8 � f3 f6 1 9 e5! with a very strong attack, Short Velikov, European Club Ch. 1 987. Velikov must have found an improvement because he repeated this line in a later game Ivanovic Velikov, Saint John Open 1 988. Unfortunately Ivanovic varied by 16 "f!th4, so we don't know what Velikov's intention was. Despite this hint, I believe White has good

Kan Variation 99 compensation for the pawn and it would require a brave Black player to take this line on. 14 g5 <2:l d7 15 f5!? 1 5 '/th3 <2:l b4 1 6 a3 Qj xd3 + 1 7 '/txd3 d5 gave Black counterplay in Georgadze-Bohlig, Halle 1 978, although White won this game too. 15 b4 16 Qje2 a5 It is almost impossible to assess positions in which the players are attacking on opposite wings. Unless one of the players is well in front it is likely that a single tempo will decide the race and obscure tactical points will often have a crucial influence on the play. 17 '/th3 Attacking e6 directly, and creating a concealed threat to h7. 17 exf5 Qj de5 18 exf5 The pin along the c8-h3 diag onal is awkward for White since he cannot move his kingside pawns, nor can he unpin by '/th4 jtxf5 Qj xd3 + and since repulses the attack. Bringing the knight to d5 is the only way to make progress. 19 <2:l f4 a4 '/tdS (83) 20 <2:l d5 Now that b3 and g5 are under attack, White is committed to the sacrificial path. .!;! hgl! 21 Qj xd3 + ? Black decides to eliminate one of the attacking pieces, but in

doing so brings the d 1 rook into the attack. The best defence was axb3! 22 <2:l f6 + gxf6 23 21 'tl!f h6! (23 gxf6 + €! g6 leads nowhere as the f5 pawn is pinned while 24 'tl!t'h6 '/txf6 25 fxg6 fxg6 26 jj_ xg6 'ltf4 + 27 '/txf4 K xf4 28 ii_ e8 + *f8 29 jtxc6 bxa2 30 * d2 lii d4 + wins for Black) <2:l g6! * h8 24 gxf6 K g8 25 14, g7 (23 .f) xd3 + 26 l! xd3 jtxf5 27 !!l h3! leads to mate at h7) 24 fxg6 fxg6 25 ii_ xg6 K a7! 26 gxf6 hxg6 (26 '/txf6 27 jtxh7 + *h8 28 � g8 + ! )i xg8 29 jtg6 + mates, *f7 28 H dfl ) 27 or 27 lii xg6 + (27 'tl!fxg6 + fi h8 leads to nothing as the d l rook cannot reach the h-file, Ji d5 being met by . Qj e5) 'f1f7 reaching a remark able position in which it appears that White must mate, but it isn't certain that he can do so. In the first edition I commented that I couldn't see a mate after 28 . 'f1 e8 29 lil g7 + 'f1 e6 (28 '/te3 + Qje5 30 '/txa7 'ltxf6 3 1 axb3 i s very good for White, with material equality but weak Black pawns and an exposed Black king) 29 '/th3 + *xf6 30 '/th6 +

100 Kan Variation *e5 (30 <IJ' f5 3 1 � g2! does mate), for example 3 1 � g3 jj_ f5 32 c4 jj_e4! I still don't see a mate, and nobody wrote to me suggest ing one, but if we return to the position after 27 <IJ'f7 White can gain a massive advantage by 28 lJ, e l � e5 29 K g7 + *e6 (29 *e8 30 l;l xa7 "tt x f6 3 1 'ith5 + <IJ'd8 32 axb3 is similar) 30 K xa7 bxa2 (30 "tt x f6 3 1 'ith3 + and "tt x b3 + ) 3 1 "tt h 3 + *d5 32 'itb3 + <IJ'c6 33 "tt xa2, with a slight material plus for White together with a raging attack. 22 K xd3 �e5 Now White wins by force. gxf6 23 �f6 + 24 "tt h6! Once again 24 gxf6 + � g6 re pulses the attack, but now 24 � g6 loses to 25 M h3 � e8 26 fxg6 fxg6 27 "tt x h7 + *f8 28 "tt h 8 + and H h7 mate. 24 *h8 25 K h3 J.. xf5 26 g6! j_ xg6 27 K xg6 is also mate, so Black must take the rook. 24. . . <£J xd3 + (84)

Black is hoping for 25 cxd3 '* h8 when 26 gxf6 X g8 27 X g7 J.. xf5 defends, as does 26 g6 fxg6 27 fxg6 "tt c 7 + and 28 "tt g 7. 25 <IJ'bl! fxg5 25 '* h8 26 g6 fxg6 27 fxg6 threatens both 28 "tt xh7 mate and 28 g7 + , while 25 "it b6 26 M g3 only makes matters worse. "tt xf6 26 f6 g4 27 "tt xf6 The only way to avoid mate. Although rook, bishop and two pawns amount to enough mater ial to balance a queen, Black still suffers from his bad king position. 28 "tt g5 + fi h8 * g8 29 "tt f6 + 30 � d4 � e5 31 h3 Intending 32 hxg4 followed by � f5 . h5 31 �g6 32 "tt g5 + gxh3 33 "tt xh5 34 "tt d5 Attacking a8 and g6. White finishes the game with the same energy he has displayed through out. 34 J.. e6 h2 35 � xe6 36 ll xg6 + Resigns 36 fxg6 37 � xf8 + and 'l!!f xa8 wins all the black pieces. .

Game 22 Nunn-Gheorghiu Hamburg 1984 1 2 3

e4 � fJ d4

c5 e6 cxd4

Kan Variation 101 4 5 6

�xd4 J.. d3 0-0 (85)

a6 � f6

85 B

d6 6 The surprising move 6 . . . e5 (6 d5 7 e5 is very bad for Black since the natural 7 � fd7 loses to 8 � xe6!) was played in Fedor owicz-Dorfman, New York 1 989, and now 7 J.. g 5! exd4 (7 h6 8 J.. xf6 -rtxf6 9 �e2 d6 1 0 � bc3 J.. e6 1 1 f4 is good for White) 8 e5 J.. e7 (8 . . . -rta5 9 J.. d 2! and 8 . h6 9 exf6 hxg5 I 0 !:! e I + are very bad for Black) 9 exf6 J.. xf6 I O J_ xf6 ttxf6 I I !:! e i + lli' f8 1 2 J.. e4 i s a little better for White since Black will still have an iso lated pawn after White regains the front d-pawn. Black has a major alternative in 6 . . . -rtc7 7 -rte2 (7 c4 �c6 8 � xc6 dxc6 is now considered sat isfactory for Black because White has spent a move on c4, which in this position only serves to weaken d4) d6 (7 . . . �c6 is bad because of 8 � xc6 and 9 e5, while 7 . . . J.. c5!? 8 � b3 is a speciality of Eingorn; the two examples 8 . .

. . . J.. e7 9 c4 d6 10 �c3 0-0 I I J.. d2 b6 1 2 �d4 j_ b7 1 3 lli' h l � c6 1 4 �xc6 j_ xc6 1 5 K ae l ttb7 1 6 a4 �d7 1 7 b4 -rtc8 1 8 f4, Govedarica-Eingorn, Novi Sad I 988 and 8 . . . j_a7 9 lli' h l d6 I O c4 � c6 1 1 J.. g5 J.. d 7 1 2 � c3 � d4 1 3 � xd4 j_ xd4 1 4 � ac i -rtc5 I 5 J_ d2, Smyslov-Eingorn, USSR Ch. I 988 were both very slightly better for White, but the idea is an interesting one and I expect to see more of it; perhaps 8 J_e3 is the best reply) 8 c4 g6 (this is the move which gives 6 . . . -rtc7 its distinctive flavour; if Black de velops his bishop at e7 we reach positions similar to those in the main line below, but after . . . g6 White's prospects of a direct kingside attack are reduced; on the other hand d6 is weakened) 9 �c3 (it is also worth considering b3 and J.. b2) J.. g 7 IO l! d 1 (White's idea is to attack d6 by jtc2, �f3 and J.. f4) 0-0 (86) and now:

( 1 ) 11 J.. c2 � bd7 (1 1 . . . b6 1 2 � f3 � fd7 1 3 J_e3 �c6 1 4 .l, ac 1 jt b 7 1 5 J_ b 1 i s good for White,

102 Kan Variation Ivanovic-Cvitan, Vrsac 1 987, while 11 . . . '2l c6 12 '2l xc6 bxc6 1 3 JH4 e5 1 4 j}_e3 J.. e6 1 5 J_b3 � fd8, Cabrilo--Gheorghiu, New York Open 1 988, gives White a slight plus after 1 6 h3 or 1 6 � d3) 12 22l f3 <E:� g4 (not 12 b6 1 3 j}_ f4! 2£l e5 1 4 2£l xe5 dxe5 1 5 J_ e3 J.. b 7 1 6 2£l a4 b5 1 7 cxb5 axb5 1 8 2£lc5 j}_c6, Ljubojevic-Hulak, Wijk aan Zee 1 987, and now 1 9 b4! followed by j}_ b3 is good for White) 1 3 J.. f4 (not 1 3 J_g5 <E:� de5 14 '2ld2 b5! and Black seized the initiative in Klinger-Jukic, Berne Open 1 988) and now 1 3 2£lde5 1 4 2£l d2 b5?! 15 h3 '2l f6?! 16 cxb5 axb5 1 7 J_ b3 b4 1 8 '2lb5 �b8 1 9 '2l xd6 was good for White in Vrnjacka Cabrilo--Kovacevic, Banja 1 9880 However Black's play was poor in this example; had he played 1 3 0 0 0 '2l ge5 White would have a slightly inferior version of line 2a below in which his bishop is on c2 instead of the better square b l . (2) 1 1 22l f3 (this may be more accurate since White often prefers to retreat his bishop to b 1 ) and now: (2a) 11 . . . '2lbd7 1 2 � f4 <2! g4 1 3 � ac l b6 1 4 A b 1 '2l ge5 1 5 b3 '2l xf3 + 1 6 'ttt xf3 l£je5 1 7 �e2 and now: (2a l ) 17 . . . jtb7 1 8 �d2 ( 1 8 j}_ e3 � fd8 1 9 �d2 � ab8?! 20 h3 j}_a8 21 f4 2£l c6 22 'ttt f2 , Armas Gheorghiu, Romania Cho 1987, was slightly better for White, but Black can try 19 Jt f8 or 1 9 '2l g4) � fd8 1 9 j}_g5! � d7 (Black

would like to play 1 9 f6 20 j}_e3 g5, but this allows 2 1 l£j a4) 20 h3, intending f4, with advan tage to White, Armas-Ionescu, Romania Cho 1 9880 (2a2) 17 . . . � b8 (this seems to be more accurate; Black first of all secures the weak b6 pawn) 1 8 h3 jt b7 19 �d2 � fd8 20 J.. g 5 (the immediate 20 j}_e3! transposes to Armas-Gheorghiu above, which was slightly better for White) f6 2 1 jt e3 g5 22 j}_d3 (now 22 2£l a4 is met by 22 jL c6) A c6 23 '2l b5 fi/f7 with equality, Wolff Hulak, Toronto 1 9890 (2b) 11 . . 2£lc6 (in many lines it makes no difference whether Black plays '2l bd7 or l£j c6 because he will continue with .'£\ f6-g4--e5 xf3 + followed by .'£)e5 in any case, but here we examine one independent line) 1 2 .ii. f4 '2ld7 1 3 � ac l <2\ de5 1 4 b3 yjfe7 ( 1 4 l£j xf3 + will trans pose to line 2a) 1 5 j}_ b 1 (White aims to expel the e5 knight by moving his 22l f3 and Jt f4 away) � d8 ( 1 5 il_d7 1 6 .'£) e l � fd8 1 7 'rl!i'd2 jt e8 1 8 jtg5 f6 19 jj_ h6 .il_ xh6 20 �xh6 also gave White an edge in Timoshchenko- Eingorn, Tallinn 1 989) 1 6 jL e3 jj_d7 1 7 '2l d2! .ii_ e8 1 8 f4 '2ld7 1 9 22lf3 );:;!, ab8 and White retains a slight advantage, Kuzmin Eingorn, M oscow 1 9890 7 c4 (87) J.. e7 Other 7th moves: (1) 7 g6 8 � c3 J.. g7 9 '2l b3! 0-0 1 0 Jt e2 i£jc6 1 1 J.. f4 <E:� e8 (Black has little choice as 1 1 0 .

. . .

Kan Variation 103 �h3, but also lining up against the b7 pawn. (3) 7 . . b6. Normally this transposes to lines considered be low. 8 '2'Jc3 White has two main attacking plans, which are distinguished by the 4evelopment of his queen's bishop. Firstly he may build up a slow kingside attack by b3, jt b2, 2£jc3, li1e2, f4, ll ae 1 and so on, with the ultimate aim of a breakthrough by f5. The other plan is to prepare for e5 by l£Jc3, 'itfe2, f4, � d2 and l;l ae l . The important point is that with the bishop on b2 the e5 plan is much less effective, because White ends up with a pawn on e5 and this would block the bishop on b2. 8 0-0 9 li1e2 M oves such as "i!!t e 2, * h l and f4 are logical because they do not commit White to one plan or the other. My view is that the e5 plan is most effective against . . . '2'1 bd7 by Black, because then the queen's knight blocks the retreat of the one on f6. Therefore it is often useful to delay committing the c l bishop until Black has moved his b8 knight. Against 2£jc6 White will take on c6, then play b3 and §l b2, and against '2'J bd7 White will play jtd2 and K ae l . 9 b6 Black also delays for as long as possible. 10 f4 .

2£j e5 1 2 c5 is very awkward) and after 1 2 li1d2 b6 1 3 ll fd 1 2E:le5 1 4 � ac l 'f!! c7 Black equalized in Nunn-Gheorghiu, Vienna 1 986. However 1 2 c5! is unpleasant for Black, for example 12 dxc5 1 3 )\'ltxd8 Qlxd8 1 4 Qla4! or 1 2 e5 1 3 � e3 � e6 1 4 Qld5. (2) 7 . . . �d7 8 Qlc3 Qlc6 9 Qlxc6 � xc6 1 0 li1e2 � e7 1 1 b3 0-0 1 2 jt b2 l1l[ e8 (12 . Q�d7 1 3 f4 is slightly better for White, while after 12 . �b8 1 3 a4 � e8 1 4 li\ ae l Q� d7 1 5 f4 � h4! 1 6 � d l A_f6 1 7 b4!? the position was unclear in Nunn-Bischoff, Dort mund 1 987; 14 f4! was more ac curate, not committing the a l rook for the moment) 1 3 f4 d5!? 1 4 cxd5 exd5 15 e5 Ql e4 1 6 l£l xe4! dxe4 1 7 jt xe4 J..b5 and now 18 ll!rg4?! �d2! 19 jt xh7 + !? was unclear in Arnason-Toshkov, Jurmala 1 987. White can draw by 18 �c2 � xfl 1 9 .�.xh7 + *h8 20 *xfl lii\ c8 2 1 �d3! g6 22 � xg6 hxg6 23 �xg6 �d2, but 18 )\'lt f3!? is the most promising, not only playing for a possible at tack by . . .

. . .

. .

104 Kan Variation Despite the above (rather sub jective) comments it is quite rea sonable to play b3 straight away, the advantage being that White can sometimes manage without 'itr>h l . After 1 0 b3 jl_ b7 1 1 jt b2 (88) there are two lines: 88 B

.!£) xb5 li;l xa2 20 jl_ xg7 !HaS 2 1 i!i xa2 !!! xa2 22 jL d4 e5, Hellers Adamski, Eeklo l 9S5, and now 23 Jtc3 gives White a won ending. ( l e) 13 � ael !i!, eS 14 f4 g6 1 5 e5! dxe5 1 6 fxe5 jl_c5 + 1 7 * h l .!£)g4 I S jl_ e4! .!£) xe5 1 9 jl_ xc6 .!£) xc6 20 )'tf1 again with a tre mendous attack for White, Prague Ermenkov-Gheorghiu, 1 9S5. .!E\ bd7 12 !!! ad l ( 1 2 f4 (2) 11 >�tc7 is also playable) � eS ( 1 2 1 3 jl_ b l li;l feS 14 f4 � acS 1 5 221 0 jl_ f8 1 6 * h i Jtc6 was less accur ate and after 1 7 e5! jl_ xf3 I S !il, xf1 dxe5 1 9 fxe5 .!£) g4 20 !il, xf7! White had a very dangerous attack in Plachetka-Ravikumar, Copenha gen 1 9SO) 1 3 Jt b l >�tbS 14 f4 A f8 1 5 * h l !il, a7 1 6 .!£) f3 ;l aS with a double-edged position, Akesson 1 9SO, Copenhagen Mestel, although I still favour White. jl_ b7 10 1 1 * h l (89) . . .

.:'£\ c6 (as the earlier (1) 11 explanation makes clear, this move plays into White's hands since we reach positions similar to the main line below, but with White having saved about half a tempo by missing out * h 1 ) 1 2 .!£) xc6 jl_ xc6 and Black has been highly unsuccessful from this position: ( l a) 13 ,g adl )'tbS?! 14 a3 (what on earth is this for?) � dS? (it doesn't matter about the tempo spent on a3 in view of the way Black plays) 1 5 f4 .!£) d7 1 6 e,d5! jl_ f8 17 B. f3! li eS IS li h3! g6 1 9 )'tg4 )'tdS 20 lil fl jl_ g7 2 1 J_ xg7 * xg7 22 f5! with a massive at tack, lvanovic-Ermenkov, Plov div 1 9S3. ( l b) 13 f4 .!£l d7 14 14, ad l b5? 1 5 cxb5! axb5 1 6 jl_ xb5 11t b6 + 1 7 J!l f2! jl_xb5 I S 'ilfxb5 >�txb5 1 9

ll Or: (1) 11 � ac l ( 1 3

.!£)c6

. • .

.:'£\ bd7 1 2 Jtd2 >�tc7 1 3 1 4 � f1 g6

l!l ae l l!l feS

Kan Variation 105 � xe5) 1 8 gxf5 exf5 1 9 j)_ d5 + 1 5 K g3 ! flh8 1 6 K h3 e5 1 7 � f3 ff h8 20 �xe5 are all good for exf4 1 8 J.. x f4 J.. f8 1 9 'ltf2 � c5 20 White. jtc2 was also good for White in jtxc6 12 � xc6 I vanovic-Peev, Ba1asiha 1 977) g6 13 b3 1 4 b4 )l ac8 1 5 a3 'ltb8 (the ad <E)d7 Or 1 3 'ltc7 1 4 J.. b2 .i! ad8 vantage of playing f4 is that the weakening of c4 created by White and now: ( I ) 1 5 � adl g6 1 6 J.. b l � h5 1 7 playing b4 cannot be exploited by K d3 ( 1 7 g4 �g7 1 8 f5! exf5 1 9 . . . � e5) 1 6 �f3 )l fe8 1 7 B, ce 1 J.. f8 1 8 �g5 h 6 ( 1 8 e 5 1 9 f5 gxf5 was probably better) e5! 1 8 gave White a strong attack at no fxe5 dxe5 1 9 M xd8 K xd8 20 �d5 material cost in Commons-Naj with just an edge for White, Matulovic-Tringov, dorf, Lone Pine 1 9 76) 19 � xf7! Vrnjacka flxf7 20 e5 �g8 21 'ltg4 �e7 22 Banja 1 986. J.. xg6 + ! � xg6 23 f5 � dxe5 24 (2) 15 K ael (intending � d5) fxe6 + ! (24 fxg6 + ff g8 is unclear) J.. b7 1 6 J.. b 1 � d7? ( 1 6 g6 is ffe7 (24 . . . * g7 25 !! xe5 dxe5 26 probably better, when 1 7 'ltd3 is only a slight plus for White) and I, f7 + mates) 25 'ltxg6! (the cli now 1 7 'lth5 H fe8 1 8 li, e3 � f6 1 9 max of a magnificent combi 'lth3 g6 20 f5! gave White a deci nation) ff d8 26 l;i xe5 dxe5 27 sive attack in Nunn-Gheorghiu, J.. x h6 11 xc4 (27 J.. x h6 28 Biel 1 983, which is annotated in H d 1 + mates) 28 'ltxe8 + (White gives up his queen after all) ff xe8 detail in Secrets of Grandmaster Play by Peter Griffiths and the 29 )l xf8 + fj e7 30 )l xb8 jtc6 3 1 � d 1 b5 32 * g l •xe6 33 K b6 present author. For some reason Resigns, Commons-Peev, P1ovdiv Gheorghiu repeated the whole line in the game Mokry-Gheorg 1 976. hiu, Prague 1 985. That game con (2) 1 1 l;l e8 (dubious as it allows White to play for e5 with tinued 1 7 'i'lf'g4 J.. f6 1 8 � e3 g6 1 9 K h3 * h8?? 20 � d5 ! winning, as out delay) 1 2 � f3 g6 (12 . . . � bd7 1 3 e5 dxe5 1 4 fxe5 <2) g4 1 5 A f4 20 exd5 is met by 2 1 'lth4. Of J.. xf3 1 6 )l xf3 and 12 . . . �c6 1 3 course 1 9 )g[ fe8 is better, but e5 dxe5 1 4 fxe5 <EJg4 1 5 J.. f4 are 20 'ltg3 intending f5 gives White a good for White) 1 3 e5! <2) h5, Sax dangerous attack in any case. Bellon, Dubai 01. 1 986 and now 14 j}_b2 g6 Sax recommends 14 A e4 � c6 1 5 14 J.. f6 leaves d6 weak and g4 � g7 1 6 f5 !, when the lines 16 after 1 5 K ad l 'i'lf'c7 1 6 J.. b l . . . exf5 17 gxf5 dxe5 1 8 fxg6 hxg6 )l fd8?! ( 1 6 � ad8 was a better 1 9 � xe5, 16 . . . gxf5 1 7 gxf5 chance) 1 7 K d3 g6 1 8 .M fd 1 �c5 �xf5 ( 1 7 exf5 1 8 jt d5) 1 8 19 � d5 ! exd5 20 J.. xf6 � xd3 2 1 J.. xf5 exf5 1 9 <E)d5 and 16 . . . exd5! �xf4 22 'lt f3 � h5 23 jt xd8 dxe5 1 7 fxg6 f5 ( 1 7 . . hxg6 1 8 � xd8 24 dxc6 White had a clear • . .

106 Kan Variation advantage in Marjanovic-Rajko vic, Yugoslavia Ch. 1 983. 15 K ad1 The purpose of this (rather than K ae 1 ) is to prevent the de velopment of Black's e7 bishop to the long diagonal. After 1 5 il.. f6, for example, 1 6 _;t b 1 �c7 1 7 l!td2 attacks d6. 15 K e8 16 _;tb1 (90) 90 B

2 1 exd5 e5 22 l!fg4 � f6 23 l!fe6 + '* g7 was unclear in Prasad Gheorghiu, Biel Open 1985. (3) 16 jj_f8 17 e5! dxe5 1 8 jj_e4! gives White a dangerous at tack. The game Vogt-Gheorghiu, E. Germany-Romania 1 984 con tinued 1 8 . . *'c7 ( 1 8 . . . jJ_ xe4 1 9 �xe4 jj_g7 20 jJ_xe5 jj_xe5 2 1 fxe5 'fi g7 22 l!tf2 wins after 22 . . . l!fe7 23 *'f6 + ! \t1 g8 24 'flixe7 !! xe7 25 � xd7 or 22 . . . � e7 23 � g5 l!f g8 24 �xf7 K f8 25 � xd7!) 1 9 l!ff3 jt xe4 20 � xe4 f5 (20 _;t g7 2 1 fxe5 �xe5 22 � f6 + 'flh8 23 jJ_ xe5 wins material) 21 X xd 7 fxe4 22 'f!fd 1 �c6 23 jj_ xe5 X ac8 24 �d4 jte7 (24 . '!A, e7 25 � d6 drops the b-pawn, while 24 . . b5 25 'f!la7! and 24 . . . jj_ c5 25 I:I g7 + 'flf8 26 j_d6 + lead to mate) 25 f5! (the immediate 25 _;th8 jj_f8 26 � g7 + 'fl xh8 27 e5, but if � xg6 + is met by exf5 this Black now plays 25 gxf5 26 line wins) K cd8 (25 jJ_ h8 J.. f8 27 M\ xf5! exf5 28 ll g7 + mates) 26 I:I xe7! (26 fxg6! � xd7 27 gxh7 + 'fl xh7 28 I� f7 + \t1h6 29 'W!'e3 + _;tg5 30 l!fh3 + also wins) exf5 (26 K xd4 27 �g7 + is mate next move) 27 !:! xe8 + � xe8 28 �xb6 Resigns. 17 llt d3 A flexible move attacking d6 and preparing l!fd4 or �h3 according to circumstance. 17 jj_f8 This move cost Black forty mi nutes, presumably checking that the line 1 8 �d4 e5 ( 1 8 . . . jj_ g7 1 9 llfxg7 + wins) 1 9 �d5 jj_xd5 20 . . .

. .

16 l!fc7 The alternatives are: ( 1 ) 16 . K a7 1 7 a4 (the direct 1 7 l!td3 was also tempting) jj_f8 1 8 jj_c2 l!fa8 1 9 lltf2 �c5? ( 1 9 � c7 was better) 20 Qld5! exd5 2 1 cxd5 jj_ xa4 22 bxa4 j_g7 2 3 !J1 xg7 'fi xg7 24 e5! and White stands well, Popovic-Kotronias, Pucar evo 1 987. (2) 16 *'b8 17 f5 ( 1 7 a4!? is possible, but I like 17 l!fd3 b5 1 8 cxb5 axb5 1 9 � e2! b4 20 Ql d4 jt b7 2 1 llth3 with dangerous threats on the kingside) b5 1 8 fxe6 fxe6 1 9 cxb5 axb5 20 � d5!? j_ xd5 (20 . . exd5 21 exd5 jj_ f6 22 *'d2 j_xb2 23 dxc6 is good for White) .

. . .

