Quantifying the value of sprints in elite football using spatial cohesive networks. (October 2020)

• Record Type: Journal Article
• Authors: Külah, Emre; Alemdar, Hande
• Is Part Of: Chaos, solitons and fractals. Volume 139(2020)
• Publication Date: 2020-10
• Subjects: Team sport -- Sprint analysis -- Collective movement -- Data mining -- Performance analysis; Chaotic behavior in systems -- Periodicals; Solitons -- Periodicals; Fractals -- Periodicals; Chaotic behavior in systems; Fractals; Solitons; Periodicals; 003.7
• DOI: 10.1016/j.chaos.2020.110306
• Languages: English
• ISSNs: 0960-0779

Abstract

Highlights: Cohesive networks between teammates created using their closeness, passes and common markings. Pass interception probability and pitch value distributions are calculated for each player. Sprint performances are quantified using cohesive networks, pass interception probabilities and pitch values. Full-backs and attackers have greater sprint averages than defensive midfielders and center backs. Teams playing possession football have less sprint averages than counter-attack style teams. Abstract: Football players are on the move during games and the sprint is one of the distinctive type of those movements. In this study, we focus on quantifying the value of the sprints using the spatial data of players and the collective movements of the teams during the game. We first propose a method to quantify the dispersion of the teams, namely, the weighted team spread. In order to find the weights of the team spread, we use individual players' interaction behavior, using spatial cohesion matrices. Spatial features of the pitch such as the pitch value and the pass probability value are also used together with the weighted team spread to quantify the value of the sprints. These models are used to understand sprint character of the players according to their role and teams' collective movements depending on their tactics. The proposed method applied on 306 Turkish first division games from 2018/2019. The sprint analysis results show that attackers have greater sprint averages than midfielders and defenders based on 5498 sprints from corresponding games. Full-backs and attacking midfielders are positions with the best sprint averages other than attacking players. Center backs and defensive midfielders are the weakest positions in sprinting. The results further show that the teams that are focused on having the possession of the ball have less average sprint value than teams playing counter-attack style.