Premier League

Model record · independent validation

How accurate are our football predictions?

On 1,140 Premier League matches it had not seen, the model's most likely result happened 51.8% of the time. On the same matches, simply backing the club higher in the table was right 49.0% of the time: the model gets about 3 more results right per 100 matches.

The online record below grows with every match played. It needs many more results before it says much on its own.

Check the published online record ↓ · Historical test and baseline ↓

Separate recent-data sample

Backtest · 10 matches

Forecasts fitted without seeing these results. Separate from the published online record.

Log loss 1.103 · Brier 0.666.

Small sample; descriptive results only.

Premier League · published online

Live prediction record

Published
50
Settled
50
Result hits
26 / 50

0 awaiting results · fixed pre-match snapshots

Manchester United vs Aston Villa

Home / draw / away: 37.9% / 23.9% / 38.1% · 4–4

Crystal Palace vs Liverpool

Home / draw / away: 23.6% / 24.3% / 52.1% · 1–3

Sunderland vs Chelsea

Home / draw / away: 39.0% / 24.6% / 36.4% · 1–0

Ipswich vs Bournemouth

Home / draw / away: 26.3% / 23.7% / 50.1% · 1–2

Brighton vs Brentford

Home / draw / away: 45.8% / 24.8% / 29.4% · 2–0

Nottingham Forest vs Manchester City

Home / draw / away: 22.6% / 26.8% / 50.6% · 3–1

Arsenal vs Hull City

Home / draw / away: 63.2% / 24.7% / 12.1% · 1–0

Fulham vs Newcastle

Home / draw / away: 41.6% / 27.9% / 30.5% · 1–1

Leeds United vs Tottenham

Home / draw / away: 38.6% / 29.2% / 32.2% · 1–0

Everton vs Coventry City

Home / draw / away: 58.5% / 24.3% / 17.2% · 3–1

Newcastle vs Everton

Home / draw / away: 46.0% / 27.4% / 26.6% · 0–1

Liverpool vs Arsenal

Home / draw / away: 33.6% / 30.2% / 36.2% · 0–1

Tottenham vs Crystal Palace

Home / draw / away: 42.0% / 28.1% / 29.9% · 2–0

Aston Villa vs Fulham

Home / draw / away: 46.1% / 26.2% / 27.7% · 1–1

Chelsea vs Manchester United

Home / draw / away: 36.2% / 23.7% / 40.1% · 2–0

Manchester City vs Brighton

Home / draw / away: 56.1% / 23.5% / 20.5% · 0–1

Bournemouth vs Leeds United

Home / draw / away: 43.7% / 28.4% / 28.0% · 2–0

Coventry City vs Sunderland

Home / draw / away: 31.2% / 27.8% / 40.9% · 3–1

Hull City vs Ipswich

Home / draw / away: 56.4% / 25.3% / 18.3% · 2–2

Brentford vs Nottingham Forest

Home / draw / away: 47.4% / 28.1% / 24.5% · 2–0

Sunderland vs Leeds United

Home / draw / away: 33.6% / 29.2% / 37.1% · 0–0

Liverpool vs Brighton

Home / draw / away: 39.1% / 26.2% / 34.7% · 2–0

Crystal Palace vs Newcastle

Home / draw / away: 31.6% / 25.9% / 42.5% · 0–2

Manchester United vs Bournemouth

Home / draw / away: 37.9% / 27.6% / 34.5% · 0–3

Hull City vs Brentford

Home / draw / away: 37.6% / 29.8% / 32.6% · 1–1

Chelsea vs Tottenham

Home / draw / away: 53.2% / 22.3% / 24.5% · 0–2

Arsenal vs Everton

Home / draw / away: 55.8% / 28.1% / 16.1% · 2–0

Coventry City vs Fulham

Home / draw / away: 28.3% / 26.6% / 45.1% · 0–1

Aston Villa vs Manchester City

Home / draw / away: 25.3% / 25.0% / 49.7% · 2–1

Ipswich vs Nottingham Forest

Home / draw / away: 25.7% / 27.0% / 47.3% · 1–1

Tottenham vs Coventry City

Home / draw / away: 44.3% / 27.5% / 28.2% · 3–1

Nottingham Forest vs Arsenal

Home / draw / away: 19.8% / 28.9% / 51.3% · 0–1

Bournemouth vs Sunderland

Home / draw / away: 49.5% / 26.7% / 23.8% · 2–0

Leeds United vs Manchester United

Home / draw / away: 40.4% / 27.3% / 32.3% · 1–1

Brighton vs Crystal Palace

Home / draw / away: 60.4% / 21.5% / 18.1% · 5–0

Newcastle vs Aston Villa

Home / draw / away: 47.2% / 24.4% / 28.4% · 0–1

Fulham vs Hull City

Home / draw / away: 38.8% / 29.1% / 32.1% · 1–0

Manchester City vs Ipswich

Home / draw / away: 84.7% / 10.6% / 4.8% · 3–2

Brentford vs Liverpool

Home / draw / away: 40.6% / 28.1% / 31.3% · 1–2

Everton vs Chelsea

Home / draw / away: 37.2% / 27.0% / 35.8% · 0–2

Coventry City vs Newcastle

Home / draw / away: 17.8% / 24.6% / 57.6% · 4–1

