Time-Series Analysis of Ball Carrier Open-Space in Association Football

Assessing team performances in association football (commonly known as football or soccer) is challenging due to its low-scoring and unpredictable nature. While space control and generating open spaces on the field have been used to evaluate strategies for team performance, the temporal nature of space available for the ball carrier has not been explored. This work proposes the temporal nature of space available for the ball carrier in the attacking third as a novel time-series performance evaluation metric to assess team performance. Furthermore, it suggests a novel approach to quantify the relatively open space available for the ball carrier in the attacking third using player information retrieved from television video footage. The machine learning model trained with the features extracted with the proposed time-series metric achieved 80% accuracy in classifying match-winning performances, indicating its importance as a performance evaluation metric. Correlation coefficients between temporal features and match results suggest that first-half performance is more crucial to the match result than second-half performance. Additionally, this work evaluates individual performances in the FIFA World Cup 2022 under two aspects: active open space creation and passive open space creation, assessing contributions to ball carrier open space creation through on-ball and off-ball actions, respectively.