Creating and sharing knowledge for telecommunications

Motion variability during small-sided games: using the dots to assessing the tactical behaviour

Clemente, F.M.C. ; Sequeiros, J. ; Correia, A. ; Martins, F. ; Silva, F.

Motion variability during small-sided games: using the dots to assessing the tactical behaviour, Proc World Conference on Science & Soccer - WCSS , Rennes, France, Vol. , pp. 119 - 120, June, 2017.

Digital Object Identifier:

Abstract
Introduction
Small-sided games (SSGs) are one of the most addressed topics in the science of soccer1. The monitoring process to assess the external load of these games has been made by using geolocation systems. However, these systems allow the analysis of more than just the physical impact. The progress of individual dots in the bi-dimensional space during the games can be used to estimate tactical behaviour and motion variability of the players2. Consequently, with these opportunities in mind, this study aims to examine the motion variability and individual tactical behaviour of soccer players during different SSGs.

Methods
Participants were ten collegiate male players (20.34.8 yrs; 175.177.5 cm; 69.313.0 kg). Five formats (Fs) (1vs.1 to 5vs.5) and two pitch sizes (PS) (S: 50 m2 and L: 125 m2 per player) were tested randomly in two different days. Players were tracked by portable GPS devices (10Hz, Accelerometer 1kHz, FieldWiz, Paudex, Switzerland). Position data was used to calculate the player’s variability in the ultimate Performance Analysis Tool (uPATO). The following measures were tested: Shannon entropy (SE: quantify the uncertainty of location of each player); Spatial exploration index (SEI: assess the player’s exploration considering the game scenarios); Kolmogorov entropy (KE: quantify the variability of a player over time). Comparative analysis between Fs and PS was performed with one-way ANOVA and eta squared (ES) for a p<0.05.

Results & Discussion
Strong effects of Fs on SEI were found (p=0.00; ES=0.82). SEI was significantly smaller on the largest format (5v5: 63.31, 59.72-66.90CI95%) comparing with 1v1 (97.28, 93.90-100.65CI95%), 2v2 (90.28, 86.22-94.34CI95%), 3v3 (85.72, 80.94-90.50CI95%) and 4v4 (67.02, 62.96-71.08CI95%). PS had no effect on the SEI (p=0.80; ES=0.00).
On SE, moderate effects of Fs on SE were found (p=0.00; ES=0.28). Smaller SE values were found on 1v1 (0.25; 0.21-0.29CI95%), when compared to 2v2 (0.33; 0.28-0.38CI95%), 3v3 (0.33; 0.28-0.39CI95%), 4v4 (0.36; 0.32-0.41CI95%) and 5v5 (0.38; 0.33-0.42CI95%). Non-significant effects of PS on SE were (p=0.52; ES=0.01).
Small and non-significant effects of Fs and PS on the KE were found (p=0.18, ES=0.10 and p=0.14, ES=0.4, respectively). Values of KE varied between 0.09 and 0.31 [CI95%].

Conclusion
Larger Fs decreased the SEI, thus contributing to a more positional play style of each player. Smaller Fs may contribute to a greater exploration of the pitch. However, this exploration does not seem to reflect a greater uncertainty. Smaller values of SE were found on the smallest Fs, thus suggesting a greater predictability to reproduce the same trajectories.