Using physiological data to predict future career progression 14 17 australia by david da silva

Journal of Sports Sciences

ISSN: 0264-0414 (Print) 1466-447X (Online) Journal homepage: http://www.tandfonline.com/loi/rjsp20

Using physiological data to predict future career progression in 14- to 17-year-old Austrian soccer academy players Christoph Gonaus & Erich MĂźller To cite this article: Christoph Gonaus & Erich MĂźller (2012) Using physiological data to predict future career progression in 14- to 17-year-old Austrian soccer academy players, Journal of Sports Sciences, 30:15, 1673-1682, DOI: 10.1080/02640414.2012.713980 To link to this article: http://dx.doi.org/10.1080/02640414.2012.713980

Published online: 23 Aug 2012.

Submit your article to this journal

Article views: 1016

View related articles

Citing articles: 13 View citing articles

Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=rjsp20 Download by: [UQ Library]

Date: 28 July 2016, At: 21:22

Journal of Sports Sciences, November 2012; 30(15): 1673–1682

Using physiological data to predict future career progression in 14- to 17-year-old Austrian soccer academy players

¨ LLER CHRISTOPH GONAUS & ERICH MU Department of Sport Science and Kinesiology, University of Salzburg, Salzburg, Austria

Downloaded by [UQ Library] at 21:22 28 July 2016

(Accepted 17 July 2012)

Abstract The crux of any soccer-specific talent identification programme is to determine early predictors of future playing success (Williams & Reilly, 2000). We compare physiological characteristics among 14- to 17-year-old soccer academy players in terms of subsequent career progression (‘drafted’ vs. ‘non-drafted’). In a longitudinal design (2001–2010), players passed through 10 fitness tests at four age levels: 14 (n ¼ 410); 15 (n ¼ 504); 16 (n ¼ 456); and 17 years (n ¼ 272). MANOVAs showed statistically significant (P 5 0.05) superior performances for drafted players in all components (‘speed’, ‘power and flexibility’, ‘coordination and endurance’) and age categories. ANOVAs revealed significantly (P 5 0.013) better performances for drafted players in almost all tests, with the largest effect sizes for shuttle sprint (Z2 ¼ 0.07–0.09), 20 m sprint (Z2 ¼ 0.04–0.05) and medicine ball throw (Z2 ¼ 0.05–0.11). Follow up discriminant analyses confirmed that a combination of three variables correctly classified 62.7 to 66.2% of the players. Soccer-specific speed and power of upper limbs best discriminated future playing status, irrespective of age category. It is concluded that physiological measurements at adolescence can provide useful information in terms of predicting future career progression.

Keywords: football, talent selection, longitudinal design, drafted vs. non-drafted, fitness testing

Introduction Recognising and promoting potentially talented soccer players at an early age are crucial aims for many top soccer clubs to ensure both sporting and financial success and/or survival. Thus, identifying early predictors of long-term success ensures that the most talented players receive high quality coaching from an early age (Stratton, Reilly, Williams, & Richardson, 2004; Williams & Reilly, 2000). Previously, cross-sectional research on youth soccer players typically revealed that several anthropometric (height, weight, body composition), physiological (speed, agility, explosive power, aerobic capacity), and psychological (ego orientation, anticipation skill) characteristics, as well as soccer-specific skills (dribbling, ball control) contribute to current playing success and team selection in youth soccer (Coelho et al., 2010; Gil, Ruiz, Irazusta, Gil, & Irazusta, 2007; Reilly, Williams, Nevill, & Franks, 2000). Furthermore, Vaeyens et al. (2006) highlighted the dynamic behaviour of discriminating characteristics during adolescence. While running speed and soccer skills were important

discriminating factors in younger (U13, U14) players, aerobic endurance and trunk strength were more relevant in older (U15, U16) players. In addition, chronologically older (Gil et al., 2007) and more mature (Malina et al., 2000) players may have advantages in the selection process. As a result, in order to avoid overlooking late maturation of highly skilled players in the selection procedure, coaches are encouraged to consider players’ potential, as well as attributes such as cognitive factors and technical and tactical skills, rather than physical size and other characteristics related to prematurity and birth date (Musch & Grondin, 2001; Williams & Reilly, 2000). However, after being admitted to a soccer academy, little influence of chronological age was found for U14 players relative to whether or not they reached professional level after academy graduation (Carling, Le Gall, Reilly, & Williams, 2009). Thus far, cross-sectional research on the actual performance level of youth soccer players dominates the literature. However, for an advanced understanding of factors contributing to later expert performance, a longitudinal design following the players into adulthood would be more profitable

Correspondence: C. Gonaus, Department of Sport Science and Kinesiology, University of Salzburg, Schlossallee 49, 5400 Hallein-Rif, Austria. E-mail: christoph.gonaus@sbg.ac.at ISSN 0264-0414 print/ISSN 1466-447X online Ó 2012 Taylor & Francis http://dx.doi.org/10.1080/02640414.2012.713980

