Full Terms & Conditions of access and use can be found athttp://www.tandfonline.com/action/journalInformation?journalCode=rjsp20

Download by: [Staffordshire University] Date: 07 March 2016, At: 02:52

Journal of Sports Sciences

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

The evaluation of small-sided games as a talentidentification tool in highly trained prepubertalsoccer players

Jonathan S J Fenner, John Iga & Viswanath Unnithan

To cite this article: Jonathan S J Fenner, John Iga & Viswanath Unnithan (2016): The evaluationof small-sided games as a talent identification tool in highly trained prepubertal soccer players,Journal of Sports Sciences, DOI: 10.1080/02640414.2016.1149602

To link to this article: http://dx.doi.org/10.1080/02640414.2016.1149602

Published online: 03 Mar 2016.

Submit your article to this journal

Article views: 53

View related articles

View Crossmark data

The evaluation of small-sided games as a talent identification tool in highly trainedprepubertal soccer playersJonathan S J Fenner a,b, John Igac,d and Viswanath Unnithanb

aMedical and Exercise Science Department; Wolverhampton Wanderers Football Club, Wolverhampton, UK; bCentre for Sport, Health and ExerciseResearch, Staffordshire University, Stoke-on-Trent, UK; cFootball Development, The Football Association, London, UK; dResearch Institute for Sportand Exercise Science, Liverpool John Moores University, Liverpool, UK

ABSTRACTThe aim of this study was to evaluate physiological and technical attributes of prepubertal soccerplayers during multiple small-sided games (SSGs), and determine if SSGs can act as a talent identifica-tion tool. Sixteen highly trained U10 soccer players participated and separated into two groups of eight.Each group played six small-sided (4 vs. 4) matches of 5-min duration. Each player was awarded totalpoints for the match result and goals scored. A game technical scoring chart was used to rate eachplayer’s performance during each game. Time-motion characteristics were measured using microme-chanical devices. Total points had a very large significant relationship with game technical scoring chart(r = 0.758, P < 0.001). High-speed running distance had a significantly large correlation with gametechnical scoring chart (r = 0.547, P < 0.05). Total distance covered had a significant and moderatecorrelation with game technical scoring chart (r = 0.545, P < 0.05) and total points (r = 0.438, P < 0.05).The results demonstrated a large agreement between the highest-rated players and success in multipleSSGs, possibly due to higher-rated players covering larger distances in total and at high speed.Consequently, multiple SSG could be used to identify the more talented prepubertal soccer players.

ARTICLE HISTORYAccepted 28 January 2016

KEYWORDSSmall-sided games; youthsoccer; identifying talent

Introduction

The majority of professional soccer clubs run their own youthacademy, in an attempt to identify and nurture talent from anearly age, with an expectation that players will progress to thesenior team (Reilly & Gilbourne, 2003). Identifying talent,therefore, is a priority within soccer clubs, as it would equateto a more effective financial investment by focusing resourceson the development of a small but talented pool of players.The identification of talented players is historically based on acoach or scouts’ subjective view of a player (Unnithan, White,Georgiou, Iga, & Drust, 2012). This evaluation technique isoften used in isolation and can result in repeated misjudge-ments (Meylan, Cronin, Oliver, & Hughes, 2010). This systemunfortunately is associated with low predictive value andtherefore its usefulness has been criticised (Vaeyens, Lenoir,Williams, & Philippaerts, 2008).

In recent years, cross-sectional models have been used topredict future success in adult competition by measuring thecurrent performance of adolescent players (Vaeyens et al.,2008) on a combination of anthropometric, physiological andsoccer-specific skills in isolation within a testing battery(Gonaus & Müller, 2012: Reilly, Bangsbo, & Franks, 2000:Vaeyens et al., 2006). A cross-sectional design is based onthe assumption that important characteristics of successfuladult performance can be extrapolated from identification of

talented young players; however the required characteristicsmay not necessarily be retained throughout maturation(Vaeyens et al., 2008). A further limitation of this design istalent development is unstable and non-linear due to differ-ences in growth, development and training, therefore, makinga long-term prediction unreliable, particularly in prepubertalparticipants (Vaeyens et al., 2008). These models predictivevalue are problematic in team sports, as excellence in sportssuch as soccer are not idiosyncratic to a standard set of skills;as superior ability can be achieved in individual and uniqueways through different combinations of technical skills andphysical attributes (Vaeyens et al., 2008). In order to identifytalent within soccer, an approach that allows simultaneousevaluation of a combination of technical skills and physicalattributes that contribute to successful soccer performance isessential (Unnithan et al., 2012), the use of small-sided games(SSGs) may play a crucial role in future talent identificationmodels.

