The Race Prediction Machine — Predictions for AMA Motocross & Supercross
Predictions Before Every Round
The Race Prediction Machine (RPM) is a data-driven model that generates predicted race outcomes before every round of AMA Motocross and Supercross.
Built from the same data-focused mindset that drives RP Moto product design, RPM uses machine learning trained on race data going back to 2013. Before each round, it ingests inputs such as qualifying results and weather conditions and produces its predicted finishing order.
The predictions shown below are generated fresh for each race week and are not manually adjusted or overridden.
Jump to: Weekly Predictions · How to Use · What It Considers · Past Results
Weekly Predictions
Predicted overall finishing positions for this week’s race.
250 Class — Predicted Finish Order
450 Class — Predicted Finish Order
Show First-Lap Leader (FFL) Probability Charts
250 Class — FFL Probabilities
450 Class — FFL Probabilities
How to Use the RPM Predictions
The short video below walks through how to interpret RPM output and use it as a reference when making picks.
What the RPM Takes Into Account
After every race, RPM retrains using the latest results to refine its understanding of which factors most influence outcomes.
The chart below shows the relative importance of the inputs used by the model. These weightings can change slightly from week to week as new data is incorporated.
Riders with limited historical data or higher incident rates are naturally more difficult to predict, which is reflected in the model’s output.