Downloaded by [UQ Library] at 21:22 28 July 2016

1674

C. Gonaus & E. Mu¨ller

(Kannekens, Elferink-Gemser, & Visscher, 2010; Williams & Reilly, 2000). Some authors (Carling et al., 2009; Franks, Williams, Reilly, & Nevill, 1999) reported that neither physical nor physiological differences at adolescence discriminate between future successful and less successful academy players. In contrast, some benefits in functional capacities such as speed, agility, maximal anaerobic power, explosive power and endurance, soccer-specific skills, and physical maturity were displayed at puberty for future elite and international players (Figueiredo, Goncalves, Coelho, & Malina, 2009; Le Gall, Carling, Williams, & Reilly, 2010). Furthermore, regarding the development of soccer-specific endurance, Roescher, Elferink-Gemser, Huijgen, and Visscher (2010) assigned the age of 15 years to be crucial with respect to reaching professional level in soccer or not. At early adolescence, the development of intermittent endurance capacity of players who reached professional level in adulthood parallels that of those that remained amateur. However, from the age 15 years on, the gap between the two groups becomes progressively larger. In terms of dribbling skill, future professionals outperform their amateur counterparts from age 14 years on (Huijgen, Elferink-Gemser, Post, & Visscher, 2009). Few researchers have examined the physiological differences of younger age groups in players who were already exposed to systematic training with respect to their career progression. Some reasons include the slightly inhomogeneous performance levels of the previously tested players or the small sample sizes typically used in longitudinal studies. Therefore, in contrast to the classical cross-sectional designs, prospective studies are preferable. In the present study, we compared the physiological characteristics of subsequently drafted and non-drafted soccer academy players from the 14 to 17 years old category. Given the reported characteristics of youth soccer players, we assumed that those drafted to play at the international youth soccer level would perform better in physiological tests than their non-drafted counterparts, irrespective of age level. Furthermore, we hypothesised that a combination of physiological measurements could discriminate future playing status at each age category. Methods ¨ FB), In 2001, the Austrian Football Association (O the Department of Sport Science of the University of Salzburg and the Institute for Sports Medicine and Science (IMSB Austria) launched a project targeting the development of the most talented Austrian soccer

players. One section of the project was the scientifically guided education and control of training in elite Austrian youth soccer players who attended one of the 12 – at least one per federal state – youth soccer academies throughout Austria (until 2008 there were 13 academies). In general, the players and their parents/guardians signed a training agreement ¨ FB, who, for their part, with the academy and the O gave their permission to the scientific processing of the test data. Ethical approval was obtained from the local university ethics committee. Talented youth soccer players are promoted systematically across Austria from the age of 10 years on. The selection criteria into academies at age 13 or 14 years were multi-factorial containing assessment of technical skills, speed, game intelligence and personality. After academy graduation at 17 to 18 years of age, the future career of the players was examined by categorising them into the following two groups: (1) ‘drafted’, players who subsequently had been drafted at least two times into a youth national team (U18 to U21); and (2) ‘nondrafted’, players who had never been drafted to play at international youth soccer level. Participants In a longitudinal design (July 2001 until December 2010), approximately 3000 players aged 14 to 17 years with dates of birth ranging from 1983 to 1993 performed a battery of fitness tests aimed at providing coaches with information on the players’ general and soccer-specific fitness. Due to the ordinary fluctuation of academy players, injuries, or other reasons for non-participation, a total of 4733 measurements remained for analysis, resulting in the following subgroups: 14 years drafted (n ¼ 205) and non-drafted (n ¼ 1160); 15 years drafted (n ¼ 252) and non-drafted (n ¼ 1089); 16 years drafted (n ¼ 228) and non-drafted (n ¼ 995); and 17 years drafted (n ¼ 136) and non-drafted (n ¼ 668). Measurements were taken twice a year each at the beginning of the summer (July or August) and winter (December or January) periods. For players tested twice in the same age category, only the best score was used for statistical analyses. Test procedure To ensure standardisation and to limit circadian fluctuation, a standard warm up was conducted, and tests were executed at the same time of day, in the same order, using the same measurement systems (Drust, Waterhouse, Atkinson, Edwards, & Reilly, 2005; Svensson & Drust, 2005). All participants passed through the same test battery at earlier stages of their career (from age 11 years on); therefore, they

Predicting career progression in youth soccer

1675

were familiar with the test proceedings. Players started with the 20 m straight-line sprint and foot tapping. Next, they conducted two attempts of sitand-reach and the 2 kg standing medicine ball throw each, and subsequently performed the second attempt of the 20 m sprint and foot tapping. Afterwards, participants were split into four groups, each passing in a random order through one trial of reaction test, and two trials of 5 6 10 m shuttle sprint, hurdles agility run, and jumps (countermovement and drop jump). Finally, players performed a 20 m multi-stage endurance run. For those tests where two trials were allowed, only the better performance was used for statistical analysis. All tests were performed on an indoor surface.

Downloaded by [UQ Library] at 21:22 28 July 2016

20 m sprint The players’ sprinting ability was measured by maximal 20 m straight-line sprint with 5 and 10 m split times. Infrared timing gates with 0.01 s accuracy (Brower Timing Systems, Utah, USA) were placed approximately at hip height 90 cm above ground, at 0, 5, 10, and 20 m. Players started in a standing position 0.5 m behind the first timing gate. Current research on youth sprint ability displayed intraclass correlation coefficients (ICC) of 0.88– 0.98, and Pearson correlation coefficient of 0.90– 0.97 for 10 to 40 m sprints (Rumpf, Cronin, Oliver, & Hughes, 2011).

Figure 1. Schematic representation of the hurdles agility run.

Table I. Body size and the corresponding hurdles height of hurdles agility run. Body size [cm] 136 141 146 151 156

– – – – –

140 145 150 155 160

Hurdles height [cm] 56 58 60 62 64

Body size [cm] 161 166 171 176 181

– – – – –

165 170 175 180 185

Hurdles height [cm] 66 68 70 72 74

5 6 10 m shuttle sprint Soccer-specific speed and agility were measured by the 5 6 10 m shuttle sprint. Participants started in a standing position 0.5 m behind the first timing gate (Brower Timing Systems, Utah, USA). In this running and turning test, players had to complete five runs of 10 m at maximal speed, comparable to the shuttle sprint test described by Verheijen (1998). Turns had to be as quick as possible with one foot – freely selectable – crossing a line while changing direction. Two trials were separated by a 5 min break. Hurdles agility run General agility was examined using the hurdles agility run. After starting in the standing position 0.5 m behind the timing gate (Brower Timing Systems, Utah, USA), players had to perform a roll forwards; then, always passing the middle pole, they had to jump over the hurdles and crawl back under them. Measurement setup and running path are displayed in Figure 1. Hurdle heights were adjusted to body size as reported in Table I. A 5 min break was given between the two attempts.

