| Score | Likelihood | Prob | Fair odds | |
|---|---|---|---|---|
| 2–0MODE | 13.4% | 7.45 | ||
| 3–0 | 12.2% | 8.23 | ||
| 1–0 | 9.9% | 10.13 | ||
| 2–1 | 8.3% | 12.05 | ||
| 4–0 | 8.3% | 12.10 | ||
| 3–1 | 7.5% | 13.29 | ||
| 1–1 | 6.1% | 16.38 | ||
| 4–1 | 5.1% | 19.56 |
FC Twente '65 scores shown first. Bivariate Poisson with team xG from recent form — most-likely score ≠ most-likely winner.
Probabilities derived from a Poisson model on expected goals. Edge % shown only when the book posts both sides of a market.
| Combo | Joint prob | Naïve (book) | Corr. | Fair odds |
|---|---|---|---|---|
Home win + BTTS home + btts | 30.8% | 34.9% | -12% | 3.25 |
Home win + Over 2.5 home + total_goals_2.5 | 57.7% | 52.5% | +10% | 1.73 |
Away win + BTTS away + btts | 3.1% | 2.7% | +15% | 32.66 |
Away win + Over 2.5 away + total_goals_2.5 | 3.2% | 4.0% | -19% | 30.96 |
Joint probability comes from the score grid (the truth). Naïve is what a book gets by multiplying legs — they undershoot correlated positive legs like "home + BTTS" and overshoot anti-correlated ones like "draw + over 2.5". Buy positive correlation; avoid negative.
Model probabilities FC Twente '65 51.7% · Draw 26.0% · Sparta Rotterdam 22.3% Why FC Twente '65 is favoured Elo essentially level (1500 vs 1500) — winner depends on context. Recent: FC Twente '65 6W-3D-1L · Sparta Rotterdam 1W-4D-5L (last 10). Goal differential per game: FC Twente '65 +0.7, Sparta Rotterdam -1.0. Form depth FC Twente '65: avg 1.8 goals scored, 1.1 conceded per game (3/5 at home, 3/5 on the road). Sparta Rotterdam: avg 0.9 goals scored, 1.9 conceded per game (1/5 at home, 0/5 on the road). FC Twente '65 on a 2-draw streak. Goals expectation Model expects ~2.9 total goals (FC Twente '65 1.9, Sparta Rotterdam 1.0) — above the 2.5-goal mid line. But — counter-arguments Elo confidence is low (5%) — ratings are based on a thin sample, so the prediction itself is wider than the headline number suggests. Head-to-head 5 prior meetings — FC Twente '65 0W, 5D, Sparta Rotterdam 0W (draws are common). Last 3: 1-5, 0-2, 1-1. Confidence note Elo data on these teams is thin (5% sample-size confidence); headline number is sharper than the underlying signal — treat it as directional, not absolute.