Kan Variation 107 '/Wxd5 exf4 2 1 � xf4 �e5 22 Z:�: dfl lli\ a7 presented no dangers . 1 8 �b5!? A shock for Black. After 1 8 axb5 1 9 'ltc3 eS 20 cxb5 B, ac8 2 1 bxc6 '/txc6 2 2 'i!ltt f3 Black cannot exchange at f4 since he has no satisfactory way to cover fl, so White gets to play 23 f5, when Black's white squares look very sickly. 18 jixb5 19 cxbS axb5 1 9 . a5 20 !:!: c l �c5 2 1 'ltc3 is much worse as Black has to play the weakening . e5. �b7 20 '/txb5 The upshot of White's mini combination is that he has the two bishops and a queenside pawn majority. Now 2 1 a4! would have been logical, relieving the b l bis hop of its defensive duty and pin ning down the b6 pawn. If Black remains passive the b I bishop can eventually move to b5. During the game I didn't like 2 1 a4! �c5 22 �c4 X ac8, but 23 f5 e5 24 ll!. f3 gives Black no way to exploit the position of White's queen (24 � e6 25 fxe6!), and consolidation by 25 'lte2 will be good for White. b5! 21 'lte2?! Black takes the chance to pre vent a4. Now . . . b4 would perma nently cripple White's queenside pawns, so 22 b4 Black cannot transfer his knight to c4 because b5 is weak, for example 22 '/ta6 (threat . <EJ b6) 23 ji d3 '(!lfxa2 24 ji xb5

'/Wa7 (24 ll: ed8 25 ji xd7 and .i, a l ) 25 jic6 followed by jid4 and b5, with an excellent position for White. fj!. a6 (91) 22 Another useful defensive move, covering d6 in preparation for ji g?. Here I thought for a long time trying to find a way to keep the advantage.

23 � f3! 24 ji xg7 25 iiii fd3 26 � h3 The point of White's ma noeuvre is that he gains a tempo by attacking b5 to transfer his rook to the h-file, reviving some threats against Black's king. 26 X b6 Not 26 . . . 'ltc4? 27 "*'b2 + and 28 X c3 trapping the queen. 27 '(!lfd2 The threat is 28 e5 d5 29 f5 and the queen gets to h6. 27 .:2)f6 Once again Black finds a good defence, relieving the pin down the d-file. 28 f5 exfS

108 Kan Variation e5 29 trh6 + \fi g8 30 g4! 28 is very unpleasant. d5! 29 exf5 All these difficult moves were very time-consuming, so that Black had only a couple of mi nutes left to reach move 40. At first sight White can win by 30 fxg6 fxg6 3 1 trh6 + \tg8 32 Axg6 hxg6 33 tyh8 + (33 trxg6 + tyg7 34 trf5 I. ce6 defends) *f7 34 K h7 + *e6 35 K e 1 + (35 trxf6 + * xf6 36 X xc7 X be6 gives Black enough for the pawn in view of his active rooks and king), but then comes * f5 ! 36 D, fl + *g5! 37 35 h4 + \t g4 and White's queen is trapped. 30 fxg6?! The immediate 30 X fl ! was much better, when the defence Black plays in the game would have been prevented. fxg6 30 xn 31 Now truly threatening 32 trh6 + *g8 33 Axg6. 31 <2l g4! Suddenly exploits Black White's weak back rank. If 32 <2lf2 + wins. trxd5? then 32 32 trd4 + H f6 (92) White cannot take the rook or the knight, nor can he play X hf3 . The move played supports the back rank and attacks the knight. h5 33 A d3 Forced, but good. The draw would now be in sight if it were not for Black's time-trouble. <2l xf6 M xf6 34

K e1 + K f3 A fl tye7?! tye5 37 trxe5 K xe5 38 36 Axb5 gives White good winning chances as the d-pawn cannot pass over the d3 square. However 36 . . . �d6! was more accurate, for example 37 *g1 (37 h3 I! e4 38 'l'!lrc3 K xb4 39 H xf6 trxf6) l!!( e4 38 trc5 (38 tyc3 d4 39 tyd3 <2l g4! 40 g3 X e 1 threatening . . . <2le5 is unpleasant) trxc5 39 bxc5 X e 1 and a draw is inevitable (40 H c3 d4). 37 \tg1 H e4 Black can also choose to wait, since progress isn't easy for White, but in time-trouble it is very natural to break the pin. d4 38 trc3 K xb4 39 X xf6 d4 Not 38 because of 40 M e6! 39 trd3 With two moves still to make Black suddenly finds himself in trouble over the b5 pawn. 39 . . . tr xb4 loses to 40 Ill xf6, so the <2ld5 40 best chance is 39 trxb5 <£! xb4. Then 41 Ac4 H e l + 42 fif2 fih6 isn't danger-

35 36

• . .

Kan Variation 109 ous, so White's winning prospects are very slight. <£)g4? 39 40 h3 and Black's flag dropped before he could make his 40th move. After 40 <£) e3 41 iifxb5 <£) xfl 42 � xfl � e2 43 �d3 iife3 + 44 iifxe3 dxe3, suggested by Black after the game, White can win by 45 a4! (but not 45 b5 � xa2 46

!I b l 1! d2! 47 '* fl

.1: f2 + 48 <f1 e l :!:! xg2 49 :!:! b 4 K d2 5 0 b 6 H d8 5 1 b7 1! b8 when Black has drawing chances) H b2 46 b5 g a2 47 b6 il xa4 48 H b 1 H d4 49 * fl ! (49 b7 e2) H d8 50 b7 .1: b8 5 1 *e2 etc. Obviously Black has other ways to play, but the two con nected passed pawns give White good winning chances in any case.

�aroczy Bind

This most commonly arises if Black plays an early g6, for example 1 e4 c5 2 � f3 � c6 3 d4 cxd4 4 .t:l xd4 g6, aiming to reach a Dragon position without having played d6. This restricts White's options quite severely, since he has to be careful not to allow Black to play d5 in one go, saving a crucial tempo. An effective way out of this dilemma is to continue 5 c4, setting up the formation of pawns on c4 and e4 known as the Maroczy bind. The asset of this formation is the auto matic restraint of . . . b5 and d5, Black's basic freeing thrusts. Black does sometimes succeed in orgamzmg b5, but this is nor mally only good when White has made a mistake. White's main asset is his space advantage, lead ing to the corollary that he should avoid exchanges which would re lieve the cramp in Black's posi tion. If Black does succeed in liquidating to an ending, White's c4 pawn and black-squared weak nesses can become a liability. Play often becomes a matter of slow manoeuvring as White tries to in crease his space advantage and force weaknesses in the Black position while his opponent re mains crouched on his back two

ranks waiting for the first sign of over-extension to launch a coun terattack. The Maroczy Bind can also occur if Black adopts an unusual move order, for example I e4 c5 2 � f3 � c6 3 d4 cxd4 4 .t:l xd4 d6 and now 5 c4 will proba bly transpose to this chapter after 5 .t:l f6 6 .t:lc3 g6. Although the Maroczy Bind is slightly passive for Black, players such as Larsen, Petursson and Velimirovic have shown that by patiently waiting for a lapse of concentration from White this line can offer winning chances for Black. The theoretical opinion is that White should maintain a slight advantage, but White players should not believe that this is a line in which they cannot lose. Game 23 Karpov-Kavalek Nice Olympiad 1974 c5 1 e4 2 � f3 � c6 3 d4 cxd4 4 � xd4 g6 5 c4 Here Black has two possibili ties. He may play 5 ll. g7 in order to force White's jj_e3, but in doing this he forfeits the chance to take on d4 at a moment when

Maroczy Bind 111 White must recapture with the queen. The alternative is 5 � f6, which will often transpose to 5 . . . !/,t g7 if Black does not take up the chance to play � xd4. The 5 !/,t g7 systems are exa mined in game 24, while in this LE� f6. game we look at 5 �f6 5 6 � c3 (93) 93 B

d6 6 If Black adopts the move order 6 � xd4 7 �xd4 d6 (possibly to avoid 7 � c2 as in the next note) then White has an interest ing alternative based on playing j_d3 rather than j}_e2, providing the e4 pawn with useful extra pro tection, for example 6 �xd4 7 'i!)fxd4 d6 8 j}_g5 !/,t g7 9 �d2 0-0 (9 !/,t e6 1 0 � c l � c8 1 1 b3 *a5 1 2 f3 h6 1 3 jte3 0-0 1 4 !/,td3 \li h7 1 5 0-0 a6 16 h3 � d7 1 7 f4 f5 1 8 exf5 J_xf5 1 9 !it e2! was good for White in Polugayevsky Belyavsky, USSR Ch . 1 975) 1 0 j}_ d 3 ( 1 0 f3 i s also interesting, for example after 1 0 !it e6 1 1 !ii c I � a5 1 2 b3 a6? 1 3 � d5! �xd2 + 1 4 \lixd2 !/,t xd5 1 5 cxd5 � fc8 1 6

K xc8! K xc8 17 g3 \lif8 18 !/,th3 K c7 19 � c l � xc l 20 * xc l White was winning i n Byrne-Gar cia Padron, Torremolinos 1 97712 K fc8 was better when 1 3 !/,te2 a6 1 4 � a4 transposes to Karpov-Kavalek) !/,t e6 ( 1 0 a5 1 1 0-0 a4 1 2 K ac l J.. e6 1 3 'lli!' c2 gave White his usual space advan tage in Portisch-Reshevsky, Petropolis 1 973) 1 1 � c l �a5 1 2 0-0 � fc8 1 3 b3 a6 1 4 � fe l * f8 ! (14 b5? 1 5 -2ld5! 'Wfxd2 1 6 !/,txd2 !/,t xd5 1 7 exd5 bxc4 1 8 !/,txc4 \li f8 1 9 b4 with a clear plus for White, Geller-L. Garcia, Bogota 1 978) 1 5 Zi c2 ( 1 5 f4 is Jansa's double-edged suggestion) b5 1 6 cxb5 axb5 1 7 � ee l b4 1 8 � b5 � xc2 1 9 � xc2 with just an edge for White, Polugayevsky Jansa, Sochi 1 974. 7 !it e2 -21 xd4 This is Black's last chance to force White to recapture on d4 with his queen. If he plays 7 !/,t g7 White should transpose to game 24 by 8 j{_e3. At one time 8 -2lc2 was thought the best reply to 7 . j_ g7, but after 8 �d7 9 j_d2 a5! (not 9 � c5?! 1 0 b4 -2l e6 1 1 � c l 0-0 1 2 0-0 f5 1 3 exf5 gxf5 1 4 f4 � ed4 1 5 c2l xd4 � xd4 1 6 J.. e3 with a positional advan tage for White, Nunn-Rind, Man chester 1 980) 10 0-0 0-0 1 1 if;\ c I �c5 Black has a much better ver sion of the Maroczy Bind than in other lines. In Nunn-Petursson, Wijk aan Zee 1 990 I made matters worse by 1 2 b3? -21 b4! and Black had a clear advantage.

112 Maroczy Bind London 1 987, so it makes sense for Black to try to force White's rook to the less active square c l ) Or: ( I ) 9 . . . � d7 10 �d2 h6 (after 1 5 a3 ( 1 5 � ac l i s also possible, 10 ;Lc6 White should continue although 1 5 . . . a6 I6 f5 �d7 1 7 with 1 1 f3) l l J,. f4 a6 1 2 0-0 J,.c6 h3 �c6 I 8 jtd3 �d7 was unclear 1 3 f3 0-0 14 a4 <E) d7 1 5 a5! <E.)c5 1 6 in Gulko-Petrosian, Biel IZ 1 976) a6 I 6 f5 �d7 1 7 b4 �e5 (Larsen Fischer, match 1 97 1 ) and now, 94 according to various analyses of B this famous match, 1 8 � ad l would have been good for White. 10 li;l c1 (95) 1 0 0-0 is possible, and after 1 0 0-0 I I 'l'td2 a6 ( 1 1 � c8 1 2 b3 b5?! is doubtful because of I 3 e5! when 13 . . . dxe5 1 4 'i!ifxd8 � fxd8 1 5 � xb5 gives White the better ending and 13 . . . b4 1 4 g a3 with a small advantage for exf6 exf6 1 5 _2_ e 3 bxc3 1 6 ')$'xc3 f5 White, Pomar-Cordovil, Malaga 1 7 .k_d4 jt xd4 1 8 'ii!{ xd4 gives White the better middlegame, 1 972. Moscow (2) 9 . . h6 1 0 �e3 0-0 1 1 �d2 Bukic-Romanishin, 1 977) 1 2 f3 (not 1 2 � c l allowing *h7 ( 1 1 �a5 12 0-0 forces . b5! with equality) �a5 1 3 *h7 in any case) 1 2 O-O J.. e6 1 3 f4 1 2 � fd l � fc8 1 4 � d5 'iti' xd2 1 5 ( 1 3 � d4 � c8 1 4 b3 a6 1 5 ·l!i!fe3 � xd2 ,k xd5 ( 1 5 � xd5? 1 6 <E.)d7 1 6 � xg7 * xg7 1 7 f4 �b6 1 8 �xb6 � xb6 1 9 f5 �d7 20 � ad l exd5 � d7 1 7 J,. xe7! J.. h 6 and only gave White a slight edge in now not 18 � c2 J.. f5 and 1 9 Timman-Ribli, Amsterdam 1 973) � e8, but 1 8 � ad1! 1i_ xd2 1 9 � xd2 � e8 20 J.. xd6 with a clear li;l c8 (after 1 3 . . . � a5 14 f5 forces 14 . � d7 since 1 4 gxf5 1 5 plus for White, Britton-Donald exf5 � xf5 allows 1 6 � xf5, so son, Rhodes 1 980) 1 6 cxd5 w f8 White avoids wasting a tempo on White can claim a very slight plus. b3) 14 b3 �a5 ( 1 4 . . . a6 1 5 � ad l In general these endings with an �a5 1 6 � d4 is good for White open c-file are very drawish unless after both 16 . . . b5 1 7 f5 jt d7 1 8 Black has weakened his queenside � xf6! exf6 1 9 <E.)d5 l!i' xd2 20 by playing . b5 (which both � xd2, Nunn-van der Sterren, allows a4 and gives White an Groningen 1 974-5 and 16 . entry point at c6), or White can �d7 1 7 J,.xf6 exf6 1 8 �d5 �c5 + quickly seize the c-file by playing 1 9 * h i a5 20 f5, Nunn-1. Ivanov, his bishop to the h3--c8 diagonal. 8 9

� g7 � xd4 � g5 (94) �e6

. .

Maroczy Bind l l3 Neither situation exists here, so White's advantage is insignificant and in practice Black would have few problems reaching the draw. 10 �aS If Black omits this move, we again face move order questions. After 1 0 0-0 1 1 '*"d2 ili c8 1 2

b3 a6 the obvious 1 3 f3 allows 1 3 b5!? 1 4 cxb5 axb5 1 5 2! xb5 g xc 1 + 1 6 �xc 1 i:ii a 5 + 1 7 'M'd2 g a8 18 �xa5 g xa5 19 a3! .�xb3 20 fTf2 jj_ a4 2 1 � b 1 , Tukmakov Vaganian, USSR 1 984, and now 21 jj_xb5 22 � xb5 g xa3 is a simple draw. Therefore White should prefer 13 0-0, when 13 . �aS 14 f3 transposes to the next note, while 13 bS 14 cxb.5 axb5 1 5 jj_xb5 '*a5 ( 1 5 g xc3 1 6 -�xc3 � xe4 1 7 1!i1e3) 1 6 jj_d3 also seems good for White, e.g. 1 6 g xc3 1 7 �xc3 � xc3 1 8 g xc3 � xe4 1 9 jj_ xe4 jj_ xc3 20 jj_xe7. 0-0 11 'i!i!t'd2 After 1 1 g c8 1 2 f3 ! B lack is in a rather awkward situation since he has the wrong rook on c8 if he wants to castle. The p oint is that with the f-rook on c8 White is . .

. . .

never threatening �d5, because after the sequence �xd2 � xe7 + 1tt f8 White just loses his knight. With the a-rook on c8, however, Black will sooner or later have to waste time meeting this threat. So 12 . . . 0-0 1 3 b3 a6 1 4 0-0, for example, is better for White than the positions in the note to 1 0 .l c l . It is also too dangerous to take the c-pawn, for example 12 . . . jtxc4 1 3 � d5! tf'xa2 1 4 0-0 .:E) xd5 1 5 g xc4! � xc4 1 6 'l!!f xd5 l! a4 1 7 jtb5 + *f8 1 8 � c l ! and in Geller-Stean, Teesside 1 975 Black resigned because of 1 8 jj_d4 + 1 9 '*'xd4! � xd4 20 Ah6 + mating. In fact White had an even more convincing win by 14 � b4, since 14 't)ltb3 1 5 Ad1 traps the queen. If Black doesn't castle he soon runs out of things to do, e.g. 1 2 a6 1 3 b3 b 5 1 4 � d5! (as men tioned earlier, b5 makes this a much better proposition) tf'xd2 + 1 5 \fi xd2 A xd5 1 6 cxd5 fTd7 1 7 a4 h6 1 8 jte3 � xc 1 1 9 � xc 1 l! b8 20 H c6 with a very good ending for White, Nunn-Reuben, Lon don 1 978. 12 f3 � fc8 13 b3 By securely defending c4 and e4 White has prevented any tricks based on an immediate b5 so Black has nothing better than to prepare this thrust with a6. 13 a6 14 � a4 White chooses a favourable

1 14 Maroczy Bind moment to exchange queens. Black's last move weakened b6 and he must waste a tempo pre venting White's knight fork. 14 �xd2 + After 1 4 iitd8 White may either play 1 5 c5 (as suggested by Karpov) based on the idea 1 5 . . dxc5 1 6 �xd8 + � xd8 1 7 ..£) b6 when 1 7 � ab8 allows 1 8 J,. f4, or continue more quietly by 1 5 _zt e3 � ab8?! ( 1 5 . . ..£) d7) 1 6 J,. a7 � a8 1 7 .zt b6 iltf8 1 8 J.. e3 � ab8 1 9 ..£) b6 � c7 20 0-0 ..£)d7 21 ..£) d5 with an excellent position, Nunn Blum, London 1 979. 15 fj> xd2 (96) 96 B

15 ]l c6 After Black lost with this move in Karpov-Kavalek attention turned to 1 5 ..£) d7, but this doesn't seem to be any better, for example 1 6 g4 fJ f8 1 7 h4 ll( c6 1 8 Mt c2 <E)c5 1 9 ..£) c 3 a 5 2 0 <E)d5 H e8 2 1 J,.e3 J,. c8 22 h5 e6 23 � c3 f5 24 hxg6 hxg6 25 exf5 exf5 26 gxf5 J.. xf5 27 � ee l , Averbakh-Popov, Polanica Zdroj 1 976, or 1 6 h4 (this is perhaps even stronger) fJ f8 1 7 h5 h6 ( 1 7 . . ll( ab8 18

hxg6 hxg6 1 9 '£) c3 is also good for White) 1 8 .zt e3 g5 1 9 g3 � cb8 20 ..£)c3 b5 2 1 �d5!, Psakhis-Pigu sov, USSR 1 980 when in both cases White had a good ending. It is curious that the exchange of queens is just what White needs to start a kingside attack by h4-h5. The explanation is that Black's counterplay by b5 would be very dangerous with queens on the board, since it would lead to an attack against the centralized white king. With queens off this counterplay is relatively harmless. 16 '£)c3 � ac8 Karpov suggested 1 6 � e8 as a possible improvement, so as to trap White's bishop in case of 1 7 '£) d5 -2\ d7 1 8 ..£) xe7 + ?! '1J, xe7 19 .iJ.. xe7 f6 20 _i d8 b6. 1 6 � cc8 was played i n Sakharov Pereira, corr. 1 976, which finished in a draw after 1 7 ..£)a4 � c6! I don't suppose Karpov would have agreed a draw if Kavalek had 'found' 1 6 � cc8! *f8 17 '£) d5 18 1l_e3 '£) d7 Defending such an ending is an unpleasant task at the best of times, doubly so against Karpov. White has the choice of expanding on the queenside by a timely b4, or of gaining space on the other flank by g4 and h4, as in the note to Black's 1 5th move. Until White shows his hand Black can only wait. 19 h4 j_xd5 Black resolves to do away with the dangerous knight. 1 9 . . h5

Maroczy Bind 1 15 was well met by 20 <EJ f4 and Kar f5 pov's suggestion of I9 would require strong nerves in view of Black's king position. 20 exd5 � 6c7 21 h5 wgS?! (97) This move is probably a mis li!\ e8 followed by . . take. 2I . e6 would have opened the posi tion up for White's two bishops, but by activating his rooks on the central files Black would have d e again st counterplay veloped White's king.

22 f4! Most players would have rejected this as it allows Black's knight to settle at e4 (supported f5). Karpov, however, is by f5, actually aiming to provoke which gives him the lever g4 by which he can prise open Black's kingside. 22 <EJ c5 23 jj_ g4 <EJ e4 + rs 24 wd3 25 jj_f3 b5 It looks as though Black's counterplay has got off the

ground at last but White defuses it adroitly. bxc4 + 26 g4 g xc4 27 H xc4 28 bxc4 <2Jc5 + Black had little choice as he could not allow the white rook to occupy the b-file, nor could he � b8 without losing a play pawn at e4. 29 j_xc5! If there are rooks on the board opposite coloured bishops tend to lose their drawish influence. Here Black runs into trouble because his king is badly placed and he will have two or even three pawns stuck on white squares, where they cannot be defended by his bishop. 29 � xc5 B lack plays for a counterattack dxc5 30 h6 � a5. 29 by jj_ d4 3 I � b I is also unpleasant, the a6 pawn being particularly weak. 30 h6 j_ f8 I t looks horrible t o bury the bishop but Black lacked a reason able alternative, for example 30 fxg4 (30 . . . J.. f6 3 I lilt b i threatens 32 gxf5 gxf5 3 3 � b8 + wfl 34 j_ h5 mate) 3 1 j_xg4 lt.f8 J.. f6 32 lt.e6 + wf8 33 (3 I � b i ) 32 j_e6 + wh8 33 f5 ill( a5 34 g b I � a3 + 3 5 * e2 � xa2 + 36 \\' fl j_xh6 37 f6 and the pawn exf6 38 slips through (37 � b8 + wins a piece). 31 w c3 Karpov also analyses 3 I g5 � a5 22 � b i as good for White,

1 16 Maroczy Bind but the variations are by no means simple and in practice it is not surprising that he chose to prevent 3 1 . . . lil a5 by simple means (32 \91 b3, and the rook has to go back). fxg4 31 32 jj_ xg4 (98) 98 B

32 Black resolves to extract his king, even at the cost of the h7 pawn. Despite its dangerous appearance, he would probably have done better to try 32 . . � c7 33 j)_ e6 + f1Jh8 34 f5 � b7! (pre venting 35 M b l j)_ xh6 36 f6 exf6 37 SI, b8 + ), when White finds it hard to make progress because of Black's attack on the h6 pawn. f1Jf6 33 Ae6 + 34 jj_ g8 11 c7 34 jt xh6 35 � xh6 \91 g7 leads to a lost rook and pawn ending after 36 � xh7 + and 37 � xe7. e6 35 j)_ xh7 'IJ fl is refuted by 36 f5 g5 35 37 f6! exf6 38 A f5 \91 g8 39 h7 + f1Jh8 40 i: b l and 4 1 l! b8.

36 jj_ g8 exd5 37 h7 Not 37 jj_xd5? � h7. 37 jj_g7? Loses by force. 37 11 xc4 + 38 fld3 jj_g7 39 jj_xd5 (39 h8( "ltf) Axh8 40 11 xh8 )!;. c8 and 4 1 . . . \91 g7) � c8 leads to the same posi tion as the game but with White having a pawn less. Black would still be worse, but he would have chances of a draw. jj_h8 38 j)_xd5 39 ftd3 \91f5 40 f1Je3 � e7 + a5 41 \91 13 42 a4 � c7 43 jj_e4 + f1Jf6 � g7 44 � h6 44 ft g7 45 � xg6 + f1J xh7 46 � g l + *h6 47 � h i + and 48 iii\ h7 + wins the rook. 45 * g4 Resigns Black is totally paralysed.

Game 24 Vaganian-lvkov Moscow 1985 1 e4 c5 2 Q} f3 Q} c6 3 d4 cxd4 4 Q} xd4 g6 5 c4 jj_ g7 6 j)_e3 (99) 6 Q} f6 Black can also play Q} h6 followed by . . . f5, either with or without d6, but this idea does not equalize: ( I ) 6 . . . d6 7 e,c3 � h6 8 jt e2 0-0 9 0-0 f5 1 0 exf5 gxf5 ( 10 . Q} xd4 1 1 j_ xd4 j_ xd4 1 2 '*xd4

Maroczy Bind 1 1 7 99 B

� xf5 1 3 '¥'d2 jj_d7 was good for White after both 1 4 1i_f3, Tal Kupreichik, Sochi 1 970 and 1 4 j_ g4, Vilela-Estevez, Cienfuegos 1 9 80) 1 1 f4 '®' b6 ( 1 1 _it d7 1 2 � d2 c2l g4 1 3 _itxg4 fxg4 1 4 c2l d5 is a little better for White, Szabo Larsen, Vinkovci 1 970) 1 2 c2l xf5 � xb2 1 3 c2lxh6 + jj_ xh6 1 4 � c l 1i.._ g7 1 5 � c2 ( 1 5 �d5 sacrificing the a-pawn was possible) -�a3 1 6 'i#d2 with an edge for White, Spassov-Nicevski, Sofia 1 976. (2) 6 . . . � h6 7 � c3 0-0 8 jl e2 f5 9 exf5 jj_xd4 1 0 Ji_ xh6 � xf5 1 1 0-0 d6 ( 1 1 ·i!lJ' b6 is met by 1 2 �d5!) 1 2 _itf3 ( 1 2 '¥'d2 'i#a5 1 3 <i1J h 1 � f7 1 4 f4 was also promis ing in Shamkovich-Vasyukov, USSR 1 965) _itg7 1 3 jj_e3 1i.._ d 7 1 4 g e l b 6 1 5 1i.._ e4 � f7 1 6 1i_ g5! 1Lf6 1 7 .ii_ xf6 � xf6 1 8 �d2 'i!lff8 1 9 g ad l g d8 20 c2l d5 with a clear plus for White, Kudrin-L lvanov, New York 1 983. In this note we have seen a pawn structure with white pawns on c4 and f4 against black d-, e and f-pawns. This structure arises frequently in the Maroczy Bind

and i t i s almost always good for White. Black's problem is that any central pawn advance leaves him with either a backward pawn or hanging pawns, while if the pawns stay where they are White can just build up pressure down the d- and e-files. Similar com ments apply in the case where Black plays f5 and recaptures on f5 with a piece. The hanging pawns are a more important fac tor than the temporary piece ac tivity Black obtains. 7 �c3 0-0 Black has a major alternative in 7 c2l g4, which has recently become more popular. This new respectability has been based partly on an original idea for Black involving kingside pawn expansion, and partly on a reali zation that the older lines are not so bad for Black as had been thought. White has his typical space advantage, but Black's position is solid and Larsen in particular has achieved quite good results for Black. After 7 c2l g4 8 � xg4 � xd4 (or 8 J., xd4 9 j_ xd4 Q) xd4 1 0 0-0-0 e 5 1 1 '¥' g 3 d 6 1 2 f4 f6 1 3 f5! '11J f7 1 4 � b5 � xb5 1 5 cxb5 with an excellent position for White, Mestel-Karlsson, Las Palmas 1 982) 9 �d I (100) Black has three possibilities: ( 1 ) 9 . . . e5 1 0 Jl,d3 (this gives White a positional advantage with no risk, but there seems nothing wrong with the older tac tical line 1 0 Q) b5 0-0 1 1 �d2!

