This paper discusses how social network analyses and graph theory can be implemented in team sports performance analyses to evaluate individual (micro) and collective (macro) performance data, and how to use this information for designing practice tasks. Moreover, we briefly outline possible limitations of social network studies and provide suggestions for future research. Instead of cataloguing discrete events or player actions, it has been argued that researchers need to consider the synergistic interpersonal processes emerging between teammates in competitive performance environments. Theoretical assumptions on team coordination prompted the emergence of innovative, theoretically driven methods for assessing collective team sport behaviours. Here, we contribute to this theoretical and practical debate by re-conceptualising sports teams as complex social networks. From this perspective, players are viewed as network nodes, connected through relevant information variables (e.g. a ball-passing action), sustaining complex patterns of interaction between teammates (e.g. a ball-passing network). Specialised tools and metrics related to graph theory could be applied to evaluate structural and topological properties of interpersonal interactions of teammates, complementing more traditional analysis methods. This innovative methodology moves beyond the use of common notation analysis methods, providing a richer understanding of the complexity of interpersonal interactions sustaining collective team sports performance. The proposed approach provides practical applications for coaches, performance analysts, practitioners and researchers by establishing social network analyses as a useful approach for capturing the emergent properties of interactions between players in sports teams.
Keywords: Cluster Coefficient; Interpersonal Interaction; Social Network Analysis; Team Performance; Team Sport.