What if? Run it.
Change the grid, put a driver in another car, retire the favourite — then play out the rest of the season. Every number comes from TELOS's race model, trained on every Grand Prix since 2015.
| Driver | Car | Grid | Win | DNF |
|---|
The starting point is the official TELOS pick for the race: a statistical model trained on every Grand Prix since 2015 (qualifying position, team form, driver form). Your changes are applied through effects the data measured, not guessed:
- Grid — moving from P2 to P1 multiplies a driver's win odds by about , the average over 248 races.
- Car — each team's car is rated by the Driver Rating model, which separates driver from machinery; putting a driver in another car applies that difference.
- DNF removes a driver; Chaos spreads the odds out, the way safety cars, crashes and rain do. It is a dial for randomness, not a weather model.
The championship plays every remaining race and sprint thousands of times from today's standings, with each driver's strength from their skill, their car and current form.
What it can't see: track-specific car strengths, weather, strategy and upgrades. Probabilities, not promises — the model's live record is on /accuracy, and every test behind these numbers is on the model cards.