Mining Tracks of Competitive Video Games

François Rioult (1) , Jean-Philippe Metivier (1) , Boris Helleu (2) , Nicolas Scelles (3) , Christophe Durand (2)

1 Equipe CODAG - Laboratoire GREYC - UMR6072
2 CesamS - Centre d'étude sport et actions motrices
3 University of Stirling

The development and professionalization of a video game requires tools for analyzing the practice of the players and teams, their tactics and strategies. These games are very popular and by nature numerical, they provide many tracks that we analyzed in terms of team play. We studied Defense of the Ancients (DotA), a Multiplayer Online Battle Arena (MOBA), where two teams battle in a game very similar to rugby or American football. Through topological measures - area of polygon described by the players, inertia, diameter, distance to the base - that are independent of the exact nature of the game, we show that the outcome of the match can be relevantly predicted. Mining e-sport's tracks is opening interest in further application of these tools for analyzing real time sport

Mots clés: Electronic sport; MOBA; Data mining; Classification

Domaines: Informatique et théorie des jeux [cs.GT]; Intelligence artificielle [cs.AI]; Modélisation et simulation

François Rioult, Jean-Philippe Metivier, Boris Helleu, Nicolas Scelles, Christophe Durand. Mining Tracks of Competitive Video Games. AASRI Conference on Sports Engineering and Computer Science (SECS 2014), Jun 2014, Londres, United Kingdom. pp.82-87, ⟨10.1016/j.aasri.2014.08.014⟩. ⟨hal-01011551⟩