Evertonv
Coventry CityHome win46.9%Draw26.3%Away win26.8%
Goal means 1.46–1.03 · Leading score 1–1Premier League 2026/27 · Experimental
The run-in, forecast.
Every remaining fixture, simulated into a final table. Actual points stay banked.
Results 5 Nov 2026, 15:46 · Stats 5 Nov 2026, 15:46 UK290 remaining matches20,000 season simulations1990 training matches · 2021/22, 2022/23, 2023/24, 2024/25, 2025/26
How did the forecast change last season? →
Fixture difficulty, Forest’s run-in and season blocks →
Forest’s chances this season
Chance of each league finish in our xG-based model, across 20,000 simulated seasons.
Top four0.7%1st–4th
Top seven6.2%1st–7th
Top half19.6%1st–10th
Bottom half80.4%11th–20th
Relegation24.6%18th–20th
Top half and bottom half cover all finishes. Top four and top seven are within the top half; relegation is within the bottom half. These chances overlap. Top seven does not guarantee European qualification.
Explore exact finishing positions
Average simulated finish14.2Average across all seasons · dashed line
Middle 80% of finishes9th–19thA range of outcomes, not a guarantee
Most-common single finish18th9.5% of seasons; other finishes combined are more likely
Forecast table rank: 16th. This is Forest’s place when clubs are ordered by their average finish. It is different from Forest’s own average of 14.2.
18th: 9.5%; 1,893 of 20,000 simulated seasons
The tallest bar is the most common single finish—not a guarantee. Several positions have very similar probabilities (15th–18th); the tallest can move between runs. The league forecast table orders all clubs by average finish, which can differ from their most common position.
Are 20,000 simulations enough?
At this sample size, the worst-case approximate 95% simulation noise is ±0.7 percentage points for a probability near 50%. This measures sampling noise only, not uncertainty in team strengths or the model. More runs smooth the bars; better data and model testing improve the forecast. Very rare outcomes are less stable; a zero count does not prove an outcome is impossible.
Next-match outcome distribution
Nottingham Forest (H) v Manchester City (A)
From Forest’s perspective: left = defeat, 0 = draw, right = win.
Draw: 24.5%
Forest defeatDrawForest win
Calculated directly from the same model goal probabilities—not from the season simulations. A draw can be the most likely exact score even when one team has the higher total win probability.
Projected points
A range of seasons, not one predicted total.
Nottingham ForestForest4436–53 pts · 80% range
Manchester CityNext opponent7263–80 pts · 80% rangeMedian pointsMiddle 80%Taller curve = more simulated seasons
Tap a club to inspect · ordered by average projected points
Nottingham ForestMedian 44 ptsMiddle 80% 36–53 pts6 points already earned · 29 matches left
20,000 simulations · curves use a 3-point moving average; markers use exact simulated totals. Middle 80% is a model range, not a guarantee.
Projected final league table
Forecast rank orders clubs by average simulated finish. It is not each club’s average finishing position. Final points are averages; the range covers the middle 80% of simulated points.
| Forecast rank | Club | Now | Final pts | 80% pts range | Title | Top 4 | Top 7 | Releg. |
|---|---|---|---|---|---|---|---|---|
| 1 |  Arsenal24 pts · 9 played | 1 | 78.5 | 70–87 | 68.8% | 98.8% | >99.9% | <0.1% |
| 2 | Manchester City21 pts · 9 played | 2 | 71.5 | 63–80 | 21.9% | 90.6% | 98.8% | <0.1% |
| 3 |  Brighton & Hove Albion19 pts · 9 played | 3 | 64.1 | 55–73 | 4.0% | 61.0% | 89.8% | <0.1% |
| 4 |  Liverpool15 pts · 9 played | 5 | 63.6 | 55–73 | 3.3% | 56.6% | 87.7% | <0.1% |
| 5 |  Chelsea16 pts · 9 played | 4 | 60.3 | 51–69 | 1.3% | 36.6% | 75.6% | 0.2% |
| 6 |  Bournemouth12 pts · 9 played | 8 | 56.9 | 48–66 | 0.4% | 20.3% | 58.8% | 0.7% |
| 7 |  Brentford14 pts · 9 played | 6 | 54.6 | 46–64 | 0.2% | 13.0% | 46.2% | 1.7% |
| 8 |  Leeds United11 pts · 9 played | 12 | 50.0 | 41–59 | <0.1% | 4.0% | 22.1% | 6.7% |
| 9 |  Newcastle United11 pts · 9 played | 13 | 50.1 | 41–59 | <0.1% | 4.1% | 21.8% | 7.0% |
| 10 |  Aston Villa12 pts · 9 played | 10 | 49.6 | 41–59 | <0.1% | 3.6% | 19.5% | 8.0% |
| 11 | Everton13 pts · 9 played | 7 | 49.4 | 41–58 | <0.1% | 2.9% | 18.1% | 8.3% |
| 12 |  Manchester United7 pts · 9 played | 17 | 48.6 | 40–58 | <0.1% | 2.9% | 17.0% | 10.1% |
| 13 |  Tottenham Hotspur12 pts · 9 played | 9 | 48.3 | 40–57 | <0.1% | 2.1% | 14.0% | 10.4% |
| 14 |  Fulham9 pts · 9 played | 15 | 46.6 | 38–56 | <0.1% | 1.4% | 10.4% | 15.3% |
| 15 |  Crystal Palace7 pts · 9 played | 18 | 44.9 | 36–54 | <0.1% | 0.7% | 6.4% | 24.2% |
| 16 | Nottingham Forest6 pts · 9 played | 19 | 44.4 | 36–53 | <0.1% | 0.7% | 6.2% | 24.6% |
| 17 |  Hull City11 pts · 9 played | 11 | 42.6 | 34–51 | <0.1% | 0.3% | 3.3% | 33.2% |
| 18 | Coventry City9 pts · 9 played | 16 | 40.8 | 32–50 | <0.1% | 0.2% | 1.9% | 45.6% |
| 19 |  Sunderland5 pts · 9 played | 20 | 39.8 | 31–49 | <0.1% | 0.1% | 1.5% | 50.9% |
| 20 |  Ipswich Town11 pts · 9 played | 14 | 39.8 | 31–48 | <0.1% | <0.1% | 1.1% | 53.0% |
Fixed team strengths as of this update. Future injuries, transfers and form changes are not simulated. Top 7 is a league finish, not guaranteed European qualification.
