List of betting outcomes for today where our Poisson model (based on xG/PredictedTotal/team strength) predicts a higher probability than the bookmaker's odds imply. Additional signals: team trends and last-10-matches form.
Model transparency
How often the model is right and how close to the market. Backtest of 19 595 picks, 2025-06-11 – 2026-06-09. The market is efficient — this is an honest picture, not a profit promise.
Calibration
| Model said | Actual | Picks |
|---|
| ~57,4% | 50,4% | 6 562 |
| ~64,3% | 51,8% | 10 216 |
| ~73,2% | 53,2% | 2 817 |
The model slightly overrates the favourite — we don't hide it.
By value-edge
| Edge | Winrate | CLV | Picks |
|---|
| <5% | 50,9% | +0,15% | 3 614 |
| 5-10% | 51,0% | +0,28% | 7 342 |
| 10-20% | 51,8% | +0,34% | 7 330 |
| 20%+ | 54,8% | +0,37% | 1 309 |
Higher edge means higher winrate and CLV — the ranking works. Average CLV over 16 280 picks: +0,29% (beating the closing line).
CLV (closing line value) — how much the pick's price beats the market close; the key honest quality metric. Analysis, not betting advice. 18+.