Jumps To determine lower body power and explosive strength, both vertical countermovement jump and drop jump were conducted on a force plate applying piezoelectric sensors (Kistler Instrument Corporation, Winterthur, Switzerland). Ground reaction force measurement was used to calculate jump height in a countermovement jump (0.1 cm), whereas flight time determined the height of the drop jump. The two attempts at the countermovement jump and the drop jump were separated by 15 s of passive recovery. Read and Cisar (2001) demonstrated that a 15 s rest interval was a sufficient amount of time for recovery during jumps at maximal effort. In the countermovement jump, participants were allowed to use arm swing and preparatory countermovement. Jaggers, Swank, Frost, and Lee (2008) reported an ICC of 0.98 for countermovement jumps conducted on a Kistler force plate. For the drop jump, players were instructed to drop down from a height of 40 cm and were encouraged to perform a maximal vertical jump as explosively and

1676

C. Gonaus & E. Mu¨ller

quickly as possible after landing two-legged on the Kistler plate. Legs were not allowed to be tucked up while landing. An ICC of 0.96 was found for a drop jump from 30 cm (Haj-Sassi et al., 2011). Both ground contact time (to the nearest 1 ms) and height (0.1 cm) were entered into the following formula to calculate a drop jump coefficient (to the nearest 0.01): Drop jump coefficient ¼

ð8=9:08665Þ ðheight=100Þ ðground contact time=1000Þ2

A low contact time in common with a high jump resulted in a better drop jump score.

Downloaded by [UQ Library] at 21:22 28 July 2016

Foot tapping Maximum speed of lower limb movement was measured with foot tapping. Players stood with the right foot on the force plate (Kistler Instrument Corporation, Winterthur, Switzerland) and the left foot on a wooden box of the same height. Participants started autonomous and performed foot tapping at maximal speed over 5 s. The sum of the contacts of the right foot was multiplied by 2 and divided by 5 to obtain the tapping cycles per second (0.1 Hz). Previously, unpublished data from a Master’s thesis at the local institute showed an ICC of 0.88 for foot tapping. Reaction test Multi choice reaction time of the lower limbs was measured by a computer based system of Fitronic (Fitronic Incorporation, Bratislava, Slovakia). Four contact switch pads were placed in the four corners inside a 35 cm square. With either the right or the left foot, players had to touch the pad as fast as possible in one of four directions in accordance with the location of a stimulus in one of the corners of the screen. The rationale for talent identification and information on measurement error (7.1%) were reported by Zemkova´ and Hamar (2004). In the present version of the test, 20 stimuli were given and the mean reaction time (to the nearest 1 ms) was used for statistical analysis. 2 kg medicine ball throw Upper limb power was examined using the 2 kg standing overhead medicine ball throw. A maximum of one step run-up was permitted and, referring to a soccer throw-in, both feet had to touch the ground. The distance of the throw was recorded to the nearest 1 cm.

Sit-and-reach Measuring whole-body flexibility in the sagittal plane, the sit-and-reach was conducted in a seated position with the legs straight out ahead and the fingers on a measuring line which was fixed on a box. With soles placed flat against the box and knees stretched, players were encouraged to reach as far as possible with their fingers. Jerky movements were not allowed and the extreme position should be held for at least 2 s (Mirkov, Kukolj, Ugarkovic, Koprivica, & Jaric, 2010). Positive values (to the nearest 1 cm) indicated that the players’ reaches extended over the toes. 20 m multi-stage endurance run Aerobic endurance was measured by a 20 m multistage endurance run. Players had to perform repetitive runs between two lines 20 m apart with the speed given by an audio signal. Each 3 min the following speeds had to be completed: 7.92 km h71; 9.72 km h71; 11.52 km h71; 13.32 km h71. At the end of each 3 min stage, a 90 s break was applied. Blood lactate concentrations were determined from capillary samples (0.02 ml) obtained from the ear lobe during the first 45 s of each break, which was within the recommended sampling time, between 15 to 45 s (Kass & Carpenter, 2009). Blood samples were further analysed using the Laktat Analyser Biosen 5040 (EKF Industrie-Elektronik, Barleben, Germany). Kohler and Boutellier (2004) reported a measuring accuracy of 98.5% at 12 mmol l71 of the Biosen 5040 analyser. Finally, the speed (0.01 km h71) corresponding to 4 mmol l71, calculated by an exponential function, served as an indicator of aerobic endurance. Data analysis Inferential statistics were obtained using PASW statistics version 18.0 (SPSS Inc., Chicago, USA). Initially, a principal component analysis with orthogonal rotation (varimax) was conducted on the 12 variables. Considering eigenvalues over Kaiser’s criterion of 1 (Field, 2009), three components, explaining 61.7% of the variance, were extracted. Based on the rotated factor loadings of the individual items (Table II), the 12 variables were categorised into the following three components: ‘speed’ (5, 10, 20 m sprint; 5 6 10 m shuttle sprint), ‘power and flexibility’ (drop jump, countermovement jump, medicine ball throw, sitand-reach), and ‘coordination and endurance’ (hurdles agility run, reaction test, foot tapping, endurance run). Principal component analyses within each age group separately displayed similar factor extractions.

Predicting career progression in youth soccer The differences between groups (i.e. drafted vs. non-drafted) were analysed using separate multivariate analysis of variance (MANOVA) for each component within each age group. Since the assumptions of homogeneity of variances and covariance matrices were violated, cases in the larger, non-drafted groups were randomly deleted to equalize the group size (Field, 2009). The remaining group sizes for further statistical proceedings were: n ¼ 205 (age 14 years), n ¼ 252 (15), n ¼ 228 (16), and n ¼ 136 (17) for each drafted and non-drafted group. The significance level for MANOVA was set at P 5 0.05. Univariate analyses of variance (ANOVAs) were conducted separately for each dependent variable to

Downloaded by [UQ Library] at 21:22 28 July 2016

Table II. Summary of principal component analysis results for the test battery (N ¼ 4733).

