Predicting Formula 1 Podium Finish Probabilities Using Machine Learning: A Comparative Study of Pre-Race and Within-Race Feature Scenarios

Sargur Yathishkumar, Likhita (2025) Predicting Formula 1 Podium Finish Probabilities Using Machine Learning: A Comparative Study of Pre-Race and Within-Race Feature Scenarios. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (2MB) | Preview
Preview	PDF (Configuration Manual) Download (1MB) | Preview

Abstract

The unpredictability of the race, the sensitivity of the results to minor variations in performance, and the effect of the weather and strategy, as well as the form of the drivers, all these makes the predicting in Formula-1 a challenging task. However, precise and transparent the forecasting is, it is very useful for the teams, analysts and broadcasters who require information, based on predictions before the race and during a race. This thesis creates a machine learning model that predicts podiums finishers based on some pre-race data and within-race summaries for the performance. Multi-season Formula-1 data are combined and tested using chronological split to prevent future information leakage. Random Forest and LightGBM are the two models, evaluated and compared and later the winner that is chosen on validation of Macro-F1 and log-loss. The quality of probabilities is evaluated with Brier and macro one-vs-rest AUC. It is demonstrated that the informative, well-calibrated pre-race probabilities can be achieved and when the within-race aggregates are added, Macro-F1 and the calibration as an upper bound which is further benefited. SHAP analysis of the Pre-race LightGBM shows that grid position and recent form gives most contribution to podium probability. The system provides practical, probability-based, and explainable team, analyst and broadcast podium prediction.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Singh, Jaswinder UNSPECIFIED
Subjects:	T Technology > TL Motor vehicles. Aeronautics. Astronautics Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning G Geography. Anthropology. Recreation > GV Recreation Leisure > Sports
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Ciara O'Brien
Date Deposited:	09 Sep 2026 08:30
Last Modified:	09 Sep 2026 08:30
URI:	https://norma.ncirl.ie/id/eprint/9904

Actions (login required)

View Item