118 Maroczy Bind /0/ B

/00 B

�h4 1 2 J.. d3 d5 1 3 cxd5 <EJ xb5 1 4 J.. xb5 �xe4 1 5 0-0 � d8 1 6 � fd l when 1 6 j_e6 fails to 1 7 f3 �xd5 1 8 'iif e2 trapping Black's queen) 0-0 1 1 0-0 d6 1 2 �d2 J.. e6 (12 f5 1 3 exf5 gxf5 14 f4 <EJ c6 1 5 � ad l �e7 1 6 ..§.e2 <EJd4 1 7 ..§.xd4 exd4 1 8 <EJd5 gave White a clear plus, Andersson-Rogers, Malta 1 980) 1 3 l;l ac l a6 14 b3 � c8 1 5 f3 and now both 1S . . . fS 1 6 exf5 gxf5 1 7 f4 �f6 1 8 <EJ e2 � cd8 1 9 <EJxd4 exd4 20 J.. f2 , Ghi tescu-Radovici, Romania Ch. 1 977, and 1S . . . �aS 16 � fd l f5 1 7 exf5 <2:1 xf5 1 8 A e4, Tal-Partos, Nice 1 974, were very good for White. (2) 9 . <EJc6 l 0 �d2 'iif a 5 1 1 � c l 0-0 1 2 J.. e2 d6 1 3 0-0 j_e6 1 4 b 3 � ac8 1 5 f4 with a good posi tion for White, Polugayevsky Suetin, Kislovodsk 1 972. (3) 9 . . . <2:Je6 (the main line) 1 0 � c l ( 101) and now: (3a) 10 . . . � aS and now it is unclear whether the bishop should be developed at e2 or d3: (3a l ) 11 J.. d3 with a further branch : (3a l l ) 1 1 . . . ..§. xc3 + 1 2 � xc3 . .

�xa2 1 3 '(l1cl �a5 1 4 c5 is ex tremely dangerous for Black. (3a 1 2) 1 1 . . . d6 1 2 0-0 (better than 1 2 ·� d2 1l_ d7 1 3 0-0 ..§.c6 1 4 � fe l 0-0 1 5 1l_ h6 �e5! 1 6 j_ xg7 ·fii xg7 1 7 � cd l <EJc5 1 8 1l_ fl a5 with equality, Nogueiras-Korch noi, Montpellier 1 985) 0-0 1 3 j'j_ b l j_d7 1 4 f4 <EJc5 1 5 <EJ d5 with advantage to White in Mednis-D. Byrne, US Ch. 1 973. (3a l 3) 1 1 . . . b6 12 0-0 1l_ b7 1 3 f4!? (probably better than 1 3 '(l1d2 g5 1 4 � fd l d6 when 1 5 f3 j)_e5 1 6 * h l J.. f4 1 7 j)_ xf4 <EJ xf4 1 8 j)_fl <EJ e6 1 9 a3 ·;we5 20 <EJ d5 h5 2 1 b4 f!' f8 22 � e l � c8 was equal in Popovic-Cebalo, Yugoslavia 1 988, and 1 5 a3 h5 1 6 � c2 j_d4 1 7 b4 �e5 1 8 <EJd5 it_ xe3 1 9 fxe3 � c8 20 � fl <2:J g7 2 1 �f2 f6 was unclear in Ljubojevic-Korchnoi, Tilburg 1 987) 0-0 14 .it. b l d6 1 5 � f2 � ac8 1 6 <EJd5 jt xd5 1 7 exd5 <2:Jc5 1 8 a3 and White is better, A. Rodriguez-Hernandez, Cuba Ch. 1 988. (3a2) 1 1 it_e2 b6 12 0-0 1l_b7 1 3 f3 g5 (Larsen's plan increases the black-squared pressure and re serves e5 for the queen, but the

Maroczy Bind 1 19 obvious danger is that Black's jt b7 1 4 �d5 �d7 1 5 il.. g4! f5 1 6 king has to stay in the centre) 1 4 il_h3 �c7 1 7 �xc7 "jJxc7 1 8 exf5 � f2 ! ( a number o f other games gxf5 1 9 c5! with a fine game for had continued with �d2, but the White, Adorjan-Larsen, Hastings 1 986/7) ji b7 14 f3 ( 1 4 <E:Jd5 <E:J c7! rook transfer to d2 appears the best way of meeting Black's dou was played in Rogers-Hernandez, ble-edged plan) h5 1 5 J.. fl '*'e5 1 6 Calcutta 1 988, and now 1 5 0-0 e6 J4 d2 d6 1 7 �d5 ft f8 1 8 b4 jj_ h6 1 6 <E:J xc7 �xc7 1 7 f3 would have 1 9 � b3 g4 20 jtxh6 + � xh6 2 1 kept an edge for White) f5 1 5 exf5 'iiJ e3 ·f!! g7 22 f4 with a distinct gxf5 1 6 <E:Jd5 l';;l f7 1 7 0-0 <E:J f8 1 8 ll: fd 1 with advantage to White, advantage for White, Short-Lar Smejkal-Radulov, Skara 1 980. sen, Hastings 1 987/8. (3b) 10 . . . b6 1 1 _!d3 ( 1 1 b4 is (3c2) 12 . . . a5 1 3 a3 axb4 1 4 also good, for example 1 1 axb4 il_d7 ( 1 4 l';;l a3 1 5 <E:Jd5) il_ b7 1 2 J:,d3 0-0 1 3 0-0 �d4 1 4 1 5 0-0 jtc6 1 6 "1Jd2 � a3 ( 1 6 j_ b l � c6 1 5 a 3 d 6 1 6 �d3 � c8 jtxc3 1 7 �xc3 jixe4 fails to 1 8 1 7 f4 and White has consolidated jt h6 g e8 1 9 lil\ ce l followed by his space advantage, Suba-Tai jtg4 with a catastrophe at g7) 1 7 manov, Bucharest 1 979, or 1 3 <E:Jd5 * h8 1 8 1t b6 �d7 1 9 f4 with � c8 1 4 f4 with attacking chances a fine position for White, Por for White) j_ b7 1 2 0-0 ·iSoi' b8 tisch-Pfteger, Manila 1 974. (3d) 10 0-0 1 1 b4 will (Black intends a variant of Lar sen's plan to dominate the black quickly transpose into 3b or 3c. squares on the kingside; normal 8 ji e2 (102) d6 development would lead to posi tions similar to Suba-Taimanov above) 1 3 �d2 �d6 14 � d5 g5 1 5 b4 h5 1 6 .!!l[ fd l j1e5 1 7 h3 .ztf4 1 8 jt fi ! and Black's attack has become bogged down while White has all sorts of threats against Black's king and queen, Mocha lov-Kapengut, USSR l st League 1 976. (3c) 10 . . . d6 1 1 b4! (when Black's knight is on e6 White Black may try to do without should in general aim to play b4 as quickly as possible, preventing this move: ( 1 ) 8 . . . a5 9 0-0 a4 1 0 c5!? (an Black cementing his knight on c5 attempt at outright refutation; 1 0 by a5) 0-0 1 2 j_ e2 and now: (3c l ) 12 . . . b6 1 3 �d2 ( 1 3 0-0 <Z� db5 would b e similar to 9 . . . a5 below) d5 1 1 cxd6 'filxd6 1 2 <E:J db5 may be better, for example 1 3 _ .

120 Maroczy Bind 'it b4 ( 1 2 \'txd I 1 3 � axd I gives White some endgame ad vantage) 1 3 a3 'lta5 1 4 f4 e5 1 5 fxe5 <2l xe5 1 6 � xf6! A xf6 1 7 <2ld5 j_ d8? ( 1 7 g a6! 1 8 � c l ! J.. d 7 1 9 � c5 �d8 is better, although White has an ominous initiative) 1 8 J,.d4! f6 19 J,.c3 \'ta6 20 <2l bc7 �a7 + 21 J.. d4 �b8 22 <2l xa8 with a clear plus for White, Nunn-Haik, Paris 1 983. (2) 8 b6 (an important alter native) 9 0-0 J,.b7 1 0 f3 (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 exten sive range of possibilities: (2a) 10 � h5 1 1 '£) xc6! 11_xc6 (1 1 dxc6 1 2 c5 is good for White) 1 2 � c l f5 1 3 exf5 gxf5 1 4 f4 <2l f6 1 5 J.. f3 � c8 1 6 b 3 �e8 1 7 'E)d5 '/iif f7 1 8 '£) xf6 + �xf6 1 9 � c2 J.. xf3?! 20 � xf3 d 6 2 1 � d2 lfJ> h8 22 j_ d4 gave White his stan dard favourable positiOn m Nunn-Ristoja, Malta 1 980. (2b) 10 d6 (this is inconsis tent with the choice of b6) 1 1 �d2 '/iil d 7 1 2 a4!? e6 1 3 � fd l !ill fd8 1 4 12Jxc6 � xc6 1 5 a5 bxa5 16 <2l b5 with unpleasant threats to d6, a7 and a5, Gheorghiu-Bel lon, Las Palmas 1 976. (2c) 10 � c8 1 1 'ii!t' d 2 12! h 5 (11 � e8 1 2 i! ac l 'W!c7 1 3 b4! <2lh5 14 <2lxc6 j_ xc6 1 5 1£)d5 '/iif b 8 1 6 f4 <2:J f6 1 7 J,.f3 d6 18 j_d4 was very good for White in Nunn Karlsson, Helsinki 1 98 1 ) 1 2 � fd I Qj e5 1 3 b3 f5 1 4 exf5 gxf5 1 5 • • •

. . .

12!d5! \'te8?!, Kir. Georgiev-Kris tensen, Saint John Open 1 988, and now 16 f4! J,.xd5 ( 1 6 . . . <2:J g4 1 7 Qj xf5 Qj xe3 1 8 12J xg7 <2:J xg7 1 9 "*'xe3 wins) 1 7 cxd5 d2lg4 1 8 <2lxf5! J,. xa l 1 9 11_xg4 !2l f6 ( 1 9 J,. f6 20 d6! e 6 2 1 IE:l e7 + JJ... xe7 22 dxe7 � f6 23 �xd7 wins and 1 9 j_c3 20 �d3 !2l g7 2 1 d6 are no better) 20 <2:J h6 + *g7 2 1 � xa l * xh6 22 f5 + '#; g7 23 j_ h6 + '\fi' h8 24 JJ... x f8 IE:l xg4 2 5 �d4 + l£) f6 26 JJ... h6 wins for White. (2d) 10 IE:l xd4 1 1 J,. xd4 d6 1 2 'ii!t' d 2 <2:Jd7 1 3 JJ... xg7 ( 1 3 JJ... e 3 is promising) '#; xg7 14 f4 � c8 1 5 � ad I IE:l f6 1 6 e 5 dxe5 1 7 fxe5 IE:lg8 18 'i!ii' e 3! with some advan tage for White, Cvetkovic Cebalo, Yugoslavia 1 985. (2e) 10 IE:l e8 1 1 "i!Wd2 'E) c7 12 B, ad l IE:l e6 13 IE:l db5 d6 1 4 IE:ld5 li!\ b8 1 5 f4 <2:Jc5 1 6 JJ... f3 a6 1 7 IE:l d4 12:! xd4 1 8 J,.xd4 jj_ xd4 + 1 9 \'txd4 b5, Agapov-Kimelfeld, USSR 1 985, and now 20 e5 would have kept some advantage for White. (2f) 10 e6 1 1 "i!Wd2 ( 1 1 IE:ldb5 is probably also good, e.g. 1 1 d5 1 2 cxd5 exd5 1 3 exd5 IE:l e7 1 4 d 6 <2:J f5 l 5 JJ... f2 IE:l e 8 1 6 d 7 IE:l f6 1 7 g4 IE:le7 1 8 ..\l h4 with advantage, A. Rodriguez-Pinal, Sagua la Grande 1 984) d5 1 2 l£) xc6 J,.xc6 1 3 cxd5 exd5 1 4 e5 IE:ld7 1 5 f4 � c5 1 6 � ad l ! f6 ( 1 6 IE:l e4 1 7 IE:l xe4 fxe4 1 8 �d6! j_ b7 1 9 lt_c4 is also unpleasant) 1 7 12J xd5 fxe5 18 j_c4 *h8 19 fxe5 � xfl + ( 1 9 j_xe5 20 j_ d4 '/iii d6 2 1 j_ xe5 + '/iil xe5 2 2 � fe I .£) e4 23 � xe4! Resigns, Kuporosov-Jak. . .

. . .

Maroczy Bind 121 ovic, USSR 1984) 20 J;i xfl �h4 2 1 � f4! �h5 22 ttJ e7 J.. b7 23 b4 ttJ e4 24 �d7 � b8 25 ttJ c8! � xc8 26 �xb7 � d8 27 � xe4 j_ xeS 28 J.. e2 Resigns, Kuporosov-Malis hauskas, USSR 1 985. 9 0-0 ( 103)

J.. d7 9 Or: aS (certainly not 9 (I) 9 ttJ g4 losing a piece after 1 0 j_ xg4 J.. xg4 1 1 tjj xc6) 1 0 f3 ttJd7 1 1 ttJdb5 tjj c5 1 2 �d2 a4 1 3 � fd l �a5 1 4 � ac l J,.e6 1 5 ttJ d 5 � xd2 � fd8 is 16 .!;! xd2 j_ xd5? ( 1 6 only slightly better for White) 1 7 cxd5 ttJ b4 1 8 .it xc5! dxc5 1 9 J.. x c5 tjj xa2 20 j_ xe7 and Black is in trouble, Andersson-Larsen, Linares 1 983. (2) 9 � e8 (another Larsen idea) and now: (2a) 10 � b 1 a6 1 1 �d2 .ii_ d 7 1 2 l! fd I � c8 1 3 f3 ttJ xd4 1 4 .ii_ xd4 j_ e6 1 5 ttJd5 ttJ xd5 1 6 cxd5 j_ xd4 + 1 7 �xd4 J.. d7 1 8 � be l �a5 with equality, Speelman Larsen, Hastings 1 988/9. (2b) 10 f3 ttJ d7 1 1 'ii(( d2 ttJc5 (taking twice on d4 is also pos. . .

. • .

sible) 12 .l fd l '*'a5 13 � ab l ttJ xd4 1 4 J.. xd4 J.. xd4 + 1 5 �xd4 ttJ e6 16 �f2 J.. d7 and White has no advantage, Andersson-Lar sen, Naestved 1 985. (2c) 10 a3 J.. d7 1 1 f3 a6 12 b4 � c8 1 3 � c l ttJ xd4 1 4 j_ xd4 J.. h 6! 1 5 � c2 j_e6 and Black is at least Short-Larsen, equal, Naestved 1 985. (2d) 10 �d2 <E\ g4, 1 0 � c l ttJ xd4 1 1 J.. xd4 J.. h6 and 1 0 f4 e5 are other points of Larsen 's move, so what is the best answer? Proba bly White should play 1 0 ttJc2, avoiding the knight exchange and preparing solid development by .!;! c l . 10 �d2 It is also possible to move the knight on d4, thereby frustrating ttJ xd4 and Black's plan of J.. c 6. In Korchnoi-Soos, Rome 1 982, White continued 10 ttJb3 a5 1 1 a4 ttJ b4 12 f3 ttJa5 (10 jj_c6 1 3 �d2 � c8 14 * h l ttJd7 15 ttJd5 was also better for White, Tarjan-Strauss, USA 1 982) 1 1 0 €) xb3 1 2 axb3 a6 1 3 b4 j_ e6 14 �d2 .k c8 15 b3 ttJd7 16 � a2, while Schmidt-Kagen, Lucerne 1 982, went 1 0 ttJ c2 j_ e6 1 1 '!li!(d2 a5 12 f4 a4 1 3 l! ab l j_ g4 14 J.. d 3 J.. c8 1 5 h3 <E\d7 16 .ii_ e 2 and in both cases White had a good posi tion. It could well be that these lines are as strong as the traditio nal continuations 10 J;,i; c l and 1 0 ·�!!i· d2, but they d o not seem to have been tested in recent years. A separate question is whether White should play 1 0 � c 1 or I 0

122 Maroczy Bind �d2. I have preferred the latter move for two reasons. Firstly, the rook on c l is sometimes vulner able to unwelcome attacks by � xd4 and � h6, and secondly White often starts a queenside ad vance by a3 and b4 later on. If Black plays a5 White needs a rook on b 1 to force through b4. but playing the rook to c l and then to b 1 loses a tempo. The lines after � c 1 and "ii!i d 2 are quite simi 14 �e3 <E) c5 (104) lar, so the material from the first 15 .;, ab1 edition of the book is also worth In order to make progress reviewing, but here we shall only White must expel the knight from examine the lines which follow 1 0 c5, and so he needs to play a3 and �d2. b4. 1 5 g ab 1 appears the most <E:J xd4 10 logical, and this intuitive assess ment is supported by the fact that 10 l!l c8 1 1 f3 a6 1 2 g ac l <E:J xd4 1 3 �xd4 � e6 14 b3 is after the alternative 1 5 )!g, ac l (this passive and White has a very com position can also arise via 1 0 g e l ) �b6 ( 1 5 f5 1 6 exf5 gxf5 fortable position, Smejkal-Diez 1 7 <E)d5 g f7 1 8 � fd l b6! 1 9 �g5 del Corral, Skopje 01. 1 972. .1. a7 was unclear in Vaganian 1 1 � xd4 �c6 Yudashin, USSR Ch. 1 988) 1 6 12 f3 a5 <E) b5 � fc8 1 7 � fd 1 �d8 1 8 <E:J d4 The move order is flexible, but White always answers . . . <E:J d7 by ( 1 8 �fl �f8 1 9 <E)c3 b6 20 <E:Jd5 .l, ab8 21 g bl �e5 22 � h6 �g7 �e3 (to avoid freeing exchanges) and a5 by b3 (or else Black 23 �g5 g b7 24 l! e l <E) e6 25 �e3 <E)c5! was level, Sax-Petursson, plays . . a4 followed by �a5, Reykjavik 1 988), Nunn-Velimir with an active position). ovic, Dubai 01. 1 986, Black can 13 b3 <E:J d7 � d7 1 9 � b l <E) e6 20 Or 1 3 . l!l e8 ( 1 3 . . . <E:J h5?! 1 4 play 1 8 <E:J xe6 �xe6 with equality (instead �e3! f5 1 5 exf5 gxf5 1 6 f4 was good for White in Kavalek-Lar of 1 8 . . . 'lt'f8?! 1 9 I;K b 1 � f6 20 a3 'W/g7 21 b4 axb4 22 axb4 <E) e6 23 sen, USA-Nordic match 1 986) 1 4 <E) xe6 fxe6 24 b5 � e8 25 f4 with a � fd l <E:J d7 1 5 � e 3 <E:J c 5 1 6 � acl clear plus for White in the game). (16 � ab 1 appears more consis 'it'b6 15 tent) 'lt'b6 1 7 <E:J b 5 li;! ec8 1 8 'lt'el Or 1 5 b6 (after 1 5 e6 1 6 ( 1 8 .:E:l d4 is better) �xb5! 1 9 cxb5 � d l ! intending <E) e2-d4, a3 and �h6 with equality, Arnason b4 Black's panic reaction 1 6 . . . f5 Karlsson, Helsinki 1 986.

Maroczy Bind 123 1 7 exf5 � xf5 1 8 -2l e2 was good for White in Tringov-Haik, Vrnjacka Banja 1 986) 1 6 iLd l (White's problem is that the im mediate a3 may be met by . . a4; the idea of iL d 1 is to be able to take the Black knight when it ar rives at b3, but White has other ways to nullify . . a4, for example with the slow preparatory plan of � fc l -c2, .:;lfl and Wf2 to line up against the weak pawn on b6) i"<i b8 1 7 a3 � c8 1 8 2!d5 .k.xd5?! (probably bad, but White is slightly better in any case) 19 exd5 a4 20 b4 2! b3 21 !:ioi e2 with an excellent position for White, Anand-Larsen, Cannes 1 989. 16 !i6 fcl After 1 6 -21 b5 (not 1 6 a3? 2! xb3) 8 fc8 17 � fd l � d8 1 8 -d d4 � f8 1 9 a3 ( 1 9 JLfl may be better, but I doubt if White has more than a tiny edge) ..ii. d7! (not 19 _li f6? 20 b4 axb4 2 1 axb4 2!e6 22 2! xe6 fxe6 23 f4 with the same advantage for White as in Nunn-Velimirovic above) 20 b4 axb4 21 axb4 2! e6 22 i±, a l J;t, xa l 23 &i x a 1 ;;,;, a8 24 � xa8 !:ioi· xa8 25 Bxe6 1L xe6 Black is completely equal, Jansa-Petursson, Naestved 1 988. £ fc8 16 17 � c2 Again this is useful preparation for a3 and b4. Now Black cannot meet 1 8 a3 by 1 8 � xb3 because 1 9 � d l ! wins material. 17 'i'ii' d8 There is no point to 1 7 � b4 because after 1 8 "M·c l Black will

be driven back by a3 with great loss of time. 18 j_ fl li.. e5 White intends iioi fl, followed by a3 and b4, so Black has to orga nize some counterplay. With 1J. e5 he hopes to become active with . e6 and 'i'i'i h4, but the exposed bishop on e5 is a target which causes White to switch plans away from his queenside pawn advance to a more aggres sive idea. 19 2! d l !? The knight transfer to h6, which gains time along the way when B g4 hits the bishop, is not a very thematic approach, but chess cannot always be played accord ing to the recipe book. 'itie8 ( 105) 19 Preparing b5 is the most natural way to counter White's slow build-up towards a kingside attack. Moreover the queen defends the f7 pawn which might come under fire after B f2-g4h6 +

20 � f2 b5?! It would have been better to

124 Maroczy Bind play 20 J.. g7, so as to meet 2 1 :£J g4 b y 21 h5. Then White could hardly venture into the lion's den by 22 .£) h6 + \\' h7, so the knight would have to return to f2. After 20 .1J... g7 White should switch plans again by 21 �d3 b6 22 � f2, followed by a3 and b4. 21 .£) g4 .1J... g7 il_ xb5 22 cxb5 23 '-2l h6 + *f8 Unfortunately Black has to go to f8 because 23 ffh8 loses a pawn after 24 8 xf7 + , while 23 .1J... x h6 24 il_ xh6 gives White a permanent advantage. 24 .1J... xb5 �xb5 25 �d5 'i'!(e8 26 e5! � d8?! In a difficult position Black fails to offer the most resistance. He should try to give up a pawn to reach an ending in which the off-

side knight on h6 gives White problems. Therefore 26 .£) e6 was the best (26 � a6 27 exd6 is bad since 27 � xd6 28 �xf7 + ! wins a clear pawn), with the idea 27 � xc8 � xc8 28 exd6 exd6 29 �xd6 + �e7 30 ¥ifxe7 + ffxe7 with excellent counterplay. White should prefer 29 :£Jg4!, when the weak pawns on d6 and a5 make Black's position unat tractive, but his chances would certainly be much better than in the game . 27 exd6 exd6 28 � el � ac8? The final collapse. 28 Ql e6 was necessary to meet the threat of 29 � xc5, when 29 .£) g4 gives White a positional advantage but nothing decisive. 29 � xc5! � xc5 30 �xc5 Resigns

Taimanov Variation

The first moves of this system run 1 e4 c5 2 -E:J f3 e6 3 d4 cxd4 4 -E:J xd4 e6). '-E:lc6 (or 2 -2)c6 and 4 In a way this resembles the Kan Variation, since Black keeps the f8-b4 diagonal open for his bish op, but here White cannot play JL.d3, so there are fewer options for the first player. The variation I am recommending for White, 5 -E:l b5, is the most obvious way to exploit the substitution of <21 c6 for a6. The main line leads to a kind of Maroczy Bind position in which White's knight is on a3 instead of d4. This provides more support for the c4 pawn, so White is not normally forced to play b3, but on the minus side the knight is distinctly offside on a3 and in many lines White tries to bring it back via b 1 or c2. Black's best strategy is watchful waiting as in the Maroczy Bind. One difference between this line and the Maroczy Bind is that Black's bishop is on e7 instead of g7, so that White may be able to play for a direct kingside attack by advancing his g- and h-pawns. As usual, such a strategy carries many risks, but with the knights on a3 and c3 holding up Black's counte rplay by b5 or d5 White may have enough time to break through.

Game 25 Chandler-Pritchett British Ch. 1985 1 2 3 4 5

c5 e4 e6 �f3 d4 cxd4 Ql xd4 .:21 c6 .f)b5 ( 106)

106 B

5 d6 Not 5 .R_c5 (5 Q} f6 6 <El l c3 transposes to Chapter 9) 6 j}_ f4 � f6 7 ·l!lf c l 'lfrf8 8 .f) l c3 when both 8 . . . � ge7 9 .R_e3! b6 1 0 �d2 j}_ xe3 1 1 � xe3 d5 1 2 0-0-0 a 6 1 3 exd5 exd5 1 4 � c7 d4 1 5 � xd4 <£) xd4 1 6 <2J xa8, Gufeld Plaskett, Hastings 1 986/7 and 8 . . . a6 9 j}_d6 + �xd6 1 0 Q} xd6 �d4 1 1 � f4 ( 1 1 •d 1 is also good) � f6 1 2 ,g d l • b4 1 3 •d2, M okry-Plaskett, Trnava 1 984, gave White a clear advantage.

