← All projects

RaceIQ WEC

FIA WEC · endurance

Experimental

World endurance on the MotorsportVerse core. Endurance is multi-class, so every round is forecast per class (Hypercar, LMP2 at Le Mans, and LMGT3), and the competitor is the car entry, not a driver. Each entry's chance is an ensemble of recency (the last race and this-season form) with a smoothed cross-season entry rating and the team's operational strength — validated by walk-forward (2021-2026) to beat both naive baselines (last-race order and season-form order) on win- and podium-Brier over the full history, while tracking the best baseline in the current era. simulation-based per-class win/podium probabilities, finishing-range intervals, and a per-class drivers' championship projection. Le Mans is included as a round (24 hours, with its extra LMP2 field). Runs on the real season pulled from the official Al Kamel timing archive into a committed, reviewable snapshot (2021-2026 corpus). Deployed site + scheduled refresh; forward-eval grades every round honestly against the last-race and season-form baselines.

Experimental. This product runs on the real season, but its forward accuracy is still accruing over live rounds — read the forecasts with that track record in mind.
endurance
Category
2
Datasets
5
Model components
8
Core modules

Datasets

  • Al Kamel FIA WEC timing archive (official, pulled to a committed snapshot)
  • 2021-2026 WEC corpus (offline, reproducible)

Model components

  • per-class-entry-elo-cross-season
  • recency-form-ensemble
  • Probability calibration
  • Championship simulation
  • Standings projection

Shared core modules

  • Probability calibration
  • Championship simulation
  • Standings projection
  • Driver-skill ratings
  • Uncertainty intervals
  • Forward evaluation
  • Leakage guards
  • Predictor interfaces

Tags

  • endurance
  • multi-class
  • le-mans
  • car-entry
  • calibrated probabilities
  • race simulation
Maintained byMotorsportVerse

Registered 2026-08-11

Related projects

View all →