Collective movement analysis reveals coordination tactics of team players in football matches

Abstract

Collective behavior is a hallmark of every living system and utilizing methods from statistical physics (such as correlation functions) could aid in our understanding of their underlying rules. We analyzed five football (soccer) matches as this game provides a unique but yet mostly unexplored example to study a system of collective cooperation and competition. The aim of our study was to analyze the collective motion patterns exhibited by football players to unfold the underlying coordination among them in order to understand collective strategies associated with team performance. By analyzing pairwise relationships among all the players using spatio-temporal correlation functions we reveal that there exist identifiable collective dynamics that characterize winning and losing teams. Using our metric we find clear and robust differences between the players, indicating a difference in their behavior and their interactions. And this enables us to assign a unique behavioral pattern - a ‘fingerprint’ - for each individual and for each team. Furthermore, we reveal there exists a relationship between the market value of the players and the metrics introduced here, suggesting that these metrics could potentially serve as valuable performance indicators in the future, with applications ranging from talent identification to player scouting. In a broader context team sports could open up new directions for quantitative analyses of human collective behavior.