Modularity in Football Passing Networks

Abstract

In recent decades, within the boundaries of complexity sciences, network science has been used to analyze many kinds ofnetworks. A complex system is formed by smaller subsystems which can be designed independently yet function together as awhole. Modules of a network can be called as groups, clusters or communities and modularity can be defined as a measure ofthe structure of networks or graphs. If it has dense connections within a network’s modules and sparse connections betweennodes in different modules, in this case networks have high modularity. At the end of each football match, successful passnetworks can be achieved. These modular structures can be thought as “independent yet function as a whole” football modules.Generally so called modules in the technical directors’ thoughts can be listed as defense, midfielders area and strikers area. Ifthey want to know how modules are generated and how their disconnection leads to functional decay they should analyze themodularity formed as a result of a football match. The aim of this study is to examine to what extent the football teams'managers have implemented their strategies in games. First, 10 matches with the e-analysis football program were analyzed.And then modularity analysis began with transforming the video of a football match into a pass network. Using this passnetwork, network metrics was to be computed and then these metrics was to be used to make a modularity analysis usingGephi. After modularity analysis using Gephi, modularity classes of these networks were found for all networks. When theresults obtained from the modularity analysis were examined, it was observed that the number of modules varied between 2and 4. Consequently, it was observed that the systems found as a result of the modularity analysis were very different than theplanned systems.