The majority of the research pertaining to SSGs has focusedon their use as a training modality as it is suggested that SSGreplicates movement demands, physiological intensity andtechnical demands of competitive match play (Hill-Haas,Dawson, Impellizzeri, & Coutts, 2011). An appropriate gamedesign can alter the physiological and technical outcomes ofan SSG (Hill-Haas et al., 2011), through manipulating factorssuch as player number (Hill-Haas, Dawson, Coutts, & Rowsell,

CONTACT Jonathan Fenner Jonathanfenner@wolves.co.uk Wolverhampton Wanderers Football Club, Sir Jack Hayward Training Ground, Douglas TurnerWay, Wolverhampton, West Midlands, WV3 9BF, United KingdomLaboratory where the research was conducted:Centre for Sport, Health and Exercise Research, Staffordshire University, Stoke-On-Trent, Staffordshire, ST4 2DE

JOURNAL OF SPORTS SCIENCES, 2016http://dx.doi.org/10.1080/02640414.2016.1149602

© 2016 Taylor & Francis

Dow

nloa

ded

by [

Staf

ford

shir

e U

nive

rsity

] at

02:

52 0

7 M

arch

201

6

2009; Little & Williams, 2006, 2007), coach encouragement(Rampinini et al., 2007), continuous and intermittent training(Hill-Haas et al., 2009), and pitch size (Dellal et al., 2011; Kelly &Drust, 2009). Technical analysis of SSGs revealed that therewas significantly more ball contacts per individual (P < 0.05)when compared to larger-sided games (9 vs. 9) (Owen, Wong,McKenna, & Dellal, 2011). This increase in ball interaction aswell as the close replication of the physical demands of com-petitive match play (Jones & Drust, 2007) demonstrates SSGsecological validity as a surrogate for successful competitivesoccer performance and, therefore, as a potentially effectiveplatform to identify talented soccer players.

SSGs have recently been used within a talent identificationmodel using post-pubertal elite soccer players and a smallrelationship that was not significant (r = 0.39, P > 0.05) wasfound between success in SSG and coaches subjective ratingof a player (Unnithan et al., 2012). Unnithan et al. (2012)concluded that the elite status of the sample led to homo-geneity of their technical skills and thus reduced the predic-tive power of SSGs. It may be postulated that if this model wasapplied to a sample of elite soccer players that had not beenexposed to as many years of systematic training, multiple SSGscould be a useful tool within a talent identification model.Furthermore, recent literature in talent identification hasassessed adolescent soccer players, whereas sport specialisa-tion occurs early in soccer (Vaeyens et al., 2008). Consequently,talent identification models may need to be applied prior tothe onset of puberty. Therefore, the aim of this study was toevaluate the physiological loading and technical attributes ofhighly trained prepubertal soccer players during multiple SSGsand to determine whether SSGs can act as a talent identifica-tion tool within youth soccer players.

Methods

Participants

Sixteen highly trained U10 soccer players (Mean ± SD; age,10.6 ± 0.3 years; stature 1.45 ± 0.07 m; body mass37.2 ± 4.1 kg) were recruited from a youth soccer academyin England. Participants were considered physically activebased upon a physical activity questionnaire (7 ± 1.6 h aweek), and had three sessions of technical and tactical traininga week (8 ± 0.8 h a week) with one competitive match a week.Players had trained at the club for 2 ± 0.4 years and weretraining for 45 weeks a year. Prior to the start of the study, theacademy manger, coaches, players and parents of the playershad an opportunity to discuss all aspects of the study with theprimary investigator. Physical activity questionnaires, writteninformed consent provided by the players’ parents, writtenassent from the players, medical questionnaires and a trainingdiary were all obtained prior to the study. Ethical approval wasobtained from a local ethics university committee. The studywas also approved by the academy manager at the club.

SSG protocol

The SSGs were played outdoors on natural turf which is thesquad’s usual training venue at the squad’s allocated training

time of 18:00. Two (2 m × 1 m) goals with no identified goal-keepers were used. Goals could only be scored in the attackinghalf. A standardised 15-min warm-up was prescribed prior toparticipating in the SSGs. A multi-ball system was applied (i.e.,balls were placed around the perimeter of the pitch) so that thegame was continuous. No verbal encouragement or feedbackwas allowed from coaches throughout the session. Only refereedecisions and the score was provided during the game.

