The field of e-sports statistics is getting more popular as e-sports become generally known. In this thesis, we discuss the adaptations of two models with proven usage in sports statistics and their possibilities in the realm of e-sports. The most significant advantage shared by both of these models, known as general autoregressive score model and common opponent model, is that they require only publicly available data in order to forecast future match results. In an empirical study, we test the... show full abstractThe field of e-sports statistics is getting more popular as e-sports become generally known. In this thesis, we discuss the adaptations of two models with proven usage in sports statistics and their possibilities in the realm of e-sports. The most significant advantage shared by both of these models, known as general autoregressive score model and common opponent model, is that they require only publicly available data in order to forecast future match results. In an empirical study, we test the predictive power of said models on the game Counter-Strike, abbreviated as CS:GO, and introduce another approach to the common opponent model, which accounts for the time-varying performance of modelled teams and slightly improves the predictive ability of the model |