How dots behave in two different pitch sizes? Analysis of tactical behavior based on position data in two soccer field sizes

¿Cómo se comportan los puntos en dos campos diferentes? Análisis del comportamiento táctico basado en los datos de posición en dos tamaños de campo de fútbol

This work is licensed under Creative Commons Attribution-NonCommercial-NoDerivs 3.0 International.

Team sports such as soccer carry a high level of uncertainty and time pressure due to the permanent relationships between players and team’s cooperation-opposition dynamics (Gréhaigne, Bouthier, & David, 1997; Serra-Olivares, García-López, & Calderón, 2016). This kind of sport contexts involve a variety of changing environments and a wide diversity of game problems which need to be solved (Davids, Button, Araújo, Renshaw, & Hristovski, 2006). It deals with continuous challenges to the players and teams, who need to develop different abilities to adapt their behaviours to each specific situation, for achieving sport performance (Araújo, Davids, & Hristovski, 2006). For that reason, diverse match analysis techniques (i.e., observational tools, positional data) have been employed to understand the nature of sport tactical behaviours (Serra-Olivares, Clemente, & González-Víllora, 2016).

Match analyses in soccer have been growing in the last decade by studying the players' relative pitch positioning data (Carling, Williams, & Reilly, 2005; Coutts, 2014; Gómez-Ruano, 2017), by using different systems such as radio frequency (W Frencken, Lemmink, Delleman, & Visscher, 2011), video tracking (Di Salvo et al., 2007), and GPS devices (Aughey, 2011). The information about the dots (the players) can be accessed at each instant by considering bidimensional data position. Such information has been contributing to the study of tactical behaviour of players and teams' performance at micro, meso and/or macro levels (Clemente, Couceiro, Martins, Mendes, & Figueiredo, 2013).

Different tactical behaviour indicators regarding inter- and/or intra-player coordination variables have been examined. Some examples are distances (length and width), teams' areas and/or teams' (Surface Area, Effective Area of Play, Territorial Domain, Defensive Play Area) and/or teams’ dyads or centroids (Clemente, Couceiro, Martins, Mendes, & Figueiredo, 2014b; Duarte, Araújo, Correia, & Davids, 2012; Travassos, Davids, Araújo, & Esteves, 2013).

Most common indicators have focused on the teams' geometrical centre from the player’s positions, the centroid (Duarte et al., 2013; Frencken et al., 2011), and wCentroid (Clemente, Couceiro, Martins, Mendes, & Figueiredo, 2014a). Findings suggest strong coordination patterns between opposing teams centroids (Duarte et al., 2013), although they do not have strong relations to score situations (Frencken, Poel, Visscher, & Lemmink, 2012). Other analyses have showed higher degrees of coordination when players/teams became more offensive, and unpredictable movements of the attackers during the attacking phase (Gonçalves, Figueira, Maçãs, & Sampaio, 2014). In the same line, no relationship has been found between teams' longitudinal and lateral centroid distances with goal scoring situations, but space and temporal reorganization became crucial after ball position changes (Frencken et al., 2012).

Teams/players tend to present regular tactical behaviours along the match in several variables such as areas, length and width (Duarte et al., 2013), although it is necessary to understand if it happens due to the fatigue or the tactical approaches used (Clemente, Couceiro, Martins, Mendes, & Figueiredo, 2013). In this sense, new performance measures such as the wCentroid and the players’ dispersion over their centre (wStretch Index) are providing relevant information about the players and teams collective organization. This procedure leads to the identification of in-phase and anti-phase patterns and the tendency to act on a specific side of the field in critical moments of the game (Bartlett, Button, Robins, Dutt-Mazumder, & Kennedy, 2012).

Tactical behaviour analyses have been developed during competitive forms and using small- sided and conditioning games (SSCGs). It is due to the fact that SSCGs may have identical tactical problems and may contribute to the coaching processes (Serra-Olivares, González-Víllora, García-López, & Araújo, 2015). Thus, tactical behaviour analysis using SSCGs supports the teaching and coaching processes (Davids, Araújo, Correia, & Vilar, 2013; Serra- Olivares et al., 2016).