Match predictions
Home win · Draw · Away win. Model goal means are pre-match predictions, not the xG recorded after a game.
EvertonvCoventry CityLeeds UnitedvTottenham HotspurHome win44.0%Draw26.1%Away win29.9%
Goal means 1.44–1.14 · Leading score 1–1ArsenalvHull CityHome win65.7%Draw21.4%Away win12.9%
Goal means 1.87–0.69 · Leading score 1–0FulhamvNewcastle UnitedHome win41.7%Draw24.9%Away win33.4%
Goal means 1.52–1.34 · Leading score 1–1Nottingham ForestvManchester CityHome win30.0%Draw24.5%Away win45.5%
Goal means 1.26–1.61 · Leading score 1–1Brighton & Hove AlbionvBrentfordHome win48.3%Draw23.0%Away win28.7%
Goal means 1.80–1.34 · Leading score 1–1Crystal PalacevLiverpoolHome win32.3%Draw24.1%Away win43.5%
Goal means 1.37–1.62 · Leading score 1–1Ipswich TownvBournemouthHome win26.3%Draw22.7%Away win51.0%
Goal means 1.28–1.86 · Leading score 1–1SunderlandvChelseaHome win32.4%Draw24.7%Away win42.8%
Goal means 1.32–1.55 · Leading score 1–1Manchester UnitedvAston VillaHome win48.8%Draw23.6%Away win27.6%
Goal means 1.74–1.25 · Leading score 1–1Manchester CityvFulhamHome win58.0%Draw21.7%Away win20.3%
Goal means 1.96–1.07 · Leading score 1–1Aston VillavSunderlandHome win46.1%Draw25.3%Away win28.6%
Goal means 1.54–1.16 · Leading score 1–1Availability and the forecast
Next game: Nottingham Forest v Manchester City. News updated 5 Nov 2026, 15:46 UK.
No numerical injury adjustment is applied yet. These are reported concerns for the next round, not confirmed lineups or a prediction of minutes. They are not carried into every remaining fixture.
A player already missing from recent games is partly reflected in team performance. A fresh absence needs a tested estimate of the replacement’s impact—not a second blanket penalty.
- Igor Jesus Maciel da CruzDoubtful · reported 75% chance of playing
Calf injury - 75% chance of playing
257 season minutes · 46 minutes in the last 6 team games - Nicolás DomínguezInjured · reported 0% chance of playing
Muscular injury - Expected back 30 Dec
74 season minutes · 74 minutes in the last 6 team games - Nicolò SavonaInjured · reported 0% chance of playing
Knee injury - Unknown return date
0 season minutes · recent minutes not confidently matched
Model checks and method
Recency-weighted xG attack and defence strengths, adjusted for opposition and home advantage, with shrinkage towards league average. One fit is used for every remaining fixture; the displayed model is the existing baseline, not an unverified new winner selected from backtests.
The stored xG archive covers 5 completed seasons. The first season is reserved for initial training in historical tests. Goals/xG blends, low-score corrections and coach effects have not yet shown a reliable improvement over the baseline.
Each test trains only on earlier calendar dates, not later matches or the same day’s results. Checks use log loss, Brier score and calibration as well as picking the most likely result. Historical source data can include later revisions. Manager moves and player availability are not yet separate numerical adjustments.
Simulations keep current deductions and banked points; remaining scores use the baseline’s independent Poisson distributions. Ties use points, GD, GF, head-to-head points and away goals; any unresolved tie is split randomly rather than alphabetically (24 unresolved tie groups across this run). Early-season points ranges have been too narrow in retrospective checks; the middle 80% is a simulation range, not a calibrated confidence interval.