Journal of Human Kinetics volume 25 2010, 93‐100 93 Section III – Sport, Physical Education & Recreation Ball Possession Strategies in Elite Soccer According to the Evolution of the Match‐Score: the Influence of Situational Variables by Carlos Lago‐Peñas1, Alexandre Dellal2 In soccer, the ability to retain possession of the ball for prolonged periods of time has been suggested to be linked to success. The accuracy of this assertion was investigated by examining 380 matches involving Spanish League First Division teams during the 2008‐2009 season. Possession of the ball, according to the status of the match (win‐ ning, drawing and losing), was recorded during the different matches using a multiple‐camera match analysis sys‐ tem (Gecasport®). The results suggest that the best classified teams maintained a higher percentage of ball posses‐ sion and that their pattern of play was more stable. The coefficient of variation, with respect to ball possession per match, was smaller for the best placed teams. Indeed, first placed F.C. Barcelona had the smallest coefficient of varia‐ tion for possession time (8.4%), while bottom placed Recreativo showed the highest values with 17.1%. Linear re‐ gression analysis showed that possession strategies were influenced by situation variables. Team possession was greater when losing than when winning (p<0.01) or drawing (p<0.01), home teams enjoyed greater possession than visiting teams (p<0.01), and playing against strong opposition was associated with a reduction in time spent in pos‐ session (p<0.01). The findings indicate that strategies in soccer are influenced by situational variables and that teams alter their playing style accordingly during the match. Key words: match analysis; possession strategies; soccer; team performance; tactical component area of great interest in performance analysis Introduction (Hughes and Bartlett, 2002). In soccer, for a goal to be scored a team usually Performance analysis refers to the objective re‐ has to have possession of the ball. Although it might cording and examination of behavioral events that be anticipated that longer periods of possession occur during sporting competition (Carling et al., Sport, Physical Education & Recreation should predict goal scoring, support for this notion 2005; Dellal, et al., 2010). The main aim of analyzing is divided. Bate (1988), for example, found that the one’s own team’s performance is to identify higher number of possessions a team had, the strengths that can be further developed, and weak‐ greater the chance they had of entering the attacking nesses that might be improved. Understanding the third of the field and consequently more goal scoring differences between the patterns of play developed opportunities were created. On the basis of this by successful and unsuccessful teams, as well as finding, Bate (1988) rejected the notion of possession those of the same team in different matches, is an soccer and advocated a direct strategy. However, Hughes and Franks (2005), Grant et al. (1999), Hook 1 - Faculty of Education and Sports Sciences, University of Vigo, Pontevedra, Spain 2 - University of Sports Sciences, Strasbourg, France and National Center of Medicine and Science in Sport, Tunis, Tunisia Authors submitted their contribution of the article to the editorial board. Accepted for pinting in Journal of Human Kinetics vol. 25/2010 on July 2010. 94 Ball Possession Strategies in Elite Soccer According to the Evolution of the Match‐Score: the Influence of Situational Variables and Hughes (2001) and Bloomfield et al. (2005a) re‐ tion found in this study and those of Sasaki et al. ported that successful teams (e.g., European Cham‐ (1999) and Tucker et al. (2005) emphasize the need to pions League, World Champions, Europa Cup) validate the models developed across numerous ad‐ maintained possession for longer than unsuccessful ditional teams. Secondly, Lago and Martin (2007) did teams. In contrast, Stanhope (2001) found that time not incorporate into their study the quality of oppo‐ in possession of the ball was not indicative of success sition as an independent variable to explain the de‐ in the 1994 World Cup. terminants of possession of the ball in soccer. Many of the equivocal findings of previous stud‐ In this context, the aim of this investigation was ies examining the possession strategies of successful to provide a large‐scale study of elite professional and unsuccessful teams may originate from con‐ soccer teams and examine the effects of situational ceptual and methodological issues. For example, variables on ball possession strategies. It was hy‐ Jones et al. (2004) indicate that many have failed to pothesized that ball possession was influenced by demonstrate the reliability of the data gathering such variables as match location, match status, qual‐ system used. In addition, selecting matches played ity of opposition and by the level of the team. On the as part of individual tournaments means that the basis of these findings, it is hoped that the informa‐ chosen teams (successful and unsuccessful) are im‐ tion will be of value in contributing