In regards to the modification of the field size, findings show that different playing areas alter the relative space per player, and it has an impact on performance behaviours (Casamichana & Castellano, 2010; Fradua et al., 2013). This results from the necessity of players and teams to adapt their tactical organization (Chow et al., 2006; Davids et al., 2013). For example, smaller playing areas decrease the relative space per player and teams constraining tactical organizations such as defensive stability and the creation of offensive opportunities (Travassos, Gonçalves, Marcelino, Monteiro, & Sampaio, 2014).

Different trajectories of the players and spatial distributions have been observed in SSCGs when the size of the pitch is altered (Silva et al., 2014). In this way, more advantageous numerical relations adjacent to each individual player and broader individual spatial distributions were observed when the field dimension is modified (Silva et al., 2016). Other studies have showed that in smaller field dimensions players and teams have fewer opportunities to maintain ball possession (Vilar et al., 2014), and in shorter dimensions different inter-team lateral and longitudinal distance values may be observed (Frencken, Van Der Plaats, Visscher, & Lemmink, 2013).

Despite of such evidences of tactical adjustments in smaller and conditioned formats, no studies have been analysing the effects of conditioned games (e.g., changing the size) in official format 11 vs. 11 on soccer, as far as we know. Such adaptation is commonly used by coaches in daily practice to increase the pressing in the team with possession of the ball and also to reduce the time to decision making time. In the foregoing, it is important to know the effects of the size modification on tactical behaviours by using new measures of tactical behaviour analysis. Therefore, the aim of this study was to analyze the variance of tactical behavior measured by the centroid and stretch index of the team between two different field sizes (full and half) in 11 vs. 11 format. We hypothesize that the centroid will stay in a forward position in the bigger field and a greater expansion of the team will be also observed.

Analysis on the variance of the centroid (goal-to-goal and lateral-to-lateral) and stretch index (total, goal-to-goal and lateral-to-lateral) between pitch sizes was tested with independent t- test. Descriptive statistics can be found in table 1.

Table 1
Descriptive statistics (mean, standard deviation and 95% Confidence interval) of centroid and stretch index for both pitch sizes (half and full).

*Significant different for a p < 0.005

C(x): centroid (goal-to-goal); C(y): centroid (lateral-to-lateral); SIt: Stretch index total; SI(x): Stretch index (goal-to-goal); SI(y): Stretch index (lateral-to-lateral); p: p-value; ES: Cohen D value.

Significant greater values of centroid (goal-to-goal) were found in the full size (p = 0.001; ES = 3.794, strong effect). Considering the middle of the pitch in both conditions, it is possible to verify that in half condition the mean of the centroid is 2.5 m (3.70% of the x size for the half pitch) behind the middle of the pitch and in the full condition it is 4.6 m (4.26% of the x size for the full pitch) beyond the middle of the pitch. Variation of the centroid throughout the time can be verified in Figure 3.

Figure 3
Centroid position (goal-to-goal and lateral-to-lateral) in both pitch sizes over the time.

Significant greater values of the centroid (lateral-to-lateral) were found in the full size (p = 0.001; ES = 0.729, minimum effect). In the half size condition, it was found that the average of centroid was 0.93 m (2.94% of the lateral-to-lateral half size) for the right side of the middle of the lateral-to-lateral and in the full size it was observed 2.8 m (4.12% of the lateral-to-lateral full size) for the left.

Significant greater values of total stretch index were observed in the full size (p = 0.001; ES = 1.185, moderate effect). As is possible to identify, the coefficient of variation of the total stretch index was 28.74% for the half size and 16.91% for the full size, thus suggesting a greater stabilization of this variable in bigger sizes. Values of the total stretch index during the matches can be observed in Figure 4.

Figure 4
Stretch index in both pitch sizes over the time.

Significant greater values of stretch index (lateral-to-lateral) were found in the full size (p = 0.001; ES = 0.628, minimum effect). A coefficient variation of 46.32% was found in the full size and 58.26% in the half size, suggesting a greater variation of lateral-to-lateral displacement in the half size. In the case of stretch index (goal-to-goal), no significant differences were found between pitch sizes (p = 0.114; ES = 0.291, no effect). A coefficient of variation of 73.47% was found in the full size and 49.04% in the half size, thus suggesting a greater dispersion of values in the full size.

The size of the pitch is one of the key conditions utilized by coaches to constrain the collective behavior during training sessions (Davids et al., 2013). For that reason, this study aimed to analyze the spatio-temporal relationships between teammates in the official format of 11 vs. 11 in two different pitch scenarios: half size and full size. The main results revealed significant differences in centroid position and stretch index between the two analyzed situations.

