Network Analysis in Basketball: Inspecting the Prominent Players Using Centrality Metrics

(ProQuest: ... denotes formulae omitted.)IntroductionThe organization between team-members it is one of the main challenges in team-sports games (Grehaigne, Bouthier, & David, 1997). The dynamic properties of the game, the cooperation-opposition relationship and the contextual factors constraints the cooperation process of the team-members (Balague, Torrents, Hristovski, Davids, & Araujo, 2013). For those reasons the tendencies and patterns of relationship between team-members it is not always the same or repetitive (Grehaigne, Godbout, & Zerai, 2011). In fact, the ability to be variable and improve the dynamic of cooperation may improve the possibility to overcome the opponent and decrease the chance to be blocked by the opponent (Couceiro, Clemente, Martins, & Machado, 2014).In the specific case of team sports there are a particular interest to recognize the cooperation tendencies between team-members (Pena & Touchette, 2012). In fact, such systematic knowledge may augment the perception of coaches to understand the dynamic of their team and to identify the weakness and strong points of opponent's team (Clemente, Martins, Couceiro, Mendes, & Figueiredo, 2014). This specific process of analyse the performance in match it is designated as match analysis in sports (Carling, Williams, & Reilly, 2005). In the field of match analysis there are different approaches to process the observation and the analysis. One of the most common is the traditional notational analysis based on codification of events (actions) or behaviours (Franks & McGarry, 1996). This analysis leads with a tendency to analyse the technical events and reduce the outputs about the inter-relationship between team-members, organization, and tactical behaviour (Clemente, Couceiro, Martins, Mendes, & Figueiredo, 2014). In other hand, recently there are been suggested some computational metrics that integrates the bi-dimensional position of players (using GPS devices, multi-cameras tracking systems, or RFID) to compute the geometric organization of teams and the coordination tendencies between team-members (Bartlett, Button, Robins, Dutt-Mazumder, & Kennedy, 2012; Duarte, Araujo, Correia, & Davids, 2012). These computational metrics shows relevant information about organizational level but do not identify the cooperation of team-members in the passing sequence (Clemente, Couceiro, Martins, Mendes, et al., 2014).Thus, a new approach based on Social Network Analysis has been used in the last few years to identify the cooperation process of players in the attacking moments (in the moments with possession of the ball) (Duch, Waitzman, & Amaral, 2010; Grund, 2012). The Social Network Analysis provide an opportunity to analyse the general level of cooperation of the team and also to identify the centrality levels of players for the overall cooperation (Clemente, Couceiro, Martins, & Mendes, 2014). Thus, the network focus more in the collective than in the individuals (Wasserman & Faust, 1994).In the specific case of team sports, the majority of reported studies have been in soccer (Clemente, Couceiro, Martins, & Mendes, 2014; Duch et al., 2010; Pena & Touchette, 2012). In the case of European Cup 2008 it was found that the midfielder of Spain was the centrality player of the winning team (Duch et al., 2010). After, in a study on FIFA World Cup 2010 it was possible to identify that the centrality players in the winning team were the midfielders and in the case of the opponent team in the final the biggest centrality levels were found in defenders and midfielders (Pena & Touchette, 2012). In the case of general cooperation, it was found in the Premier League case that the greatest levels of density and smallest levels of heterogeneity were associated with the teams that had best performance in matches (Grund, 2012). In a more recent study it was found that in a top team in Portuguese Premier League the centrality players in the attacking process were the external defenders and the midfielders (Clemente, Couceiro, Martins, & Mendes, 2014). …