1677

follow up significant results of MANOVAs with future playing status as the between-participant variable. In order to improve the strength of the analysis and to control Type I error rate caused by multiple testing, Bonferroni correction was applied, setting the significance level at P 5 0.013 (Vincent, 2005). Furthermore, eta squared (Z2) was calculated according to the formulas for multivariate and univariate testing described in Sink and Stroh (2006). Finally, all 12 test variables were pooled together to determine which variable(s) best predicted group membership in each age group. Therefore, age specific stepwise discriminant analyses using Wilks’ lambda as selection criterion were calculated with future playing status as the dependent variable. Results

Rotated factor loadings

Item/variable

Speed

Power and flexibility

Coordination and endurance

5m 10 m 20 m SS CMJ DJ MBT SR HAR RT FT ER Eigenvalues % of variance

0.88 0.93 0.90 0.60 70.47 70.10 70.33 0.03 0.36 0.11 70.23 70.02 3.36 28.00

0.15 0.20 0.22 0.50 0.02 70.41 70.18 70.04 0.67 0.61 70.60 70.66 2.16 18.00

70.04 70.13 70.22 70.25 0.58 0.53 0.64 0.74 70.20 70.23 0.22 70.18 1.88 15.70

5/10/20 m ¼ 5/10/20 m sprint; SS ¼ 5 6 10 m shuttle sprint; CMJ countermovement jump; DJ ¼ drop jump; MBT ¼ 2 kg standing medicine ball throw; SR ¼ sit-and-reach; HAR ¼ hurdles agility run; RT ¼ reaction test; FT ¼ foot tapping; ER ¼ 20 m multi-stage endurance run.

The results of the MANOVAs are displayed in Table III. In general, future playing status significantly affected all three components at each age category (P 5 0.05). The effect sizes for ‘speed’ ranged from Z2 ¼ 0.08–0.09 and remained fairly stable from the ages of 14 to 17 years. For ‘power and flexibility’ and ‘coordination and endurance’, MANOVAs reported effect sizes of Z2 ¼ 0.11 and Z2 ¼ 0.07 at age 14 years and a slight decrease of effect sizes with increasing age. In 15- to 17-year-old players, effect sizes ranging from Z2 ¼ 0.06–0.08 for ‘power and flexibility’ and Z2 ¼ 0.05–0.06 for ‘coordination and endurance’ were demonstrated. Follow up ANOVAs at age 14 years presented no significant differences within future playing status for drop jump and sit-and-reach. All other measures displayed statistically significant superior performances for future national team players (P 5 0.013). The largest effects were reported in

Table III. Results of the MANOVAs within age group: differences by future playing status. P

(1–b)

405 499 451 267

0.000 0.000 0.000 0.000

0.087 0.086 0.079 0.083

1.000 1.000 1.000 0.985

4, 4, 4, 4,

405 499 451 267

0.000 0.000 0.000 0.000

0.113 0.067 0.060 0.081

1.000 0.999 0.996 0.983

4, 4, 4, 4,

405 499 451 267

0.000 0.000 0.000 0.015

0.073 0.056 0.050 0.045

0.998 0.997 0.985 0.813

Age

Wilks’ lambda

Speed

14 15 16 17

0.913 0.914 0.921 0.917

9.654 11.731 9.684 6.024

4, 4, 4, 4,

Power and flexibility

14 15 16 17

0.887 0.933 0.940 0.919

12.961 9.003 7.180 5.908

Coordination and endurance

14 15 16 17

0.927 0.944 0.950 0.955

8.012 7.462 5.942 3.129

Downloaded by [UQ Library] at 21:22 28 July 2016

1678

C. Gonaus & E. Mu¨ller

medicine ball throw, Z2 ¼ 0.11, shuttle sprint, Z2 ¼ 0.09, hurdles agility run, Z2 ¼ 0.06, and 10 and 20 m sprint, Z2 ¼ 0.04 (Table IV). ANOVAs at age 15 years displayed no significant differences for 5 m sprint, sit-and-reach and foot tapping. Drafted players performed significantly better in all other measurements with the largest effects in shuttle sprint, Z2 ¼ 0.08, medicine ball throw, Z2 ¼ 0.05, hurdles agility run, Z2 ¼ 0.04, and 10 and 20 m sprint, Z2 ¼ 0.04 (Table V). In 16-year-old players, follow up ANOVAs showed no significant difference within future playing status for sit-and-reach and foot tapping. In all other variables, drafted players performed significantly better than non-drafted ones. The largest effect sizes were found in shuttle sprint, Z2 ¼ 0.07, 20 m sprint, Z2 ¼ 0.05, medicine ball throw, Z2 ¼ 0.05, and 5 and 10 m sprint, Z2 ¼ 0.04 (Table VI). Finally, ANOVAs at age 17 years displayed significant superior performances for future national team players in shuttle sprint,

Z2 ¼ 0.07, medicine ball throw, Z2 ¼ 0.07, and 10 and 20 m sprint, Z2 ¼ 0.05 (Table VII). In 14-year-old players, stepwise discriminant analysis showed that the combination of medicine ball throw, shuttle sprint and hurdles agility run best discriminated future playing status, with 63.4% of players correctly classified. The medicine ball throw contributed most to group separation with a negative standardised discriminant function coefficient of 70.67 followed by the positive coefficients of 0.36 for shuttle sprint and 0.31 for hurdles agility run. The drafted players’ centroid was negative (70.42) compared with the positive non-drafted centroid (0.42). Results for 15- and 16-year-old players demonstrated that the combination of shuttle sprint, medicine ball throw and endurance run best discriminates between future playing status with 62.7% (15) and 63.6% (16) of correctly classified players. The standardised discriminant function coefficients were 70.68 (15) and 70.64 (16) for

Table IV. Results of the ANOVAs at age 14 years: means and standard deviations within future playing status (n ¼ 410).