The stretch index can be used to measure the dispersion of the players around the geometrical center based on the instantaneous position of players (Bartlett et al., 2012; Bourbousson et al., 2010). This measure has also been associated with collective patterns during attacking and defensive moments, representing the capacity of the team to play closer during defensive pressing and to expand while attacking (Clemente et al., 2013). In our study, it was found that there was a significant increase in stretch index (total) values in the case of full pitch. The average expansion of the team in the full size game was almost 29% (~10 m) more than in half pitch game, thus suggesting that the size of the pitch may constrain the way how dots (players) interact and rationally occupy the space. In a previous study, conducted in 4 vs. 4 format, it was found that the surface area (area of the polygon capable of contain all dots of a team, formed by the minimum amount of dots) was minimally affected by reducing width and length of the field considering a stabilization of interactive behavior (Frencken, van der Plaats, Visscher, & Lemmink, 2013). However, our stretch index in the goal-to-goal axis revealed no significant differences between pitch sizes, thus suggesting that the strategy of the team did not increase dispersion between sectors (length), but raised the dispersion in terms of the width. This can be a specific option of the team to attract opponents from the middle for the wingers to be able to exploit the possibility of unbalance in the defensive ‘block’ of the opponents. Our results only found significant differences in the lateral-to-lateral axis, thus suggesting the use of the option to spread the team through its wings. This option can be partially justified by the mobility principle that aims to unbalance the defensive organization of opponent through a fast shifting and good ball circulation (Gréhaigne, Richard, & Griffin, 2005).

The geometrical center of the team (centroid) in both axes was also analyzed in this study. The centroid represents the middle point of all the dots of a team and can provide information about the zone of balance of a team (Clemente et al., 2014a). In a study that compared the relationship between the teams’ centroid, it was found that there is a strong association between the two teams in goal-to-goal and lateral-to-lateral axes and also that smaller pitches resulted in a closer relationship between the teams’ centroid (Frencken et al., 2013). Our results revealed significant differences of the centroids between both pitch sizes. In the goal-to-goal direction of the centroid, it was found that, in average, the team occupied a region behind the middle line of the pitch in the half size, and a zone beyond in the full pitch. Our study did not consider the status of possession of the ball, and for that reason, it is possible to justify the region of centroid being close to the middle line of the pitch. However, these results were similar to previous studies that compared the centroid position in different status of possession of the ball and score (Clemente et al., 2013; Clemente, Couceiro, Martins, Mendes, & Figueiredo, 2014c). A tendency to occupy the right side in half size and the opposite side in the full pitch was observed in the case of lateral-to-lateral centroid. This may be constrained by the specific interactions that emerged from the game (Vilar, Araújo, Travassos, & Davids, 2014).

This study has some limitations. The absence of analysis of the possession of the ball reduces the analysis of the collective patterns due to not having data on the phase of the game. Moreover, future studies must consider the inter-teams distance (centroids) and the relationship between the centroids. Such analysis will improve the perception on how teams may be interactively constrained by the same variable.

Despite the limitations of the study, and to the best of our knowledge, this study was the first to analyze the influence of playing in a full size and half size in a 11 vs. 11 format, and one of the few studies that have analyzed the impact of the size of the pitch in the tactical behavior of teams measured by data position metrics. As practical implications, it can be possible to recommend using the full pitch size to increase the lateral-to-lateral space and to increase the space occupied by the team, thus being in line with the tactical principle of mobility. Future studies must consider extending such analysis to smaller formats of the game, different task conditions and to add more variables associated with the status of ball possession and spatio-temporal relationship between teams.

Results of this study revealed a greater lateral-to-lateral dispersion of the team in full size of pitch when compared to half size. Moreover, the centroid was beyond the middle line of the pitch in full size, and behind it in the half size. It is possible to conclude that the full pitch size promotes the adjustment of the team in order to spread to the wings and to be in line with the tactical principle of mobility. Half pitch can be more adequate to act proximally and to promote the principle of acting in unit. Future studies must consider the analysis of the spatio-temporal relationships between teams in different sizes and in other formats.

This study was carried out within the scope of R&D Unit 50008, financed by UID/EEA/50008/2013. This study was conducted in the aim of the granted project: uPATO from Instituto de Telecomunicações. We would like to thank to António Barbosa and Lillian Gonçalves for their help in data collecting.