Balázs Tobak
This thesis examines the prediction of the outcomes of Formula 1 overtaking attempts based on data from 2018 to 2025. I developed and compared three models (logistic regression, XGBoost, and an Entity Embedding neural network) in terms of predictive accuracy and real-time decision support effectiveness. The results show that XGBoost is the most accurate and stable. The research highlights the importance of interpretability and the dominance of dynamic factors.