Cheat Detection in a Multiplayer First-Person Shooter Using Artificial Intelligence Tools

Abstract

The use of cheating software in video games to gain an unfair advantage has required the use of anti-cheat software and deterrents such as account bans. Anti-cheat software is, however, always a step behind the opposition and as such new and innovative solutions are required. This paper considers AI driven tools as one such approach and compared decision trees, SVMs and Naïve Bayes classifiers in an attempt to classify cheating and honest player behaviour. The results of the research highlighted the potential for mouse dynamics as a measure of player behaviour, and decision trees as the most accurate detector of honest player behaviour.