Player Behavior Analysis for Predicting Player Identity Within Pairs in Esports Tournaments: A Case Study of Counter-Strike Using Binary Random Forest Classifier

Abstract

This research utilizes a binary random forest classifier to predict individual players based on their in-game behavior, analyzing and distinguishing within player pairs through a comprehensive set of in-game features. The analysis is based on a dataset from 119 "Counter-Strike: Global Offensive" (CS:GO) esports tournament matches. The classifier achieves a testing accuracy of up to 87%, highlighting its ability to effectively differentiate between players. A key contribution of this paper is the demonstration of the potential to predict player identities through in-game behavior data from CS:GO. This has implications for the gaming industry such as mitigating security issues. Additionally, the study pinpoints a detailed set of behavioral features that can uniquely identify players in a competitive esports setting.