The need for weighting indirect connections between game variables: Social Network Analysis and eigenvector centrality applied to high-level men’s volleyball

ABSTRACT Performance analysis in volleyball has seldom analysed the interrelationships of game actions under the systemic view of distinct game complexes, and how different patterns of game flow emerge. In this study, we used Social Network Analysis with eigenvector centrality to weight direct and indirect relationships between game actions and to assess the similarities and differences between six game complexes in high-level men’s volleyball. The study sample comprised 10 matches of the final phase of the 2015 World League (1600 game actions). Results indicated that dividing the game into six complexes and analysing game actions as nodes offers a more detailed understanding of the game and highlights the distinct constraints that typify each game complex. Specifically, the use of eigenvector centrality afforded a more accurate weighting of the variables for each complex. Because off-system situations were predominant in several game complexes, i.e.: Setting Condition C in Complex I (0.36), Complex II (0.55), Complex III (0.80) and Complex V (0.58); Attack Zone 4 and 2 in Complex I (0.30 and 0.28), Complex II (0.48 and 0.51), Complex IV (0.55 and 0.48) and Complex V (0.37 and 0.36); and Attack Tempo 3 in Complex I (0.33), Complex II (0.55) and Complex III (0.66). Our results suggest that coaches should prioritise these situations in training.