Saturday 22 Aug, 15:00 · model dc-v1 · doesn't know lineups or injuries
Everton
44Index37Attack53Defense
LLDDL
38%EVE WIN
1–1
Most likely score
Crystal Palace
52Index56Attack47Defense
LDLDL
EVERTON 38%DRAW 28%CRYSTAL PALACE 34%
Over 2.5 goals
54%
Both teams score
59%
Expected goals λ
1.5 · 1.4
Why
- The model makes this 38% vs 34% — closer than the table suggests.
Everton — projected scorers
Crystal Palace — projected scorers
Projections assume recent minutes patterns — the model doesn't know lineups.
Score probabilities
| EVE ↓ · CRY → | 0 | 1 | 2 | 3 | 4 | 5 |
|---|---|---|---|---|---|---|
| 0 | 7% | 7% | 6% | |||
| 1 | 7% | 13% | 8% | 4% | ||
| 2 | 6% | 9% | 6% | |||
| 3 | 3% | 4% | ||||
| 4 | ||||||
| 5 |
Show the math
Home attack / defense-0.023 / 0.108
Away attack / defense0.166 / 0.063
Home advantage γ0.200
Dixon-Coles ρ-0.110
Decay ξ (per year)1.375
lambda_home = exp(c + atk_h - def_a + gamma); 10x10 Poisson grid with Dixon-Coles correction. Full method: /model.