Track record
Accuracy
How the forecasts have actually scored against real results — measured across every scored class-round, with nothing left out. Endurance grids are large, so a car landing within a few places of its prediction is a strong result.
Evidence
27 Sept 2026, 12:21 UTCOver 6 paired rounds, the model no difference demonstrated (last race order).
| Baseline | Race | Rounds | Model | Baseline | Gain | Verdict |
|---|---|---|---|---|---|---|
| Last race order | HYPERCAR | 6 | 3.83 | 4.20 | +0.37 | no difference demonstrated |
| Last race order | LMGT3 | 6 | 5.29 | 5.09 | -0.20 | no difference demonstrated |
| Last race order | LMP2 | 1 | 5.11 | 2.11 | -3.00 | too few rounds to say |
| Season form order | HYPERCAR | 6 | 3.83 | 3.91 | +0.08 | no difference demonstrated |
| Season form order | LMGT3 | 6 | 5.29 | 4.71 | -0.58 | does NOT beat the baseline |
| Season form order | LMP2 | 1 | 5.11 | 4.11 | -1.00 | too few rounds to say |
Paired bootstrap on the round-by-round difference: 0.368, 95% CI [-0.137, 0.882] in positions gained. Positive means the model is closer to the real finishing order than last race order is; the interval covers zero, so no difference has been demonstrated.
forward evaluation — each round was forecast before it ran and scored afterwards; not a backtest. Historical replays and the forward record are computed separately and never merged.
- The model does not beat season form order on every measured comparison. That is stated rather than hidden.
- Some comparisons straddle zero: a difference has not been demonstrated, in either direction.
- Metrics are only comparable within this series. A position error over this field size means nothing next to another series' number.
By class
Mean err.
4.37
Podium
22%
Winner
0%
Mean err.
5.90
Podium
17%
Winner
0%
Mean err.
5.33
Podium
0%
Winner
0%
Round by round
| Round | Cars | Mean error | Podium | Exact | Within 3 | Winner |
|---|---|---|---|---|---|---|
| R1Imola | 17 | 4.35 | 2/3 | 3/17 | 7/17 | ✗ |
| R2Spa Francorchamps | 13 | 4.77 | 0/3 | 1/13 | 5/13 | ✗ |
| R3Le Mans | 14 | 3.50 | 1/3 | 1/14 | 7/14 | ✗ |
| R4Sao Paulo | 17 | 4.24 | 1/3 | 1/17 | 10/17 | ✗ |
| R5Circuit Of The Americas | 16 | 5.25 | 0/3 | 0/16 | 4/16 | ✗ |
| R6Fuji Speedway | 15 | 4.13 | 0/3 | 1/15 | 9/15 | ✗ |
Does the forecast beat guesswork?
Combined win + podium forecast error across 13 scored class-rounds — lower is better — measured against two naive baselines: simply repeating the last race’s order, and ranking cars by their season points.
Last-race form
baseline error 0.228
Season form
baseline error 0.215
“Matches it” means the model is statistically no worse than that baseline over the rounds scored so far. On a young season with few rounds, matching a strong season-form baseline while beating last-race form is an honest, expected result.
Calibration
Probabilities are calibrated on 6 real rounds of results, per class. Calibrated on real WEC results, per class stratum.
Continuous evaluation
Every round is re-scored against simple baselines as results come in, so the track record above updates honestly over the season — no cherry-picking, nothing dropped.