THE MODEL, ON THE RECORD
Every predictionand how it went
Before each Grand Prix we run it hundreds of times from the real starting grid and publish the odds. Afterwards we put the real classification next to them and count the damage. Nothing here is written after the fact: every figure is computed from the published predictions.
The favourite won 4 of 6 races. With this few races these numbers are an anecdote, not a track record: they are here because they grow on their own every other Sunday, and because publishing them before the race is the only version of this that means anything.
Race by race
| Grand Prix | We said | It was | Error |
|---|---|---|---|
| Bahrain GP in Malaysia 2026Round 16 · Sepang | Max Verstappen23.5% | Max Verstappen ✓9/10 in the points | 3.1places |
| Azerbaijan GP 2026Round 15 · Baku City Circuit | George Russell48.3% | George Russell ✓6/10 in the points | 4.0places |
| Spanish GP 2026Round 14 · Circuit de Madrid | Kimi Antonelli21.5% | Kimi Antonelli ✓9/10 in the points | 2.4places |
| Italian GP 2026Round 13 · Monza | George Russell22.8% | Kimi Antonelli ✗7/10 in the points | 2.6places |
| Dutch GP 2026Round 12 · Zandvoort | George Russell25.0% | Lando Norris ✗8/10 in the points | 2.2places |
| Hungarian GP 2026Round 11 · Hungaroring | Lando Norris22.7% | Lando Norris ✓8/10 in the points | 2.7places |
Where it gets it wrong
The biggest misses from every race so far. They are not spread evenly: the front of the grid is where the car decides, and the model is good at that. The mistakes pile up in the midfield, where strategy and judgement decide instead.
- Oscar PiastriAzerbaijan GP 2026P3.7 → P139.3
- Nico HulkenbergDutch GP 2026P15.7 → P87.7
- Arvid LindbladAzerbaijan GP 2026P14.4 → P77.4
- Esteban OconAzerbaijan GP 2026P14.3 → P86.3
- Nico HulkenbergAzerbaijan GP 2026P17.3 → P116.3
- Arvid LindbladBahrain GP in Malaysia 2026P16.2 → P106.2