Home win
50.89%
Erzurumspor
TUR1 · Süper Lig · Match Analysis
Erzurumspor vs
Eyüpspor
Model-based match probabilities derived from multi-season xG data, team strength, venue performance, recent form, Elo ratings and league context.
Model output
Match probabilities.
Probabilities are model estimates, not certainties or betting guarantees.
Draw
26.95%
X
Away win
22.16%
Eyüpspor
Score model
Most likely scorelines.
Exact-score probabilities from the Poisson + Dixon-Coles goal model.
Additional scorelines
1:25.95%
0:15.78%
3:15.31%
3:05.19%
2:24.93%
Team context
Strength and recent form.
Venue-adjusted team strength, Elo rating and recent xG performance.
Home
Erzurumspor
Elo1501.0
Overall attack0.83
Overall defense1.47
Home attack0.74
Home defense1.59
Form signal-0.221
Last 5 xGF / xGA1.01 / 1.13
Last 10 xGF / xGA0.91 / 1.33
Last 10 xGD-0.42
League statusNEW
Away
Eyüpspor
Elo1437.3
Overall attack0.97
Overall defense1.70
Away attack0.68
Away defense1.97
Form signal-1.741
Last 5 xGF / xGA1.13 / 3.00
Last 10 xGF / xGA1.00 / 2.51
Last 10 xGD-1.51
League statusSTABLE · from TUR1 2025/26
Elo difference
+63.7
Elo is one input among several and does not determine the final 1X2 probability on its own.
Head-to-head
Historical matchup context.
H2H values provide context and are not used as a standalone prediction.
Goal distribution
Total goals profile.
The model's probability distribution for the total number of goals.
08.12%
115.90%
225.88%
321.99%
414.72%
57.89%
63.52%
7+1.98%
Technical detail
How the model reached the estimate.
Detailed diagnostics are kept below the headline probabilities so the public report remains readable.
League baseline and model setup
LeagueTUR1
Loaded teams30
Loaded historical matches1433
Target season2026/27
Previous seasons in history4
Season window2022/23 → 2023/24 → 2024/25 → 2025/26 → 2026/27
League tierTier 1
Historical league matches1433
xG matches710 / 1433
Effective xG matches476.93
Observed home xG1.57
Observed away xG1.21
Model home xG baseline1.53
Model away xG baseline1.24
Total league xG baseline2.78
Home-advantage weight0.806
Model configdata/models/TUR1.json
Team attack half-life90.0 days
Team defense half-life240.0 days
Attack strength weight0.806
Defense strength weight1.081
Venue prior matches4.94
DistributionPoisson + Dixon-Coles
Dixon-Coles rho-0.108 (effective -0.107)
Dixon-Coles decay0.016
Goal overdispersion0.000
Implied goal variance H / A1.65 / 1.02
Relative strengths and form diagnostics
Erzurumspor
Home attack strength0.861
Home defense weakness1.065
Base lambda home1.61
Attack form xGF level0.978
Defense form xGA level (lower is better)1.199
Opponent-adjusted attack residual-0.361
Opponent-adjusted defense residual (higher is better)+0.240
Opponent-adjustment weight0.000
Model attack signal-0.410
Model defense signal+0.189
Opponent blend0.022
Form matches 5 / 105 / 7
Last 5 xGF / xGA / xGD1.01 / 1.13 / -0.11
Last 10 xGF / xGA / xGD0.91 / 1.33 / -0.42
Eyüpspor
Away attack strength0.827
Away defense weakness1.168
Base lambda away1.14
Attack form xGF level1.084
Defense form xGA level (lower is better)2.826
Opponent-adjusted attack residual-0.306
Opponent-adjusted defense residual (higher is better)-1.436
Opponent-adjustment weight0.000
Model attack signal-0.304
Model defense signal-1.438
Opponent blend0.022
Form matches 5 / 105 / 10
Last 5 xGF / xGA / xGD1.13 / 3.00 / -1.87
Last 10 xGF / xGA / xGD1.00 / 2.51 / -1.51
Historical data coverage
Erzurumspor
Total matches7
Home / away matches3 / 4
xG matches7 / 7
xG home / away coverage3 / 3 · 4 / 4
Effective total5.19
Effective attack / defense5.19 / 6.24
Effective home matches2.18
Effective home attack / defense2.18 / 2.66
Eyüpspor
Total matches77
Home / away matches39 / 38
xG matches77 / 77
xG home / away coverage39 / 39 · 38 / 38
Effective total10.02
Effective attack / defense10.02 / 27.13
Effective away matches4.62
Effective away attack / defense4.62 / 13.11
Historical data coverage
26.83%
This measures usable historical evidence. It is not the probability that the prediction is correct.
