Using statistical methods to predict results in combat sports

The bachelor thesis deals with the use of statistical methods for the prediction of results in mixed martial arts (MMA), especially in UFC fights. The aim of the thesis was to contribute to a better understanding of the factors influencing match outcomes, to implement predictive models and thus provide useful insights into the sport. The thesis is structured in five parts, which include understanding the sport, obtaining, processing and cleaning the input data, analyzing the key variables affecting the outcomes and the way fights are finished, and implementing predictive models. The analysis examined, among other things, the physical characteristics of the fighters, their previous matches, external match conditions, and the historical development of match finishing methods. Selected machine learning models and an Elo model were created to predict match outcomes and finishing methods. This paper presents a comprehensive approach to the analysis and prediction of sporting outcomes using modern statistical techniques and python programming language libraries. The main contribution of the work is seen by the author in particular in the accuracy of the prediction of one of the models for the match finishing method.