KPI: Collective Behavior in Football

Abstract

This chapter presents the need to explore advanced methodologies that can process positional data in football and potentially provide information about how the players’ movements are related to each other. The players’ high-resolution trajectories were used to calculate spatio-temporal correlation-based metrics with other players (teammates and opponents) and the ball, in order to identify highly correlated segments (HCS). This metric seems to be promising to identify differences between the players and, thus, bringing up the concept that each player and team has a unique behavioral pattern – a ‘fingerprint’. Therefore, 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.