Participants were separated into two groups of eight. Withintheir groups, two teams of four players were made. Each teamplayed six, four versus four matches, which were 5 min induration, with 3 min of passive recovery, on a pitch 18.3 m x23 m in dimension. The players were reorganised into differentcombinations after each game within their group of eightplayers, so no player played with the same three teammateson two occasions. All players had been habituated to SSGs, asthey were often used in their regular training sessions. Thispitch size, game duration and player number was selected asit is the players’ normal SSG methodology for SSGs in training attheir club. A pilot study demonstrated that the current pitchsize and game duration in comparison to a larger pitch size andshorter game durations did not change the technical outcomesof the SSGs and provided a consistent physiological stimulus.Test–retest reliability measured over six SSGs was assessed andproduced similar physiological responses in most of the time-motion characteristics (typical error <9%) with the exception ofhigh-speed running distance in the last three games of theprotocol, which demonstrated greater variability with a typicalerror ranging from 16% to 69%.

Game technical scoring chart (GTSC) and total points (TP)protocol

During each individual game of the SSG protocol, each playerwas awarded total points (TP) for the outcome of each match,4 points for a win, 2 points for a draw and 0 points for a loss.Players were also awarded one point each for their teamscoring regardless of result. Two coaches with F.A. qualifica-tions (level 1) technically evaluated the players throughout thegames using a game technical scoring chart (GTSC) and boththe coaches had previous experience using the GTSC. TheGTSC is a tool that mimics the perception of a coach orscout when they are identifying talented players or making adecision on whether to retain or release a player. All players’performance was evaluated on 10 football elements and weregiven a score between 0 and 5. Each point described theplayers’ performance using the following criteria: 1 – poor, 2– below average, 3 – average, 4 – very good and 5 – excellent.The criteria in the GTSC were: Cover/Support, Communication,Decision-making, Passing, First touch, Control, One versusOne, Shooting, Assist and Marking. Thus, during a game, if aplayer was perceived by the coach rating them to be a poorpasser of the ball during that SSG, then they would be given ascore of 1 in the passing element.

The inter-tester reliability of the GTSC has previously beenestablished through a pilot study that found no significantdifferences (P > 0.05) between two coaches (F.A. level 1) ratingthe same players across three games. Reliability of GTSC hasbeen established in literature previously and yields similar

2 J. S. J. FENNER ET AL.

Dow

nloa

ded

by [

Staf

ford

shir

e U

nive

rsity

] at

02:

52 0

7 M

arch

201

6

results to the present study, with an inter-tester reliability of0.83 and 0.782 for Cronbachs alpha between a highly qualifiedcoach (F.A. B licence) and two research assistants with coach-ing qualifications (F.A. level 1) (Unnithan et al., 2012).Construct validity of the GTSC was examined in a pilot study.An FA qualified (F.A. B licence) coach assessed players usingthe GTSC via video feedback, in comparison to a coach asses-sing the same players live in the field during three SSGs. Theresults demonstrated significant (P < 0.05) inter-observer dif-ferences in technical skills for only one out of the three games.Content validity was established via a panel of highly qualifiedand experienced coaches (F.A. A licence), detailing the impor-tant performance elements that needed to be scored during asoccer performance to accurately identify talented players.

Time-motion analysis

The micromechanical device (MEMS; MinimaxX S4, CatapultInnovations, Melbourne, Australia) contains a 10 Hz globalpositioning satellite (GPS) chip that was used to record time-motion data. Total distance covered (TDC, meters) and high-speed running distance covered (HSRD, meters) were used asmeasures of locomotor activities. High-speed running distancecovered was defined as any distance covered above 60% ofthe individual player’s maximum velocity attained during theSSG; this threshold was used in accordance with Harley et al.(2010). These researchers concluded categorising playermotion into speed zones should be normalised relative tothe individual’s speed capabilities, specifically, in youngplayers. Data was included if the number of satellitesexceeded 6 and there was a horizontal displacement of posi-tioning (HDOP) that was less than 1.5. The number of satelliteswas 7 and HDOP was 1.02 which met these requirements.

