Premier League 2026/27 · the Onside supercomputer

We already played
all 10,000 seasons — here's who lifts it.

380 fixtures. 10,000 full seasons. Every kick priced by the engine that went 82.9% at the World Cup — with this summer's transfers, new managers and European campaigns baked into the numbers. Think it's wrong? Beat the model, match by match.

Champions
Arsenal56.1%
Most likely down
Hull City65.8%
Best survival shout
Ipswich Town54.6% stay up

All 20, position by position

BAR = 5%–95% POINTS RANGE
#ClubxPtsRangeTitleTop 4Relegated
1ArsenalUCL84.37098
56.1%95.3%
2Manchester CityUCL76.66092
20.4%79.2%
3LiverpoolUCL73.15689
11.2%67.4%
4Manchester UnitedUCL685184
5.3%46.7%0.1%
5Chelsea65.34982
3.3%35.3%0.2%
6Aston VillaUCL62.54579
1.8%25.9%0.4%
7Tottenham Hotspur60.34377
1%19.2%0.5%
8Newcastle United56.74073
0.5%11.4%1.3%
9BrightonUECL54.33771
0.1%7.7%2.4%
10BournemouthUEL51.53569
0.1%4.7%3.7%
11Brentford48.33166
0.2%3.3%7.8%
12Everton46.83163
1.5%9.1%
13Fulham43.32861
0.8%16.1%
14Leeds United42.42659
0.7%19.5%
15Crystal PalaceUEL41.22658
0.4%21.5%
16SunderlandUEL412657
0.4%21.7%
17Nottingham Forest36.32252
0.1%39.1%
18Ipswich Town352151
45.4%
19Coventry City34.92051
45.5%
20Hull City30.51745
65.8%

What's in the numbers

Champions (85 pts), Arteta year five, same spine — the market's clear favourite.

Guardiola gone, Maresca in; double cup winners; Elliot Anderson £116m club record.

Slot out, Iraola in from Bournemouth after a 5th-place season; squad rebuild flagged.

Carrick made permanent after the 3rd-place surge from January.

Xabi Alonso arrives after a 10th-place mess; market prices a big rebound.

Emery continuity; 4th + Europa League winners; back in the UCL.

Survived on the final day, then De Zerbi + £200m spent including Tonali (£100m).

Howe stays but Tonali sold for £100m and Guimarães wanted — market cooled.

Hürzeler, 8th two years running; Conference League adds midweeks.

Lost Iraola to Liverpool; Marco Rose in; first-ever European campaign (UEL).

Andrews; equalled their best-ever PL finish (9th).

Moyes continuity, 13th.

Arbeloa in — untested at PL level; market prices genuine relegation risk (15%).

Farke, safe 14th in season one back.

Won the Conference League, then lost Glasner to Forest; Sage in — upheaval priced.

7th-place fairy-tale + Europa League load; Le Bris; market sees regression risk.

Glasner in but Elliot Anderson sold for £116m; 16th last season.

84-pt runners-up with the best defence of the trio, but McKenna left; Gary O'Neil in.

Championship winners with 95 pts and 97 goals; Lampard extended to 2029 — best-backed of the trio alongside Ipswich.

Promoted via the play-offs off the weakest underlying record (73 pts, +4 GD); 80% relegation-implied — the market's clearest call.

How these numbers are made — and where we differ from the bookies

Team strengths start from three seasons of expected-goals data with time decay and Dixon–Coles low-score correction, then get calibrated so full-season simulations agree with July's bookmaker season markets — the cleanest available price on transfers, managerial changes, European load and promoted-club quality. Each simulated season draws correlated team-strength noise, so a club that breaks good or bad is good or bad all season — that is why the promoted clubs survive in a genuine share of simulations rather than being rubber-stamped down. One honest divergence: our title number for the favourite runs above the betting market's, because title books carry longshot bias; we publish ours and let the season grade it. Every match that feeds this table is called at /predictions and graded in public at /model-record. Full write-up: /methodology. Free to cite with a link (CC BY 4.0). Last refreshed 14 July 2026.