Machine Learning-Based Estimation of Pit Stop Probability During Safety Car in Formula 1

In the context of Formula 1, strategic decision-making plays a crucial role in determining the outcome of both individual races and the overall championship. One of the most critical scenarios during a race is the deployment of the Safety Car, which introduces a narrow time window for teams to decide whether to call their drivers into the pits. Making the right call in these moments can significantly influence race results. This thesis proposes the development of a machine learning model that leverages historical race data to estimate the probability of each driver making a pit stop during a Safety Car period. By quantifying these probabilities, the strategy team is equipped with a data-driven tool to support rapid and informed decisions on whether their own drivers should pit or stay out, ultimately enhancing competitive performance.