Podium Predictor

A machine-learning model for Formula 1 podium finishes, trained on real FastF1 data (2021–2024).

Predicted podium

Full grid — podium probability

Predicted constructors’ points

What the model learned

SHAP feature-importance summary

How it works. For every driver the model reads the factors that decide a race — starting grid, qualifying pace, recent form, race-pace trend, reliability, constructor strength and weather — and an XGBoost classifier estimates the probability of a top-three finish.

What it learned, in order of impact: grid position, then qualifying pace, then championship form, then constructor strength. The ordering matches how a race engineer would rank them — learned entirely from data.