to more tactical balanced in term of the strength of opposition and knowledge for the prescription of specific exercises number of matches played. Such factors are likely to within the training regimen and for analyzing match influence a team´s performance and might explain performance. the discrepancies seen among studies. According to Taylor et al. (2008), effective Materials and Methods evaluation of soccer performance requires knowl‐ edge of the contextual factors that can potentially af‐ Study Design fect behaviour incidence and outcomes (Carling et al., 2005; Dellal, 2008). Existing literature (Jones et al., Although it might be anticipated that longer pe‐ 2004; Shaw and O´Donoghue, 2004) suggests that the riods of possession should predict increased goal variables match location (i.e., playing at home or scoring opportunities, support for this notion is di‐ Sport, Physical Education & Recreation away) (Tucker et al., 2005), match status (i.e., vided. To verify this, all 380 matches of the Spanish whether the team was winning, losing or drawing) soccer League played throughout the 2008‐2009 sea‐ (Bloomfield et al., 2005a, 2005b; Lago and Martin, son were assessed. The dependent variable was the 2007; Taylor et al., 2008; Tucker et al, 2005) and the proportion of time (%) during matches in which the quality of the opposition (strong or weak) (Taylor et team had possession of the ball when the ball was in al., 2008), which all require consideration when play. Empirical evidence suggests that the variables evaluating soccer performances. Unfortunately, match location, match status (winning, drawing and these findings are still inconclusive given that most losing) and the quality of the opponent can affect of these studies were based on small sample sizes, soccer performances. These factors were included in and with the exception of Lago and Martin (2007) the study as independent variables. and Taylor et al. (2008), the existing performance analysis literature has examined situation factors in‐ Match Sample dependently, thereby neglecting to account for the The examined sample consisted of 380 Spanish complex and dynamic nature of soccer performance Soccer League First Division matches played (MacGarry and Franks, 2003; Reed and O´Donoghue, throughout the 2008‐2009 season. The performance 2005; Taylor et al., 2008). of one team obviously impacts upon the second (i.e., However, the results of Lago and Martin (2007) the frequency and duration of possession is depend‐ and Taylor et al. (2008) have two limitations in their ent on the opposition). As a consequence, data were findings. First, Taylor et al. (2008) adopted a fine‐ analyzed using one team from each match. The grained approach to soccer analysis by considering number of observations was therefore 380. The col‐ the performances of a single team over a sustained lected data (possession of the ball, especially ac‐ period (two seasons). An obvious limitation of case cording to the match status: winning, drawing and studies designs is that generalization of findings is losing during a match) were provided by a multiple‐ precluded. The contradictory effects of match loca‐ camera match analysis system (Gecasport®), a pri‐ Journal of Human Kinetics volume 25 2010, http://www.johk.awf.katowice.pl by C. Lago‐Peñas and A. Dellal 95 vate company dedicated to assessing the perform‐ of time each team was winning (Minutes Winning: ance of teams in the Spanish Soccer League. The ac‐ MW), drawing (Minutes Drawing: MD) and losing curacy of the Gecasport® System has been verified (Minutes Losing: ML) during a match were included. by Gómez et al. (2009). Written permission to analyze This means that the panel match status in the regres‐ data was provided by Gecasport®. Ethics approval sion model presents two coefficients from the com‐ for all experimental procedures was granted by our parison of Drawing to Losing and from the compari‐ institute’s Human Research Ethics Committee. son of Winning to Losing. Match location was re‐ corded as “home” or “away”, depending on whether Reliability Testing or not the sampled team was playing on its own ground or that of its opponent: 0 = playing at home, Reliability of the data was assessed through inter‐ 1= playing away (Match Location: ML). Quality of and intra‐ observer test procedures. Inter‐observer opposition was the distance in the end‐of‐season‐ reliability was assessed by the authors, coding five ranking between competing teams (Quality of Op‐ matches randomly selected, with data being com‐ position: QO). The 20 teams were divided into four pared with those provided by Gecasport®. Intra‐ob‐ groups according to their final league ranking server reliability was completed by the authors, (Team: TE). Group 1 contained the top five teams, coding five random matches selected from the data Group 2 contained those that finished 6th through sample. Following a six‐week period, to avoid any 10th, and Group 3, those that finished 11th through possible negative learning effects, the matches were 15th, and Group 4, those that finished 16th through recoded and the two data sets compared.