Probabilistic approaches to cheating detection in online games

Cheating is a key issue in online games, as it reduces the satisfaction of honest players with a game and can result in reduced revenue for game providers. Cheat prevention techniques used today typically aim to identify that a particular known technique for cheating is in use, and then prevent that. In this paper we present two techniques for identifying cheating based purely on the results of games. These techniques are based on the law of large numbers and the Bradley-Terry model, and aim to identify statistical behaviour that suggests a player is cheating. Preliminary simulation studies suggest that such techniques are a promising way to identify cheaters.