AI Strategy Engine

Predictive Strategy

Machine Learning model predicting GP outcomes based on telemetry features, historical qualifying pace, and real-time tire degradation.

ML Pipeline Status: Standby

No trained model matching current season telemetry is loaded.

Strategy Predictions Architecture

The BoxUp AI predictor relies on an XGBoost classifier combined with SHAP feature explanations. It evaluates sector times from Qualifying, stint times from FP2 long runs, and ambient session weather conditions to compute the winning probability of each driver on the grid.

HOW TO TRAIN THE MODEL

To enable ML predictions, you need to execute the training script locally. This will process the historical Ergast / Jolpica database and dump model parameters into the backend.

# 1. Navigate to the backend directory
cd backend

# 2. Run the training command using uv
uv run python -m ml.train
Once the model is trained, the API will automatically activate predictions on the frontend.