Home win
35.95%
Everton
ENG1 · Premier League · Match Analysis
Everton vs
Chelsea
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
27.44%
X
Away win
36.62%
Chelsea
Score model
Most likely scorelines.
Exact-score probabilities from the Poisson + Dixon-Coles goal model.
Additional scorelines
0:07.32%
0:26.10%
2:05.97%
2:25.73%
1:33.86%
Team context
Strength and recent form.
Venue-adjusted team strength, Elo rating and recent xG performance.
Home
Everton
Elo1516.5
Overall attack1.28
Overall defense1.42
Home attack1.36
Home defense1.28
Form signal+0.308
Last 5 xGF / xGA1.62 / 1.24
Last 10 xGF / xGA1.51 / 1.33
Last 10 xGD+0.18
League statusSTABLE · from ENG1 2025/26
Away
Chelsea
Elo1529.3
Overall attack1.67
Overall defense1.42
Away attack1.51
Away defense1.54
Form signal-0.609
Last 5 xGF / xGA1.12 / 1.79
Last 10 xGF / xGA1.18 / 1.69
Last 10 xGD-0.51
League statusSTABLE · from ENG1 2025/26
Elo difference
-12.8
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.
07.32%
115.58%
225.09%
322.17%
415.28%
58.42%
63.87%
7+2.27%
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
LeagueENG1
Loaded teams27
Loaded historical matches1580
Target season2026/27
Previous seasons in history4
Season window2022/23 → 2023/24 → 2024/25 → 2025/26 → 2026/27
League tierTier 1
Historical league matches1580
xG matches1580 / 1580
Effective xG matches796.59
Observed home xG1.60
Observed away xG1.32
Model home xG baseline1.58
Model away xG baseline1.33
Total league xG baseline2.92
Home-advantage weight0.900
Model configdata/models/ENG1.json
Team attack half-life240.0 days
Team defense half-life240.0 days
Attack strength weight0.769
Defense strength weight1.006
Venue prior matches3.81
DistributionPoisson + Dixon-Coles
Dixon-Coles rho-0.080 (effective -0.080)
Dixon-Coles decay0.025
Goal overdispersion0.000
Implied goal variance H / A1.37 / 1.39
Relative strengths and form diagnostics
Everton
Home attack strength0.864
Home defense weakness0.973
Base lambda home1.37
Attack form xGF level1.580
Defense form xGA level (lower is better)1.271
Opponent-adjusted attack residual+0.138
Opponent-adjusted defense residual (higher is better)+0.203
Opponent-adjustment weight0.003
Model attack signal+0.122
Model defense signal+0.187
Opponent blend0.050
Form matches 5 / 105 / 10
Last 5 xGF / xGA / xGD1.62 / 1.24 / +0.38
Last 10 xGF / xGA / xGD1.51 / 1.33 / +0.18
Chelsea
Away attack strength1.134
Away defense weakness0.968
Base lambda away1.43
Attack form xGF level1.143
Defense form xGA level (lower is better)1.752
Opponent-adjusted attack residual-0.330
Opponent-adjusted defense residual (higher is better)-0.308
Opponent-adjustment weight0.003
Model attack signal-0.314
Model defense signal-0.294
Opponent blend0.050
Form matches 5 / 105 / 10
Last 5 xGF / xGA / xGD1.12 / 1.79 / -0.66
Last 10 xGF / xGA / xGD1.18 / 1.69 / -0.51
Historical data coverage
Everton
Total matches158
Home / away matches79 / 79
xG matches158 / 158
xG home / away coverage79 / 79 · 79 / 79
Effective total31.37
Effective attack / defense31.37 / 31.37
Effective home matches15.54
Effective home attack / defense15.54 / 15.54
Chelsea
Total matches158
Home / away matches79 / 79
xG matches158 / 158
xG home / away coverage79 / 79 · 79 / 79
Effective total31.37
Effective attack / defense31.37 / 31.37
Effective away matches15.63
Effective away attack / defense15.63 / 15.63
Historical data coverage
96.08%
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 1560 historical matches.
Positive improvement means the model performed better than the league baseline.
