An agent-based modeling approach for football performance analysis

In football, off-ball movement is a crucial component of player performance due to the scarcity of ball possession and the influence it has on other players’ actions. However, traditional performance analysis approaches like notational analysis fall short when applied to off-ball movement as it fails to consider the movement’s context and variability. From a complex systems perspective, this research aims to develop an agent-based model for players’ off-ball movement analysis. In a football game simulation, football players are agents with behaviors and attributes based on sports analytics literature and expert interviews. The model was validated through focus group discussions, wherein football experts evaluated the model\u27s simulation of a football game, and empirical validation, a comparison between real-world and the model\u27s match data. Off-ball movement are movements the football players perform when they do not have the ball but still influence the play. The players\u27 attacking off-ball movement were affected by the formation of their opponents, the amount of open space available in the field, and the player\u27s own role and positioning. These factors also contribute to where within the field these off-ball movements were performed. Off-ball movement is an important tool for attacking teams to create shots and score goals