Prediction history

Preserved forecasts across race weekends.

Prospective predictions are kept separate from legacy descriptive records and historical backtests, so public performance claims use the right evidence.

Prospective events: 1Settled: 1Aggregate gap MAE: 0.679s

Coverage note

Why some race weekends may be missing

This archive is intentionally not a complete Formula 1 calendar. A past Grand Prix may have no prediction—or may not appear here at all—because Sprint weekends are not currently supported, because monitoring was introduced during the season, or because ongoing model, data-pipeline and reliability updates—or a software issue later identified and fixed— prevented a trustworthy forecast at the time. Missing predictions are never recreated after qualifying.

Valid prospective evidence

Past predictions

Open an event to inspect its preserved forecast and, where available, official comparison.

2026

Spanish Grand Prix 🇪🇸

Qualifying completedPrediction after FP3Coverage 20/20

Gap MAE

0.679s

Open event +

Preserved prediction

1Mercedes logo

ANT

Mercedes

+0.238s

↓ 1 over

2McLaren logo

NOR

McLaren

+0.457s

↑ 1 under

3McLaren logo

PIA

McLaren

+0.515s

↓ 4 over

4Mercedes logo

RUS

Mercedes

+0.538s

↓ 2 over

5Red Bull Racing logo

VER

Red Bull Racing

+0.587s

↑ 2 under

6Ferrari logo

LEC

Ferrari

+0.616s

↑ 1 under

7Red Bull Racing logo

LAW

Red Bull Racing

+1.198s

↓ 1 over

8Audi logo

HUL

Audi

+1.273s

↓ 3 over

9Ferrari logo

HAM

Ferrari

+1.582s

↑ 5 under

10Racing Bulls logo

LIN

Racing Bulls

+1.656s

Exact

11Audi logo

BOR

Audi

+1.658s

↓ 1 over

12Haas logo

OCO

Haas

+1.780s

↓ 1 over

13Alpine logo

COL

Alpine

+1.812s

↑ 4 under

14Alpine logo

GAS

Alpine

+1.901s

Exact

15Racing Bulls logo

TSU

Racing Bulls

+2.292s

Exact

16Williams logo

SAI

Williams

+2.392s

↓ 1 over

17Aston Martin logo

ALO

Aston Martin

+2.412s

↓ 1 over

18Cadillac logo

BOT

Cadillac

+2.418s

↓ 2 over

19Cadillac logo

PER

Cadillac

+2.676s

Exact

20Williams logo

ALB

Williams

+4.214s

↑ 4 under

Historical backtest context

Model development evidence

Backtests simulate earlier race weekends for model development. They are useful, but are not counted as prospective public predictions.

44 historical events39 successful foldsWalk-forward evaluation