Home/Track record
Historical match forecast track record: every forecast checked against the result
PitchTensor stores every forecast at the moment it is made and compares it with the real result afterwards. This is the forecast verification, with the numbers updated daily.
// forecasts played 2026-10-10 to 2026-10-30 · figures as of 2026-10-31 · refreshed daily
Do the percentages hold?
Outcomes are grouped by the chance they were given. A group given about 60% should happen about six times in ten.
| Chance given | Outcomes | Happened |
|---|---|---|
| 0-20% | 138 | 17% |
| 20-35% | 408 | 20% |
| 35-50% | 170 | 52% |
| 50-65% | 90 | 64% |
| 65%+ | 46 | 70% |
By competition
How often the most likely outcome happened, by competition.
| Competition | Forecasts | Most likely outcome happened |
|---|---|---|
| Serie A | 40 | 70% |
| UEFA Champions League | 36 | 64% |
| UEFA Europa League | 36 | 50% |
| UEFA Conference League | 36 | 56% |
| Premier League | 30 | 43% |
| Ligue 1 | 28 | 64% |
| Bundesliga | 27 | 63% |
| Süper Lig | 27 | 78% |
| La Liga | 24 | 67% |
Forecast verification
Predictive data verification: three rules that keep the record honest
Only forecasts made in advance count
A forecast enters the record only if it was made before match day. Revisions after the fact never count, so the record cannot be flattered.
Compared with a flat guess
The forecast error is compared with giving every match the same average chances. A smaller error means the forecasts carry real information.
Wins and misses alike
Every checked forecast is in the numbers, not a selection. The number of forecasts behind each figure is shown, so small samples are visible.
Reading the numbers
What to look for
Most likely outcome happened is how often the outcome with the highest chance was the one that happened. In football, even a good forecast is often wrong about a single match.
Forecast error lower than a plain baseline shows whether the forecasts say more than the average does. A positive gap means they do.
Do the percentages hold? is the most telling check: outcomes given about 60% should happen about 60% of the time. With few forecasts in a group, the figure is rough.
Request access
Try the forecasts on your own pages
The record shows how the forecasts have done. To see them on your own pages, or to look closer at the numbers, leave your email: PitchTensor will set up a free 14-day pilot or answer your questions. No commitment.
Thanks, you're on the list. You will hear from PitchTensor soon.
// your email is used only to contact you about PitchTensor access. No spam!