Manchester City v Everton
Manchester City vs Everton
Premier League- Model pick
- Manchester City win 62%
- Likely score
- 1-1 (12%)
- Kickoff
- Wed 3 Mar, 20:00 UK · Manchester City home venue
- Absences
- Manchester City: 1 out, 2 doubtful · Everton: no absences reported
Why
- Manchester City rated 190 Elo points higher, home advantage included
- Manchester City stronger on head-to-head ratings
“Elo alone” is the plain rating model before form, attack, defence and the goals model are added; the gap shows what those add.
What the model weighed
- Manchester City rated 190 Elo points higher, home advantage included
- Manchester City stronger on head-to-head ratings
- Manchester City score more (1.3 goals a game more)
- Manchester City have the better goal difference (1.1 a game)
- Manchester City are at home
- Manchester City concede less (0.3 goals a game fewer)
Goals
Total goals
Both teams to score
Most likely scores
- 1-111.8%
- 1-011.1%
- 2-010.8%
- 2-19.8%
- 0-07.6%
- 3-06.5%
Score by result
- Manchester City win · pick1-011.1%
- Draw1-111.8%
- Everton win0-15.1%
From the goals model (expected goals with the Dixon-Coles low-score correction), so these can differ slightly from the calibrated win/draw/win forecast. A single score is rarely right (about one match in eight); the actual score is among the top three about a quarter of the time. Model estimates, not betting prices.
- Semenyo MIDinjuredKnee injury - Unknown return date
- Guéhi DEFdoubtfulUnspecified injury - 75% chance of playing
- Gvardiol DEFdoubtfulUnspecified injury - 75% chance of playing
No team news on the club's site in the last 7 days.
mancity.comNo team news on the club's site in the last 7 days.
evertonfc.comAbsences from the league's injury list; news from each club's own website, scanned Tue 17 Nov, 15:01 UK. The forecast does not use this news.
Goalkeepers
Defenders
Midfielders
Forwards
- Allan
- Antoine Semenyo2 goalsinjured
- Erling Haaland9 goals
- Iliman Ndiaye
- Jeremy Doku
- Ryan McAidoo
Goalkeepers
Defenders
Midfielders
Forwards
Season goals from the competition's top scorers; absences from the injury list.
Lineups are published about an hour before kickoff (Wed 3 Mar, 20:00 UK) and will show here then.
Weather at kickoff
No forecast yet: weather is forecast about a week ahead.
Rest
- Manchester City
- 4 days' rest · plays first Hull City AFC (Premier League), Sat 27 Feb, 15:00
- Everton
- 4 days' rest · plays first Nottingham Forest (Premier League), Sat 27 Feb, 15:00
Prices
| Result | Model | Model fair odds | Bookmaker |
|---|---|---|---|
| Manchester City win | 61.8% | 1.62 | – |
| Draw | 21.7% | 4.61 | – |
| Everton win | 16.6% | 6.04 | – |
No bookmaker prices for this match yet. Picks use the model's chances alone, so prices are only ever shown for comparison.
Model
- Version
- football-outcome-ensemble-v1+dc-mle-xi0.0019+goalcal
- Confidence
- 70.5% (54.8% after data quality)
- Data quality
- 77.7%
- Simulations
- 3,000
- Basis
- Upcoming Premier League fixture (pre-match Elo 1629 v 1495). Manchester City are favoured at 62% from 480 completed matches (2025-08-15 to 2026-11-08). No bookmaker prices are available for this fixture yet, so no value signals are shown.
- Sources
- Fixtures and results: openfootball and football-data.org · weather: Open-Meteo · injuries: league injury list
How The Model Does In Premier League
100 gradedAre its chances honest?
- Said 20–30%happened 27%143 calls
- Said 30–40%happened 33%78 calls
- Said 40–50%happened 42%50 calls
- Said 50–60%happened 53%19 calls
On average its chances are 1.6% away from what happened. Full record
How Fresh It Is
updated by the worker- Results and fixtures15m agoon schedule
- Injury list21m agoon schedule
- Weather1h agoon schedule
- Bookmaker prices21m agoon schedule
Sources
- openfootball (fixtures, results)
- football-data.org (results, squads)
- ESPN (live scores, lineups)
- Open-Meteo (weather)
- Premier League injury list
- football-data.co.uk (prices, display only)
Checks The Model Passes
6 of 7- reviewPatterns are not assumed to be causes
- passedTested on matches it was not trained on
- passedOverlapping signals kept apart
- passedUses only data from before each match
- passedRelegated and promoted clubs kept in
- passedEvery fixture forecast, not a chosen few
- passedNo hindsight in the features
Hover a check for how it is enforced.