5 m (s) 10 m (s) 20 m (s) SS (s) CMJ (cm) DJ (coeff.) MBT (m) SR (cm) HAR (s) RT (ms) FT (Hz) ER (km h

–1

)

Drafted

Non-drafted

1.06 + 0.06 1.82 + 0.08 3.16 + 0.13 11.53 + 0.43 35.8 + 5.5 7.12 + 1.97 9.6 + 1.4 10.9 + 6.0 11.71 + 0.66 598 + 65 13.0 + 1.1 12.14 + 0.79

1.08 + 0.08 1.86 + 0.10 3.21 + 0.15 11.80 + 0.46 34.1 + 5.5 6.63 + 2.25 8.6 + 1.4 9.6 + 6.1 12.04 + 0.69 623 + 82 12.6 + 1.2 11.90 + 0.91

11.239 15.961 18.403 38.313 8.993 5.587 49.209 4.270 23.980 11.410 7.574 8.282

1,408 1,408 1,408 1,408 1,408 1,408 1,408 1,408 1,408 1,408 1,408 1,408

0.001 0.000 0.000 0.000 0.003 0.019 0.000 0.039 0.000 0.001 0.006 0.004

0.027 0.038 0.043 0.086 0.022 0.014 0.108 0.010 0.056 0.027 0.018 0.020

Statistically significant at P 5 0.013; 5/10/20 m ¼ 5/10/20 m sprint; SS ¼ 5 6 10 m shuttle sprint; CMJ ¼ countermovement jump; DJ ¼ drop jump; MBT ¼ 2 kg standing medicine ball throw; SR ¼ sit-and-reach; HAR ¼ hurdles agility run; RT ¼ reaction test; FT ¼ foot tapping; ER ¼ 20 m multi-stage endurance run.

Table V. Results of the ANOVAs at age 15 years: means and standard deviations within future playing status (n ¼ 504).

5 m (s) 10 m (s) 20 m (s) SS (s) CMJ (cm) DJ (coeff.) MBT (m) SR (cm) HAR (s) RT (ms) FT (Hz) ER (km h

–1

)

Drafted

Non-drafted

1.04 + 0.07 1.78 + 0.08 3.07 + 0.11 11.27 + 0.41 38.8 + 5.4 7.95 + 2.03 10.4 + 1.3 12.4 + 6.0 11.37 + 0.62 562 + 58 13.5 + 1.1 12.36 + 0.75

1.05 + 0.06 1.81 + 0.08 3.12 + 0.12 11.51 + 0.41 36.5 + 5.6 7.39 + 2.03 9.8 + 1.3 11.2 + 5.8 11.62 + 0.67 584 + 76 13.2 + 1.2 12.16 + 0.74

6.131 18.126 23.255 42.362 14.283 9.506 27.422 4.766 18.907 13.173 6.070 9.192

1,502 1,502 1,502 1,502 1,502 1,502 1,502 1,502 1,502 1,502 1,502 1,502

0.014 0.000 0.000 0.000 0.000 0.002 0.000 0.029 0.000 0.000 0.014 0.003

0.012 0.035 0.044 0.078 0.028 0.019 0.052 0.009 0.036 0.026 0.012 0.018

Predicting career progression in youth soccer

1679

Table VI. Results of the ANOVAs at age 16 years: means and standard deviations within future playing status (n ¼ 456).

5 m (s) 10 m (s) 20 m (s) SS (s) CMJ (cm) DJ (coeff.) MBT (m) SR (cm) HAR (s) RT (ms) FT (Hz) ER (km h

)

Drafted

Non-drafted

1.01 + 0.06 1.75 + 0.07 3.02 + 0.11 11.14 + 0.36 39.3 + 5.7 8.69 + 2.15 11.2 + 1.4 13.3 + 5.8 11.23 + 0.58 543 + 53 13.9 + 1.1 12.40 + 0.74

1.04 + 0.07 1.78 + 0.07 3.07 + 0.11 11.35 + 0.40 37.7 + 5.7 8.05 + 2.27 10.6 + 1.4 12.4 + 5.9 11.43 + 0.67 561 + 61 13.6 + 1.3 12.18 + 0.81

18.547 18.532 23.918 33.638 9.496 9.599 21.744 3.022 12.421 10.586 4.741 9.546

df 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,

454 454 454 454 454 454 454 454 454 454 454 454

0.000 0.000 0.000 0.000 0.002 0.002 0.000 0.083 0.000 0.001 0.030 0.002

0.039 0.039 0.050 0.069 0.020 0.021 0.046 0.007 0.027 0.023 0.010 0.021

Downloaded by [UQ Library] at 21:22 28 July 2016

Table VII. Results of the ANOVAs at age 17 years: means and standard deviations within future playing status (n ¼ 272).

5 m (s) 10 m (s) 20 m (s) SS (s) CMJ (cm) DJ (coeff.) MBT (m) SR (cm) HAR (s) RT (ms) FT (Hz) ER (km h

)

Drafted

Non-drafted

1.00 + 0.06 1.74 + 0.07 2.99 + 0.10 11.07 + 0.35 40.2 + 5.5 8.56 + 1.88 11.6 + 1.4 12.6 + 7.0 11.10 + 0.47 543 + 62 13.8 + 1.1 12.44 + 0.68

1.03 + 0.06 1.77 + 0.07 3.03 + 0.11 11.28 + 0.40 39.0 + 5.7 8.36 + 2.39 10.8 + 1.3 12.5 + 6.2 11.30 + 0.62 558 + 58 13.8 + 1.3 12.37 + 0.88

