Win, lose or cheat: The analytics of player behaviors in online games

Analysis of game telemetry data is fundamental for understanding common player behaviors and combating cheating in massively multi-player online games (MMO). The challenge of such analysis is data contamination resulted from system and network latency, data tempering and bugs, In this paper we are analyzing race results from a racing MMO. We are addressing the challenge of identifying true race winners and detecting cheaters. Our approach is based on statistical analysis techniques for detecting outliers. Outliers are classified using heuristics of multiple measures. We identify potential cheaters by analyzing outlier results in multiple races. Resulting player classification identifies suspicious player behaviour and provides insight into potential improvements to game tuning and cheat prevention measures.