25 October 2026 · Circuit of The Americas · Data cutoff: 20 October
US Grand Prix probabilities
The main forecast uses 3 million joint race simulations. The same rules apply to every driver. Practice and qualifying had not taken place at the data cutoff, so the starting grid is also uncertain.
All drivers
Sorted by win probability. DNF/DNS and DSQ overlap with result positions; they are not additional categories to add to 100%.
Probability of each result position
Each row and column sums to 100%. The results list covers all 22 entrants, with retirements ordered by completed laps and time. Hover over a cell for its exact probability and observation count.
How to interpret these results
This is not an engineering simulator based on clean-air telemetry. Pace proxies inferred from official qualifying and race results are combined with lap, tyre, pit-stop, traffic, weather and reliability processes. The three form models have equal weights in the main forecast. A large sample reduces Monte Carlo error; it does not make the model assumptions certain.
The wet-race scenario is not a rain-probability forecast. The McLaren scenario tests an assumed 0.15-second-per-lap benefit from the announced upgrade; this gain has not been measured. No confirmed Austin grid penalty was verified at the data cutoff.
The PDF report explains the method, retrospective race tests, sources and limitations. The Python scripts in the data/code package can be rerun using the recorded random seeds.
Official race schedule · Official results · Pirelli compound selection