9.120 12.827 13.613 21.391 3.356 .581 19.878 .008 8.793 4.272 .017 .591

df 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,

270 270 270 270 270 270 270 270 270 270 270 270

0.003 0.000 0.000 0.000 0.068 0.447 0.000 0.930 0.003 0.040 0.897 0.443

0.033 0.045 0.048 0.073 0.012 0.002 0.069 0.000 0.032 0.016 0.000 0.002

shuttle sprint, 0.49 (15) and 0.55 (16) for medicine ball throw, and 0.28 (15) and 0.38 (16) for endurance run with the corresponding group centroids of 0.35 for drafted, and 70.35 for non-drafted players. Finally, in 17-year-old players, the combination of shuttle sprint, medicine ball throw and foot tapping best discriminated between the two groups with 66.2% of correctly classified players. According to the appropriate standardised discriminant function coefficients, shuttle sprint (0.75) and medicine ball throw (70.68) contributed more than foot tapping (0.37) to group separation. The corresponding group centroids were 70.39 for drafted, and 70.39 for non-drafted players (Table VIII). Discussion We compared the physiological characteristics of Austrian youth soccer academy players between 14 to 17 years of age. The superior performance of players, who had subsequently been drafted to play

in a youth national team, was demonstrated on several physiological measures across all age categories. In previous research, physiological performance measures were reported to differentiate between elite versus non-elite (Vaeyens et al., 2006) and future professional versus amateur level players (Le Gall et al., 2010). Of particular interest remains the question of whether physiological characteristics also distinguish an already preselected, homogeneous group of advanced soccer players. Franks et al. (1999) found no significant physical and physiological differences between subsequent successful and unsuccessful academy players. However, the present findings reveal a trend towards superior fitness performance in subsequent drafted versus nondrafted players, corresponding to some extent to the findings of professional versus international players reported by Le Gall et al. (2010). While differences between elite and non-elite players may already occur in early adolescence, distinguishing

1680

C. Gonaus & E. Mu¨ller Table VIII. Summary of stepwise discriminant analyses by age group: variables entered/removed*. Wilks’ lambda Exact F

Age

Step

Entered

Statistic

df1

df2

df3

Statistic

df1

df2

1 2 3 1 2 3 1 2 3 1 2 3

MBT SS HAR SS MBT ER SS MBT ER SS MBT FT

0.892 0.862 0.852 0.922 0.899 0.891 0.931 0.905 0.891 0.927 0.879 0.865

1 2 3 1 2 3 1 2 3 1 2 3

1 1 1 1 1 1 1 1 1 1 1 1

408 408 408 502 502 502 454 454 454 270 270 270

49.209 32.712 23.533 42.362 28.145 20.344 33.638 23.839 18.490 21.391 18.514 13.932

1 2 3 1 2 3 1 2 3 1 2 3

408 407 406 502 501 500 454 453 452 270 269 268

0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000

Downloaded by [UQ Library] at 21:22 28 July 2016

At each step, the variable that minimizes the overall Wilks’ lambda is entered. * Maximum number of steps is 24; minimum partial F to enter is 3.84; maximum partial F to remove is 2.71; F level, tolerance, or VIN insufficient for further computation. MBT ¼ 2 kg standing medicine ball throw; SS ¼ 5 6 10 m shuttle sprint; HAR ¼ hurdles agility run; ER ¼ 20 m multi-stage endurance run; FT ¼ foot tapping.

sub-elite and elite players would probably become more apparent in later stages of development (Vaeyens et al., 2006). Current data from previously selected academy players indicated superior fitness performances in favour of future drafted players compared to their non-drafted counterparts (Table III). All in all, the extent of any significant differences in fitness testing diminished in older age groups. Whereas 10 out of 12 variables showed significantly superior fitness performance in favour of drafted players at age 14 years, only 6 out of 12 test items showed statistically significant differences by future playing status at age 17 years (Table IV–VII). The possibility exists that multiple selection procedures in pre-adolescence and systematic training during adolescence may result in a ‘physically’ more homogeneous group of players in late adolescence. Thus, the differentiating potential of fitness characteristics may decrease with age, indicating that in late adolescence, when the latematuring players caught up with the early-maturing players, other aspects such as psychological matters or technical and tactical skills would probably become more powerful in distinguishing between future drafted and non-drafted players (Gil et al., 2007; Ro¨sch et al., 2000; Williams & Reilly, 2000). Discriminant analyses showed that 62.7 to 66.2% of players could be correctly classified by a combination of only three variables across all age categories. The dynamic behaviour of discriminating characteristics during adolescence (Vaeyens et al., 2006) could only partially be confirmed with the present results, since medicine ball throw and shuttle sprint remained very stable in discriminating, by a substantial amount, between drafted and non-drafted players over all age levels. However, the third highest discriminating variables changed within age

category: whereas agility was an important discriminating factor at age 14 years, endurance capacity gained weight at age 15–16 years, and finally, foot tapping discriminated at age 17 years. The discriminating power of shuttle sprint across all age categories, as well as hurdles agility run at age 14 years was not surprising given the crucial impact of speed and agility on talent indication (Mirkov et al., 2010; Reilly, Williams, et al., 2000) and on adult soccer playing level (Kaplan, Erkmen, & Taskin, 2009). Both the intermittent pattern of activity in soccer, with changes in activity approximately every four seconds (Rienzi, Drust, Reilly, Carter, & Martin, 2000), as well as the essential impact of speed and anaerobic activities on matchwinning situations in senior soccer (Reilly, Bangsbo, & Franks, 2000) underline the importance of speed and agility measurements in terms of the selection process in youth soccer. Similar suggestions were made by Hoare and Warr (2000) and Gil et al. (2007), who recommended the importance of speed and acceleration when analysing the outcomes of talent selection tests in youth soccer. The present results would, to some extent, be a suitable fit into the ‘window of accelerated adaptation to motor coordination’ in the early teens (Balyi & Hamilton, 2004). In this critical period during the development of performance capacity through childhood and adolescence, enhancements of muscle strength, speed and power were attributed to an improved coordination of muscle activity (Viru et al., 1999). Like shuttle sprint, the medicine ball throw substantially discriminated players by future playing status across all age levels. In general, advantages in strength and power of the upper limbs may help to resist the physical aspects of the game, such as pushing and pulling activities, as well as to withstand

