Measuring the pitch control of professional football players using spatiotemporal tracking data
-
Published:2023-05-09
Issue:2
Volume:4
Page:025008
-
ISSN:2632-072X
-
Container-title:Journal of Physics: Complexity
-
language:
-
Short-container-title:J. Phys. Complex.
Author:
Higgins LewisORCID,
Galla Tobias,
Prestidge Brian,
Wyatt Terry
Abstract
Abstract
We study pitch control in football, using data from six complete seasons of the English Premier League. Our objective is to investigate features of pitch control in the data. We process the data to ensure consistency of the tracking and event datasets. This represents the largest coherent dataset analysed in the literature and allows the observation of consistent patterns across several seasons’ data. We demonstrate that teams playing in front of a crowd at home control on average
2.9
±
0.2
%
more of the pitch than teams playing away, which reduces to
1.5
±
0.6
%
in matches played behind closed doors. We observe that match by match the difference in pitch control between the teams has a weak, positive correlation with the difference in expected goals (Pearson correlation R = 0.38). As a further manifestation of home advantage we find that in games which the two teams have equal pitch control, on average the home team accumulates greater expected goals (
0.16
±
0.03
). The concept of weighted pitch control is introduced, by assigning a weight to regions of the pitch. We demonstrate that pitch control of the penalty box of the out-of-possession team is negatively correlated with expected goals in each of the six seasons, and interpret this apparently counter-intuitive result.
Funder
Ministerio de Ciencia e Innovación
Engineering and Physical Sciences Research Council
Agencia Estatal de Investigación
Subject
Artificial Intelligence,Computer Networks and Communications,Computer Science Applications,Information Systems
Reference36 articles.
1. An examination of expected goals and shot efficiency in soccer;Rathke;J. Hum. Sport Exerc.,2017
2. Opta analytics—stats perform
3. Soccer data
4. Tracab technologies
5. Second spectrum—our work;Spectrum,2019