Validation
Historical model performance.
Leakage-safe backtesting compares the fitted model with a league baseline over 1413 historical matches.
Positive improvement means the model performed better than the league baseline.
Total goals log-loss vs baseline+0.69%
1X2 log-loss vs baseline+6.65%
BTTS Brier vs baseline-0.19%
Over 2.5 Brier vs baseline+1.62%
Full backtest metrics
Model
MAE total goals1.3470
Exact score log-loss3.0002
Total goals Brier0.8314
Total goals log-loss1.9173
1X2 Brier0.5985
1X2 log-loss1.0021
BTTS Brier0.2491
Over 1.5 Brier0.1698
Over 2.5 Brier0.2467
Over 3.5 Brier0.2132
League baseline
MAE total goals1.3665
Exact score log-loss3.0942
Total goals Brier0.8341
Total goals log-loss1.9306
1X2 Brier0.6486
1X2 log-loss1.0735
BTTS Brier0.2486
Over 1.5 Brier0.1715
Over 2.5 Brier0.2507
Over 3.5 Brier0.2177
Calibration diagnostics
Home win · ECE 5.66%
0- 10% n=8 avg= 7.9% actual= 25.0% 10- 20% n=48 avg= 14.6% actual= 12.5% 20- 30% n=133 avg= 25.6% actual= 18.0% 30- 40% n=454 avg= 36.4% actual= 36.8% 40- 50% n=506 avg= 44.0% actual= 51.6% 50- 60% n=148 avg= 54.3% actual= 64.9% 60- 70% n=59 avg= 63.9% actual= 72.9% 70- 80% n=40 avg= 74.9% actual= 82.5% 80- 90% n=16 avg= 83.7% actual= 62.5% 90-100% n=1 avg= 92.6% actual=100.0%
Draw · ECE 2.66%
0- 10% n=5 avg= 8.0% actual= 0.0% 10- 20% n=80 avg= 16.4% actual= 21.2% 20- 30% n=1268 avg= 27.5% actual= 25.1% 30- 40% n=60 avg= 31.0% actual= 26.7%
Away win · ECE 4.06%
0- 10% n=52 avg= 6.9% actual= 3.8% 10- 20% n=141 avg= 16.0% actual= 12.8% 20- 30% n=455 avg= 26.1% actual= 20.4% 30- 40% n=548 avg= 33.6% actual= 32.5% 40- 50% n=127 avg= 44.3% actual= 52.8% 50- 60% n=51 avg= 54.2% actual= 64.7% 60- 70% n=21 avg= 64.7% actual= 71.4% 70- 80% n=17 avg= 72.7% actual= 76.5% 80- 90% n=1 avg= 89.6% actual= 0.0%
BTTS · ECE 1.43%
30- 40% n=9 avg= 38.7% actual= 77.8% 40- 50% n=150 avg= 47.3% actual= 54.0% 50- 60% n=1197 avg= 54.8% actual= 54.8% 60- 70% n=57 avg= 61.7% actual= 50.9%
Over 2.5 · ECE 4.62%
20- 30% n=1 avg= 29.7% actual=100.0% 30- 40% n=29 avg= 36.8% actual= 41.4% 40- 50% n=772 avg= 48.4% actual= 52.8% 50- 60% n=501 avg= 53.3% actual= 58.3% 60- 70% n=88 avg= 63.6% actual= 67.0% 70- 80% n=20 avg= 73.6% actual= 75.0% 80- 90% n=2 avg= 83.8% actual= 50.0%
Walk-forward - latest periods
| Period | N | Model 1X2 LL | Baseline 1X2 LL | Improvement |
|---|---|---|---|---|
| 2025-09 | 31 | 0.9927 | 1.1163 | +11.08% |
| 2025-10 | 30 | 0.9993 | 1.1113 | +10.08% |
| 2025-11 | 33 | 1.0018 | 1.1243 | +10.90% |
| 2025-12 | 29 | 0.9805 | 1.1075 | +11.47% |
| 2026-01 | 24 | 1.0013 | 1.1478 | +12.76% |
| 2026-02 | 36 | 0.9076 | 1.0454 | +13.17% |
| 2026-03 | 28 | 0.9634 | 1.0645 | +9.49% |
| 2026-04 | 38 | 0.9983 | 1.0348 | +3.53% |
| 2026-05 | 27 | 1.1044 | 1.0203 | -8.24% |
| 2026-08 | 27 | 1.0459 | 1.0716 | +2.40% |
| 2026-09 | 27 | 1.0542 | 1.0368 | -1.68% |
| 2026-10 | 9 | 0.8923 | 1.1785 | +24.29% |
Football is data.
NILYTIX turns match data into transparent model probabilities and measurable analytical signals.