Analysis of the limitations of machine learning in predicting sports results

Abstract

This paper presents an experimental evaluation of the effectiveness of a predictive model based on machine learning algorithms. The key objective was to analyze the reasons for the discrepancies between predicted and actual results. The results of the experiments revealed a significant impact of dynamic and unpredictable factors (sudden injuries, changes in composition, team motivation, weather conditions) on the accuracy of predictions. The main limitations of existing statistical approaches were systematized and priority areas for further improvement of sports analytics methods were identified. The main result of the work is the identification of key challenges and justification of the need for new approaches to forecasting in the context of highly dynamic sporting events.