The MinimaxX S4 contains a triaxial piezoelectric linearaccelerometer (Kionix: KXP94) sampling at a frequency of100 Hz. PlayerLoad™ (anterior-posterior PlayerLoad, medial-lateral PlayerLoad and vertical PlayerLoad) was recordedusing this accelerometer. PlayerLoad™ is expressed in arbitraryunits (au). PlayerLoad™ per meter was a ratio of PlayerLoad™divided by total distance covered; this was used to account forinter-individual variation in distance covered within games.PlayerLoad™ (PL) alongside the locomotor activities was ana-lysed using the Catapult software (Sprint 5. 9. 2, CatapultSports, Melbourne, Australia).

Physiological testing

Physiological testing was conducted two weeks post SSG pro-tocol as part of the clubs periodic testing battery. Testing wascompleted at the same time of day as the SSG to control forcircadian variation in performance (Chtourou et al., 2012). Thetesting was conducted on a 3G artificial turf pitch, in an indoorfacility to ensure weather conditions did not affect the testingbattery. The players followed a standardised warm up lasting15 min prior to testing. The players then performed the testsin the following order: countermovement jump (CMJ) withoutthe use of their arms, 10 m sprint and 30 m sprint.Unfortunately, one player could not attend the testing sessiondue to illness.

Jumps

Lower body power was determined with a CMJ without theuse of the arms. All the jumps were recorded using a jumpmat (Jump Mat, Perform Better, Warwick, UK). Each player hadthree attempts with 2-min recovery in between each attempt.The highest score was recorded and used for analysis.

Countermovement jump protocol

The players performed the CMJ without the use of arms bystanding on the jump mat and placed their hands on their hips.Following a countdown, the players then went down as fastand as deep as felt natural to them and jumped up as high aspossible, landing on the jump mat with a cushioned landing.

Speed protocol

The players sprinting ability was measured with a maximal 30-metre straight-line sprint with 10 m and 30 m split times.Infrared timing gates (Brower Timing System, Utah, USA) wereplaced approximately at hip height above the ground, at 0, 10and 30 m. Players started in a standing position, 1 m behind thefirst timing gate. Each player was instructed to run as fast asthey could to a cone placed 5 m beyond the last timing gate.This was done to ensure the players were maximally sprintingwhen passing the last gate rather than slowing down. Eachplayer performed three sprints with 2-min rest in-between eachsprint. The fastest time was recorded and used for analysis.

Statistical analysis

The data is reported as mean and standard deviation (mean ±SD). Prior to conducting any statistical analyses, assumptions ofnormality were checked via Kolmogorov–Smirnov test and noviolations of normality were found in any testing variables.Homoscedasticity was also checked through P-P plots, linearitywas identified in the data. To assess if there were any differ-ences between the winning and losing teams in the SSGs,Mann-Whitney tests were used, due to the small sample size(4 vs. 4), to test for differences between total distance covered,high-speed running distance, PlayerLoad™ and PlayerLoad™ permeter between the winning and losing teams for each match.No differences (P > 0.05) were found between the winning andlosing teams in any of the time-motion variables in every game.Therefore, the individual player’s load per game was pooled toprovide a total load for the entire SSG session. A Pearsoncorrelation coefficient test was used to examine the relationshipbetween GTSC and TP and between GTSC, TP, fitness testsresults and the total load for all six games for each of thetime-motion variables. Magnitudes for thresholds were set at0.1 for small, 0.3 for moderate, 0.5 for large, 0.7 for very largeand 0.9 for extremely large correlation coefficients (Hopkins,Marshall, Batterham, & Hanin, 2009). Statistical significancewas set at P < 0.05. All statistical analyses were performedusing SPSS version 21 (SPSS Inc., Chicago, IL).

JOURNAL OF SPORTS SCIENCES 3

Dow

nloa

ded

by [

Staf

ford

shir

e U

nive

rsity

] at

02:

52 0

7 M

arch

201

6

Results

Technical evaluation

Figure 1 illustrates the results from the technical evaluations(GTSC) along with each individual’s TP for the six SSGs. Therewas a significant and large relationship between TP and GTSC(r = 0.758, P < 0.001, R2 = 57%).

Time-motion analysis

Tables 1 and 2 present the mean and standard deviationbetween the winning and losing teams for each game forthe external physiological variables. GTSC had a significantand large correlation with high-speed running distance(r = 0.547, P < 0.05, R2 = 30%) (Figure 2). Total distancecovered had a significant and moderate correlation withGTSC (r = 0.545, P < 0.05, R2 = 21%) (Figure 3) and TP(r = 0.438, P < 0.05, R2 = 19%) (Figure 4).

Physiological testing

No physiological testing variable was correlated to either GTSCor TP.

