Edinburgh Napier University

  Online gaming is very popular and has gained some recognition as the so called e-sport over the last decade. However, in particular First Person Shooter (FPS) games suffer from the development of sophisticated cheating methods such as aiming robots (aimbot), which can boost the players ability to acquire and track targets by the illicit use of internal game states. This not only gives an obvious unfair advantage to the cheater, but has negative impact on the gaming experience of honest players.

In this paper we present a novel supervised method based on distribution comparison matrices that shows very promising performance in the identification of players that use such aimbots. It extends our previous work in which two features were identified and shown to have good predictive performance. The proposed method is further compared with other classification techniques such as Support Vector Machines (SVM). Overall we achieve true positive and true negatives rates well above 98% with low computational requirements.