Profiling Successful Team Behaviors in League of Legends

Despite the increasing popularity of electronic sports (eSports), there is still a scarcity of academic works exploring the playing behavior of teams. Understanding the features that help to discriminate between successful and unsuccessful teams would help teams improving their strategies, such as determine performance metrics to reach. In this paper, we identify and characterize team behavior patterns based on historical matches from the very popular eSpor League of Legends web API. By applying machine learning and statistical analysis, we clustered teams' performance and investigate for each cluster how and to what extent these features have an influence on teams' success and failure. Some clusters are more likely to have winning teams than others, the results of our study helped to discover the characteristics that are associated with this predisposition and allowed us to model performance metrics of successful and unsuccessful team profiles. At all, we found 7 profiles in which were categorized into four levels in terms of winning team proportion: very low, moderate, high and very high.

[1]  R. Edge,et al.  Predicting Player Churn in Multiplayer Games using Goal-Weighted Empowerment , 2013 .

[2]  Brent E. Harrison,et al.  Identifying patterns in combat that are predictive of success in MOBA games , 2014, FDG.

[3]  Hannes Werthner,et al.  Ranking factors of team success , 2013, WWW.

[4]  Marc Herrlich,et al.  Classification of Player Roles in the Team-Based Multi-player Game Dota 2 , 2015, ICEC.

[5]  Víctor Codocedo,et al.  What Did I Do Wrong in My MOBA Game? Mining Patterns Discriminating Deviant Behaviours , 2016, 2016 IEEE International Conference on Data Science and Advanced Analytics (DSAA).

[6]  Jacqui Taylor,et al.  A Qualitative Exploration of Factors Affecting Group Cohesion and Team Play in Multiplayer Online Battle Arenas (MOBAs) , 2016, The Computer Games Journal.

[7]  Kaushik Kalyanaraman,et al.  To win or not to win ? A prediction model to determine the outcome of a DotA 2 match , 2015 .

[8]  Jean-Philippe Métivier,et al.  Mining Tracks of Competitive Video Games , 2014 .

[9]  Diego Klabjan,et al.  Skill-based differences in spatio-temporal team behaviour in defence of the Ancients 2 (DotA 2) , 2014, 2014 IEEE Games Media Entertainment.

[10]  Nicholas Kinkade,et al.  DOTA 2 Win Prediction , 2015 .

[11]  Young Bin Kim,et al.  Efficiently detecting outlying behavior in video-game players , 2015, PeerJ.

[12]  Anders Drachen,et al.  Esports Analytics Through Encounter Detection , 2016 .

[13]  Yun Huang,et al.  Team vs. Team: Success Factors in a Multiplayer Online Battle Arena Game , 2015 .

[14]  Trupti M. Kodinariya,et al.  Review on determining number of Cluster in K-Means Clustering , 2013 .

[15]  Mohammed J. Zaki.  Data Mining and Analysis: Fundamental Concepts and Algorithms , 2014 .

[16]  Chunming Rong,et al.  Using Spearman's correlation coefficients for exploratory data analysis on big dataset , 2015, Concurr. Comput. Pract. Exp..

[17]  Filip Johansson,et al.  Result Prediction by Mining Replays in Dota 2 , 2015 .

[18]  Hao Yi Ong,et al.  Player Behavior and Optimal Team Composition for Online Multiplayer Games , 2015, ArXiv.