Track Record
Every prediction we publish is scored against reality. This page is our permanent, public accountability record. No retroactive adjustments. No cherry-picking.
Every prediction is scored against the actual outcome and surfaced on the chart below. We share both hits and misses — the regulation reset taught the model new patterns and our rookie predictions are still calibrating, so each scored race shifts the picture.
0.035
Brier score
Lower is better, 0 is perfect
+26%
Skill score
vs grid baseline
4
Races scored
2026 season
Brier = Reliability - Resolution + Uncertainty
Reliability
Calibration error — how well probabilities match reality
Resolution
How much predictions vary from the base rate
Uncertainty
Inherent unpredictability of F1 outcomes
| Model | Brier | Skill Score |
|---|---|---|
| The Data Driver | 0.035 | reference |
| Grid baseline | 0.047 | +25.7% |
| Championship baseline | 0.060 | +41.8% |
| Random uniform | 0.054 | +35.3% |
Skill Score = improvement over baseline. BSS = 1 - (model / reference).
When we predict a 30% chance, it happens ~30% of the time. Points on the diagonal indicate perfect calibration.
Log loss penalises confident wrong predictions more heavily than Brier score
log loss
brier
log loss
brier
log loss
brier
log loss
brier
Mean favourite probability
Higher sharpness = model is more decisive
32.3%
Trend: increasing — model confidence grows as season data accumulates
Walk-forward backtest
Validated on 114 races (2021-2025)
Model retrained before each race using only past data — no future leakage
0.041
Brier score
0.558
Spearman rank
| Race | Question | Prediction | Result | Brier |
|---|---|---|---|---|
Australian GP R01 | Russell wins the Australian Grand Prix | 26% | Correct | 0.548 |
Australian GP R01 | Safety car at Melbourne | 55% | Correct | 0.203 |
Australian GP R01 | Antonelli on the podium in Australia | 40% | Correct | 0.360 |
Australian GP R01 | Mercedes 1-2 in Australia | 15% | Correct | 0.723 |
Australian GP R01 | Hamilton top 5 in Australia | 62% | Correct | 0.144 |
Australian GP R01 | Verstappen on the podium in Melbourne | 42% | Wrong | 0.176 |
Australian GP R01 | Leclerc on the podium in Melbourne | 35% | Correct | 0.423 |
Australian GP R01 | More than 3 retirements in Australia | 40% | Correct | 0.360 |
Chinese GP R02 | Russell wins the Chinese Grand Prix | 31% | Wrong | 0.096 |
Chinese GP R02 | Safety car deployed in China | 62% | Correct | 0.144 |
Chinese GP R02 | Antonelli on the podium in China | 45% | Correct | 0.303 |
Chinese GP R02 | Hamilton podium at Shanghai | 30% | Correct | 0.490 |
Chinese GP R02 | Same winner in Sprint and Race in China | 38% | Wrong | 0.144 |
Chinese GP R02 | More than 3 retirements at the Chinese GP | 42% | Correct | 0.336 |
Chinese GP R02 | Mercedes 1-2 in China | 18% | Correct | 0.672 |
Chinese GP R02 | Verstappen on the podium in China | 28% | Wrong | 0.078 |
Australian GP
Russell wins the Australian Grand Prix
Australian GP
Safety car at Melbourne
Australian GP
Antonelli on the podium in Australia
Australian GP
Mercedes 1-2 in Australia
Australian GP
Hamilton top 5 in Australia
Australian GP
Verstappen on the podium in Melbourne
Australian GP
Leclerc on the podium in Melbourne
Australian GP
More than 3 retirements in Australia
Chinese GP
Russell wins the Chinese Grand Prix
Chinese GP
Safety car deployed in China
Chinese GP
Antonelli on the podium in China
Chinese GP
Hamilton podium at Shanghai
Chinese GP
Same winner in Sprint and Race in China
Chinese GP
More than 3 retirements at the Chinese GP
Chinese GP
Mercedes 1-2 in China
Chinese GP
Verstappen on the podium in China
Transparency
Predictions published before each race. Track record computed automatically after each race using Brier score. We never modify predictions retroactively. Every probability is timestamped and immutable.
Scoring methodology: Brier score = mean squared error of probabilistic predictions. Skill score = 1 - (model Brier / baseline Brier). Grid baseline uses qualifying order.