Prior record: 2-0-0 across 2 since 2025. Most recent 2026, 1-0 in comp-fd-DED. The present numbers are 1417 against 1566, and the model turns that into 25.7% / 25.1% / 49.2%.
The ratings above are current values, not a pre-match snapshot. Rounded, that is 1-2. Between best and worst outcome there is 24.1 points of daylight. Neither figure is thin: 34 and 34 matches respectively sit behind them. Raw profiles are 0.85 and 1.38 against 1.21 and 0.91. Outcomes are scattered: top of the list is 1-2 at 9.7%. From the 1500 baseline they currently stand at -83 and +66. 149 points separate the two ratings — a wide margin. The higher rating belongs to the away side. Someone wins 74.9% of the time by this model. Scorelines above 8%: 3. The model's expected-goals figures are 1.15 and 2.24. Top of the distribution: Utrecht, 49.2%. A 23.5-point margin separates the top two outcomes. The match finished 3-3. By expected share of the points: 38.1% and 61.9%. Neither component contradicts the other: Utrecht leads 1.21-0.85 forward and 0.91-1.38 back.
Computed by elo-poisson-v1 from the whole corpus rather than republished. Analytical only.