126 Taimanov System 6 c4 �f6 Black can also play 6 . . . a6 7 � 5c3 � f6, when White can transpose to the main line by playing 8 � a3, but in practice White has usually chosen to try exploiting Black's move order by developing his knight to d2 in stead. After 6 . . . a6 7 � 5c3 � f6 8 J. e2 J. e7 9 0-0 0-0 10 J. e3 Black has tried: ( I ) 10 ijc7?! (usually wrong in this system, since after White plays :i, c l Black will be in danger from tactics based on !E) b5 or � d5) 1 1 �a3 (reasonable now that Black is committed to a bad line) b6 1 2 ij e l ll b7, Dieks-Mar janovic, Manila 1 974, and now 1 3 l;l c l followed by f3 and itf2 would be good for White. (2) 10 lld7 1 1 f4 ijb8 1 2 IE)d2 b6 1 3 a 3 � a7 1 4 llf3 J.c6 1 5 � b3 �d7 1 6 � d4 J.. b 7 17 f5 with a slight plus for White, Kar pov-Taimanov, USSR Ch. 1 976. (3) 10 � e8 1 1 a3 J.. f8 1 2 ijb3 �d7 1 3 � d2 b6 1 4 � f3 J. b7 1 5 X fd I and again White is slightly better, Bronstein-Lom bardy, Teesside 1 975. (4) 10 b6 1 1 ijb3 (the im mediate 1 1 �d2 is also good, and after 1 1 1, b8 1 2 Ji c l J.b7 1 3 ijb3 � d7 1 4 .i. fd l � ce5 1 5 f4! IE)g6 1 6 � f3 ! ijc7 1 7 f5! �ge5 1 8 IE) d4 � c5 1 9 ijc2 White was doing well in Gufeld-Hort, Dort mund 1 983) K b8 1 2 l, d l ijc7 1 3 a3!? Jld7 1 4 �d2 !I fc8 ( 1 4 . . . �a5 is just slightly better for White) 1 5 !;! de l ! ijd8 1 6 ijd l . . .

K c7, Psakhis-Holm, Plovdiv 1 983 and now 1 7 b4! intending !E) b3 is good for White according to Psakhis. 7 � 1c3 Now 7 � 5c3 makes no sense, for Black gains a tempo by missing out a6. 7 a6 After 7 . . . Jle7 8 ll f4 e5 9 J. g5 a6 1 0 Jlxf6 gxf6 1 1 � a3 � d4 1 2 � c2 � xc2 + 1 3 ijxc2 White has an evident positional advantage, Chiburdanidze-Taborov, USSR 1 979, so Black must push the knight away. 8 � a3 ( 107)

. . .

Jle7 8 The speculative pawn sacrifice 8 d5 9 cxd5 exd5 1 0 exd5 � b4 was played a few times, but is now considered bad after 1 1 J. e2, for example: ( I ) 11 Jl c5 1 2 Jle3! J. xe3 13 '!fl'a4 + � d7 ( 1 3 b5 1 4 '!fl'xb4 J.. b6 1 5 0-0 Jl a 5 1 6 jl xb5 + ! axb5 1 7 K fe 1 + \fid7 1 8 iifxb5 + \fi c7 1 9 d6 + ! ijxd6 20 K ac 1 wins for White, while 1 3 . . . . . .

. . .

Taimanov System 127 j_d7 14 �xb4 �b6 1 5 �xb6 ( I ) 10 £ b8 l l J.. e 3 �a5 1 2 J.. xb6 1 6 �c4 j_d4 1 7 !l; d l leaves � c2 j_d7 l 3 f4 b5 1 4 cxb5 axb5 Black with nothing for the pawn) 1 5 a3 !! fc8 16 g4 l!l;d8 1 7 g5 � e8 14 �xb4 ( 1 4 fxe3 �h4 + 1 5 g3 1 8 jj_xb5 �a5 1 9 a4, Mednis �e7 16 !l; d l is also good) lL_c5 1 5 Schmidt, Nice 1 977, and Black �e4 + <ll' f8 1 6 0-0 b5 1 7 �c2 and does not have enough for the Black has no real compensation pawn. for the pawn, Karpov-van der (2) 10 J.. d7 1 1 J.. e 3 �b8 (1 1 Wiel, Brussels 1 986. . . . � e5 1 2 f4 �g6 l 3 l!l; e l jj_c6 (2) 11 . . . J.. e7 12 j_O k_ f5 l 3 14 jj_O �e8 1 5 �c2 !! c8 1 6 a4 a5 0-0 0-0 1 4 �c4 j_c5 1 5 J.. e3 J.. d6 1 7 g3 �d7 1 8 h4 gave White 1 6 � xd6 'tifxd6 1 7 'i!]td2 � fd8 1 8 attacking chances in Yudashin !l; ad l �d3 1 9 jj_ e2 � xd 5 20 Osnos, USSR 1 986, while 11 . . . jj_ xd3 � xc3 2 1 jj_xf5 � xd l 22 � b8 1 2 f4 �a5 1 3 � c2 b5 14 !! xd l �xd2 23 !! xd2 and White cxb5 axb5 15 a3 is also slightly won the ending in Lawton-Gay better for White) 12 0 � a7, Aga pov-Taimanov, USSR 1 987, and son, British Ch. 1 987. Some Black players have exper now l 3 �d3!? j_c6 14 � fd l � d7 imented with delayed castling, but 1 5 jj_ fl �e5 1 6 J%' e2 gives White this increases the danger that an edge according to Agapov. If White will gain time by missing Black cannot force through . b5 out jj_ e3, for example 8 b6 9 then his piece deployment looks J.. e2 jj_ b7 10 f4 ( 1 0 0-0 is less very strange. 11 j_e3 ( 108) accurate, because after I 0 � b8! White has no good way to defend the e4 pawn, since the re ply 1 1 0 cuts out the plan based on f4 recommended in this game) � b8 l l jj_O � bd7 1 2 0-0 j_e7 l 3 'tife2 0-0 1 4 g4!? d 5 1 5 exd5 exd5 1 6 g5 � e8 1 7 �g2 � e4, Geller-1. Sokolov, Panchevo 1 987, and now 1 8 cxd5 is good for White according to Geller. 9 jj_ e2 0-0 b6 10 0-0 By far the most popular move. 11 �e5 The alternative is to play for This is a major branch. Black b5 directly in order to save time, but the problem is that a prema has a number of possibilities, but ture . b5 j ust leaves Black with a since this line is governed largely by general principles rather than weak b-pawn, for example: • . .

• • .

. . •

128 Taimanov System specific variations, we do not waste time considering numerous very similar lines. Instead we only seriously examine two lines, 1 1 J.. b7 and 1 1 . . . 2JeS, which represent quite different plans. In the latter case Black intends repositioning his knight to d7, opening the diagonal for the b7 bishop. This line plans dS at some stage. 1 1 iil.b7 is a more restrained line; Black concen trates on developing his pieces and waits to see what White in tends. After 1 1 1L b7 (or 1 1 J.. d7 1 2 f] -� b8 1 3 '1\oi e l 8, a7 14 ·M'f2 );£ b7 l S � fd l � d8 1 6 £ ac l JL e8 1 7 <iW h l �d7 1 8 ·iiii fl JL f6 1 9 �abl �cS 20 f4 j_ e7 2 1 b3 h6 22 � d2 � c8 23 � cd l � d8, Matulo vic-W. Schmidt, Vrnjacka Banja 1 983, and now 24 8 b2! is good for White, while 1 1 J.i e8 1 2 � c l � e S 1 3 f4 2J ed7 1 4 it_ f] will probably transpose to the main line of the game) 12 'ii! b3 <dd7 ( 1 2 .£) aS?! 1 3 'jojlf xb6 � xe4 1 4 � xe4 .il..xe4 l S w xd8 Jt. xd8 1 6 � ad 1 left Black under unpleasant pressure in Karpov-Kasparov, Moscow match 1 984/S; nobody has cared to repeat this line) 1 3 � fd l Black may play: ( 1 ) 13 . � c8 14 )4 ac l � aS l S 'l!!' a4 fS 1 6 exfS )el xfS 1 7 b4 J.. g S, Chandler-Kurajica, Sarajevo 1 98S, and now 1 8 .zl g4! iL.xe3 1 9 ji xfS J.. xc l 20 J.. xe6 + 1fr h 8 2 1 J.. xd7 is good for White accord ing to Chandler, the point being that 21 � gS 22 �dS ;g xc4 23 � xc4 � xc4 is met by 24 J.. h3!, . . .

. . .

both defending g2 and threaten ing i!ff e8 mate. (2) 13 . � e8 (a bit irrelevant) 14 � ac l 1J.f8 l S ji_ fl � c8 1 6 ·;..« c2 '-2l ceS 1 7 h 3 'lii c7 1 8 f3 with a typical slight plus for White, Kar pov-Romanishin, USSR 1 98 1 . (3) 13 . J.i a7!? 1 4 £t; d2 ( 1 4 <2\ a4 <2\aS doesn't achieve any thing since l S �c2 .£)c6 and l S J.. x b6 <2\ xb6 1 6 �xb6 �xb6 1 7 .£) xb6 jl xe4 are not better for White) jl a8 l S 'l!¥fd l '*b8 1 6 'i!ir fl 22! f6, Nunn-Cebalo, Biel 1 986, and now 1 7 f4 intending � f2 would have created awkward pressure against the b6 pawn. (4) 13 . � b8 14 � ac l j_ a8 l S � d2 � b7 1 6 �d l �b8 1 7 '* fl <2\ f6?! (Black's plan is similar to that in Nunn-Cebalo above, but this time White finds the right reply; 1 7 J.i\ c7 was more solid) 1 8 f4 � c7, A. Rodriguez-Kirov, Havana 1 986, and now 19 'i!iff2 � fc8 20 � cd 1 followed by J.. f3 would, by defending b2 a second time, create a genuine threat to take on b6. (S) 13 .£) c5 (the most im portant line) 14 'l!tc2 (109) and now: (Sa) 14 .ii. f6 l S � ac l and now: (Sa l ) 15 . . . J.. e5 1 6 <2\ ab l 'l!¥fh4 1 7 g3 �f6 1 8 f4 J.. d4 19 � d2 eS 20 �dS �d8 (20 i!th6 2 1 �El bc3 fS 2 2 exfS � xfS 2 3 J_g4 is good for White) 21 � bc3 J.. x e3 + (2 1 ltl' h8 22 fS � d7 23 J_ f3 jl cS 24 ltl'g2 f6 2 S <2\ e2 gave White the advantage in Karpov.

. .

. . .

Taimanov System 129 White the advantage in Yudas hin-Dzhandzhava, Simferopol 1 988. (5b2) 15 � ac1 � ac8 ( 1 5 J.. f6 1 6 c2\ ab l � ac8 1 7 a3 � b8 1 8 b4 �d7 1 9 'i!-1 d2 .iJ... e 7 2 0 f4 � fd8 21 'ii-J e l gave White his usual slight plus in Tseshkovsky Hulak, Dubai 01. 1 986) 1 6 :d ab ! ( 1 6 JL. fl )4 fd8 1 7 � b l c2l b4 1 8 � c2 �xc2 was roughly equal in Olafsson, Amsterdam 1 9 76) 22 Lobron-Liberzon, Ramat-Hash �xe3 exf4 23 gxf4 � e8 24 b4 aron 1 982) .t.:� e5 17 �d2 :2Jcd7 1 8 '2\d7 25 J.. f3 with a slight plus for a3 1£ fe8 1 9 b4 � f6 20 h3 �g6 White, Westerinen-Liberzon, was played in Jadoui-Karpov, Brussels 1 986. This position is Geneva 1 977. (5a2) 15 . . . J.. xc3 1 6 �xc3 slightly unusual in that Black's l£] xe4 1 7 �d3 l£]e5 18 �d4 and knight has moved away from the White will regain the pawn while queenside to g6. The logical res ponse would be for White to start keeping the two bishops. (5a3) 15 LtJ b4 16 'l!!l' d2 (if action on the relatively bare left White wants to avoid complica flank by 21 � b3 followed by a4tions then 1 6 �bl is a safe op a5. (5b3) 13 i-ii d2 (this appears the tion) J.. xc3 1 7 bxc3 l£] xe4 18 'M b2 most logical because it allows the '2\c6 1 9 f3 l£] f6 20 J_ xb6 ·1!t!f e7 21 '2\c2 � fc8 22 *a3 l£] d7 23 1i_f2 a3 knight to return to the game l£] c5 24 � d2 � d8 25 � cd l f6 26 via c2) �e5 ( 1 5 � ad8 1 6 � c2 �e5 1 7 f3 � cd7 1 8 � ac t � f6 1 9 l£] d4 l£]e5 27 l£] b3 l£] xb3 28 axb3 � d4 was slightly better for White with an edge for White, Mokry in Karpov-Small, Lucerne 01. Lobron, Reggio Emilia 1 984/5. 1 982) 1 6 f3 with an edge for White (5b) 14 . �c7 and now there Erevan is plenty of flexibility about move in Akopian-Semkov, 1 988. It is worth giving the rest of order, for example: this game in full because it is an (5b I ) 15 f3 (the start of a slightly unusual plan based on excellent example of how Black preparing b4 by � ab I ) ll, fe8 1 6 can be tortured in this variation. •d2 X ac8 1 7 � ab ! l£] d7 1 8 f4 White goes round and round, an noying Black with one little threat '2\cb8 1 9 M be l (having persuaded after another, always being care the knights to retreat the rook may be repositioned) l£]f8 20 � c2 ful not to allow Black to free himself. Finally White adopts the l£] bd7 2 1 l£]d4 .b8 22 b3 ;Ld8 23 plan of a q ueenside breakthrough, J.. f3 and Black's passive play gave . . .

. .

130 Taimanov System which eventually proves decisive: 16 ;s_ ac8 1 7 � ac l £ fd8 1 8 i'ii e l -2J g6 1 9 iii f2 k f6 20 b4 �d7 21 -2Ja4 £ b8 22 Ji.fl ka8 23 'if�> h i h6 24 � d2 JJ.. e7 2 5 � b l i);l b7 26 a3 -2J ge5 27 � f2 � dc8 28 ..t:Jd2 d c6 29 "21 b3 k!, bc8 30 � b2 � 6c7 3 1 ii! d2 -d f6 32 i>i d4 2\ed7 33 :s. b l � b8 34 .k. f2 -d e8 35 i<id2 _k g5 36 ..k.e3 .k.xe3 37 m-xe3 -2)e5 38 a be l -dd7 39 i<id2 :&_b7 40 a4 :s. d8 41 a5 _:&a8 42 -2J a4 � dc8 43 jj_ xa3 18 bxa3 Q)xd5 19 cxd5 axb6 -2J xb6 44 -2J xb6 ·i'tJxb6 45 c5 � xc3 20 j_ d4 ,g, xa3 21 g5 hxg5 dxc5 46 -2J xc5 iLb7 47 E, a l � a8 22 fxg5 e5!, A. Rodriguez-Polu 48 -2Jd7 iil a7 49 b5 1J..c8 50 b6 gayevsky, Moscow 1 985 and 1 5 8 Xd7 5 1 bxa7 8 xd2 52 � xd2 I!! ad 1 fi/c7 1 6 g4 d5! 1 7 cxd5 'itr f8 53 � d8 'itre7 54 E:; xc8 j_xa3 1 8 bxa3 �xc3 19 g5 Q) xd5 20 exd5 j_xd5 2 1 j_xd5 exd5 22 Resigns. j_d4 �xa3, Hernandez-Renet, 12 f4 -2Jed7 The knight transfer from c6 to Thessaloniki 01. 1 988, with ad d7 improves the position of vantage to Black in both cases. 15 �c7 Black's pieces, since he may exert 16 ,g, fd1 pressure against e4 by J_ b7 It is an open question as to and .:dc5, or he may play for d5. The defect is that it gives whether this further preparatory White extra time, and this allows move is really necessary. The him the chance to start a kingside alternative is the immediate 1 6 g4 ( 1 1 1) and now: attack. 13 jj_ f3 b7 11_ 14 ·l.l!fe2 li( c8 ( 1 10) Once again the move order is a matter of personal choice, but this move is bound to be played sooner or later, so Black may gain some flexibility by playing it first. 15 ,g, acl It is unwise not to play this move, because a poorly defended knight on c3 is an invitation for Black to start a tactical storm by playing d5. This happened ( 1 ) 16 Q) c5 1 7 �g2 (after 1 7 after 1 5 g4 h6 1 6 h4 d5! 1 7 exd5 jtd2 h6 1 8 h4 Q) h7 1 9 b4 jtxh4 . . .

Taimanov System 131 20 frg2! <iJ d7 2 1 £il h 1 j_ e 7 22 g5 hxg5 23 � xh7! frxh7 24 £il h l + ltr g8 25 frgl f5 the complications should have led to a d raw in Gufeld-Georgadze, USSR 1 985) d5 (Kasparov suggested 1 7 g5 18 fxg5 <iJ fd7, but not surpris ingly nobody has cared to try this) 1 8 exd5 ( 1 8 e5 <2l fe4 1 9 cxd5 exd5 20 b4 <iJ xc3 21 � xc3 d4 was unclear in Tseshkovsk y-Kas parov, USSR Ch. 1 979) �d3 1 9 ;;,( cd l ! .t:l xf4 20 _k xf4 JL.c5 + (20 �xf4 2 1 d6 wins a piece) 2 1 l\' h l �·xf4 2 2 g5 -2) e8 (22 . .t:ld7 23 dxe6 wins a piece) and now White has a pleasant choice: he can win a pawn for insufficient compensation by 23 dxe6, but 23 d6! appears even stronger to me. (2) 16 . . . ·i'i.i' b8 (this appears too passive) 1 7 g5 -2J e8 1 8 ii_ g2 g6 19 .il.. h3 -2Jg7 20 f5 exf5 21 exf5 gxf5 22 � xf5 <iJ xf5 23 iL xf5 -2l e5 24 ·;>;r h5 -2J g6 25 ..l!Ld4 �;;;; fe8 26 s fl il..f8 (26 g c5 27 -EJd5! is good for White after 27 . . . 1t_xd5 28 1L xg6 fxg6 29 ·rti h6 or 27 . . . � xd5 28 cxd5 li_f8 29 j_ e6! fxe6 30 � f7!) 27 li_ xc8 'i'ti' xc8 28 � xf7 'ifi' xf7 (28 .R h6! 29 .!:;( xh7! w f5 30 .!:;i, h8 + -2Jxh8 3 1 iifxe8 + 1t_ f8 i s the critical line, but i t i s doubt ful if Black has enough compensa tion for the two pawns since White's king can flee to the queen side) 29 � xh7 + 'i\' e6 30 'iii' xg6 + 'lt,ld7 3 1 ·�f5 + �c7 32 i'i\' xc8 + *xc8, Chandler-Quinteros, Vienna Open 1 986, and now simply 33 h4 would have been good for White.

(3) 16 . . . h6! 17 h4 -2Jc5 1 8 �g2 d5 19 exd5 -2Jd3 ( 1 9 exd5? 20 g5 hxg5 21 hxg5 .t:ld3 22 gxf6 jl xf6 23 ..Rxd5 <iJ xc l 24 <2l e4 wins for White, Geller-Franzoni, Berne 1 987) with unclear compli cations. Note that the continua tion of line 1 is ineffective here, since after 20 &i\ cd 1 (20 g5 lL c5 is unclear) .t:l xf4 21 1t_xf4 li_c5 + 22 'iti' h l &xf4 the h4 pawn is attacked. � fe8 16 h6 17 g4 If Black plays 1 7 -E) c5, then 1 8 !if g2 leaves him much worse off than in the last note. There are two reasons for this; firstly the rook on e8 takes away a flight square from the f6 knight, so that there is an immediate threat to win a piece by g5 and b4, and secondly there is no chance that d5 will work when the rook is on d l . 18 h4 .t:l h7 Black is trying to hold up g5 for as long as possible. Now 1 9 g5 hxg5 20 hxg5 e5 is fine for Black, so White must take time out to defend h4. 19 �h2 ( 1 12) 'l!! b8 19 This passive move looks wrong, but Black has not had much luck with the alternatives: ( l ) 19 . . . <iJc5 20 �h3 (defend ing the bishop on f3, for the im mediate 20 g5 is met by 20 . . . f5!) JL. f6 2 1 <iJ ab l (meeting the threat to the e4 pawn; White is finally ready for g5!) g6 22 £!l c2 j_g7 23

132 Taimanov System *xf3 exf5 25 Ql d5 followed by A d4 was also good. hxg5 21 fxe4 22 hxg5 d5 23 Qlxe4 Ac5! 24 Qlf2 Black defends well. By ex changing Black-squared bishops he exposes the weakness of f4. 25 Axc5 Ql xc5 (1 13) � cd2 §L f8 24 g5 h5 25 §Lf2 §Lc6 26 Ql a3 Qld7? (26 *b7 was better, although the out-of-play knight at h7 gives White some advantage), Nunn-Cramling, Zurich 1 984, and now 27 e5! is very good for White. (2) 19 * d8 20 jlf2 g5 (a drastic way to prevent g5 by White; although this gains Black the e5 square he still has problems along the h-file) 2 1 hxg5 hxg5 22 f5 Qle5 23 *g2 §L f6 24 � h l Qlf8 25 jld4! A a8 26 i, cfl *e7 27 jl e2, Hellers-Wahls, World Junior Ch. 1 986, and the threat of � f3-h3 causes serious problems for Black. 20 g5 White decides to play g5 straight away, even though Black can reply with . . f5. The alterna tive was 20 * h3 as in line I of the last note. f5 20 21 *g2 There is a choice of promising lines for White. Chandler decides to keep the h7 knight locked out, but the direct 2 1 gxh6 gxh6 22 exf5 jl xf3 23 *g3 + \trh8 24

� f8? 26 l!;! d4 This move has disastrous consequences because it gives White the time he needs to revive his kingside attack. Black should have continued his policy of liqui dation by 26 dxc4; after 27 Ql xc4 A xf3 28 '!!lt xf3 E. cd8 White still has some advantage because of the h7 knight, but with each exchange White's attacking chances are reduced and he may soon have to look to the safety of his own king. 27 g6 .£l f6 28 �h3 Black has no threat, and the free move enables White to create his own unpleasant threats of 29 b4 and 29 Qlg4.

Taimanov System 133 28 £ fe8 Relatively best. 28 d:xc4 29 �g4 j_xf3 (29 �cd7 30 .\;;; xd7) 30 -2lxf6 + � x f6 3 1 'i,t(h7 + <;tr f8 3 2 W,f h8 + ve e7 33 � xg7 + wins, as does 28 � ce8 29 <;tr g2 followed by � h 1 . 29 B, el Stepping up the pressure on e6 renews the threat of b4. 29 -2l g4 Qjcd7 was less convincing. 29 �d6 30 R g4 . . .

•

Black cannot bring any further support for e6 and he has no answer to the threat of b4 (30 a5 31 -2l b5). 30 e5 � xe5 31 fxe5 'iif xe5 32 � xe5 33 j}_ xc8 � xd4 34 lL xb7 Resigns. White has netted a clear piece since the recapture on b7 allows mate in 3 .

Sicilian Four Knights

This rather antiquated system is or 7 a4. Since we are recommend not currently in favour, but ing the main line with 7 j_g5 this fashions can change and it is transposition is not a worry and d6 just leads to the advisable to be prepared even for therefore 6 less common variations. It exper earlier chapter on the Pelikan. d5 loses to 7 ienced a brief surge of popularity After 6 l£) db5, 6 when Chandler adopted it for a exd5 exd5 8 ..;L f4 and 6 ..;Lc5 7 couple of years, but the generally ji_ f4 followed by JJ.. d6 is un passive nature of Black's position pleasant for Black, so we need j_ b4 in this proved unattractive to other only consider 6 players and it has now virtually chapter. The tactical line recom disappeared again. Black plays I mended in the first edition has e4 c5 2 1£) 0 e6 3 d4 cxd4 4 1£) xd4 suffered a serious setback in re l£) f6 5 l£)c3 l£) c6 (of course this cent years, so this time we only can arise from other move orders, analyse the positional continua in particular via 2 �c6). In tion 7 a3 JJ.. xc3 + 8 � xc3 d5 9 some ways this is akin to the Kan exd5, which either gives Black an exd5 or and Taimanov systems since isolated pawn after 9 Black leaves the f8-b4 diagonal gives White a lead in development open for his bishop, but instead of after 9 . � xd5 1 0 J.,d2. This is a playing . . . a6 he develops a piece. safe line for White in which he is Naturally this is in Black's favour likely to secure a small but perma unless White has some direct nent advantage. In practice it is method of exploiting the omission easy for White to allow the posi of a6, so 6 � db5 is the only tion to slide towards a draw, and move to cause Black problems. in some ways it is an annoying Black then very often continues 6 line to meet because instead of the d6 and after 7 ..;L f4 e5 8 J.,g5 sharp struggle typical of most we have transposed to the Sicilian lines, White is trying to Pelikan, considered in Chapter 4. exploit a slight positional edge. The point of this move order is Nevertheless it is even more un that Black avoids the possibility pleasant for Black, who can only that after 1 e4 c5 2 1£) 0 � c6 3 d4 win if White takes exceptional cxd4 4 €) xd4 €) f6 5 l£)c3 e5 6 risks, and so this line is relatively �db5 d6 White might play 7 €)d5 unpopular.

