Betting
Model performance
Two questions, answered separately: whether the model's probabilities meant what they said, across every selection it priced on a played match, and how the selections that met the published qualification rule have done at an equal stake of one unit. The first is read over every forecast and the second over the settled qualified selections only, each assessed at its last evaluation before kickoff, so neither answers the other.
Every forecast
Model calibration
When the model gives an outcome 60%, does it happen about 60% of the time? Every selection it priced on a played match, grouped by the probability it gave, against how often that selection happened.
685 matches 1,370 forecasts
Scored from the last projection made before kickoff. A projection written after a match started is never counted.
- Base rate
- 0.245
- Against base rate
- 1.1% worse
- Base rate
- 0.683
- Against base rate
- 0.8% worse
Lower is better for both. The base rate is a forecast that gave every selection how often it happened over this same sample, which it could only have known with hindsight.
Model probability → · ↑ Happened
| Model probability | Forecasts | Model avg | Happened | Gap |
|---|---|---|---|---|
| 0–10% | 0 | — | — | Too few |
| 10–20% | 0 | — | — | Too few |
| 20–30% | 48 | 26.7% | 33.3% | +6.7 pp |
| 30–40% | 276 | 35.8% | 43.5% | +7.7 pp |
| 40–50% | 361 | 44.8% | 43.2% | −1.6 pp |
| 50–60% | 361 | 55.2% | 56.8% | +1.6 pp |
| 60–70% | 276 | 64.2% | 56.5% | −7.7 pp |
| 70–80% | 48 | 73.3% | 66.7% | −6.7 pp |
| 80–90% | 0 | — | — | Too few |
| 90–100% | 0 | — | — | Too few |
- Within 2 pp
- 2–4 pp off
- More than 4 pp off
Qualified selection record
All competitions, all seasons
- Settled qualified selections
- 876
- across 412 matches
- Won–lost–void
- 295 won, 572 lost, 9 void
- 34.0% of decided selections won
- Units · ROI
- −94.36u
- −10.8% ROI · equal stake of one unit
- Avg quoted edge · avg price
- +16.5%
- at evaluation · average price 3.41
Qualification definition updated on 2 October 2026. Historical figures were recalculated using the same rules as the live board. See methodology history
- Sample
- 876 settled qualified selections on 412 matches, winners and losers alike.
- Correlation
- Selections from the same match may be correlated. The selection count is not a count of independent experiments; the match count is the smaller figure to read it against.
- Price and moment
- Historical selections use the price paired with the last qualifying pre-match evaluation. Prices were current when evaluated and may have moved before kickoff.
- Excluded pairs
- 343 settled rows whose price was captured after the evaluation that was meant to price it are left out and counted here.
- Odds source
- One bookmaker's price, the first of Bet365 → Paddy Power → Pinnacle to quote the market. Not the highest price on the market and not a consensus.
- Availability
- Prices as captured, not as matched. A price may have moved, been limited or been withdrawn before a bet could be placed.
Calibration of qualified selections
The record's own selections, scored as the calibration above scores every forecast: the model's average probability against how often the selections happened. Voids are counted and not scored; pending selections never enter.
876 settled qualified selections · 867 decided · 9 void
Over the decided selections. The base rate is a forecast that gave each selection how often it happened over this same subset, which it could only have known with hindsight; lower is better.
Qualified selections have occurred less often than their model probabilities implied over the recorded sample, which helps explain why a positive quoted edge has not translated into positive realised returns.
This is a different population from the calibration above: only settled selections that met the qualification rule at their last pre-match evaluation. It describes the recorded sample, not a fixed property of the model.
How the record is kept
Every settled selection that was a qualified edge at its last evaluation before kickoff, in every competition the site prices, at the bookmaker's price paired with that evaluation. Nothing is re-priced, and nothing written after kickoff is counted.
An equal stake of one unit per qualified selection: a win returns the price less the stake, a loss costs the stake, and a void returns it and still counts as staked. ROI is quoted only from 30 settled selections.
Pending selections, rows without a price, the excluded pairs and the daily combinations are left out. The full rules, how a selection settles and the history of the qualification rule are on the methodology page.
Detailed breakdowns
The same record cut by market and by season, every row with the same ROI floor. They are historical breakdowns in a fixed order: nothing here is sorted by return.
Performance by market Show Hide
One row per selection of a market, in the registry's order. A historical breakdown, not a recommendation to prefer the markets with a positive return. Measured means the model showed forecasting skill against the base rate in that market when its skill was measured; it is not a measure of profit.
| Market | Model status | Settled | W–L–V | Units | ROI |
|---|---|---|---|---|---|
| Home team to score | Measured | 55 | 33–21–1 | −6.65u | −12.1% |
| Home team not to score | Measured | 70 | 11–59–0 | −4.50u | −6.4% |
| Away team to score | Measured | 67 | 34–32–1 | −11.08u | −16.5% |
| Away team not to score | Measured | 72 | 11–60–1 | −17.67u | −24.5% |
| Home win | Measured | 130 | 28–100–2 | −42.81u | −32.9% |
| Away win | Measured | 118 | 32–85–1 | +18.32u | +15.5% |
| Home team 2+ goals | Measured | 75 | 22–52–1 | −16.91u | −22.5% |
| Home team at most 1 goal | Measured | 106 | 56–50–0 | +11.21u | +10.6% |
| Away team 2+ goals | Measured | 75 | 15–59–1 | −16.68u | −22.2% |
| Away team at most 1 goal | Measured | 108 | 53–54–1 | −7.61u | −7.0% |
Performance by season Show Hide
One row per season, newest first, with the same ROI floor. No season is ranked above another.
| Season | Settled | W–L–V | Units | ROI |
|---|---|---|---|---|
| 2026/2027 | 876 | 295–572–9 | −94.36u | −10.8% |
A dash in the ROI column means fewer than 30 settled selections.