Discussion

The aim of this study was to evaluate if multiple SSGs couldact as a talent identification model in highly trained prepu-bertal soccer players. The main finding of this study was that avery large relationship between GTSC and TP was identified(r = 0.758, P < 0.001). The evidence suggests that it is possibleto identify the most talented player, according to coach’ssubjective scoring system, simply by examining if they wonthe most number of games in a multiple SSG model.Successful performances in soccer are as a result of multiplefactors interacting (Unnithan et al., 2012), the most talented

5

10

15

20

25

30

35

40

45

50

120 140 160 180 200 220 240

To

ta

l P

oin

ts (

TP

) (

Au

)

Game Technical Scoring Chart (GTSC) (Au)

r = 0.758, P < 0.001, R2= 57%

Figure 1. Relationship between total points and game technical scoring chart.

Table1.

Group

1tim

e-motioncharacteristicsacross

allsixmatches.

Gam

e1

Gam

e2

Gam

e3

Gam

e4

Gam

e5

Gam

e6

AB

AB

AB

AB

AB

AB

Match

Score

13

04

03

33

03

22

TDC(m

)359±21

373±36

359±60

341±22

346±47

336±44

294±57

334±35

339±54

380±44

345±47

348±43

HSRD(m

)22

±15

56±23

23±15

27±15

46±14

19±18

7±7

16±5

12±6

21±11

14±11

7±5

PlayerLoad™(Au)

56±10

69±7

55±10

66±16

57±11

52±10

49±11

62±10

58±10

66±16

56±12

58±5

PlayerLoad™permetre

(Au)

0.16

±0.03

0.18

±0.02

0.15

±0.02

0.20

±0.06

0.17

±0.02

0.15

±0.02

0.16

±0.01

0.19

±0.02

0.17

±0.03

0.17

±0.04

0.16

±0.03

0.17

±0.01

4 J. S. J. FENNER ET AL.

Dow

nloa

ded

by [

Staf

ford

shir

e U

nive

rsity

] at

02:

52 0

7 M

arch

201

6

players in this study were those that scored the higher resultin TP from six SSG. To differentiate between what made theseplayers more successful than others we examined multiplefactors; physiological fitness, time-motion characteristics andtechnical characteristics, to examine if there was a particularfactor that contributed to their success.Ta

ble2.

Group

2tim

e-motioncharacteristicsacross

allsixmatches.

Gam

e1

Gam

e2

Gam

e3

Gam

e4

Gam

e5

Gam

e6

AB

AB

AB

AB

AB

AB

Match

Score

34

51

42

51

31

01

TDC(m

)310±48

393±40

342±44

345±29

339±37

374±88

310±37

356±36

380±26

337±38

393±41

352±36

HSRD(m

)26

±10

24±7

23±9

25±8

31±13

27±10

31±19

40±37

30±5

34±16

28±15

30±23

PlayerLoad™(Au)

51±4

58±11

53±3

51±6

51±5

52±13

47±8

54±6

51±8

50±8

53±7

54±6

PlayerLoad™permetre

(Au)

0.17

±0.02

0.15

±0.02

0.16

±0.01

0.15

±0.01

0.15

±0.01

0.14

±0.00

0.15

±0.02

0.15

±0.01

0.13

±0.01

0.15

±0.02

0.14

±0.01

0.15

±0.00

120

140

160

180

200

220

240

20 70 120 170 220 270 320

Ga

me

Te

ch

nic

al

Scro

ing

Ca

hrt (

GT

SC

) (

Au

)

High Speed Running Distance (HSRD) (m)

r = 0.547,P < 0.05 R2 = 30%

Figure 2. Relationship between game technical scoring chart and high-speedrunning distance.

120

140

160

180

200

220

240

1500 1700 1900 2100 2300 2500

Ga

me

Te

ch

nic

al

Sco

rin

g C

ha

rt (

GT

SC

) (

Au

)

Total Distance Covered (TDC) (m)

r = 0.545, P < 0.05, R2 = 21%

Figure 3. Relationship between game technical scoring chart and total distancecovered.