Sicilian Four Knights 135 Game 26 Mokry-8. Stein Gausdal 1988 1 2 3 4 5 6 7 8 9

e4 c5 e6 22! f3 d4 cxd4 �xd4 �f6 �c3 � c6 �db5 .� b4 a3 j_xc3 + �xc3 d5 exd5 ( 1 14)

9 .. exd5 This has been almost the only move played during the past decade, but it is interesting to note that in a very recent game Ulf Andersson preferred the knight recapture. The analysis runs 9 � xd5 1 0 jtd2 and now: ( 1 ) 10 . . . '2!xc3 1 1 j_ xc3 'i'ii' xd l + 1 2 � xd l f6 ( 1 2 e5 1 3 .� d3 .il e6 1 4 0-0 f6 1 5 f4 i s simi lar, lvkov-Giigoric, Amsterdam 1 964) 1 3 f4 (for some reason Ehl vest preferred the passive 1 3 f3 and after 1 3 jt.d7 14 ..l\l.d3 0-0-0 1 5 � d2 e5 16 jt e4 j_e6 1 7 .

;tt e2 � xd2 + 18 'lftxd2 ljj e7 1 9 jj_ b4 tjj d5 20 il_ xd5 j_xd5 2 1 'lfte3 the players agreed to a draw in Ehlvest-Andersson, Skelleftea 1 989) l\l.d7 14 ..itc4 0-0-0 1 5 0-0 'Yfl;c7 1 6 � de l � he8 1 7 � f3 j_c8 with a very unpleasant position for Black, Fischer-Addison, USA Ch. 1 962-3. (2) 10 . . . �h4 1 1 J!lll f3 0-0 ( 1 1 �e5 1 2 "iioi' g3 and 1 1 . . �d4 . 1 2 �d3 just make matters worse) 1 2 0-0-0 � xc3 1 3 Jtxc3 e5 14 _il_d3 Jlg4 (14 'il'Jg4 15 Jte4 -rt� xf3 16 j_ xf3 with the typical favourable ending for White, Minic-Gerusel, Halle 1 967) 1 5 �e4 'i'\lfh6 + 1 6 il._d2 lf(g6 1 7 f3 ii_e6 1 8 '¥fxg6 hxg6 1 9 Jte3 and again White has a promising end ing, Tai-Matulovic, Kislovodsk 1 966. (3) 10 . . . �b6 1 1 � b5 �d4 1 2 exd4 �xd4 1 3 j_b5 + jj_d7 14 ]lxd7 + \tr xd7 15 0-0 left Black's king badly placed in Kaplan-Sia peras, Siegen 1 970. (4) 10 . . . �f6 1 1 �h5 0-0 1 2 0-0-0 ljj xc3 ( 1 2 . . . ti!fxf2 1 3 ljj xd5 exd5 14 Jtd3 is no better) 1 3 jlxc3 "�t f4 + 1 4 � d2 e 5 1 5 il._b5 with advantage, Matulovic-Kok koris, Athens 1 969. (5) 10 . . . 0-0 1 1 � h 5 tjj f6 1 2 )1Hh4 �d4 1 3 j_g5 � d8 1 4 �xd4 lit xd4 1 5 il. d3 b6 1 6 0-0-0 is the same story as in all the other lines, Gufeld-Hasin, USSR Ch. 1 966. 10 j_d3 0-0 Or 1 0 . . d4 (after 1 0 �e7 + 1 1 �e2 �xe2 + 1 2 � xe2 ljj e 5 1 3 A,b5 + j_d7 1 4 j_ xd 7 + * xd7, . .

136 Sicilian Four Knights Liberzon-Bronstein, USSR 1 972, White could have played 1 5 j_ e3 � he8 16 j_d4 with a slight ad vantage) and now: ( 1 ) 1l �e2 + j_e6 1 2 <2:! e4 <2:! xe4 1 3 �xe4 �d5 14 J.. f4 0-0-0 1 5 0-0 g5, Frolov-Maliutin, Jur mala 1 989, and now 16 it d2 <2:! e5 17 �xd5 lt xd5 18 lt f5 + is a little better for White according to Maliutin and Kimelfeld. (2) 1 l <2:! e2 lt f5 ( 1 1 0-0 1 2 0-0 transposes into the main line) 1 2 0-0 lt xd3 1 3 �xd3 0-0 1 4 ltg5 h6 1 5 lt h4 H e8 16 Jii ad 1 � c8 1 7 c3! (more dynamic than 1 7 a fe 1 � e6 1 8 *fl �c7 1 9 j_g3 'ii1' b 6 20 b4 with just a microscopic advan tage for White, Karpov-Kuzmin, Leningrad 1 977) dxc3 18 'i!,!(h3 'fje7 19 <2:! xc3 'fje6 ( 1 9 'iil' e 5 20 f4 �e3 + 21 'fjxe3 � xe3 22 j_xf6 gxf6 23 <2:!d5 is very good for White) 20 'fjxe6 fxe6 (20 . . . � xe6 2 1 j_ xf6 Ji:ii. xf6 22 )!4 d7 b6 23 H e 1 favours White) 2 1 j_xf6 gxf6 22 <2:!e4 with an endgame advantage for White, Estevez-Chaviano, Santa Clara 1 983. 1l 0-0 (1 15) d4

This is the most logical because it forces White to decide where his knight is going immediately. To avoid liquidation it seems that 1 1 d4 should be met by the rela tively passive <2:! e2; after other 1 1 th moves White can usually ar range to meet d4 by the more active <2:!e4, for example: (1) 1l a6 1 2 JJ.. f4 ( 1 2 j_g5 is also promising) d4 1 3 <2:! e4 <2:!d5 (13 il_ f5 1 4 ltc7! illustrates why Black should not have delayed) 1 4 j_d6 K e8 1 5 j_g3 f5? (suicide, but even 1 5 . . . il_ f5 1 6 <2:! d6 ll_ xd3 1 7 �xd3 '!4 e7 1 8 <2:! f5 is very awkward) 1 6 <2:!d6 � f8 1 7 ltc4 lt e6 1 8 .l;ll. e 1 'ltd7 1 9 <2:! xb7 'ltxb7 20 K xe6 <2:!a5 21 lt a2 Resigns, Vukcevic-Ervin, USA 1 976. (2) 1 l h6 12 il_ f4 d4 1 3 <2:! b5 ( 1 3 'E} e4 as in line 1 is also pos sible) <2:!d5 14 'ltf3! j_e6 1 5 K ad 1 'ltd7 1 6 h3 K ad8 1 7 lth2 "fke7 1 8 'tlrg3 and White has a clear advan tage, Ciric-Rossolimo, Vrsac 1 969. (3) 11 ltg4 12 f3 j_e6 1 3 ll.. g5 h6 ( 1 3 K e8 1 4 'Wt'd2 d4 1 5 <2:!e2 a6 1 6 <2:! g3, Planinc-Anders son, Sombor 1 970 and 1 3 'Wt'b6 + 1 4 fl' h 1 <2:!d7 1 5 f4! f5 1 6 "f!lf3, Matulovic-Benko, Vrnjacka Banja 1 973 were also bad for Black) 1 4 lt h4 g5 1 5 il.. f2 <2:! h5 1 6 <2:! b 5 and according to Taimanov White has a clear plus. 12 <2:!e2 The available evidence suggests that this offers the best chances for an advantage. After 1 2 <2:! e4 . . .

. . .

Sicilian Four Knights 137 J.. f5 1 3 j_g5 j_xe4 1 4 j_ xe4 h6 1 5 J.. h4 ( 1 5 J.. xf6 trxf6 L 6 X e 1 ll, ad8 1 7 ty d3 <2)e5 1 8 tyb3 b6 1 9 I!I e2 g6 20 K ae 1 * g7 2 1 trb5 � fe8 22 Jl d3 <2) xd3 23 ttxd3 X xe2 24 K xe2 K d5 was equal in Kudrin-Chandler, London 1 987) g5 1 6 j_xc6 bxc6 1 7 j_g3 itd5 1 8 f4! <E:�e4 ( 1 8 . . . g4? 1 9 j_ h4 wins) 1 9 fxg5 hxg5 ( 1 9 . . . <2) xg3?! 20 hxg3 hxg5 2 1 tyd3 f5 22 m ae l is dangerous) 20 ty d3 f5 21 K ad 1 � ad8 22 J.. f2 c5 23 c3 tyb3! White's advantage was infinitesi mal in Kir. Georgiev-Chandler, Leningrad 1 987. 12 J.. g4 (1 16) Or 12 h6 ( 1 2 . . . itd5 1 3 <2) g3 gives White an edge) 1 3 h3 (probably best, although 1 3 J.. f4 .£)d5 1 4 j_g3 ty f6 1 5 � e 1 <2J de7 1 6 .£) f4 j_ f5 1 7 J.. c4 � ac8 1 8 tyd2 � fd8 1 9 <E:�d3 gave White an edge in Geller-Winants, Amster dam 11 1 987; not, however, 1 3 J.. b 5 j_g4 1 4 f3 tyb6!) a6 (because now j_ b5 is a real threat) and White may try: ( I ) 14 J.. f4 !;!! e8 ( 1 4 . . . <2)d5 1 5 j_ h2 tyf6 1 6 <2) g3 gives White an edge) 1 5 � e 1 tyd5 1 6 <2) g3 J.. d 7 17 tyd2 (threat j_xh6) g5 18 J.. c7 b5? (Black should have started hacking rooks oft) 1 9 f4 1! ac8 20 fxg5 hxg5 2 1 <2) f5! 1: e3 22 e xe3 dxe3 23 tyxe3 11 xc7 (Black has material equality but too many loose pieces) 24 lt ad 1 'lt'ffi 25 � fl * g7 26 j_ f5 and White wins, Carlier-Winants, Wijk aan Zee 11 1 987. (2) 14 l!l et <2)d5 15 e, f4 ( 1 5

J.. e4 <2)de7 16 �d3 �b6 17 � d l � d8 1 8 �g3 'flh8 1 9 j_ f4 tyb5 20 J.. d 3 �d5 2 1 j_c7 liK eS 22 c3 dxc3 23 <2)xc3 was good for White in Ernst-Prasad, Subotica 1 987) <2) xf4 16 J.. xf4 J.. e6 1 7 � h5 �d7 18 .i, e2 f5 19 � ae l J.. f7 20 �f3 g ae8 2 1 � xe8 � xe8 22 � xe8 + J.. xe8 23 j_ g3 with a solid positio nal plus for White, Lobron Gobet, Biel 1 984.

. • .

13 f3 This appears best, because the exchange on e2 often simplifies Black's defensive task. After the alternative 1 3 j_g5 ( 1 3 J.. f4 � e8 1 4 � e l � b6 1 5 b4 22! e5 1 6 j_ xe5 � xe5 1 7 tyd2 j_xe2 18 li( xe2 � xe2 1 9 j_xe2 g6 20 K d l K d8 2 1 J.. f3 gave White a small but lasting advantage in Lobron Chandler, Biel l 987) tyd6 14 }.! e l ( 1 4 tt d2 J.. xe2 1 5 trxe2 K fe8 1 6 tr d l .£)e5 offers White nothing, Esbjerg Short-Wiedenkeller, 1 984) K e8 (after 1 4 . . . a5 1 5 �d2 <2)d5 1 6 h3 j_xe2 17 j_ xe2 h6 1 8 J.. h4 Black tried to imprison White's bishop by 1 8 f5 19 c4 .£) [4 20 J.. f3 <2) g6 2 1 j_g3 f4 but

138 Sicilian Four Knights this rebounded after 22 J.. d 5 + ltl> h7 23 � e6 )t(d7 24 �d3 � ce7 25 J.. h 2 � f5 26 � ae 1 � ab8 27 �e4 b6 28 g3 'l!ltd8 29 h4 h5 30 )'!i!lf3 w h6 31 � xf4 + <ff h7 32 'W!e4 � c8 33 'i!il{ f3 w h6 34 � l e5 Resigns, R. Mainka-B. Stein, Dortmund 11 1 987) 1 5 �d2 ( 1 5 f3 jth5 1 6 � f4 h6 1 7 � xh5 hxg5 1 8 � xf6 + �xf6 1 9 'ii\' d 2 .:£)e5 20 � e4 Draw, Short-Chandler, Hastings 1 987/8) J.. xe2 (better than 1 5 � ac8 1 6 � g3 with a clear edge for White) 16 � xe2 E xe2 1 7 �xe2 lii( e8 1 8 �f3 .:2)e5 1 9 'ii\' f4 �b6 20 J..xf6 .:£) xd3 21 �g3 �xf6 22 �xd3 )'!i!lb6 Black drew easily in Kudrin-Rogers, London 1 988. J.. h5 13 "Wtd6 14 J.. g5 The position is the same as in Short-Chandler above, except that the moves � e I and E fe8 have been omitted. This dif ference favours White, as the main line of the game proves. 15 �e1 15 'ii\' d2 � ad8 16 i; ad l !:[ fe8 1 7 J.. h4 1l..g6 1 8 J.. xg6 hxg6 1 9 J.. fl .:£)d5 was level i n Zapata Chandler, Amsterdam 1 987. The move �e1 has the immediate threat of �h4, but White also intends to step up the pressure on d4 by 'itfl and � ad 1 . 15 J.. g6 16 � dl !Ue8 Black must not exchange on d3 as this gives White a free tempo to increase the pressure on d4 by � fl and H fd l . However Black

might have tried to exploit the fact that White's queen is no longer defending the c2 pawn by putting a rook on c8 . 17 'l!i!ffl � adS 18 H d2 � d7 Black decides to meet White's plan passively, even though being forced on the defensive is usually a sign that an isolated pawn posi tion has gone wrong. However Mokry's suggestion of 1 8 !:[ e5 appears no better after 1 9 J.. h4 threatening J.. g3. 19 i; fd1 )!;!( ed8 In the line 1 9 .it xd3 20 � xd3 'ite5 2 1 J.. x f6 �xe2 22 �xe2 � xe2 23 J.. xd4 � xc2 24 J.. c3 Black succeeds in exchang ing his isolated pawn, but only at the cost of giving White a domi nant bishop and good chances of penetrating to the seventh rank. 20 j_ b5! Removing a vital defender in creases the pressure on d4 intoler ably. Black's reply leads to a fatal material loss, but even 20 �e5 21 j_xc6 bxc6 22 J.. xf6 gxf6 23 f4 'itb5 24 .:£) g3, threatening f5, is very unpleasant. 20 h6 ( 1 1 7) 21 .:£) xd4! hxg5 Black's moves are all forced, .:£) xd4 loses to 22 since 2 1 J.. x d7. 22 � xc6 '!lfxd2 23 � xd2 � xd2 � xfl 24 .:£) xd8 25 <ff xfl J.. xc2 26 .:£)xb7 White is a pawn up and his

Sicilian Four Knights 139 26 J.. b3 27 '£\aS AdS j)_xc6 28 L£)c6 29 j)_xc6 '*f8 30 *e3 *e7 31 .d4 *d6 32 j_ bS Resigns Black did not wish to see the technical phase of the game.

active king makes the task of con verting his material plus into a point relatively simple.

Lowenthal Variation

This line starts 1 e4 c5 2 e, f3 e, c6 3 d4 cxd4 4 e, xd4 e5 and i s slightly akin t o the Pelikan in i t s use of a n early e5. The 'old' Lowenthal runs 5 e, b5 a6 6 e, d6 + jl xd6 7 '/i!l'xd6 �f6 (7 flte7 8 '/i!l'xe7 + is worse) and Black hopes that his lead in de velopment will compensate for his black square weaknesses and lack of the two bishops. Current theory suggests that this is a vain hope and White should be able to maintain an advantage. Black has an interesting alternative which has been pioneered by Sveshnikov and other Soviet players. This runs 5 e, b5 d6, and here White has the choice between 6 c4, aim ing for a firm grip on d5, or 6 e, l c3 as in the Pelikan. If White plays 6 e, l c3 Black plans to reach a superior type of Pelikan varia tion in which e,ge7 is played instead of e, f6, thus avoiding the doubling of Black's f-pawns. If, on the other hand, White plays 6 c4 Black will either aim for counterplay by f5 or try to exchange his bad bishop by jl e7-g5. This line is still in a state of flux, so we give a more com plete survey than usual, since at this stage it is impossible to judge which line is best for White.

Game 27 Liberzon--Franzoni Biel (Open) 1980 c5 e4 :£l c6 e,o cxd4 d4 e5 e, xd4 a6 €) b5 Or 5 d6 ( 1 18) (5 . . . � f6 6 e, l c3 transposes to the Pelikan, 5 . . . J.. c5 6 €) 1 c3 !£) f6 transposes to page 63 and 5 h6 6 2J d6 + J.. xd6 7 'i!t!xd6 �e7 8 �d I <E:l f6 9 €Jc3 is good for White) and now there are two lines: 1 2 3 4 5

. . .

( 1 ) 6 €) 1c3 a6 (6 €l f6 is a Pelikan) 7 €J a3 b5 8 e,d5 €) ge7 (after 8 €) f6 White may trans pose back to the Pelikan with 9 J.. g 5, but unfortunately the resulting line is not part of our recommended repertoire; I there fore suggest 9 e, xf6 + '\irxf6 10 c4

Lowentha/ Variation 141 b4 1 1 <E:� c2 Ab7 1 2 A d3 '(Qt d8 1 3 0-0 jl e7 1 4 a3 and White had an edge in Vito1insh-Ambarcumian, USSR 1 988; 8 J... e7 9 c4 b4 1 0 <E:� c2 a5 1 1 jl e 3 � b 8 1 2 J... e 2 � f6 1 3 iitd3 ®d7 1 4 tjj xe7 r�J x e7 1 5 ll( d 1 gave White an edge in Geo. Timoshchenko-Sveshnikov, Moscow GMA 1 989) 9 c4 tt:J d4 and now: ( l a) 10 cxb5 <E:� xd5 1 1 exdS J... e7 1 2 J... c4 axb5 1 3 tjj xb5 j_a6 1 4 tjj a3 0-0 1 5 0-0 J... f6 1 6 j_ e 3 left Black struggling to find compen sation for the pawn, Klovans Kiselev, Frunze 1 988. ( l b) 10 jle3 tt:J xd5 1 1 cxdS jl e7 1 2 J... d 3 0-0 1 3 0-0 f5? ( 1 3 A f6 is only very slightly better for White) 14 jl xd4 exd4 1 5 exf5 J... xf5 1 6 tjj c2 and Black has no compensation for his weak pawns, A. Rodriguez-Estevez, Cuba Ch. 1 988. (2) 6 c4 j_e7 (6 jl e6 generally leads to a transposition) 7 � 1 c3 a6 8 ® a3 and now: (2a) 8 . . . h6 9 j_e2 jl e6 10 0-0 J... g 5 11 tjj c2 transposes to line 2c below, but not 1 1 jl g4 e.f6 1 2 J... x e6 fxe6 1 3 J... xg5 hxg5 1 4 tt:Jc2 \ti f7! and Black is at least equal, Nikolenko-Sveshnikov, USSR 1 987. (2b) 8 . . . tt:Jf6 (Black simply develops, abandoning the idea of . . h6 and J... g 5) 9 J... e2 0-0 1 0 j_e3 j_ e6 1 1 0-0 l;! c8 1 2 iitd2 tt:J a5 ( 1 2 h6 1 3 l;! ac l was slightly better for White in Dol matov-Guseinov, Klaipeda 1 988) 1 3 tjj d 5 J... xd5 14 exd5 b6 1 5

l;! ac l ®b7 16 f4 tt:Jd7 1 7 tt:J c2 a5 1 8 tjj a3 with advantage to White, USSR Dolmatov-Minasian, 1 988. (2c) 8 . . . jle6 9 .£)c2 h6 1 0 j_ e2 jl g5 1 1 0-0 � f6 ( 1 1 tt:J ge7 1 2 �d3! ®g6 1 3 g3 jl xc l 1 4 !;l axcl iitg5 1 5 �d5 jlxd5 1 6 cxd5 with an edge for White, Geller-Lpu tyan, Moscow 1 987) 1 2 *d3 J... xc l ( 1 2 iitc7 1 3 K d l lll[ d8 14 J... xg5 hxg5 1 5 K ac l �b6 1 6 b 3 i s a little better for White, Geller-Lputyan, USSR 1 987) 1 3 i4, axc l 0-0 1 4 K fd l �b6! 1 5 '!i'xd6!? (or 1 5 a3 � a5! 1 6 �xd6 .£:! xc4 1 7 J... xc4 �xd6 18 l;l xd6 jl xc4 19 K b6 l!l ab8 20 tt:J e3 jl e6 2 1 !! d 1 li fe8 22 b3 \fi f8 with Dvoiris-Tiviakov, equality, USSR 1 988) 'ii!i xb2 16 ·�d3 � b6 1 7 � b l � a7 1 8 �d5 .£)d7 1 9 C?J c7 <E:� c 5 2 0 *' e 3 ii;[ ab8 is unclear according to Tiviakov. In practice 6 c4 has been the most popular response, but 6 .£J i c3 may be better, even though this allows Black the chance to transpose into the Pelikan. jl xd6 6 tt:Jd6 + 7 'f!l'!'xd6 'i'itf6 ( 1 19) 7 �e7 8 '!i' d l <E:� f6 9 tt:Jc3 threatening JJ.. g5 is good for White. 8 �d1 White has a wide variety of queen moves and most of them are good! There seems little doubt that 8 'i'itc7, which has always been highly regarded theoreti cally, gives White a good game but I have not recommended it

142 Lowenthal Variation

here because White must always be careful that his queen is not trapped, so the simpler �d I seems preferable. One should note that 8 �xf6 is also quite good, e.g. 8 �xf6 <£) xf6 9 <£) c3 and now: ( I ) 9 . . . d5 10 Jl.. g5 d4 (10 . . . ld b4 1 1 J.. xf6 gxf6 1 2 <£) xd5 <£) xc2 + 1 3 lt'd2 8 xa l 14 � c7 + 'I!J e7 1 5 <£) xa8 J..e6 16 �b6 JL xa2 1 7 fr c3 and 10 . . . -d xe4 1 1 � xd5 0-0 12 ii_ e3 are also good for White) 1 1 ll_ xf6 dxc3 1 2 .1J.. xg7 � g8 1 3 J.. h6 c.2) b4 14 0-0-0 <£) xa2 + 1 5 fr b l .§_e6 1 6 1i d6 h g6 1 7 jj_e3 <£) b4 1 8 jlc5 with advantage for White according to Gligoric. (2) 9 . . . <£) b4 and now: (2a) 10 ll_d3 l2\ xd3 + ( 1 0 h6 1 1 b3 d6 1 2 j}_a3 <£)xd3 + 1 3 cxd3 f,;e7 14 f4 * e6 1 5 f5 + lt; e7 1 6 � d l � e8 1 7 d4 exd4 1 8 � xd4 'l!f f8 1 9 jj_ xd6 + *g8 20 0-0 b5 2 1 e 5 Resigns was a drastic finish, Byrne-Evans, USA Ch. 1 98 1 ) 1 1 cxd3 h6 1 2 b3 with an edge for White. (2b) 10 lt' d2 d5 1 1 a3 d4 1 2 axb4 dxc3 + 1 3 fr e3 <£)g4 + 1 4 'I!J e2 f5 1 5 bxc3 l2! f6 1 6 .!ill a5

<£) xe4 17 f3! <£)d6 1 8 � xe5 + and Black has very little for the lost pawn, Velimirovic-Ristic, Yugos lavia 1 979. 8 ... �g6 8 <£) ge7 9 � c3 0-0 (9 �g6 transposes to the next note) 1 0 j}_e3 b5 1 1 �d2 �g6 1 2 f3 d6 1 3 0-0-0 li!( d8 14 * b l J.. b 7 1 5 g4 f6 16 <£)d5 QJ xd5 1 7 �xd5 + is also good for White, Gligoric Benko, Dublin 1 957. 9 � c3 ( 120)

9 d5!? For a time this move caused a revival of the Lowenthal, but now White has found a way to defuse the complications and liquidate to a favourable ending. The older line runs 9 . . . � ge7 10 h4! h5 (10 . . . d5 1 1 h5 �d6 12 h6! g6 1 3 exd5 and now 13 . . . � d4 and 13 <£)b4 allow 1 4 <£J e4, while 10 . . . h6 1 1 h5 � f6 12 j}_e3 0-0 1 3 �d2 b 5 1 4 0-0-0 b4 1 5 � a4 a 5 1 6 � b6 � b8 1 7 �d6 gave White an excellent ending in Boleslavsky Sakharov, USSR 1 957) 1 1 J_ g5 d5 (the only move that makes sense, for example 1 1 . . . b5 1 2

Lowenthal Variation / 43 �d3 .i:..b7 1 3 0-0-0 l,;;l d8 14 'i!i!(d6 1 5 l,;;l xd6 f6 16 ..:d_ e3 � c8 1 7 l,;;l d2 <2:! 6e7 1 8 j_d3 d6 1 9 l! hd 1 was very good for White in Hazai Csom, Warsaw 1 987) 1 2 exd5 (the tempting 1 2 J.. xe7 is met by 1 2 . d4!) <2:! b4 ( 1 2 <2:! d4 1 3 J.. d 3 J.. f5 14 J.. xf5 -EJ exf5 1 5 �d3 f6 1 6 j_e3 i s very good for White) 1 3 JJ... xe7 f'i xe7 1 4 JJ... d3! (much better than the often recom mended 14 d6 + since White reaches the same type of ending, but with his d-pawn securely defended) <2:! xd3 + 1 5 �xd3 �xd3 1 6 cxd3 and now: ( I ) 16 . . . b5 17 a3 and Black cannot recover his pawn, for ex ample 17 . . . J.. f5 1 8 f'ld2! llil, h6 1 9 iiii he 1 'l; d6 20 � ac 1 , Svesh nikov-Panchenko, USSR 1 977 or 17 . . . J.. b7 1 8 0-0-0 'l;d6 (sug gested by Sveshnikov) 1 9 d4, or finally 1 7 lliL b8 (suggested by Baumbach) 1 8 0-0-0 b4 1 9 axb4 � xb4 20 .i. he! Jii! xh4 (or else H e4) 2 1 d4!, and in all cases White has a good ending. (2) 16 . . H h6 1 7 0-0-0 .i g6 1 8 Jiii he1! li;1 xg2 (18 . . . f6 1 9 d4 and 18 . . . 'l; d6 19 d4 are also good for White) 1 9 .i, xe5 + ( 1 9 d4!? is interesting) 'l;d6 ( 1 9 wd8 20 lll[ del j_d7 2 1 d6 threatens lll( xh5) 20 d4 followed by -EJ e4 + and again White has the advantage. 10 � xd5 �xe4 + l l j_e3 �d4 This move, which is the only reasonable reply to the threat of -EJc7 + , is the idea behind 9 d5!?

