Results update the state
Completed matches are fixed into every future season path and exponentially weighted residuals update team attack and defence.
Premier League predictions built as probabilities, not hot takes
The current HORIZON 6.2 baseline runs every unresolved fixture through 100,000 matched season paths. Inspect the projected table, uncertainty and the evidence behind each club.
SEASON PATHS100,000
FIXTURES MODELLED380
POSTERIOR STATES24
RUN HASHf7087d41
CURRENT BASELINE
Projected 2026/27 Premier League table
Mean points determine the displayed order. Probability columns are direct counts from the published simulation.
| Pos | Club | Pts | 90% range | Title | Top four | Top six | Relegation |
|---|---|---|---|---|---|---|---|
| 1 | Arsenal | 73.6 | 60 to 87 | 48.6% | 89.2% | 95.7% | 0.0% |
| 2 | Manchester City | 69.1 | 55 to 83 | 23.9% | 75.6% | 88.0% | 0.0% |
| 3 | Chelsea | 64.1 | 49 to 79 | 11.4% | 53.5% | 70.7% | 0.4% |
| 4 | Liverpool | 59.9 | 46 to 74 | 4.1% | 34.5% | 54.8% | 0.7% |
| 5 | Manchester United | 58.6 | 44 to 73 | 3.3% | 30.0% | 48.8% | 1.6% |
| 6 | Brentford | 57.1 | 43 to 71 | 1.9% | 22.3% | 41.1% | 1.9% |
| 7 | Brighton & Hove Albion | 55.2 | 41 to 69 | 1.6% | 17.3% | 32.8% | 2.7% |
| 8 | Aston Villa | 55 | 41 to 69 | 1.3% | 16.4% | 31.6% | 3.2% |
| 9 | Newcastle United | 54.1 | 40 to 68 | 1.2% | 14.6% | 28.7% | 4.2% |
| 10 | Leeds United | 51.8 | 37 to 66 | 0.9% | 9.6% | 20.1% | 7.2% |
| 11 | AFC Bournemouth | 51.4 | 38 to 65 | 0.4% | 8.5% | 18.8% | 6.2% |
| 12 | Everton | 50.2 | 36 to 64 | 0.3% | 6.2% | 14.9% | 8.4% |
| 13 | Nottingham Forest | 49.9 | 36 to 63 | 0.3% | 5.7% | 13.6% | 8.1% |
| 14 | Tottenham Hotspur | 49.8 | 35 to 65 | 0.4% | 7.0% | 15.3% | 11.5% |
| 15 | Fulham | 47 | 33 to 61 | 0.2% | 3.2% | 8.3% | 16.2% |
| 16 | Ipswich Town | 46.9 | 31 to 63 | 0.2% | 4.7% | 11.1% | 20.8% |
| 17 | Hull City | 42.5 | 28 to 57 | 0.0% | 1.1% | 3.5% | 34.5% |
| 18 | Crystal Palace | 40.4 | 28 to 53 | 0.0% | 0.3% | 1.2% | 42.1% |
| 19 | Sunderland | 38.8 | 26 to 52 | 0.0% | 0.3% | 0.9% | 50.6% |
| 20 | Coventry City | 32.2 | 19 to 45 | 0.0% | 0.0% | 0.1% | 79.9% |
Form is regressed
Underlying xG direction can carry forward. Finishing streaks and shot-stopping variance are deliberately pulled back toward normal.
Transfers change both clubs
A confirmed arrival can strengthen the buyer while the same movement removes contribution from the selling club.
Counterfactuals rerun the league
Ask what happens if a key player leaves, becomes unavailable or joins another team, then compare matched baseline and scenario paths.
CLUB FORECASTS
Explore every team
QUESTIONS ANSWERED
How to interpret the forecast
How are the Premier League predictions calculated?
The Football Simulator models attack, defence, venue, sustainable recent form, confirmed transfers and verified results, then samples 100,000 complete season paths.
Are these predictions guaranteed?
No. They are probability estimates with uncertainty ranges. Football outcomes remain uncertain and the model is updated as verified evidence changes.
Do transfer rumours affect the table prediction?
No. Reporter and social-media signals can enter a discovery queue, but team strength changes only after a transfer is officially confirmed.
Does last season's form affect the forecast?
Yes, but only after regression. The final 25 matches are compared with xG and xGA, while finishing and shot-stopping overperformance are reduced before carry-over.