JOURNAL OF SPORTS SCIENCES 5

Dow

nloa

ded

by [

Staf

ford

shir

e U

nive

rsity

] at

02:

52 0

7 M

arch

201

6

A player’s technical ability is the predominant componentof successful soccer players, as it can consistently determinethe difference between non-elite, sub-elite and elite players(Meylan et al., 2010). Therefore, the use of a competitive SSGformat, which would induce a high amount of technicalactions, should discriminate between talented and lesstalented players. The present study evaluated technical abil-ity through the use of GTSC which demonstrated a significantand very large relationship with success in multiple SSGs viathe accumulation of TP (r = 0.758, P < 0.001). This datademonstrates that the more talented players, as identifiedby the GTSC, emerge within multiple SSGs simply by winninggames and 57% of the variability in TP is explained by thetechnical skill level of the players as measured by the GTSC.This ability to win regardless of team combination indicatesthat these players have superior technical ability and are ableto work effectively within a team to ensure a victory. Thesefindings are in agreement with previous research (Unnithanet al., 2012) that technical skill level is a strong characteristicthat distinguishes between the more talented soccer players.Unnithan et al. (2012) demonstrated a moderate relationshipbetween GTSC and TP in a group of highly trained adolescentsoccer players, however, it was not significant (r = 0.39,P > 0.05). These researchers theorised that due to the num-ber of years that this cohort had been involved in systematictraining, a relative homogeneity of their technical ability hadoccurred; thereby, reducing the predictive power of the SSGs.The present study used a cohort of players that had beenexposed to minimal systematic training whilst still beingconsidered highly trained; this, therefore, has resulted in alarger relationship between TP and GTSC and, therefore,increasing the predictive power of SSG and has establishedthat SSGs potentially, could act as a talent identificationmodel.

The present study pooled the data across the six games, asthere was no significant difference (P > 0.05) between thewinning and losing teams, and analysed it as a total load forall six games. High-speed running distance was found to be adetermining factor in the coaches’ scoring a player with ahigher GTSC score, as high-speed running distance had alarge correlation which was significant with GTSC (r = 0.547,P < 0.05). This demonstrates that players that had exhibitedgreater high-speed movements also displayed better technicalskills. The R2 value demonstrates that high-speed runningdistance explains 30% of the variability technical level of theplayer, as measured by the GTSC. A recent study, however,reported no difference between a retained and released groupfor high-intensity running during a 11-a-side match play forU11’s (pitch dimension 78.7 x 54.1 m) (Goto, Morris, & Nevill,2015). These researchers found that the retained group spenta lower proportion of the match in the speed zone of 0.0–1.5 m·s−1 and during standing and walking compared to thereleased group. These two findings suggest that the retainedgroup could cover the required amount of high-intensity run-ning for match play with less recovery time than the releasedgroup. Consequently, it would appear that talented playershave the ability to recover quickly from bouts of high-speedrunning (Goto et al., 2015); this ability has been demonstratedin the present study, with the more talented players able tocover more distances at greater speeds than their less talentedpeers with the same recovery period between games.

The present study also demonstrated that the moretalented players covered greater distance than their lesstalented peers as total distance covered was significant andmoderately correlated to both GTSC (r = 0.545, P < 0.05) andTP (r = 0.438, P < 0.05) with 21% and 19% of the variabilityshared with total distance covered, respectively. A recentstudy demonstrated similar findings as, a retained group ofplayers covered significantly (P < 0.05) greater total distanceduring a match than a released group of players in U11competitive matches (Goto et al., 2015). Goto et al. (2015)suggest that total distance covered may help to identifyplayers that could progress through an academy, as withtotal distance covered is one factor that potentially coachesmay knowingly or unknowingly consider when retaining orreleasing players. Similar findings to the present paper arefound throughout literature up to senior level that indicatestotal distance covered during match play is shown to distin-guish between standards of soccer players (Mohr, Krustrup, &Bangsbo, 2003). The present study findings suggest that totaldistance covered is a possible factor in identifying talentedplayers.

Physiological characteristics have been seen as importantdiscriminating factors between elite and sub-elite players, par-ticular running speed in younger players (U13 and U14), andaerobic endurance, power, and flexibility in older players (U15and U16) (Gonaus & Müller, 2012). The present study demon-strated that there were no significant correlations betweenany of the physiological variables and either the GTSC or TP.This suggests that power and speed attributes may not havean effect at the prepubertal age group in identifying talentedplayers. As players start to enter puberty, this finding changesas previous literature found that future drafted players were

0

5

10

15

20

25

30

35

40

45

50

1500 1700 1900 2100 2300 2500

To

ta

l P

oin

ts (

TP

) (

Au

)

Total Distance Covered (TDC) (m)

r = 0.438, P < 0.05, R2 = 19%

Figure 4. Relationship between total points and total distance covered.