Downloaded by [UQ Library] at 21:22 28 July 2016

Predicting career progression in youth soccer being pushed and pulled (Bloomfield, Polman, & O’Donoghue, 2007; Reilly, Bangsbo, & Franks, 2000). The discriminating power at age 14 years was highest and decreased slightly at later age levels, yet still was remarkably apparent at age 17 years. Gil et al. (2007) showed that 14-year-old players who were selected to play in next year’s team, of 15-yearolds, were taller and heavier than their non-selected counterparts. Furthermore, the impact of body size and stage of puberty on speed and explosive power measures were greater than on soccer-specific skills in players aged 13 to 15 years (Malina et al., 2005; Malina, Eisenmann, Cumming, Ribeiro, & Aroso, 2004). Admittedly, body size and weight data were measured in the present study; however, several of these data were lost over the study period, which is undoubtedly a limitation. Analysing these data with notably smaller sample sizes, n ¼ 67 for age 14, n ¼ 86 (15), n ¼ 100 (16) and n ¼ 52 (17) per group, revealed statistically significant smaller and lighter non-drafted players compared with the drafted ones. Thus, differences in power of upper limbs could possibly be explained by benefits in physical maturity, as demonstrated by Jones, Hitchen, and Stratton (2000) who highlighted the large influence of sexual maturity on upper and lower body strength measures in 11- to 16-year-old boys. The results of the present study agree with those of Vaeyens et al. (2006), revealing that endurance capacity becomes more important in discriminating future playing status at later stages of adolescence, potentially because of an enhanced trainability of aerobic endurance during adolescence (Viru et al., 1999). Additionally, Buchheit, Mendez-Villanueva, Simpson, and Bourdon (2010) observed a trend of increasing total running distance from U13 to U17 level in young elite soccer academy players. Due to increases in fat free mass and haemoglobin content, development of the cardiovascular system, and hormonal changes (Philippaerts et al., 2006), there was a ‘window of accelerated adaptation to aerobic and strength training’ (Balyi & Hamilton, 2004) between the ages of 12 to 16 years. Moreover, according to Roescher et al. (2010), the age of 15 years seemed to be crucial with respect to reaching professional level in soccer. The progressively larger gap from age 15 years on between those two groups was assigned to both soccer-specific and additional training. In summary, several differences in fitness characteristics were found between future drafted and nondrafted youth soccer players at ages 14 to 17 years. Soccer-specific speed and power of upper limbs most discriminated future playing status, independently of age level. Furthermore, at age 14 years, agility discriminates between the two groups, whereas endurance performance becomes more crucial at later

1681

stages of adolescence. Consequently, trainers and coaches can consider integrating those discriminating tests into both the talent identification and development process at the specific age level. It appears that measuring fitness characteristics can provide useful information in terms of predicting future career progression (Le Gall et al., 2010; Reilly, Bangsbo, & Franks, 2000); however, using these measures exclusively would probably not be sensitive enough to compare an already selected, homogeneous group of soccer players (Williams & Reilly, 2000). For this latter purpose, a multidimensional approach with assistant information on physical and psychological characteristics, as well as soccer-specific technical and tactical skills would be preferable. Acknowledgements This work was facilitated by cooperation between the ¨ FB), the DepartAustrian Football Association (O ment of Sport Science of the University of Salzburg and the Institute for Sports Medicine and Science (IMSB Austria). In addition, the authors would like to thank all the players and staff of the Austrian youth soccer academies for their participation.

References Balyi, I., & Hamilton, A. (2004). Long-term athlete development: trainability in childhood and adolescence. Windows of opportunity. Optimal trainability. Victoria: National Coaching Institute British Columbia & Advanced Training and Performance Ltd. Bloomfield, J., Polman, R., & O’Donoghue, P. (2007). Physical demands of different positions in FA Premier League soccer. Journal of Sports Science and Medicine, 6, 63–70. Buchheit, M., Mendez-Villanueva, A., Simpson, B. M., & Bourdon, P. C. (2010). Match running performance and fitness in youth soccer. International Journal of Sports Medicine, 31, 818–825. Carling, C., Le Gall, F., Reilly, T., & Williams, A. M. (2009). Do anthropometric and fitness characteristics vary according to birth date distribution in elite youth academy soccer players? Scandinavian Journal of Medicine & Science in Sports, 19(1), 3–9. Coelho, E. S. M. J., Figueiredo, A. J., Simoes, F., Seabra, A., Natal, A., Vaeyens, R., Philippaerts, R. M., Cumming, S. P., & Malina, R. M. (2010). Discrimination of u-14 soccer players by level and position. International Journal of Sports Medicine, 31, 790–796. Drust, B., Waterhouse, J., Atkinson, G., Edwards, B., & Reilly, T. (2005). Circadian rhythms in sports performance – an update. Chronobiology International, 22(1), 21–44. Field, A. (2009). Discovering statistics using SPSS (3rd ed.). Los Angeles: Sage Publications Inc. Figueiredo, A. J., Goncalves, C. E., Coelho, E. S. M. J., & Malina, R. M. (2009). Characteristics of youth soccer players who drop out, persist or move up. Journal of Sports Sciences, 27, 883–891. Franks, A. M., Williams, A. M., Reilly, T., & Nevill, A. (1999). Talent identification in elite youth soccer players: Physical and physiological chracteristics. Journal of Sports Sciences, 17, 812. Gil, S., Ruiz, F., Irazusta, A., Gil, J., & Irazusta, J. (2007). Selection of young soccer players in terms of anthropometric and physiological factors. Journal of Sports Medicine and Physical Fitness, 47(1), 25–32.