Liverpool vs Manchester City

Home / draw / away: 27.3% / 28.3% / 44.4% · 0–3

Hull City vs Everton

Home / draw / away: 40.3% / 31.8% / 28.0% · 2–2

Crystal Palace vs Nottingham Forest

Home / draw / away: 38.6% / 26.1% / 35.3% · 3–1

Manchester United vs Tottenham

Home / draw / away: 48.3% / 25.4% / 26.3% · 1–1

Sunderland vs Brighton

Home / draw / away: 24.4% / 25.8% / 49.8% · 0–2

Chelsea vs Bournemouth

Home / draw / away: 49.5% / 26.9% / 23.6% · 5–1

Ipswich vs Fulham

Home / draw / away: 23.0% / 25.7% / 51.3% · 2–1

Aston Villa vs Brentford

Home / draw / away: 37.4% / 24.3% / 38.3% · 2–2

Arsenal vs Leeds United

Home / draw / away: 60.6% / 24.2% / 15.2% · 2–1

How this record is scored

Predictions are stored from published pre-24h snapshots and never rewritten. Result hits count the most likely 1X2 outcome. This online sample is separate from the historical test.

Prediction accuracy by league

Each league has its own model, tested on matches it had not seen. The hit rate is how often the most likely result happened; log loss scores the full probabilities, and lower is better.

LeagueHeld-out matchesMost likely result happenedBacking the higher clubLog loss
Premier League1,14051.8%49.0%0.993
Bundesliga1,14051.7%48.8%0.992
Championship1,14044.3%43.5%1.056
La Liga1,14053.6%49.7%0.972
Ligue 11,14052.4%47.1%1.000
Serie A1,14052.7%52.0%0.981

European competitions · shared model

Walk-forward log loss 0.942 on 2,817 European matches (2016-17 to 2026-27); historical outcome baseline 1.051. This is one shared test, not three independent validations.

How to check a prediction record

Check timestamps before kickoff, an unedited record of hits and misses, and probability scores against a baseline on the same matches. Hit rate alone does not measure probability quality.

When the model says 60%, does it happen 60% of the time?

Every match has three possible results, each with a probability. Grouping all 3,420 of them from 1,140 held-out matches by what the model said shows how often each group actually happened. Well-calibrated probabilities sit close to the figure the model gave.

Said 0%–40%avg 25% · 2,436 results25%
Said 40%–50%avg 45% · 450 results44%
Said 50%–60%avg 55% · 277 results53%
Said 60%–70%avg 64% · 167 results65%
Said 70%–100%avg 76% · 90 results82%

Bar length is how often results in each group actually happened. Groups with fewer than 30 results are left out.

Historical test · 1,140 matches

Validated 1 Nov 2026 · forecasts made without seeing the results

ForecastMost likely result happened
Our model51.8%
Back the club higher in the table49.0%

On the same matches, simply backing the club higher in the table was right 49.0% of the time: the model gets about 3 more results right per 100 matches.

Test method

None of these matches were used to choose the model's settings. The table rule uses the league table as it stood before each kickoff; when two clubs are level, as on the opening weekend, it backs the home side. It never predicts a draw, which is part of why it is hard to beat by much.

The model also gives probabilities, not just a pick. Its log loss was 0.993 and Brier score 0.593 (lower is better); see the calibration above for whether those probabilities can be taken at face value.

Probabilities were produced by fits refreshed every 28 days using only results available before each match.

About the accuracy scores

Log loss and Brier score measure how much probability a forecast gave to what actually happened; lower is better for both. They reward honest probabilities rather than picking winners, so a model can pick the favourite every time and still score badly if it is overconfident. See how log loss and Brier score work.

Results up to
2023-24, 2024-25, 2025-26, 2026-27
Matches used
1,140
Method
Walk-forward Dixon–Coles, checked for calibration and compared with backing the club higher in the league table.
Data source
Loaded match database
Limitations
Score-based model only. Historical accuracy does not guarantee future accuracy.