Part I
Match analysis 1
Changing the ratio of area per player during small-sided soccer games shapes performance
M. S. Azlia,b, K. Davidsc,d, R. Duartee, D. AraĂşjoe, C. Buttonf and A. Shielda
aSchool of Exercise and Nutrition Science, Queensland University of Technology, Australia
bSports Centre, University of Malaya, Malaysia
cCentre for Sports Engineering Research, Sheffield Hallam University, United Kingdom
dFiDiPro Programme, University of Jyväskylä, Finland
eCIPER , Faculdade de Motricidade Humana, Universidade de Lisboa, Portugal
fSchool of Physical Education, University of Otago, New Zealand
Introduction
The role of small-sided and conditioned games (SSCGs) for learning and performance enhancement in team sports like soccer has been widely investigated in the last decade (e.g. Davids, AraĂşjo, Correia, & Vilar, 2013; Hill-Haas, Dawson, Impellizzeri, & Coutts, 2011). However, current investigations of SSCGs have been restricted to the study of game actions and conditioning effects. For instance, decreasing the number of players in SSCGs typically increases physiological demands and the number of individual game actions performed. Previous studies have demonstrated that altering the ratio of area per player or number of players on the pitch can enhance the performance intensity of players (Hill-Haas, Rowsell, Dawson, & Coutts, 2009). This increasing trend of using SSCGs in order to achieve desirable physiological adaptations suggests that SSCGs might also provide benefits for technical (e.g. ball control) and tactical (e.g. space management) skill development (Owen, Twist, & Ford, 2004). Therefore, research is needed on how to use SSCGs to create complex practice environments which provide opportunities for players to improve and adapt their performance behaviours.
Performance analysis in team sports has emphasized the use of notational and time-motion analysis as methods to examine movement patterns during match-play and practice settings (Frencken & Lemmink, 2008). However, these methodological approaches have mainly focused on a playerâs motion when in possession of the ball and/or on the activity pattern imposed on individual players (e.g. player running distances and speeds). Moreover, in these approaches team performance is operationally conceived as the sum of the individual playerâs performances or frequency of actions. However, Duarte, AraĂşjo, Correia and Davids (2012) argued that studying team behaviours as a whole needs a synthesising approach as teams are more than the sum of all individual playersâ performances. Indeed, tactical performance indicators in soccer must be able to reveal important aspects such as the playersâ interactions, pace, fitness and movement patterns (Hughes & Bartlett, 2001), which implies methods suitable to capture how the on-field playersâ interpersonal interactions create certain patterns of play at the team level.
In the past years, theoretical and experimental evidence has emphasized the need for a complex systems approach to investigating team sports behaviours (Davids, AraĂşjo, & Shuttleworth, 2005; McGarry, Anderson, Wallace, Hughes, & Franks, 2002). The complex systems approach has been revealed as useful for understanding spatial-temporal patterns in both individual and team sports (Davids et al., 2013; McGarry et al., 2002). For example, recent studies have employed team dispersion variables to capture complex group behaviours that express emergent collective performance patterns in team sports (Duarte et al., 2013; Folgado, Lemmink, Frencken, & Sampaio, 2012; Frencken, Lemmink, Delleman, & Visscher, 2011). For example, Duarte et al. (2013) investigated changes in the complexity of the stretch/contraction shape patterns of two professional teams over the course of a competitive match. For that, authors used variables like the surface area, geometrical centre, stretch index, length and width of the teams. That study showed that teams displayed different absolute values expressing their idiosyncratic behavioural tendencies, though both teams became more regular and predictable (i.e. less complex) throughout the match. Despite these data there is an absence of evidence concerning the impact of changing task constraints during practice. Theoretically, through manipulating key task constraints in training, players can be encouraged to explore the performance environment for effective coordination solutions (Davids, AraĂşjo, Seifert, & Orth, 2015).
Therefore, the aim of this study was to investigate the influence on team dispersion behaviours of changing the ratio of area per player in SSCGs during attacking and defending game phases. This study was expected to provide empirical data for coaches and sport practitioners to aid understanding of the tactical effects of manipulating task constraints (i.e. number of players/pitch area) in SSCGs.
Methods
Twenty-two male, developing U14 soccer players (M = 13.36 years, SD = 0.49) from a Portuguese Soccer Club in Lisbon voluntarily agreed to participate in the study. They were intermediate club level participants with between three to four years playing and training experience (undertaking four 60-min practice sessions with one competitive game per week). We avoided experienced soccer players (i.e. national or regional players) who may have developed idiosyncratic behaviours, or inexperienced players who could not successfully perform basic actions (e.g. passing or dribbling). All players and parents were informed about the procedures of this study and provided informed consent to participate, according to the guidelines of the American Psychological Association. The study was approved by a university ethics committee.
