Work in progress.

This page will cover the XGBoost models behind qualifying and finish position, the walk-forward season splits that keep training honest, and the backtest harness that grades every round against the oracle and two baselines. Still being written, check back soon or follow along on GitHub.

View on GitHub