Total goals log-loss vs baseline+0.39%
1X2 log-loss vs baseline+8.77%
BTTS Brier vs baseline+0.20%
Over 2.5 Brier vs baseline+1.47%
Full backtest metrics
Model
MAE total goals1.3214
Exact score log-loss2.9763
Total goals Brier0.8267
Total goals log-loss1.8983
1X2 Brier0.5816
1X2 log-loss0.9769
BTTS Brier0.2450
Over 1.5 Brier0.1565
Over 2.5 Brier0.2406
Over 3.5 Brier0.2225
League baseline
MAE total goals1.3162
Exact score log-loss3.0981
Total goals Brier0.8288
Total goals log-loss1.9058
1X2 Brier0.6478
1X2 log-loss1.0708
BTTS Brier0.2454
Over 1.5 Brier0.1573
Over 2.5 Brier0.2442
Over 3.5 Brier0.2253
Calibration diagnostics
Home win · ECE 1.88%
0- 10% n=23 avg= 7.6% actual= 4.3% 10- 20% n=124 avg= 16.0% actual= 11.3% 20- 30% n=255 avg= 25.4% actual= 26.7% 30- 40% n=307 avg= 35.1% actual= 38.8% 40- 50% n=308 avg= 44.9% actual= 44.8% 50- 60% n=242 avg= 55.0% actual= 55.4% 60- 70% n=180 avg= 64.5% actual= 67.8% 70- 80% n=90 avg= 74.0% actual= 74.4% 80- 90% n=31 avg= 84.0% actual= 87.1%
Draw · ECE 1.60%
0- 10% n=9 avg= 8.8% actual= 0.0% 10- 20% n=202 avg= 16.9% actual= 13.9% 20- 30% n=1298 avg= 25.3% actual= 26.2% 30- 40% n=51 avg= 31.0% actual= 19.6%
Away win · ECE 1.94%
0- 10% n=98 avg= 7.2% actual= 8.2% 10- 20% n=287 avg= 15.3% actual= 12.5% 20- 30% n=371 avg= 25.0% actual= 22.6% 30- 40% n=311 avg= 34.6% actual= 34.4% 40- 50% n=250 avg= 44.5% actual= 42.4% 50- 60% n=153 avg= 54.4% actual= 56.2% 60- 70% n=61 avg= 64.8% actual= 68.9% 70- 80% n=24 avg= 74.2% actual= 79.2% 80- 90% n=5 avg= 83.7% actual= 80.0%
BTTS · ECE 2.23%
30- 40% n=15 avg= 37.4% actual= 33.3% 40- 50% n=297 avg= 46.4% actual= 50.8% 50- 60% n=857 avg= 55.2% actual= 56.8% 60- 70% n=385 avg= 63.3% actual= 61.6% 70- 80% n=6 avg= 70.7% actual= 66.7%
Over 2.5 · ECE 1.98%
20- 30% n=1 avg= 25.7% actual= 0.0% 30- 40% n=44 avg= 37.3% actual= 40.9% 40- 50% n=329 avg= 46.3% actual= 48.6% 50- 60% n=702 avg= 54.9% actual= 57.3% 60- 70% n=417 avg= 63.9% actual= 63.1% 70- 80% n=64 avg= 72.4% actual= 73.4% 80- 90% n=3 avg= 80.5% actual=100.0%
Walk-forward - latest periods
| Period | N | Model 1X2 LL | Baseline 1X2 LL | Improvement |
|---|---|---|---|---|
| 2025-09 | 30 | 0.9867 | 1.1044 | +10.66% |
| 2025-10 | 30 | 0.9685 | 0.9879 | +1.96% |
| 2025-11 | 40 | 0.9440 | 1.0317 | +8.51% |
| 2025-12 | 56 | 0.9968 | 1.0997 | +9.36% |
| 2026-01 | 49 | 1.1465 | 1.1210 | -2.28% |
| 2026-02 | 42 | 1.0314 | 1.1383 | +9.40% |
| 2026-03 | 32 | 1.1157 | 1.0927 | -2.11% |
| 2026-04 | 30 | 1.0022 | 1.0548 | +4.98% |
| 2026-05 | 41 | 1.0433 | 1.0775 | +3.17% |
| 2026-08 | 20 | 0.9402 | 1.0435 | +9.90% |
| 2026-09 | 30 | 1.1023 | 1.1547 | +4.53% |
| 2026-10 | 10 | 1.0262 | 1.0587 | +3.08% |
Football is data.
NILYTIX turns match data into transparent model probabilities and measurable analytical signals.