Downloaded by [UQ Library] at 21:22 28 July 2016

1682

C. Gonaus & E. Mu¨ller

Haj-Sassi, R., Dardouri, W., Gharbi, Z., Chaouachi, A., Mansour, H., Rabhi, A., & Mahfoudhi, M. E. (2011). Reliability and validity of a new repeated agility test as a measure of anaerobic and explosive power. Journal of Strength and Conditioning Research, 25, 472–480. Hoare, D. G., & Warr, C. R. (2000). Talent identification and women’s soccer: an Australian experience. Journal of Sports Sciences, 18, 751–758. Huijgen, B. C., Elferink-Gemser, M. T., Post, W. J., & Visscher, C. (2009). Soccer skill development in professionals. International Journal of Sports Medicine, 30, 585–591. Jaggers, J. R., Swank, A. M., Frost, K. L., & Lee, C. D. (2008). The acute effects of dynamic and ballistic stretching on vertical jump height, force, and power. Journal of Strength and Conditioning Research, 22, 1844–1849. Jones, M. A., Hitchen, P. J., & Stratton, G. (2000). The importance of considering biological maturity when assessing physical fitness measures in girls and boys aged 10 to 16 years. Annals of Human Biology, 27(1), 57–65. Kannekens, R., Elferink-Gemser, M. T., & Visscher, C. (2010). Positioning and deciding: key factors for talent development in soccer. Scandinavian Journal of Medicine & Science in Sports, 1–7. Kaplan, T., Erkmen, N., & Taskin, H. (2009). The evaluation of the running speed and agility performance in professional and amateur soccer players. Journal of Strength and Conditioning Research, 23, 774–778. Kass, L., & Carpenter, R. (2009). The effect of sampling time on blood lactate concentration ([Bla]) in trained rowers. International Journal of Sports Physiology and Performance, 4, 218–228. Kohler, G., & Boutellier, U. (2004). Bestimmung der Genauigkeit von Blutlaktatmessgera¨ten am Beispiel des Laktatanalyser Biosen 5040. Schweizerische Zeitschrift fu¨r Sportmedizin und Sporttraumatologie, 52, 149–153. Le Gall, F., Carling, C., Williams, M., & Reilly, T. (2010). Anthropometric and fitness characteristics of international, professional and amateur male graduate soccer players from an elite youth academy. Journal of Science and Medicine in Sport, 13(1), 90–95. Malina, R. M., Cumming, S. P., Kontos, A. P., Eisenmann, J. C., Ribeiro, B., & Aroso, J. (2005). Maturity-associated variation in sport-specific skills of youth soccer players aged 13-15 years. Journal of Sports Sciences, 23, 515–522. Malina, R. M., Eisenmann, J. C., Cumming, S. P., Ribeiro, B., & Aroso, J. (2004). Maturity-associated variation in the growth and functional capacities of youth football (soccer) players 13-15 years. European Journal of Applied Physiology, 91, 555–562. Malina, R. M., Pena Reyes, M. E., Eisenmann, J. C., Horta, L., Rodrigues, J., & Miller, R. (2000). Height, mass and skeletal maturity of elite Portuguese soccer players aged 11–16 years. Journal of Sports Sciences, 18, 685–693. Mirkov, D. M., Kukolj, M., Ugarkovic, D., Koprivica, V. J., & Jaric, S. (2010). Development of anthropometric and physical performance profiles of young elite male soccer players: a longitudinal study. Journal of Strength and Conditioning Research, 24, 2677–2682.

Musch, J., & Grondin, S. (2001). Unequal competition as an impediment to personal development: a review of the relative age effect in sport. Developmental Review, 21, 147–167. Philippaerts, R. M., Vaeyens, R., Janssens, M., Van Renterghem, B., Matthys, D., Craen, R., Bourgois, J., Vrijens, J., Beunen, G., & Malina, R. M. (2006). The relationship between peak height velocity and physical performance in youth soccer players. Journal of Sports Sciences, 24, 221–230. Read, M. M., & Cisar, C. (2001). The influence of varied rest interval lengths on depth jump performance. Journal of Strength and Conditioning Research, 15, 279–283. Reilly, T., Bangsbo, J., & Franks, A. (2000). Anthropometric and physiological predispositions for elite soccer. Journal of Sports Sciences, 18, 669–683. Reilly, T., Williams, A. M., Nevill, A., & Franks, A. (2000). A multidisciplinary approach to talent identification in soccer. Journal of Sports Sciences, 18, 695–702. Rienzi, E., Drust, B., Reilly, T., Carter, J. E., & Martin, A. (2000). Investigation of anthropometric and work-rate profiles of elite South American international soccer players. Journal of Sports Medicine and Physical Fitness, 40, 162–169. Roescher, C. R., Elferink-Gemser, M. T., Huijgen, B. C., & Visscher, C. (2010). Soccer endurance development in professionals. International Journal of Sports Medicine, 31, 174–179. Ro¨sch, D., Hodgson, R., Peterson, T. L., Graf-Baumann, T., Junge, A., Chomiak, J., & Dvorak, J. (2000). Assessment and evaluation of football performance. The American Journal of Sports Medicine, 28(5 Suppl), S29–39. Rumpf, M. C., Cronin, J. B., Oliver, J. L., & Hughes, M. (2011). Assessing youth sprint ability-methodological issues, reliability and performance data. Pediatric Exercise Science, 23, 442–467. Sink, C. A., & Stroh, H. R. (2006). Practical significance: the use of effect sizes in school counseling research. Professional School Counseling, 9, 401–411. Stratton, G., Reilly, T., Williams, A. M., & Richardson, D. (2004). Youth soccer: from science to performance. London: Taylor & Francis. Svensson, M., & Drust, B. (2005). Testing soccer players. Journal of Sports Sciences, 23, 601–618. Vaeyens, R., Malina, R. M., Janssens, M., Van Renterghem, B., Bourgois, J., Vrijens, J., & Philippaerts, R. M. (2006). A multidisciplinary selection model for youth soccer: the Ghent Youth Soccer Project. British Journal of Sports Medicine, 40, 928–934. Verheijen, R. (1998). The complete handbook of conditioning for soccer. Michigan: Data Reproductions. Vincent, W. J. (2005). Statistics in kinesiology (3rd ed.). Champaign: Human Kinetics. Viru, A., Loko, J., Harro, M., Volver, A., Laaneots, L., & Viru, M. (1999). Critical periods in the development of performance capacity during childhood and adolescence. European Journal of Physical Education, 4(1), 75–119. Williams, A. M., & Reilly, T. (2000). Talent identification and development in soccer. [Review]. Journal of Sports Sciences, 18, 657–667. Zemkova´, E., & Hamar, D. (2004). The agility test in functional diagnostics of athletes. Acta Universitatis Palackianae Olomucensis, Facultas Culturae Physicae. Gymnica, 34, 61–66.