12 -E:lc7 + f'le7 12 . . . *d8? allows White to take the a8 rook, while after 12 . . . * f8? White can either play 13 � cl or take the exchange by 13 'ii( d3 <E� xc2 + 14 '*'d2 �xd3 + 1 5 jl. xd3 -EJxe3 1 6 � xa8 �d5 1 7 liii ac I or take the rook -a pleasant choice! 13 iiii cl! Until this move was discovered Black had been doing rather well against 1 3 21 xa8?! and 1 3 'i'ii' d 3. 13 _itg4 If Black moves the rook on a8 then 14 c3 is very strong. 'i'ii' xd3 14 'i'ioid3 15 ]Lxd3 8 d8 I6 h3 (121 )

16 j_c8 The alternative is 1 6 j_h5 1 7 f4 f6 1 8 'f1 f2 't1d6 1 9 c3 and now: ( 1 ) 19 . . . *xc7 is bad after 20 cxd4 + <If b8 21 fxe5 fxe5 22 lii!, c5. (2) 19 . . . <2:! b3 20 axb3 'f1xc7 2 1 J.. e4 �e7 2 2 Jiil. hel <2:! d 5 2 3 fxe5 -EJ xe3 24 <If xe3 fxe5 25 llit fl 1il deS 26 l,;;l f5 J.. g6 27 liii g5 j_ xe4 28 ki\ xg7 + f'lc6 29 'llr xe4 with a winning position for White, Mar-

144 Lowenthal Variation janovic-Simic, Yugoslavia 1 983 . (3) 19 . . . Qj c6 20 J,. b6 exf4 2 1 c4 <2:l ge7 22 J.. e4 Qjc8 2 3 c5 + ltr d7 24 <2:ld5 <2:l xb6 25 cxb6 frd6 26 Qjxf4 J.. f7 27 !;!\ hdI + lfie5 28 lfle3 with a clear plus for White, Winsnes-Hillarp, Rilton Cup 1 988. 17 f4 exf4 18 jtxf4 Qje6 Otherwise White castles and Black is unable to develop his king's rook while e8 is covered. jtxe6 19 Qj xe6 20 0-0 The outcome of the opening is very favourable for White. He has two active bishops supporting a queenside pawn majority and while so many pieces remain on the board Black's king is not well placed on e7. <2:l f6 20 Qjd5 21 a3 22 .td2 Ji;l d7 !;!l c8 23 � eel b5 24 � f3 h6 25 b3 bxc4 26 c4 27 .t xc4?! It was more important to drive away Black's centralized knight than to keep the queenside pawns intact. After 27 be <2:l f6 28 .t b4 + White has a passed pawn and an attack against Black's king. 27. . . !4 aS?! ( 122) A passive and nervous move. 27 11. c6 is better. 28 a4

Preparing both J.. c l -a3 + and b4-b5. 28 frd6 An attempt to bring the king over to help in the fight against White's queenside majority, but two rooks and two bishops are a formidable attacking force and the king soon runs into trouble. 29 b4 fr c7 axb5 30 b5 � d6 31 J,. xb5 32 M dl With the sneaky threat 33 J,. f4 <2:l xf4 34 Jil c3 + winning the exchange. 32 \fi b7 � dd8 33 J_ b4 34 Jil fd3 '\1rc 7 Trying to unpin the knight. 35 � cl + '\1r b6 35 fr b8 36 )!l c5 J! a7 37 J.. c6 wins material. 36 J,.c5 + '\1r a5 37 l!il cdl Jil ac8 � e8 38 J,.e7 liil xe8 39 J.. xe8 Resigns 40 l!I xd5 +

Pin Variation

There is no generally accepted name for this variation, which runs I e4 c5 2 � f3 e6 3 d4 cxd4 4 � xd4 � f6 5 �c3 j_ b4. There is certainly a pin involved, so 'Pin Variation' is a reasonable name. Until ten years ago this was thought to be a very poor line for Black, but round about 1 979 it suddenly reappeared with Black's play being based on a new idea involving an exchange sacrifice. After a few years during which it was used in occasional Grand master games it entered a decline and is now very rarely seen. How ever it is worth studying because there are a lot of tricky tactics in the Pin Variation, and White players who do not know the cor rect antidote may well find them selves in trouble.

Game 28 Wagman--Barle Biel (open) 1981 1 e4 2 �f3 3 d4 4 � xd4 5 � c3 6 e5 ( 123) The only move to any difficulties. 6

c5 e6 cxd4 � f6 _it b4 cause Black f) d5

Black's two alternatives are close to losing by force: (I) 6 "iif a5 7 exf6 j_ xc3 + 8 bxc3 "iif xc3 + 9 'i!i!\'d2 'i!i!\'xa l 1 0 c3 (threat 1 1 � b3 '* b l 1 2 �d3) 'i!i!\' b 1 1 1 ltd3 'i!i!\' b6 12 fxg7 li g8 1 3 �h6 and wins. (2) 6 � e4 7 'i!i!\'g4 'i!i!\'a5 (7 � xc3 8 �xg7 liK f8 9 a3 and now the lines 9 lt a5 1 0 lt h6 "i'#fe7 1 1 � b3 and 9 "i'#fa5 10 � b3 'i!i!\'d5 1 1 J.. d 3 are winning for White so Black must try 9 � b5 + 1 0 axb4 � xd4 1 1 j_g5 �b6 1 2 j_ h6 'l!l!'xb4 + 13 c3 � f5 1 4 cxb4 � xg7 1 5 j_ xg7 .!J( g8 1 6 lt f6 but White's black square pressure gives him a very favour able ending) 8 'i!i!Jxe4 j_xc3 + 9 bxc3 'i!i!Jxc3 + 10 f8i d 1 >etxa l 1 1 � b5 d5 1 2 'ltb4 .Q)a6 ( 1 2 'i!i!\'xe5 1 3 f4 � c6 1 4 fxe5 � xb4 1 5 � c7 + f;>d8 1 6 � xa8 b6 1 7 j_a3 • • .

. . .

. . •

. • .

146 Pin Variation is winning for White) 1 3 o£) d6 + wd7 14 .2_ xa6 bxa6 1 5 o£) xf7 � g8 16 lff d2 d4 (or else j_a3) 1 7 j_ b2 �Wxa2 1 8 � a l �d5 1 9 !;;;[ a5 and White has a decisive attack (analysis by Euwe). 7 il..d2 Originally theory gave 7 ·�1· g4 as best, but after 7 . 0-0 (the new idea mentioned above) 8 kh6 g6 9 j_ xf8 'i'fi·xf8 Black has reasonable compensation for the exchange with play against c3 and e5. 7 � xc3 Or 7 .2_ xc3 8 bxc3 0-0 9 lt_d3 d6 1 0 'i'S' h5 (10 exd6 'M xd6 1 1 0-0 was at least slightly better for White, Geller-Tseitlin, Moscow 1 982, while 10 f4 dxe5 1 1 fxe5 o£) d7 12 'i'ti h5 g6 13 rt, e2 i't:r C7 14 c4 'ii& b6 15 �f3 J?Je7 1 6 iil_ c3 Bc5 17 �d2 proved good for White in Epishin-Uiybin, Tbilisi 1 989) g6 1 1 � e2 dxe5 1 2 i�'i' xe5 -::Jd7 1 3 � d6 w f6 14 0-0 b 6 1 5 J?Jc6 J?J c5 1 6 c4 e5 1 7 ·r.;; xe5 with a clear plus for White, Vogt-Ermenkov, Ber lin 1 982. 8 bxc3 iL a5?! A major decision point for Black. 8 :ill._ a5 keeps the pres sure against c3 but leaves the kingside dangerously bare. Black should adopt the alternative var iation 8 !ii e7 9 ·i£i g4 0-0 (9 . . . g6 1 0 h4 h5 1 1 i'l\ g3 J?Jc6 12 J?J b5 was good for White in Rabar Fuster, Munich 1 942, while the remarkable 9 . . . g5 I 0 h4 h5 1 1 hxg5 !? hxg4 1 2 � xh8 + .k f8 was played in Grosar-De Waal, Sas

van Gent 1 986, and now 1 3 g6 fxg6 1 4 j_h6 wf7 1 5 1! xf8 + * xf8 1 6 j_xf8 wxf8 1 7 j_e2 is good for White) 10 J.,h6 g6, but the main line still favours White after 1 1 h4! ( 124) and now: 124 B

( I ) 1 1 . . . d6 1 2 h5 �a5 ( 1 2 dxe5 1 3 Ad3! exd4 14 hxg6 fxg6 1 5 jl xg6 wins) 1 3 0-0-0!? dxe5 (13 . . . 'ii!i' xc3 1 4 .!it h3 ..W a l + 1 5 wd2 'iii1 xa2 1 6 J.,d3 � a5 + 1 7 c3 'tlifxe5 18 hxg6 fxg6 1 9 j_ xg6 � xf2 + 20 * c l and 13 . . . �xe5 14 hxg6 fxg6 1 5 J.. d 3 � f6 1 6 � de l 't!l!fd5 17 j_g5 Z! f7 18 j_xe7 � xe7 1 9 J.,xg6 are winning for White) 1 4 <£) b 5 a6 1 5 hxg6 fxg6 1 6 _;txf8 J.,xf8 1 7 � xh7! 'l.r xh7 1 8 j_d3 wins for White. (2) 1 1 . . . �a5 1 2 'l\ti'g3 M. d8 (after 1 2 . d6 1 3 exd6 Black may play 13 . . . J..xd6 1 4 �xd6 Ji d8 1 5 *' b4 'l!lol'xb4 1 6 cxb4 )i xd4 1 7 c3 with the better ending for White or 13 J.. f6 14 0-0-0 �xa2 1 5 h5 <£)c6 1 6 <£) xc6 'l!lol'a3 + 1 7 1ti'd2 bxc6 1 8 j_ xf8 wxf8 1 9 hxg6 hxg6 20 J.. c4 and White has the advantage) 1 3 h5 d6 14 hxg6 fxg6 1 5 .ZI.f4 dxe5 1 6 J\l. xe5 )i d5 • . •

Pin Variation 147 1 7 f4 <£) d7 1 8 1l.. c4 � xe5 (after 1 8 .!! xe5 + 1 9 fxe5 'l'#xe5 + 20 �xe5 � xe5 21 J,Lb3 the e6 pawn is too weak) 1 9 1l.. xd5 �xd5 20 fxe5 �e4 + 2 1 'tff d2 1l.. d 7 22 � ae 1 '¥'d5, Wedberg-Pokojowc zyk, Copenhagen Open 1 984, and now 23 � xh7! 1fr xh7 24 £ h l + 'tff g7 25 �h2 1trf7 26 'i!!J h 7 + 1tr e8 27 'i!lixg6 + 1trd8 28 g h8 + should win for White. 9 �g4 White can also use the move order 9 ll d3 d6 10 'i!\f' g4 after which Black has nothing better than 1 0 0-0 transposing into the game. 0-0 9 d6 10 j_d3 11 �f3 Byrne and Mednis suggest 1 1 J,Lg5 with the lines 1 1 . . . J,L xc3 + 1 2 lfr fl f5 1 3 exf6 � xf6 1 4 � d 1 and 1 1 . . . 'W:fc7 1 2 J.. f6 �xc3 + 1 3 * e2 'ii\'{ d 2 + 1 4 1tr fl g6 1 5 £ d I . This has never be.en tried in practice but looks good to me. g6 11 Or else ll_xh7 + is crushing, for example 1 1 dxe5 1 2 j_ xh7 + \\> xh7 1 3 �h5 + 1tr g8 1 4 ®g5 £ e8 1 5 )\i1xf7 + *h8 1 6 �h5 + followed by 'l)ith7 + , �h8 + , � xg7 + and ® f7 + picking up the queen. 12 h4! The latest twist-White just plays for mate. 12 � g5?! h5! (not 12 dxe5? 1 3 � xh7! lfr xh7 14 1l.. g 5 J..xc3 + 15 *e2 �d4 1 6 'i!\f' h3 + *g8 1 7 ll_ f6) 1 3 �g3 dxe5 14 ® e4 �d7 1 5 0-0 1tr g7

enabled Black to defend in Peters-Arnason, New York 1 980, while 1 2 0-0 dxe5 (Sigurjonsson gives the attractive line 12 . . . � c6 1 3 J.,g5 �c7 1 4 il, f6 j_ xc3 1 5 'i!\f'h4 <£)xe5 1 6 �g5 h5 1 7 � xh5) 13 � xe5 followed by f4 only gave White an edge in Sigurjonsson Arnason, Iceland 1 980. 12 . . . dxe5 1 2 . . . h5 is met by 1 3 'i!\f'g3 or 1 3 'i!\f' f4. 13 h5 f5 (125)

14 1L xf5! exf5 14 a xf5 1 5 hxg6 h5 1 6 ;g xh5 ·ii-J f6 ( 1 6 £ xh5 1 7 �xh5 '!i1c7 1 8 0-0-0 wins, while other wise Black can hardly meet the threat of � h8 + ) 1 7 0-0-0 fol lowed by � dh l with a decisive attack. 15 "ii'J' C4 + £ f7 15 W; g7 1 6 hxg6 '1!r xg6 1 7 � xe5 + wins. hxg6 16 hxg6 17 � g5 w c7 If the queen defends the rook from any other square White wins by 1 8 � xf7 '@' xf7 19 � h8 + 'lt; f8 18 ·i'i)' h4 19 � xt7?

148 Pin Variation Having conducted the attack so well up to here it is surprising that White should miss 1 9 �h8 + 'f!J e7 20 <E:! xf7 when he is material up with a mating attack. Fortunately White is still winning even after 1 9 <E:! xf7? 19 '1J xf7 19 'l'fxf7 20 �h8 + �g8 2 1 tt- f6 + wins the queen. 20 �h7 + '1ie6 20 'f!J f6 2 1 j_g5 + 'ffj e6 22 �xg6 + 'ffj d5 23 0-0-0 + 'ffj c 5 24 j_ e3 + 'ffi b5 25 � d5 + is even worse. 21 'jifx g6 + 22 R h6 23 � g8 + 24 � bl!

White correctly 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. b5 24 25 j_e3 + <Eld4 exd4 26 j_xd4 + 27 � f8 + White misses it the first time round 27 28 �g8 + 29 l!I xb5 + ! but spots it the second! 'f!J xb5 29 30 �d5 + �c5 31 a4 + Resigns It is mate next move.

� f6

Although this move surfaces from time to time, its appearances at the Grandmaster level are very rare. White has a range of pos sible lines against 2 .:2) f6, which have all achieved good practical results. The variation 3 ..£J c3, which was analysed in the first edition of this book, has per formed well in the intervening years, while the more recent idea 3 e5 .:2)d5 4 .:2)c3 e6 5 .:2) e4 has also acquired a good reputation. How ever, in this second edition, we will return to the 3 e5 .:2)d5 4 .:2) c3 e6 5 .:2)xd5 exd5 6 d4 variation which was recommended in the first edition. Concentrating on this one line I will give a more detailed coverage and show that White can be sure of at least a slight advantage against 2 B f6.

Game 29 Rhine-Sprenkle USA 1981 c5 e4 .:2) f6 .:2) f3 .:2) d5 e5 3 .:2) g4 4 h3 .:2) h6 may be met by 5 d4 or 5 c3, with advan tage to White. e6 4 .:2)c3 Or 4 . . .:2) xc3 (4 .:2)b4 5 1 2 3

J.. c4 and 4 . . . .:2)c7 5 d4 cxd4 6 �xd4 Qlc6 7 '¥'e4 g6 8 ll. c4 ll.g7 9 0-0 0-0 I 0 lii e I , Kindermann Ostl, Bundesliga 1 987, are good for White) 5 dxc3 ( 126) and now: 126 B

( I ) 5 . . . d5 6 exd6 �xd6 (6 exd6 7 J.. c4 ll.e7 8 ll.f4 0-0 9 �d2 followed by 0-0-0 puts severe pressure on the backward pawn) 7 �xd6 exd6 8 §t f4 ll.g4 (8 . . d5 9 0-0-0 §te6 1 0 .:2) g5 followed by g3 and either ll.g2 or <E:l xe6 and ith3, when Black's central pawns will be fortunate to survive) 9 0-0-0 <El d 7 I 0 11. c4 (even better than j}_ xd6, since . . . 0-0-0 is pre vented) followed by j}_xd6 win ning a pawn. (2) 5 . . . <E:lc6 6 ll. f4 h6 (6 . . . e6 7 �e2 � a5?! 8 <E:l d2! 'f!lc7 9 <E:l c4 f6 1 0 � d6 + j}_ xd6 1 1 exd6 � a5 1 2 h4 was good for White in van der Wiei-Bjelaj ac, Novi Sad 1 982)

150 2

!iJ/6

7 'i!ife2 flfc7 8 0-0-0 b6 9 �e3 e6 1 0 lE:! d2 lE:J e7 1 1 lE:Jc4 lE:! f5 ( 1 1 lE:!d5? 1 2 lii( xd5 exd5 1 3 lE:! d6 + \frd8 1 4 e6 wins) 1 2 '!Wh3 Ab7 1 3 lii( g l intending g4 and White has some advantage, van der Wiel Murei, Moscow 1 982. (3) 5 g6 6 Ac4 Ag7 7 A f4 0-0 8 'i!itd2 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 7 lE:Je5!) 7 �xd8 + \fr xd8 lE:Je5 *e8 9 Ab5 + Ad7 10 lE:Jxd7 lE:J xd7 1 1 A f4 and White is close to winning already. 5 � xd5 exd5 6 d4 lE:Jc6 If Black doesn't like to sacrifice a pawn he can try 6 d6, but after 7 A b5 + (127) Black can not equalize, as the following analysis shows: . . .

. . .

127 B

(1) 7 A d7 8 A xd7 + 'l!lr xd7 (not 8 . . . lE:J xd7 9 dxc5) 9 0-0 Q�c6 10 exd6 Axd6 ( 1 0 . . 'l!\rxd6 1 1 dxc5 'i!ifxc5 1 2 A e3 is also good for White) 1 1 l;il e l + lE:J e7 1 2 dxc5 Axc5 1 3 Ag5 0-0 14 �d3 f6 ( 1 4 . . . h 6 1 5 A xe7 Axe7 1 6 l! ad l . . .

)l! ad8 1 7 c4 Af6 1 8 cxd5 Axb2 1 9 d 6 and White's passed pawn is very dangerous) 15 Ae3 and Black has a poor isolated pawn position in which he has no active pieces to compensate for the static weakness. (2) 7 lE:Jc6 8 0-0 Ae7 (8 A e6 is also met by 9 c4 when Black has nothing better than to transpose by 9 Ae7) 9 c4 Ae6 (9 dxc4 1 0 exd6 tltxd6 1 1 d5 a6 1 2 Axc4 and 9 a6 10 Axc6 + bxc6 1 1 cxd5 cxd5 1 2 exd6 �xd6 1 3 dxc5 'i!ifxc5 14 A e3 are very good for White) 10 Ae3 �b6 (White threatened exd6) 1 1 a4 a6 1 2 a5 �c7 1 3 exd6 'l!\rxd6 14 dxc5 �d8 1 5 Axc6 + bxc6 1 6 lE:Je5 'f�Jc7 1 7 'il!Ya4 with a horrid posi tion for Black, Unzicker-Pomar, Bad Aibling 1 968. 7 dxc5 Axc5 8 '{lt xd5 'i!itb6 Here there is an important alternative: 8 d6 9 exd6 'i!if b6 ( 128) (giving up another pawn to allow Black's pieces to come into play more rapidly) and now: ( 1 ) 10 Ae3!? (for brave players . . .

. . .

2 only) lWxb2 ( 1 0 A:xe3 1 1 fxe3 � xe3 + 12 �e2 � e6 13 �g5 is good for White) 1 1 2 b5!? and now 11 . . . 'ii!l' x a1 + 1 2 lfi' e2 ·iiofc3 1 3 Axc6 + bxc6 (not 13 . . \fi'f8? 14 '*xc5 nor 13 lfi' d8? 1 4 � xf7) 1 4 � xc6 + <�r f8 1 5 'ii!l' xa8 �xc2 + 1 6 �d2 j_ xd6 1 7 � b l leads to an advantage for White. 11 . . . �xb5 1 2 iWxc5 is depress ing for Black, while 1 1 . . . 0-0 1 2 0-0 �xe3 1 3 fxe3 Ae6 1 4 'ii4' c 5 was good for White in Boll-Lanz, corr. 1 982, so the best line is 11 JLb4 + 1 2 'itr e2 ·l:l!fxc2 + 1 3 '-2:! d2 JLe6 with a total mess. (2) 10 ·i* e4 + � e6 ( 1 0 'itrd8 1 1 j_g5 + f6 1 2 0-0-0! 8 e8 1 3 �h4 is good for White) 1 1 !#h4! ( 1 1 ll_c4!? is an interesting recent idea, when van der Wiei-Short, Wijk aan Zee 1 990 continued 1 1 'i!irb4 + 1 2 '-2:!d2 0-0 1 3 0-0 � ae8 14 c3 'i'11 b6 1 5 � f3 h6 1 6 b4 ll_xd6 1 7 11..e3 'i/ii c 7 1 8 � xe6 � xe6 1 9 �c4 and Black has in sufficient compensation for the pawn) with the variations: (2a) 1 1 . . . f6 1 2 JLd3 0-0-0 ( 1 2 � b4 1 3 d7 + ! ll_xd7 1 4 �h5 + followed b y 0-0 i s good for White) 1 3 0-0 and there is a further branch: (2a l ) 13 JL xd6 14 j_ e3 �xb2 1 5 � ab l ili'fa3 16 J..c4 gives White good attacking chances for no sacrifice. (2a2) 13 � xd6 1 4 j_e3 (I am not sure that it was necessary to return the pawn since Black has no immediate threats; 1 4 a3 intending b4 appears promising) .

• • •

. . .

• • .

• • •

CZJ.f6 151

11.. xe3 15 fxe3 �xe3 + 16 fi h 1 '/f(c5? ( 1 6 �b6 i s just slightly better for White) 1 7 � ae l JLd7 1 8 '/f(g3 g6 19 �d2! f5 20 '-2:!b3 � b4 21 a3 winning material, Bundesliga Chandler-Arnold, 1 987. (2a3) 13 . . . h5 14 j_e3 � xe3 1 5 fxe3 �xe3 + 1 6 '\fi' h 1 j_g4 ( 1 6 � xd6 1 7 !:( ae l � b6 1 8 ;ii xe6 � xe6 19 1.L f5) I 7 itf g3 ( 1 7 Jii ae I �c5 1 8 j,Le4 is good for White according to Gutman) �c5 1 8 � ad I ltr b8 1 9 � e4 �e5 20 � xe5 J. xd I , Hansson-Fernandes, Lon don 1 984, and although this game has appeared a number of times in print, nobody seems to have noticed that after 2 1 ili'fxg7 Black can quite reasonably resign. (2b) 11 . . . .f.L xd6 1 2 j,L e2 (even 1 2 �d3 '-2:! b4 1 3 0-0 '-2:J xd3 1 4 cxd3 0-0 1 5 'l:l!fe4 h 6 1 6 .f.L e 3 'ii\' xb2 1 7 M fb I was slightly better for White in Hellers-Ivarsson, Swe den 1 985) _11.__ e7 ( 1 2 . '-2:!b4 1 3 0-0 is good for White after 13 0-0 1 4 c3 or 13 il.e7 1 4 '&{e4 f5 1 5 �e5 8 xc2 1 6 jj_ g5) 1 3 w e4 0-0-0 14 0-0 .k d5 (after 1 4 �d4 1 5 f) xd4 a xd4 1 6 � e3 H_c5 1 7 'i'l\' C3 White stands very well) 1 5 w g4 + 'iti> b8 ( 1 5 jJ_ e6 1 6 '£!a4), Chandler-Bartsch, Bundesliga 1 985, and now 16 c4! is good for White. 9 .k xf2 + k c4 10 '\ti>e2 0-0 11 � fl H_c5 Black has regained the sacri ficed pawn, but now f7 is exposed to attack. . . .

. • •

152 2

!iJ/6

129 B

12 � gS ( 129) �d4 + Transferring the knight to e6 in order to shield fl. 12 -d xe5? (12 d6? 1 3 � xfl! � d4 + 14 <tt d l j_g4 + 15 g f3 + 1Ji> h8 1 6 -�g8 + and mate) 1 3 -� xe5 d 5 1 4 � xd5 Jg, e8 + ( 1 4 lll._ g4 + 1 5 )4 f3 _r& g l 1 6 \fr fl ! g ad8 1 7 �e4 � d l + 18 1Ji> e2 1_xf3 + 1 9 gxf3 � fd8 20 R.xfl + 'itff8 2 1 fj xh7 + lJi>xfl 22 � g5 + 'l\> f8 23 � f5 + 'l;e7 24 'i>'\' fl + wd6 25 §1.[4 + <�rc6 26 'i<i' c4 + ·;;.t c5 27 ffl'xc5 + §J.. xc5 28 � xd l � e8 + 29 �e4 Resigns was Prokopchuk-Kuz netsov, USSR 1 972) 1 5 <tt f3 1£r f6 + 1 6 'l\> g3 lll._ d6 + 1 7 g f4! .§J.. e6 18 fj xe6 � xe6 19 'ii'i xd6 ·� g6 + 20 8 g4 � e3 + 21 .Jl.xe3 ·if( xd6 + 22 w f2 l;t\ e8 23 � f4 � e7 24 J[. b3 ·i>�i e5 25 8 e l g5 26 � f3 'it;g7 27 � d l f6 28 <�r g l g4 29 .,Rd4 Resigns, Spassky-Ciric, Marianske Lazne 1 962. fje6 13 'it;d1 14 2J e4 Here White has various possibi lities, but this move attacking c5 and restraining d6 looks best. d6 14 . . .