Data collection
Participants were randomly assigned by their coaches into two formats of SSCGs (GK+4v4+GK and GK+5v5+GK). Data collection was performed on two different days for each format. Participants played on the same field dimensions (46 Ă 33m pitch) for GK+5v5+GK with individual playing area values of 150m2 (Hill-Haas, Dawson, Coutts, & Rowsell, 2009) and 189m2 for GK+4v4+GK. The session took place on an artificial turf soccer pitch with junior goalpost size (6m Ă 2m). Before performing, all players were asked to do a warm-up of ten minutes, similar to normal training sessions. Two SSCGs of four minutes each were performed with a passive recovery period of three minutes between each. Players were also encouraged to rehydrate during breaks. This procedure was carried out prior to the clubsâ normal training sessions on weekdays. The SSCGs were performed with four support players, one coach and one assistant coach positioned outside the playing area to provide a new ball every time the one in play left the playing area, to allow for continuity of play. Minimal encouragement and primary instruction was given to players to maintain collective possession of the ball and try to score goal.
Data analysis
Raw individual positional data were collected using global positioning system (GPS) devices (SPI Pro, GPSports, Canberra, Australia) with a sampling frequency rate of 15 Hz to capture player movement displacement during play (Silva et al., 2014). The pitch area used for this study was calibrated using four of the GPS devices positioned at the corners of the pitch for approximately two minutes. The longitude and latitude coordinates time series data that were retrieved from GPS devices were converted into 2D Cartesian coordinates using a Universal Transverse Mercator (UTM) system. Data re-sampling using the interpolation method was employed for the missing data gaps to obtain the same length of the time series data for all players within each trial (n = 6 in GK+4v4+GK and n = 6 in GK+5v5+GK). The positional data were converted into metres and then the rotation matrix process was applied for the alignment of playing pitch with x-axis (length) and y-axis (width). Next, this process was applied to all player positional data for alignment with the pitch as the bottom left corners of the pitch were allocated a coordinate value of zero (see Folgado, Duarte, Fernandes, & Sampaio, 2014). All calculations routines were developed and processed in MATLABÂŽ R2012a (The MatchWorks Inc., Natick, MA, USA).
Four team spatial and two speed dispersion variables were measured during the SSCGs using data from all the outfield players of each team (excluding goalkeeper) and then each trial was selected as the values of the sequences of play leading to shots on goal (i.e. the moment the ball entered or missed the goal or the defensive team blocked/intercepted after the shot on goal). Team variables were: i) team width, calculated as the teamâs maximum and minimum position values of x and y in lateral player movements on the pitch in each time frame; ii) team length, calculated as the teamâs maximum and minimum position values of x and y in longitudinal player movements; iii) surface area, calculated as a polygon (occupied/covered area) by linking each outfield playersâ position in each pattern in each time frame; iv) stretch index, calculated as the mean vectorial distance of each player to the geometrical centre (i.e. mean dispersion value of the players around the centre of each team); v) contraction-expansion speed (CES), calculated as the speed of approaching/deviation from the geometrical centre: and vi) team average speed, calculated as the average of the individual speeds of the players. These variables have already been proposed in previous investigations to capture expert team behaviours in soccer (Duarte et al., 2013; Folgado et al., 2012). All dependent variables were analysed using a two-way mixed-design ANOVA. Inferential statistics were performed using the non-parametric Mann-Whitney U test. All statistical analyses were conducted in IBMÂŽ SPSSÂŽ Statistics Version 21.0 for Windows (IBM Corp., New York, USA).
Results
Table 1.1 presents mean (Âą SD) data from spatial and speed variables according to the game phases and game formats. Analysis of the magnitude of variation (SD values) revealed a significant effect of the game phase in width direction (U = 32.00, z = â2.309, p = 0.20, r = â0.67/Éł2 = 0.509). In terms of game formats, the GK+5v5+GK format displayed a higher magnitude of variation (SD values) in the covered areas than the GK+4v4+GK format (U = 23.00, z = â2.829, p = < 0.05, r = â0.82/Éł2 = 0.859). This is probably due to the higher number of players and small ratio per player. However, we did not observe any differences in other spatial nor team speed variables.
Table 1.1 Mean and standard deviation data for the game phases and game formats
Discussion and conclusion
Overall, our results from spatial variables during attacking and defending situations revealed that players from both formats adjusted their behaviours to achieve the performance objectives of either scoring a goal or defending. According to Frencken et al. (2011), during attacking and defending situations, the attackers increasing distances between players by open-up spaces (the attackersâ surface area is large) while the defenders decreasing the distances between players by close-up spaces (the defendersâ surface area is small). However, our study showed different results where the defending teamâs movement tended to change their teamâs width more often and deeply in order to defend their goal compared to the attacking team. This finding enhanced performance variability of the defenders in exploring lateral movement by opening- and closing-up spaces of the pitch. This behaviour probably emerged as an effort to remain closer to an attacker in order to block or intercept the ball. Vilar, AraĂşjo, Davids, Correia and Esteves (2013) suggested that the defendersâ actions to decrease the distances between them and attackers when approaching the goal is to prevent passing or shooting opportunities for the attackers. In contrast, the attacking team might adopt a direct style of play by performing short and fast passes to their teammates in order to destabilize or break the defending teamâs symmetry and achieve their performance goals.
Figure 1.1 Exemplar of a time series of surface area for both formats during attacking and defending situations near the goal post
A...