6 J. S. J. FENNER ET AL.

Dow

nloa

ded

by [

Staf

ford

shir

e U

nive

rsity

] at

02:

52 0

7 M

arch

201

6

significantly more powerful in a CMJ than future non-draftedplayers at 14-years old age (P < 0.05) (Gonaus & Müller, 2012).This suggests that chronologically older and more matureplayers may gain advantage in selection process during matur-ity and, therefore, get more opportunities to become an elitesoccer player; however, it is evident from the present study’sfindings that in prepubertal soccer players technical skill leveland not any physical attributes may be the determining factorin becoming identified as talented.

Some limitations of the present study exist firstly; samplesize, an increase in the participant population may haveincreased the strength of the associations. This, however,was not possible due to the squad size available. Secondly,the results of this study are commensurate with the design ofthe SSG if there are changes to this design this could in turnnot yield the same results as a small change in SSG designcould have an effect on SSG intensity and, therefore, theplayer’s ability to demonstrate superior skill. Thirdly, whilethe SSG model in the present study demonstrates the poten-tial to accurately identify talented players simply by measuringthe number of games won, longitudinal studies are requiredto establish if the successful players at this age continue to besuccessful in their future playing career.

Conclusion

This is the first study to evaluate prepubertal players’ talentpotential in multiple SSG, based simply on the win to lossratio. Our results demonstrate that the more talented playersare more successful during SSG regardless of their team com-binations. These more talented players were also the sameplayers rated highly by coaches and played the SSGs at ahigher speed and covered greater distances than the lesstalented peers. SSGs, therefore, could be used within a talentidentification model as they provide an environment in whichtalented players can showcase their superior ability simply bywinning games.

Practical applications

This study demonstrates the potential usefulness’s of SSGswithin a talent identification model, coaches, scouts andclubs could utilise the model described in this paper to helpidentify talented players. This model has the potential to beused on a large-scale talented identification day. Whereby,there are large number of players, which would normally betoo difficult for a coach or scout to sufficiently observe aplayer for long enough to identify if they are talented or not.This model allows the scout or coach to specifically look at theplayers with the highest TP and, therefore, narrow their searchand make an efficient use of their time. Furthermore, thismodel could also be utilised within a club as a marker oftechnical and tactical proficiency in the same vein clubs per-iodically assess players fitness they could periodically assessplayers talent to establish if the same player throughout theyear and subsequent years is continually the best player intheir group and, therefore, provide that player with theresources to become successful. Further research needs toassess the longitudinal variability in players’ technical

performance as well as time-motion characteristics to provideevidence as to the stability of the data being collected.

Acknowledgements

We would like to thank the players who participated in this study. Wewould also like to thank Tom Hulme, Aaron Sharpes, Jonathan Piper andChris West for their assistance in data collection. The input and guidancefrom Professor Barry Drust (Liverpool John Moores University), and Mr.Steve Barrett (Hull City FC) has been greatly appreciated. Lastly, we wouldlike to thank Wolverhampton Wanderers Football Club for their coopera-tion and support.

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

Jonathan S J Fenner http://orcid.org/0000-0002-7356-593X

References

Chtourou, H., Hammouda, O., Souissi, H., Chamari, K., Chaouachi, A., &Souissi, N. (2012). Diurnal variations in physical performances related tofootball in young soccer players. Asian Journal of Sports Medicine, 3(3),139–144. doi:10.5812/asjsm

Dellal, A., Chamari, K., Owen, A. L., Wong, D. P., Lago-Penas, C., & Hill-Haas,S. (2011). Influence of technical instructions on the physiological andphysical demands of small-sided soccer games. European Journal ofSport Science, 11, 341–346. doi:10.1080/17461391.2010.521584

Gonaus, C., & Müller, E. (2012). Using physiological data to predict futurecareer progression in 14- to 17 – year – old Austrian soccer academyplayers. Journal of Sports Sciences, 30, 1673–1682. doi:10.1080/02640414.2012.713980

Goto, H., Morris, J. G., & Nevill, M. E. (2015). Motion analysis of U11 to U16elite english premier league academy players. Journal of Sport Science,33, 1–11. doi:10.1080/02640414.2014.999700

Harley, J. A., Barnes, C. A., Portas, M., Lovell, R., Barrett, S., Paul, D., &Weston, M. (2010). Motion analysis of match-play in elite U12 to U16age-group soccer players. Journal of Sports Sciences, 28, 1391–1397.doi:10.1080/02640414.2010.510142