14 . j e7 is too passive and in Savkin-Tseitlin, corr. 1 972 White obtained a strong attack after 1 5 c3 ( 1 5 2J d6 is also good) d6 1 6 exd6 i;; d 8 1 7 lfrc2 jlxd6 1 8 � xfl! ft>xfl 19 2Jg5 + lt>e8 ( 1 9 'it;g8 20 ·ffl'e4 h6 21 :ZJ..e3 i:iif a5 22 J1. xe6 + 1L xe6 23 ·;i,f h7 + and � fl + wins) 20 fj xe6 � f2 + 2 1 *b3 )i(b6 + 22 il.. b5 + lll._ d 7 23 fjc7 + ! and Black resigned with out waiting to see one of the lines 23 . . . \if XC7 24 ·i'fi· g8 + _Rf8 25 ,A f4! or 23 . . . 1L xc7 24 -m g8 + 'lrJ e7 25 Jl.. g5 + '1id6 26 � d l + £ d8 15 exd6 15 Jl.. xd6? is a miscombination which rebounds after 1 6 � xd6 � d8 1 7 J.. f4 d Xf4 (Black saw the danger in Zaretdinov Pugachevsky, USSR 1 977 but still lost after 17 h6 1 8 J.e5 2J g5 1 9 l;ll xfl etc) 1 8 � xfl + 'lrJh8 1 9 �g8 + Resigns, Unzicker-Sar apu, Siegen 1 970. j_xd6 16 j}_d3 fS 17 ·�hS 'l!!J xd6 18 � xd6 After 1 8 . £ xd6 19 'l!!J xf5 the threats to f7 and h7 force 1 9 Jii: xd3 + , but Black does not have enough compensatioTh 19 'iif xfS ( 130) 19 . 'ill! xh2 Or 1 9 � f8 ( 1 9 g6 20 � fl + 'itr'h8 2 1 � f6 + '\t>g8 22 ,k f4 liquidates to an ending in which White has a clear extra pawn) 20 '!!!1 f7 + '\fih8 2 l ll!f f4 and now: ( 1 ) 21 � cS 22 j_e3 'i!l!Jh5 + (22 j}_g4 + 23 '\fid2 � a5 + 24 . •

. • .

W6 153

After this move White can gain a clear endgame advantage with no risk. The critical move is 22 . �xg2! 23 'lth5! ( 131) (I gave 23 ll f2 in the first edition, but 23 �g4 + 24 'iti'd2 'i!tb4 + 25 rfi d l � g5! i s good for Black) and now:

130 B

b4 �h5 25 �g5 exchanges queens) 23 *d2 ;l e6 24 �g5 (24 ;1 d4! looks very good to me since 24 . . � g6 allows 25 jl xg7 + fl/xg7 26 �f6 + and 24 �dS 25 �e5! �xg2 + 26 � f2 �g4 27 h3! is a disaster) �e8, A. Rodri guez-Diaz, Cuba Ch. 1 983, and now 25 � f4 intending J.. d4 gives White a clear advantage. (2) 21 . . . �e7 22 �g5 �e8, Short-Minic, Banja Luka 1 985, and now Minic gives the line 23 J.. d2 J.. e6 (23 � a4 24 ;lc3! lt;l xd3 + 25 f#j c l � g6 26 ll;! f8 + mates) 24 ;lc3 �g6 2 5 fl/ d2 l! d5 26 �g3 �d7 27 g ae l , assessing the final position as slightly better for White. I suspect that White's advantage is considerably greater than this; he is a pawn up with the two bishops, and if he consoli dates with f#j c l he must be win ning. Therefore Black should play 27 J.. f5, but after 28 .i, e3 followed by * c l White is a pawn up for nothing. fl/hS 20 �ti + 21 J.. gS � gS � dS 22 J.. e3! .

. . .

( I ) 23 � f8?! 24 � f4! (intending !;:;\ h4 and mate on h7) �h3 (seemingly forced) 25 i\!i xh3 ..@_xh3, Odeev-Varlamov, corr. 1 987, and now 26 *d2 il. g2 (26 � d8 27 i:i h l JL.g2 28 i;i. h2 J.. c6 29 Ji;\ fh4 h6 30 .i.d4 wins the exchange) 27 g g 1 .ll c6 28 � h4 (intending i;i. g3-h3) i;i. d8 29 i;i. g3 J.. d7 30 J..d4 and White has a large advantage. (2) 23 . . . g6 24 'i'li h4! (24 ,@. d4 + <2) g7 25 J.. xg7 + 'l/xg7 26 �e5 + <�ih6 is a draw) � g7 (24 Ji:i: d8 25 'il!f f6 + * g8 26 'ii1 f7 + lfJh8 27 J.. d2! and 24 . ll..d7 25 .li, f7 .i;i. g7 26 J.. e4! are very good for White) 25 i;i. g l ·� f3 + (25 'l!l1d5 26 .i;i. xg6 threatening 27 {i!(xh7 + �·xh7 28 � h6 mate is crushing) 26 *d2 .Rf5 reaching a position in which White has a . . .

• • •

154 2

t}jf6

very strong attack for the pawn. I have not been able to find a forced win for White, but Black has a difficult defence in prospect, e.g. 27 1i g3 �d5 28 g h i h5 29 _:&d4 (29 fii g5? Jii gd8!) _axd3 (the threat was 30 'iii xh5 + ) 30 cxd3 'lt'h7 (to defend against both J;i xg6 and J;i g5 followed by � xh5 + ) 3 1 � g5 ·� e6 (3 1 � f7? 32 ·�e4 � f5 33 ;g; hxh5 + gxh5 34 g xf5 wins) 32 ;,;; e5 'M c6 33 li;\ c l � g2 + 34 fi, e2 'M'd5 (34 �g4 35 ·M- xg4 hxg4 36 '£ e7 wins) 35 � c5 iird6 36 g e7 ;.:;. ad8 37 �c3 and Black is in big trouble since 38 fi, cc7 is threatened and 37 � c8 fails to 38 8 xg7 + while 37 li;\ d7 loses after 38 J;i xh5 + 23 'ii? f2?! As Rhine correctly points out, 23 'i'tl' f4! i'li' xf4 (23 'i'f'i xg2 24 i;;;\ g l '!!i¥ d 5 25 �d2 and &:; h i ) 24 8 xf4 would have given White a very favourable ending at no risk. 23 o.?Jc6 %'d6? 24 �d2 The bishop on d3 is the main enemy and Black should have tried to eliminate it by 24 . o.?Je5, when 25 _a xh7 (25 Jii hi -2) xd3) o.?J c4 + 26 l\>c3 8 d8 + ! (25 we5 + 27 _:&d4 �a5 + 28 l\> xc4 _&e6 + 29 l\>d3 and White evades the checks) 26 iL.d3 -2) xd3 27 cxd3 �d6 28 ·i"f h4 + l\> g8 29 #e4 leaves White with some advan tage, but in view of the opposite coloured bishops it isn't clear how many winning chances he has.

25 J1! hl h6 26 _lieS! ( 132) White's ambition is to gain f4 for his queen, when a sacrifice on h6 will be inevitable.

26 "¥ d5 26 'I!J!ic7 is spectacularly refuted by 27 ·;i1 f6! 26 . . . )ii e5 27 � ae l ·�g5 + 28 JL.e3 followed by .;i xh6 also loses quickly. 27 '1%'f4 � d8 27 ¥·xg2 + 28 'ltfc3 doesn't help Black in his efforts to combat Jiit xh6 + 28 Jiit xh6 + '!fi g8 29 � h8 + ! w xh8 30 i$'h4 + '\ff g8 31 �h7 + f;; f7 32 �g6 + f;; g8 33 �h7 + fr f7 34 �n + .ars 34 'l; e6/e8 35 � g6 + ltrd7 36 li;\ f7 + mates. 35 .!;;:. xf5 + !ii xf5 36 �xf5 + fr g8 37 ft'cl Resigns Quite apart from his material disadvantage there is no defence to the threat of � c4 + .

This is often called the O'Kelly Variation after the late Belgian Grandmaster who played it with some regularity. Since a6 is almost universal in the Sicilian Black gets it over with as soon a � possible, reserving his options as to which Sicilian system to adopt. White should not play 3 d4? cxd4 4 � xd4 � f6 5 � c3 e5 when Black has a favourable version of the Najdorf in which his king's bishop can emerge actively at c5 or b4. 3 c3 is a sensible reply, which tends to lead to 2 c3 Sicilian positions in which Black has played the unusual move a6 which is perhaps not the best wa to spend a tempo. H owever the strongest reply of all is 3 c4, which either leads to Maroczy Bind positions or to a sort of hedgehog. In view of the rare occurrence of 2 a6 in practice, it perhaps does not rate a chapter of its own, but this did give me the excuse to include another of my own games in the book!

;

Game 30 Nunn--Surtees Basingstoke Open 1977 1 2 3

e4 �f3 c4 ( 133)

c5 a6

3 .dc6 Or: ( ! ) 3 . . . d6 (this may lead to a type of hedgehog) 4 d4 cxd4 (4 1t g4 is an interesting move, since 5 dxc5 1l._xf3 6 'ii-1 Xf3 dxc5 leaves Black with a grip on d4 to com pensate for the two bishops; 6 gxf3! dxc5 7 �xd8 + lf!> xd8 8 2J c3 is possible, but the simplest line is 5 d5 with a positional edge for White) 5 ,f) xd4 2J f6. 6 -?Jc3 b6 (6 e6 7 j_d3 leads to positions from Chapter 6, so we concen trate here on Black's attempt to develop early pressure against e4 which is unique to 2 . a6) 7 k d 3 lt b7 8 0-0 (it is more accurate to play 8 *" e2 :£J bd7 9 b3-see next -?Je5 ?� te) -?J bd7 9 i't\' e2 e6 (9 ts Interesting since I 0 lL c2 ;:; c8 1 1 k a4 + -2l fd7 is unclear, so White would have to allow Black to take on d3) 1 0 b3 ( 1 0 f4 l\'t c7 1 1 lf!> h l

156 2

§J_e7 1 2 §J_d2 is also good, when Nunn-Franklin, London 1 985, continued 12 . . h5 13 � ae 1 h4 14 f5 '-i:lf8 1 5 fxe6 fxe6 16 '-i:l d5! �d8 17 e5 dxe5 18 "tl!l' xe5 �xd5 1 9 �xe6 �d6 20 �xd6 .ZL xd6 2 1 cxd5 Resigns) l¥fc7 1 1 §J_ b2 "*c5?! (Black plays too ambitiously with his king stuck in the centre- I ! .ZLe7 followed by . 0-0 is better) 1 2 � ae l ! (exploiting the tactical point 1 2 . l¥fxd4 13 '-2l a4 White prepares a breakthrough by '-i:l d5) b5 ( 1 2 �h5 1 3 'i'l!fd2 g6 14 f4 §J_h6 1 5 '!Wf2 g5 16 j}_e2 g4 1 7 �g3 M g8 1 8 §J_ d l ! was also good for White in Nunn-Frank lin, London 1 977, since e5 is imminent, while Franklin's later suggestion of 1 3 g5 allows 1 4 �d5! with added effect a s the 15 square is now available) 1 3 cxb5! �xd4 14 bxa6 j}_ c6 1 5 '-21 b5 1¥1 b6 16 §J_d4 'l'l!fb8 1 7 � c l '-i:lc5 18 a4! (this nullifies the threat of � xd3 and prepares to break open the c-file by b4) e5 1 9 §J_e3 Ae7 20 b4 '-i:l xd3 2 1 li xc6 � xb4 22 a7 'll!( b7 23 � b6 (heading for b8) � xe4 24 '-i:l c7 + '1t>d7 25 l¥t b5 + 'f:;xc7 26 li e ! + '-i:l c2 27 M xc2 + Resigns, Nunn-Franklin, Not tingham 1 979, as 27 *'xc2 28 li b7 + 'it' d8 29 li b8 + 'f:;c7 30 §L b6 is mate. (2) 3 . . . e6 4 '-i:lc3 � c6 5 d4 cxd4 6 � xd4 A b9 (6 '-21 f6 7 � c2! is good for White) reaching an unusual position which does not seem to be considered by theory. 7 '-i:l c2 j}_xc3 + 8 bxc3 is one possibility, but I like 7 '-i:l xc6.

Then 7 bxc6 8 "iii" d4 looks very awkward since 8 . . . � f6 and 8 . . . "iii" f6 are both met by e5, so 7 . . . dxc6 8 "iii" xd8 + '1J xd8 is best. Then White plays 9 jl_ f4, intend ing 0-0-0 + and � a4 with good play against the weak black squares at c5 and b6. If Black exchanges at c3 White had the dream square d6 for his bishop . cxd4 4 d4 '-21 f6 5 �xd4 Or 5 . . e5 (5 . . . e6 6 '-i:lc2 is still good for White, and 5 . . . g6 allows 6 '-i:l xc6 and 7 l¥td4) 6 '-i:l f5 d5 (6 � f6 7 '-i:lc3 transposes to Nunn-Surtees) 7 cxd5 A xf5 8 exf5 '-i:ld4 9 '-i:l c3 �e7 (again 9 .:£) f6 transposes) 10 .ztd3 ( 1 0 f6 is also promising) � exf5 1 1 0-0 jl_d6 1 2 f4 with a dangerous initiative for White. e5 6 '-i:l c3 7 '-i:l f5 d5 After 7 . . . d6 8 §J_g5 (8 '-i:l e3 controlling d5 is also good) ll_ xf5 9 exf5 �d4 1 0 ll_d3 White's con trol of e4 and d5 gives him an excellent position. 8 cxd5 jl_ xf5 9 exf5 �d4 � xd5 10 jl_ d3 1 1 0-0 ( 134) jl_b4 Black has tried a variety of moves in this position, but with out coming close to equality, for example 1 1 . . � xc3 1 2 bxc3 '-i:l c6 (once the knight has to move from d4 the only asset of Black's position vanishes) 13 � b l � b8 14 "it f3 fl!c7 15 jl_ e4, Ravinsky Kliascicki, U SS R 1 966, 1 1 . . . . . •

2 134 B

'-E.) f6 1 2 � e 1 '-E.)c6 1 3 � b3 Ji. b4 14 l! d 1 �e7 15 Ji.g5, Rogacovski Konova1ov, corr. 1 972 or 1 1 . . . !tJ... e7 1 2 !tJ... e4 �xc3 1 3 bxc3 '-2\c6 1 4 � b 1 �c8 1 5 'I'Mg4, Matano vic-Perez, Belgrade 1 96 1 with a clear plus for White in every case. 12 !tJ... e4! '-2\ xc3 13 bxc3 jj_ xc3 14 � b1 0-0 In Altshuler-Fink, corr. 1 960 Black tried to hold on to the pawn but after 14 . . � b8 1 5 �g4 g6 1 6 !tJ... g5 gxf5 1 7 A xf5 f6 1 8 �h5 + White had a winning position. 15 � xb7 This simple move was sug gested by Gligoric and Sokolov as an improvement over the unclear continuation 1 5 �g4 �d6 1 6 � d l � ac8 (but not 1 6 )g. ad8? 1 7 � d3 A b4 1 8 f6!) 1 7 � d3 � c4 in which White lacks a knock-out blow. � d6 ( 135) 15 White's main threat was 1 6 jj_a3, driving the rook away from the defence of f7, followed by � h5 and if Black manages to defend f7 White still has the crushing blow f6 in reserve.

a6 157

Black's �d6 is of course designed to prevent jj_ a3, but un fortunately the move loses by force. He had to try 1 5 � b8 although 16 � xb8 �xb8 1 7 f6 gives White a strong attack with no material investment.

16 ;,:, b3! White utilizes the undefended queen to threaten li.xa8, � xc3 and i_a3. Black's reply is forced. 16 . . . i,f ac8 17 ,& a3 �d8 If the queen moves anywhere else 1 8 j}_ xf8 .f) xb3 1 9 j}_ xg7 wins a pawn and demolishes Black's kingside. ,g e8 ( 136) 18 � b7

158 2

Ji!\ xti 19 A piece of rather unnecessary flashiness since 1 9 �h5 ·£ f6 (or 19 '8. c7 20 f6) 20 R_d5 wins quite easily. 19 -;n g5 19 l!txf7 20 ·i'ti" h5 + <tt g8 (20 lfi> f6 2 1 iYi' Xh7 � f7 22 i\-f h5 + forces the king to g8 in any case) 2 1 f6 g6 22 1/. xg6 'i'ii' d7 23 f7 + wins, but Black can hardly hope to survive long after losing the vital f7 pawn. 'it;h8 20 jtd5

21 21 ial.

g6 f6 gxf6 22 !1J..e7 costs mater-

e4 22 � a4 Losing at once, but Black's king would have succumbed soon m any case. 23 if:\ xh7 + Resigns After 23 lfr xh7 24 � d7 + lt>h6 25 "!Wg7 + * h5 26 'i'l!/h7 + lfrg4 (26 'ii\' h6 27 g4 + lti> g5 28 f4 + exf3 29 .& c l + ) 27 ·ll'f h3 + ltr f4 28 ..&. d6 + Ji!\ e5 29 "i'rl xc8 the position speaks for itself.

Unusual Lines

The material in the first 1 3 chapters will be sufficient to pre pare the reader for the vast majority of the games he will have as White against the Sicilian, but there remain a substantial number of unusual variations which Black players might adopt. Only a few have any pretentions to respecta bility and we concentrate on these few in this chapter. Wilder eccen tricities are usually best dealt with by an application of common sense and straightforward de velopment. The following break down of lines considered in this chapter will aid the reader in locating the variation he is looking for. A

The pseudo-Dragon 1 e4 c5 2 .S f3 � c6 3 d4 cxd4 4 � xd4 cEJ f6 5 � c3 g6. Unusual lines involving 2 � c6 apart from the pseudo Dragon: I e4 c5 2 � f3 � c6 3 d4 cxd4 (3 d5) 4 � xd4 � f6 (4 a6, 4 d5, 4 fif c7, 4 . 'i!ii' b6) 5 � c3 l'l!b6. Unusual lines involving 2 d6: 1 e4 c5 2 �f3 d6 3 d4 cxd4 3 � f6) 4 � xd4 � f6 5 �bd7) 6 � c3 e5 (5 lt.. b5 + Unusual lines involving 2 e6: I e4 c5 2 � f3 e6 3 d4 cxd4

d5) 4 � xd4 iJ.. c5 or 4 j_b4 + Unusual Black second moves: I e4 c5 2 �f3 g6 (2 ·liif c 7, 2 b6). A 1 2 3 4 5

e4 � f3 d4 �xd4 � c3

c5 �c6 cxd4 � f6 g6 (137)

137 w

This is an attempt by Black to reach accelerated Dragon posi tions without allow White the op tion of playing the Maroczy bind. It has been played a few times in Grandmaster chess, but White can obtain a clear advantage with accurate play. 6 � xc6 bxc6 Or 6 dxc6 7 �xd8 + * xd8 8 .f,tc4 *e8 (not 8 b5? 9 JJ.. xf7 e6 1 0 JJ.. g5 J.... e 7 1 1 0-0-0 + win. . .

1 60 Unusual Lines ning, while 8 . . . f;J_ g7 9 JL. f4 * e8 1 0 0-0-0 l£!d7 1 1 A c7 ! is good for White since Black is not allowed to castle) 9 e5 .!E) g4 1 0 f4 h5 (after 10 . . . A f5 1 1 h3 .'£!h6 1 2 g4 Black has to go back since 1 2 jL xc2 13 H h2 j_a4 14 � xa4 b5 1 5 jlb3 bxa4 1 6 f;J_ xa4 K c8 17 )ii c2 *d7 1 8 Ae3 is winning for White) 1 1 Ad2 h4 (or 1 1 JL.f5 1 2 h3 €:) h6 1 3 0-0-0 Ag7 1 4 K he l with ad vantage to White) 1 2 � e4 �h6 1 3 j_c3 h3 1 4 e 6 K g8 1 5 ex£7 + l£l xf7 1 6 0-0-0! with a clear plus for White, M aus-Schlick, B un desliga 1 987. l£! g8 7 eS After 7 . . . � d5 8 l£! xd5 cxd5 9 �xd5 K b8 1 0 e6! (with 'il!J'e5 if the pawn is taken) f6 1 1 JL.f4 li b4 1 2 Ad2 � b6 1 3 11.b5 K d6 1 4 "�tc4 White has a winning position. 8 jt c4 ( 138) 138 B

8 jt g7 Other moves are no better: ( I ) 8 . . . dS 9 exd6 �xd6 (9 exd6 10 i!ff3 d5 1 1 � xd5 cxd5 1 2 JL.xd5 �e7 + 1 3 j_e3 � b8 1 4 0-0 jtg7 1 5 A f4 with a decisive at-

tack) 10 0-0 'l'!fxd 1 1 1 l{ xd 1 A h6 1 2 Axh6 �xh6 1 3 K d2 � f5 1 4 �e4 i s very pleasant for White, Geller-Stein, USSR Ch. 1 966--7. (2) 8 . . . 'l'!fa5 9 0-0 and now: (2a) 9 . . . 'i'!fxeS 10 K e 1 'l'!f a5 (or 10 . . . 'l'!fb8 1 1 "ijd4 f6 1 2 � e4 11.g7 1 3 Af4 'l'!tb6 1 4 l£!d6 + * f8 1 5 'l'!td3 A b7 1 6 f;J_ xg8 K xg8 1 7 'l'!fc4 Resigns, Tiviakov-Muger man, Pinsk 1 989) 1 1 b4 'l'!td8 leaves Black in a dreadful mess. In Karaklajic-Ivanovic, Yugoslavia 1 974, White won Black's queen by 1 2 l£l e4 e6 1 3 ll_ b2 f6 1 4 ll_ xe6 dxe6 1 5 � xf6 + i!fxf6 1 6 ll_xf6 l£! xf6, which proved sufficient in the end, but I would not be sur prised if White had an even stronger continuation. (2b) 9 . . . jj_ g7 10 !#f3 f5 ( 10 . . . e6 1 1 1L.f4 and 10 . . . f6 1 1 g e l are also good for White) 1 1 it. f4 leads to the note to Black's I Oth move below. 9 'i\'r f3 fS Relatively best, for example 9 e6 10 ]L f4 %'a5 1 1 0-0! J..xe5 1 2 b4 "ijc7 1 3 � b5 i!f b8 1 4 JJ.. xe5 �xe5 1 5 H ad 1 d5 1 6 H fe 1 'l'!tb8 1 7 ll.. xd5 cxd5 1 8 'il!J'xd5 *f8 1 9 iird8 + *g7 20 � c7 � f6 2 1 'W!e7 and White wins. e6 10 J1.f4 Or 1 0 H b8 ( 1 0 'W!a5 1 1 0-0 Jl. xe5 1 2 b4 i!fc7 1 3 l£! b5 �b8 14 J.. xe5 'l'!fxe5 1 5 � fe 1 'il!J'b8 1 6 'l!lfc3 is now instantly decisive) 1 1 0-0 e6 1 2 � ad 1 followed by K fe 1 and j ust a s i n the main line White has an unpleasant bind, Anders son-Bilek, Teeside 1 972. Black

Unusual Lines 161 has no way to solve the problem of his backward d-pawn and the g7 bishop is inactive . White players must be on the lookout for the exchange sacrifice . � b4xf4, which can be good for B lack if he can get the e5 pawn, but provided White keeps his bind on Black's position he can be op timistic about the future . 11 0-0 The correct choice; in other games White played 0-0-0, but this gives Black counterplay down the b-file. -2) h6 11 We give the rest of the game Short-Sosonko, Wijk aan Zee 1 986, which is a model example of how to play such positions. Black is never allowed to free himself and is finally overcome by the problems resulting from the back ward d-pawn: 1 2 � ad l 1/ic7 1 3 � fe l � f7 1 4 �g3 0-0 1 5 h4 l:fr h8 1 6 � a4 a5 1 7 b3 � e8 1 8 '!We3 h6 1 9 g4! � g8 20 J,.g3 J.. f8 2 1 Ytib6 )!! a7 22 f3 1/ixb6 + 23 � xb6 j_c5 + 24 J.. f2 J,. xf2 + 25 fr xf2 fxg4 26 fxg4 * g7 27 -2) a4 g5 28 h5 l!! f8 29 * g3 � d8 30 � c5 � f4 3 1 a4 * f8 3 2 J.. d3 *e7 3 3 j_ g6 l!! a8 34 � e3 .1ii( b8 35 � ed3 � bb4 36 � xd7 + j_ xd7 37 .lii( xd7 + Resigns. c5 l e4 B � c6 2 � f3 cxd4 3 d4 3 d5 4 exd5 \ltxd5 5 � c3 �e6 + (or 5 1/ih5 6 d5) 6 J.. e3 cxd4 7 � xd4 �d7 8 � d b5 .1ii( b8 9 �e2 and White is probably win-

ning already, Boleslavsky-Gurge nidze, USSR 1 960. 4 �xd4 1/ib6 4 . . . a6 c4 transposes to Chapter 1 3 , 4 . . . d5 5 -2)c3 dxe4 6 � xc6 �xd l + 7 * xd l bxc6 8 � xe4 j}_ f5 9 J.d3 0-0-0 10 *e2 e6 1 1 J.. f4 *b7 1 2 .lii( ad l isjust good for White and 4 . . . � f6 5 ,.£) c3 )li!J b6 transposes to the main line. That leaves 4 tfHc7, which can be met by 5 ,.£) b5 � b8 6 c4 � f6 7 2\ 5c3 e6 8 f4 (8 j}_ e3 allows either 8 . . . j_d6!? or 8 . . . b6 followed by .A_c5) d6 (8 1J.. c5 9 e5 -2:J g8 I 0 2le4) 9 J\i_ e2 1J.. e7 10 ..! e3 0-0 1 1 0-0 b6 (White has a favour able version of Chapter 8 in which Black has lost time with his queen) 1 2 -E:� d2 j_b7 1 3 j_O � d8 14 a3 (better than 1 4 � c l i f8 1 5 ·¥' e2 d5!? with unclear play, Chandler-Barlov, Haninge 1 988) 1J.. f8 15 j}_f2 -E:�d7 16 b4 with advantage to White, Karpov Kurajica, Hastings 1 97 1 /2. 5 -E:� b3 ,.£)f6 6 -E:� c3 e6 7 J.. e3 'it!J c7 8 j}_d3 ( 139)

