This study investigated changes in the complexity （magnitude and structure of variability） of the collective behaviours of association football teams during competitive performance. Raw positional data from an entir...This study investigated changes in the complexity （magnitude and structure of variability） of the collective behaviours of association football teams during competitive performance. Raw positional data from an entire competitive match between two professional teams were obtained with the ProZone tracking system. Five compound positional variables were used to investigate the collective patterns of performance of each team including： surface and geometrical centre. Analyses involve the coefficient （ApEn）, as well as the linear association between both area, stretch index, team length, team width, of variation （%CV） and approximate entropy parameters. Collective measures successfully captured the idiosyncratic behaviours of each team and their variations across the six time periods of the match. Key events such as goals scored and game breaks （such as half time and full time） seemed to influence the collective patterns of performance. While ApEn values significantly decreased during each half, the %CV increased. Teams seem to become more regular and predictable, but with increased magnitudes of variation in their organisational shape over the natural course of a match.展开更多
基金supported by a grant of the Portuguese Foundation for Science and Technology(SFRH/BD/43994/2008)
文摘This study investigated changes in the complexity （magnitude and structure of variability） of the collective behaviours of association football teams during competitive performance. Raw positional data from an entire competitive match between two professional teams were obtained with the ProZone tracking system. Five compound positional variables were used to investigate the collective patterns of performance of each team including： surface and geometrical centre. Analyses involve the coefficient （ApEn）, as well as the linear association between both area, stretch index, team length, team width, of variation （%CV） and approximate entropy parameters. Collective measures successfully captured the idiosyncratic behaviours of each team and their variations across the six time periods of the match. Key events such as goals scored and game breaks （such as half time and full time） seemed to influence the collective patterns of performance. While ApEn values significantly decreased during each half, the %CV increased. Teams seem to become more regular and predictable, but with increased magnitudes of variation in their organisational shape over the natural course of a match.