Hill-Haas, S. V., Dawson, B., Coutts, A., & Rowsell, G. (2009). Physiologicalresponses and time-motion characteristics of various small-sided soccergames in youth players. Journal of Sports Sciences, 27, 1–8. doi:10.1080/02640410802206857

Hill-Haas, S. V., Dawson, B., Impellizzeri, F. M., & Coutts, A. J. (2011).Physiology of small-sided games training in football: A systematicreview. Sports Medicine, 011-1642/11/0003-0199. 41, 199–220.doi:10.2165/11539740-000000000-00000

Hopkins, W. G., Marshall, S. W., Batterham, A. M., & Hanin, J. (2009).Progressive statistics for studies in sports medicine and exercisescience. Medicine & Science in Sports & Exercise, 41, 3–12. doi:10.1249/MSS.0b013e31818cb278

Jones, S., & Drust, B. (2007). Physiological and technical demands of 4 v 4and 8 v 8 in elite youth soccer players. Kinesiology, 39, 150–156.doi:796.332:576.8.095

Kelly, D. M., & Drust, B. (2009). The effect of pitch dimensions on heart rateresponses and technical demands of small-sided soccer games in eliteplayers. Journal of Science and Medicine in Sport, 12, 475–479.doi:10.1016/j.jsams.2008.01.010

Little, T., & Williams, A. G. (2006). Suitability of soccer training drills forendurance training. Journal of Strength and Conditioning Research, 20,316–319. Retrieved from SPORTDiscus

Little, T., & Williams, A. G. (2007). Measure of exercise intensity duringsoccer training drills with professional soccer players. Journal of

JOURNAL OF SPORTS SCIENCES 7

Dow

nloa

ded

by [

Staf

ford

shir

e U

nive

rsity

] at

02:

52 0

7 M

arch

201

6

Strength and Conditioning Research, 21, 367–371. Retrieved fromSPORTDiscus

Meylan, C., Cronin, J., Oliver, J., & Hughes, M. (2010). Reviews: Talentidentification in soccer: The role of maturity status on physical, physio-logical and technical characteristics. International Journal of SportsScience and Coaching, 5, 571–592. doi:10.1260/1747-9541.5.4.571

Mohr, M., Krustrup, P., & Bangsbo, J. (2003). Match performance of high-standard soccer players with special reference to development of fatigue.Journal of Sports Sciences, 21, 519–528. doi:10.1080/0264041031000071182

Owen, A. L., Wong, D. P., McKenna, M., & Dellal, A. (2011). Heart rateresponses and technical comparison between small- vs. large-sidedgames in elite professional soccer. Journal of Strength and ConditioningResearch, 25, 2104–2110. doi:10.1519/JSC.0b013e3181f0a8a3

Rampinini, E., Impellizzeri, F. M., Castagna, C., Abt, G., Chamari, K., Sassi, A.,& Marcora, S. M. (2007). Factors influencing physiological responses tosmall-sided soccer games. Journal of Sports Sciences, 25, 659–666.doi:10.1080/02640410600811858

Reilly, T., Bangsbo, J., & Franks, A. (2000). Anthropometric and physiologi-cal predispositions for elite soccer. Journal of Sports Sciences, 18, 669–683. doi:10.1080/02640410050120050

Reilly, T., & Gilbourne, D. (2003). Science and football: A review of appliedresearch in the football codes. Journal of Sports Sciences, 21, 693–705.doi:10.1080/0264041031000102105

Unnithan, V., White, J., Georgiou, A., Iga, J., & Drust, B. (2012). Talentidentification in youth soccer. Journal of Sports Sciences, 30, 1719–1726. doi:10.1080/02640414.2012.731515

Vaeyens, R., Lenoir, M., Williams, A. M., & Philippaerts, R. M. (2008). Talentidentification and development programmes in sport: Current modelsand future directions. Sports Medicine, 38(9), 703–714. doi:10.2165/00007256-200838090-00001

Vaeyens, R., Malina, R. M., Janssens, M., Van Renterghem, B., Bourgois, J.,Vrijens, J., . . . Philippaerts, R. M. (2006). A multidisciplinary selectionmodel for youth soccer: The Ghent Youth Soccer Project. British Journalof Sports Medicine, 40, 928–934. doi:10.1136/bjsm.2006.029652

8 J. S. J. FENNER ET AL.

Dow

nloa

ded

by [

Staf

ford

shir

e U

nive

rsity

] at

02:

52 0

7 M

arch

201

6