162 Unusual Lines 8 'ke7 a6 (8 ,k b4 9 0-0 0-0 Or 8 10 2:!b5 io� b8 1 1 f4 was good for White in Gheorghiu-Forintos, Monte Carlo 1 968) 9 f4 d6 I 0 � f3 (10 � e2 is also possible, as in some examples below, while White may start his kingside pawn advance immediately, e.g. 10 g4 b5 1 1 g5 2:!d7 12 M- d2 ii_b7 1 3 0-0-0 -2:)c5 14 '!r,) f2! with some advantage to White, Belyavsky Gufeld, Suhumi 1 972) b5 (if Black k_e7 we transpose to the plays main line below) 1 1 0-0-0 ii_ b7 1 2 'i11i> b l 2\a5 1 3 2\ xa5 ii'fxa5 14 g4 0-0-0 1 5 g5 2\ d7 1 6 a3 lfi b8 1 7 !l'i f2 1i.e7 1 8 k d4 e 5 1 9 fxe5 dxe5 20 .Jo. a7 + 1Wa8 21 2:!d5 with a clear plus for White, King-Wirth ensohn, Berne 1 988. 9 f4 d6 10 f'>'if3 It is useful for White to delay committing his king, since he can reasonably castle on either side. However, there is an argument for developing the queen to e2, for example 1 0 ·ri'i e2 a6 and now l 1 0-0-0 0-0 1 2 g4 ii;; e8 1 3 g 5 � d 7 1 4 ld f3 g6 1 5 )"( afl b5 1 6 8 h 3 b4 1 7 Rodriguez-Carlier, A. -2:\ d I , Amsterdam 1 987 or l 1 0-0-0 b5 1 2 1W b l 2:! b4 1 3 g4 2:!d7 14 g5 .Ja b7 15 a3 2\ xd3 16 cxd3 g6 1 7 ;,;; c l .2J c5 1 8 2l d4 ·M d8 1 9 h4, Todorovic-Bosic, Novi Sad Open attacking good with 1988, chances for White in both cases. In Wedberg-Benko, New York Open 1989, the continuation I 0 0-0 0-0?! (castling into the storm is

a6 was better) 1 1 g4 wrong; 1 0 (White is even better off than in the main line, as his queen may go directly to h5) �d7 12 g5 ii;;� e8 1 3 ii;;l f3 a6 14 ii;;� h3 gave White a massive attack; the finish was 1 4 � f8 1 5 'i'i1h5 � b4 1 6 ii;;l fl jj_ d8 17 a3 � xd3 18 cxd3 b5 1 9 f5 exf5 20 Qjd5 �d7 2 1 ji_d4 fxe4 22 ii;;� h4 e3 23 � xe3 :g e6 24 � hf4 � g6 25 � xf7 'tl!f xf7 26 :g xf7 � xg5 + 27 !¥'xg5 jL xg5 28 lll: xg7 + Resigns. 10 a6 ( 140) Black also delays castling since 0-0 1 1 g4 l;:;! e8 1 2 g5 .z:,d7 10 13 h4 � b4 14 h5 jLf8 1 5 0-0-0 a6 1 6 g6 gave White an immense Jansa-Martinovic, in attack Vrnjacka Banja 1 982.

l 1 g4 White has a choice of good lines. After 1 1 0-0 0-0 (not 11 . . . b5 1 2 e5, but castling invites the kingside pawn storm, so 1 1 .�d7 may be best) 1 2 g4 (in some games White played ii;;\ ae l , but this preparation is not necessary) Black has fared very badly in practice, for example 12 . . . lli\ e8

Unusual Lines 163 1 3 g5 �d7 1 4 �h5 � f8 1 5 f5 �e5 1 6 f6 �d8 1 7 � d4 b5 1 8 � £2 l.:. b7 1 9 � afl b4 20 � ce2 � fg6 21 .£l g3, Hawelko-Sznapik, Slupsk 1 988 or 12 . . . � b4 1 3 g5 -2)d7 1 4 'i>jh5 g6 1 5 �- h6 � e8 1 6 � ad l b 5 1 7 a 3 � xd3 1 8 � xd3 1.:. f8 19 �h4 lt_b7 20 j;_ d4, G. Mainka-Martinovic, Dortmund 11 1 988, and White's attack I S overwhelming in both cases. b6 11 This rather odd move is designed to support c5 in anti cipation of the manoeuvre � d7--c5. 1 1 . . . h6 1 2 0-0-0 b5 1 3 � hg l '-2\ d 7 1 4 w f2! (this possi bility explains Black's preference for b6 in the main line, since once he has played b5 the knight on d7 is hard to redeploy) l.:. b7 1 5 <ttr b l l.:. f6? (Black's 6 renders the kingside too dangerous for 0-0-he should have played 1 5 2l b4) 1 6 e5! dxe5 1 7 j_ xb5 0-0 1 8 g5 hxg5 1 9 fxg5 � e7, Estrin- Kopylov, USSR 1 973, and now 20 -�- h4! axb5 2 1 � g 3 would have given White a decisive attack according to Estrin. Perhaps Black should try 11 . . . b5 but White still has the advantage. 12 g5 2! d7 1 3 0-0-0 .£l c5 1 4 w b l (to answer .£l xd3 by cxd3) iJ_d7 15 h4 i'<1 b7 1 6 JJ_e2! (now the c5 knight isn't doing much) 2\a7 1 7 f5 2\ b5 1 8 li_ d4 kc6 1 9 fxe6 -2l xc3 + 20 ji_ xc3 2\ xe6 2 1 � hfl 0-0 22 lt.d3 b5 23 a3 2\c5 24 2\ xc5! dxc5 25 i-i f5 h ae8 (or 25

c4 26 ioii e5 f6 27 gxf6 _& xf6 28 i» e6 + ;..: f7 29 ii xf6!) 26 l.:. xg7! kd7 (26 '/ixg7 27 e5) 27 ·ffl" e5 il_d8 28 i<i d6 rk; xg7 29 � h6 + lft h8 30 e5 f5 3 1 exf6 .i_e6 32 £ de I c4 33 li_g6 � f7 34 JJ_xf7 � xf7 35 g6 i>i' xg6 37 f7 Resigns, Kavalek-Hiibner, Buenos Aires 1 978. c5 l e4 C d6 2 2\f3 3 d4 cxd4 Black quite often plays 3 -2l f6 in order to avoid the line 3 cxd4 4 � xd4. White should reply 4 2\c3 when Black is obliged to play 4 cxd4 5 -2) xd4 trans posing to normal lines. 4 -2l xd4 2\ f6 e5 ( 141) 5 2\c3 Or 5 -2l bd7 6 JL c4 21 b6 ( Black's development is not easy because e6 and .1J. .e7 will allow a il_xe6 sacrifice, while 6 g6 7 f3 kg7 8 ji_e3 0-0 9 'iti' d2 is a Dragon in which Black has de veloped his knight to the inferior square d7) 7 k b3 e5 8 � de2 it_ e6 9 .1J. g5 .i_ e7 10 ..&.xf6 ji_ xf6 1 1 2\ d5 2\ xd5 1 2 it_ xd5 ·;,;;- b6 1 3

141 w

164 Unusual Lines jtb3 0-0 14 c2Jc3 and White's con trol of d5 gives him a clear advan tage, R. Byrne-Cuellar, Siegen 1 970. 6 ll_b5 + One of the points of 5 a6 is to prepare e5 by preventing this move, so it is the only logical reply to 5 e5. 6 !El bd7 After 6 ll_d7 7 jt xd7 + �xd7 8 !El f3 (8 !El f5 allows Black to complicate the issue by 8 � xe4) the exchange of white squared bishops enhances the weakness of d5. 7 � f5 a6 8 j'j_ xd7 + 'Wfxd7 (142)

The critical moment. White has a number of possible plans, but it is not clear which is the best: ( 1 ) 9 ji gS � xe4 10 � xg7 + jtxg7 1 1 � xe4 and now: ( l a) 1 1 d5?! 1 2 � f6 + jtxf6 1 3 jtxf6 0-0 ( 1 3 � g8 1 4 0-0 is also very bad) 1 4 '/iYd3 e4 1 5 llt'd4 � e8 16 g4! �d6 1 7 0-0-0 'V!ff4 + 1 8 lfl b l � e6 1 9 g5 � e8 20 j_ h8 f6 2 1 gxf6 'V!ff5 22 � hg l + fi f7 23 � g7 + lfj e6 24 it b6 + Resigns, . . .

Camacho-Cruz Lima, Cuba 1 986. ( l b) 11 0-0 (this pawn sacri fice is the point of the variation, but it may not be correct) 1 2 �xd6 f6 ( 1 2 � f5 1 3 <El f6 + Axf6 1 4 �xf6 'lj'e4 + 1 5 lfl fl jth3 1 6 f3 �c4 + 1 7 *e l �b4 + 1 8 jtd2 is good for White) and I doubt if Black has enough for the pawn. After 13 jte3 �g4 1 4 �c3! jt f5 15 �d5 + � f7 16 h3 �g6 1 7 0-0-0! H c8 ( 1 7 jt xc2 1 8 H d2 jtf5 1 9 g4 is good for White) 1 8 � d2, K lovan-Mocalov, USSR 1 9 8 1 or 13 �xd7 jtxd7 14 j_d2 A f5 15 f3 Jlg6 16 0-0-0 � ac8 1 7 jtc3 b5 1 8 a3 H c6 1 9 H d7 jt f5 20 1i;!t a7 jth6 + 2 1 jtd2 jtxd2 + 22 lfl xd2, Perenyi-Bielczyk, Ber lin Open 1 988, White had the ad vantage and although Black managed to draw the first of these games the ending cannot be pleasant for him. (2) 9 <El e3 �c6 and now: (2a) 10 �d3 (this gives White a small but safe advantage) ll_e6 1 1 0-0 1i;!t c8 1 2 Ad2 ( 1 2 a4 jte7 1 3 � cd5 jtxd5 1 4 exd5 'i!lfc7 1 5 a5 g6 1 6 b3 0-0 1 7 � c4 H fe8 18 .il_ e3 A f8 was equal in Popovic-Rajko vic, Vrsac 1 987) jt e7 ( 1 2 . . . g6 is possible, but White is still slightly better) 1 3 � cd5 jtd8 (now 1 3 . . j}_xd5 1 4 exd5 f!fc7 1 5 <El f5 is good for White; jt d2 is much more useful than a4) 1 4 c4 0-0 1 5 !;[ ac l � e8 1 6 b3 �d7 1 7 � b4 with some advantage for White, Ehlvest-Kupreichik, Moscow TV 1 987. . . .

Unusual Lines 165 (2b) 10 �ed5 � x: d5 1 1 �xd5 il_e6 0-0 �c8 13 c4!? �·xc4 14 � b6 �xe4 15 �xc8 il_ xc8 1 6 £ e l �g4 1 7 f3 with a small plus for White, Krnic-Jovanovic, Yugoslavia 1 982. (2c) 10 0-0 .£\ xe4 ( 10 1Le6 is probably better, when White may have nothing better than 1 1 �d3 transposing to line 2a) 1 1 .£\ xe4 �xe4 1 2 �d5 � b8 1 3 b3 J. f5 14 c4 f6 1 5 il.. a3 ltr f7 1 6 �d2 � d8 1 7 � fe l 'i!l(g4 1 8 � e3 �·g5 1 9 �d5 + with advantage to White, L. Schneider-Bator, Sweden Ch. 1 986. (2d) 10 )li!ff3!? b5 1 1 0-0 j_ b7 1 2 � cd5 -2\ xd5 1 3 -2\xd5 iifxc2 1 4 ji gS il_xd5 1 5 exd5 f6 1 6 � ac l 'i!{(g6 1 7 k d2 .zt e7 1 8 � c 7 with good compensation for the pawn, Kudrin-Conquest, London 1 986. D 1 e4 c5 e6 2 2! f3 3 d4 cxd4 3 d5 4 exd5 exd5 5 1t. b5 + -2\c6 gives Black an uncomfor table isolated pawn position after 6 -2\ c3 or 6 0-0. _k c5 4 -2\ xd4 Or 4 k b4 + , when White can transpose to chapter 1 1 by 5 -2\c3 -2:1 f6, but it is also possible to play 5 c3 .k. e7 6 c4, with a Mar oczy bind position. The idea of 4 .k.c5 is to reach a position similar to that after 1 e4 c5 2 -2:1£3 e6 3 d4 cxd4 4 2J xd4 a6 5 ]Ld3 .�. cS (see game 2 1 ), but without wasting a tempo on a6. Naturally this exposes Black to the possibility of -2\ b5 at

some point, but attempts to exploit this directly don't work. White has to be a bit more careful to gain the advantage against 4 .k.c5. 5 2l b3 kb6 ( 143) 143 w

6 .EJc3 Simple development guarantees at least a slight advantage. The ambitious 6 c4 is also promising, for example 6 c4 -2\ e7 7 -2\c3 (White must prevent d5) 0-0 (or 7 -2\ bc6 8 ]L f4 e5 9 _kg5 f6 1 0 _.ik d2 d6 1 1 ·@ h5 + 'i\> f8 1 2 .ik.d3 ..k..e6 1 3 0-0 and White is better, Howell-S. Arkell, London 1 986) 8 il.f4 (it is essential to reach d6 with the bishop before Black prevents it with . f5, e.g. 8 jie2?! f5 and _k. f4 is impossible) f5 9 .1L d6 -2\ bc6 10 1/._ e2 and now if Black plays 10 . . . fxe4 1 1 c5 1Lc7 1 2 -2\ xe4 White's hold on d6 cannot be broken, while after 10 . . . f4 aiming to play .d g6-e5 (after a rook move, of course) White might even consider 1 1 il. h5!? 6 J"ij e7 7 1L.d3

166 Unusual Lines Or 7 ji_ g5!? f6 8 jj_ h4 0-0 9 ·M' h5 o.d bc6 10 0-0-0 o.de5 l l jj_ g3 o8 7g6 1 2 lti> b l f5 1 3 f4 o.d c6 1 4 ..2. c4 'f:& f6 1 5 e5 ii-ie7 1 6 � b5 a6 1 7 o.dd6 lti h8 18 h4 -?) a5 19 -2) xa5 Resigns, Sibarevic-G . Welling, Lugano Open 1 989. Weak play by Black, but this direct plan could be dangerous. 0-0 7 8 0-0 o.d bc6 9 jj_ f4 f5 (9 d5 1 0 exd5 <d xd5 l l .£) xd5 '!f(xd5 1 2 c4 'l!il(d8 1 3 j_ d6!) 1 0 Jt.d6 f4 l l �h5 f3 1 2 g3 JrLc7 1 3 e5 g6 1 4 �g5 §txd6 1 5 exd6 .£) f5 1 6 'iilf xd8 Q) xd8, Wedberg-Nunn, Helsinki 1 983, and now 1 7 11_ xf5 iiil xf5 1 8 Z! d4 l\i f8 1 9 a4! a 5 20 � a 3 would have been good for White. E l e4 c5 g6 2 -tl f3 One of the most important lines in this chapter, since it has oc curred many times in Grandmas ter chess and White can probably only secure an edge against it. Other second moves are very unusual and can be met by nor mal development, e.g. 2 b6 3 d4 cxd4 4 o.d xd4 j_ b7 5 Z!c3 Qj c6 (or 5 . . . a6 6 j_d3 g6 7 f4 !JJ... g7 8 <d O d6 9 0-0 followed by 'il'.\'e l -h4 with attacking chances) 6 _1 f4 � c8 7 o.d xc6 dxc6 8 "i¥f0 'l!iif d4 9 � d l 'l!l1c5 1 0 e5 i!j\ d8 l l � xd8 + 'ltrxd8 1 2 j_e2 *e8 1 3 0-0 f5 14 e6 Z! f6 1 5 � d 1 �d5 16 jj_e5 Resigns, Belyavsky-Quinteros, Vienna Open 1 986, or 2 'f!tlc7 3 c3 (it is probably not a good idea to play 3 d4 since 3 cxd4 4 .£) xd4 Z! f6 5 o.dc3 a6 followed by . . .

. . .

e5 gives Black a type of Naj dorf position in which his king's bishop can still be developed ac tively at c5 or b4) .:2) f6 4 e5 .:2)d5 5 d4 cxd4 6 cxd4 d6 7 4Q a3 a6 8 !JJ... d 3 e6 9 4Q c4 dxe5 10 dxe5 b5 1 1 4Qe3 � b7 1 2 0-0 followed by a4 when Black's queenside pawn structure will be weakened. 3 d4 The attempt to reach a Mar oczy bind position by 3 c4 !1.t g7 4 d4 can be met by 4 � a5 + , when it is not at all easy for White to maintain any advantage. 3 !JJ... g7 3 . . cxd4 4 .:2) xd4 transposes to lines examined earlier, for example 4 .:2)c6 5 c4 and 4 !JJ... g7 5 c4 end up in Chapter 7, 4 4Q f6 5 .:2)c3 d6 is chapter 5, and 4 . . . .:2) f6 5 .:2) c3 .:2) c6 leads to line A in this chapter. 4 dxc5 *a5 + (144) .:2) a6 5 J.. xa6 �a5 + 6 c3 4 'i'llf xa6 7 � e2 -tWc6 8 J.. e 3 'il'l' xe4 9 .:2) bd2 '/itc6 1 0 0-0 4Q f6 1 1 .:2)d4 'l!il(c7 1 2 .£) b5 'i'ir'd8 1 3 !JJ... f4 was good for White in Rajna-Nagy, Hungary 1 960.

Unusual Lines 167 5 -2J c3 Natural, but 5 c3 may be stronger, for example 5 "" xc5 6 _ke3 )5dc7 7 .ik.d4 e5? (7 -2l f6 8 e5 -2\g4 9 -2l a3 ! is better, but White can still claim a modest advantage) 8 li_e3 -2l f6 9 -2J a3 ! 0-0 10 -2l b5 i<dc6 I I -2) xe5 -;.Ii xe4 1 2 -2J xf7! with a large plus for White, Maric-Tringov, Bar I 977. -2)f6 5 Or 5 JL xc3 + (5 � xc5 6 B d5 e6 7 b4 )i-; f8 8 -2)c7 + ltr d8 9 -2J xa8 _kxa l 1 0 .:;ig5 + it.J6 1 1 JL xf6 + -2J xf6 1 2 � d4 i'tle7?! 1 3 1L. b5 b6 14 -2l xb6! axb6 1 5 iii xb6 + t\' e8 1 6 0-0 was very good for White in Mohrlok Breazu, corr. I 987) 6 bxc3 "i'<> XC3 + 7 li_d2 i>i" XC5 8 _kd3 (8 1L e2 -21 f6 9 e5 -21 g4 10 0-0 -2l xe5 1 1 11_e3 -2) xf3 + 1 2 j_xf3 fi, c7 1 3 ·i>i" d4 was also promising in Petrov-Limonikov, corr. I 974) -2) f6 (8 d6 9 0-011_g4 1 0 £ b l ·if[c7 1 I g b3 -2:! d 7 1 2 ;,; c 3 -2lc5 1 3 h3, Ambrosz-Petran, Czechoslo-

vakia I 979 gave White more than enough for the pawn) 9 0-0 0-0 1 0 e5 2} g4 I I g b 1 2} xe5 1 2 g b5 -2:! xf3 + 1 3 �xf3 ·iif c7 14 JL. h6 � e8 15 g f5 g f8 (New in Chess gave . f6 as unclear, but 1 6 g e 1 ! appears crushing after 1 6 tii c3 I 7 .k c4 + ! ·ii'i' xc4 1 8 g xf6) I 6 JL. xf8 gxf5 1 7 JL xe7 d5 1 8 � e 1 .k.e6 1 9 J.:. f6 -2) d7 20 .it.. d4 ii'i'd6 2 1 k xf5 k xf5 22 -� xf5 -2:! f8 23 i!i' g5 + -21 g6 24 h4 g f8 25 h5 Resigns, Frivaldszky-Monostori, corr. 1 986. "ffl' XC5 6 it.. d3 7 _ke3 ·;.<; a5 7 'if! h5 is possible, but I cannot find any practical ex amples of it. 8 iii d2 -2\ c6 9 0-0 0-0 1 0 h3 d6 1 I a3 ..k .e6 I 2 -2)g5 d5 ( 1 2 1J. d7 1 3 f4) 1 3 exd5 kxd5 1 4 b4 i>ii d8 1 5 � ad ! and White has a slight advantage, Sveshnikov-Romanishin, USSR eh. 1 977.

Index of Variations 1 2 Now: A 2 B 2 c 2 D 2

e4 � f3

ci)c6 d6 e6 others

g6 159 5 5 0 0 0 e5 6 � db5 d6 (6 0 0 0 others j}_e6 63) 8 63) 7 ;Lg5 a6 (7 � a3 b5 (8 0 !1L e7 63; 8 0 d5 64; 8 Jte6 64) 9 §J.. xf6 gxf6 ( 1 9 �xf6 68) 1 0 � d 5 f5 ( 1 0 jj_g7 68-70) 1 1 j_d3 !1L e6 12 0-0 !1Lg7 (12 Jtxd5 70) 1 3 �h5 71 e6 6 ;Lg5 Jtd7 7 Jl_e2 59 6 6 g6 5 7 �a5 58 6 6 '/!!f b6 58 a6 58 6 7 �d2 a6 h6 40 7 � xd4 41 7 7 1J.. e 7 8 0-0-0 0-0 (8 � xd4 41; 8 a6 49) 9 <2J b3 a5 (9 d5 43; 9 �a5 43; 9 h6 43; 9 a6 43-4; 9 t'rl b6 44-5) I 0 a4 d5 1 1 iL b5 46 8 0-0-0 h6 8 $t e7 49 8 .1d7 9 f4 b5 (9 h6 50; 9 k e7 51) 10 _k xf6 gxf6 52 1J.. d7 9 :;ie3 9 others 54 10 f4 55 0

A 2 �c6 3 d4 cxd4 3 d5 161 4 � xd4 �f6 4 *' b6 161 4 d5 161 4 ·r!fc7 161 4 a6 5 c4 156 4 0 0 0 e5 5 � b5 a6 ( 1 5 � f6 6 � l c3 63; 5 d6 140-1; 5 j}_c5 6 � I c3 .t) f6 63; 5 0 h6 140) 6 �d6 + 1t. xd6 7 tif xd6 141 4 g6 5 c4 � f6 (5 j}_g7 6 � xd4 j}_e3 1 1 7) 6 � c3 d6 (6 I l l) 7 j}_e2 j}_g7 (7 ci) xd4 I l l ) 8 j}_e3 ( 8 �c2 I l l) 1 19 4 e6 5 � b5 d6 (5 � f6 6 � l c3 135; 5 1t.c5 125) 6 c4 � f6 (6 a6 126) 7 � l c3 a6 8 � a3 j}_e7 (8 0 0 d5 126) 9 jj_ e2 0-0 1 0 0-0 b6 ( 1 0 others 127) 1 1 ;Le3 �e5 ( 1 1 others 127-30) 1 2 f4 �ed7 130 5 Q}c3 d6 5 0 0 0 e6 6 .£�db5 J.. b4 (6 d6 7 J.. f4 e5 8 j}_g5 -see 5 e5) 7 a3 ;L xc3 + 8 � xc3 135 . . •

B 2 3 4 5

d4 :£:) xd4 .£� c3

d6 cxd4 .£�f6

Index of Variations 169 12

h4 j.i·a5 85 a5 85 -2Jc4 86 13 k. g5 13 a6 87 13 -2Jh7 87 13 -2Jc4 87 14 '!irb1 90

Now : e6 B1 5 g6 B2 5 B3 5 a6 Bc6 see A an d for 5 For 5 -2\ bd7 see p. 1 63. e5 and 5 81

6 6 6 6 7 7 8 8

e6 5 ... h6 6 g4 e5 32 -2J c6 7 g5 Bd7 8 h4 33 a6 7 g5 -2\ fd7 8 h4 36 1J..e7 7 g5 -2\ fd7 8 h4 37 -2Jc6 7 h4 k. e7 20 a6 8 k g2 21 h5 8 )4 g1 j_d7 26 d5 26 i!1 Xh5 28 9 gxh5

82 g6 5 1J.. g7 6 .1J.. e3 6 a6 78 0-0 7 f3 7 a6 78 -2\c6 78 7 -2\c6 8 •i>i-!d2 d5 79 8 9 j_c4 ii. d7 9 a6 80 9 a5 80 -2\ xd4 80 9 �¥· a5 80 9 �d7 80-1 9 � c8 10 0-0-0 ftiC7 82 10 10 ·;;.;- b 8 82 �a5 1 1 .k.b3 83 10 11 -2\ e5 JJ.. b3

12 12 12

;g_ c5

83 5 a6 e5 6 f4 6 <2Jc6 2 6 g6 2 6 e6 7 'i!i;f3 2 � bd7 7 � e2 3 6 6 . 'i!i;c7 7 <-2l f3 <2J bd7 (7 8 �d3 6) 8 jj_d3 7 � bd7 7 �f3 7 fijc7 8 a4 9 8 a4 JJ.. e7 8 f;tc7 9 8 d5 12 9 $td3 0-0 10 0-0 12

2 e6 3 d4 cxd4 3 d5 165 4 � xd4 a6 Ac5 165 4 4 $t b4 + 165 4 � c6-see A 4 � f6 5 � c3 jt b4 (5 � c6 134) 6 e5 d6-see B 1 ; 5 145 � f6 5 Ad3 5 � c6 94 5 g6 95 5 . . . :t:J e7 96

1 70 Index of Variations 5 5 5 5 5 6 6

�b6 96 �c7 96 b5 96 A_c5 6 � b3 96 d6 6 0-0 � f6 10/ d6 6 0-0 'ft(c7 101-2 e5 101 7 c4 102

D 2 2 2 2 2

� f6 149 a6 155 g6 166 b6 166 'fife? 166

BATSFORD publish the world's best

CHESS B O OKS written by the world's best

CHESS PLAYERS

KASPAROV - KARPOV - SHORT NUNN - SPEELMAN - CHANDLER FLEAR - SMYSLOV - KORCHNOI ADOR]AN - KOTOV - KEENE PLASKETT - HODGSON - RETI NIMZOWITSCH - TAIMANOV LASKER - SVESHNIKOV POLUGAYEVSKY - SUETIN STAl JNTON