Premier League · Gameweek 2 Verdict
Gameweek 2 Verdict: 5/10 Correct as Model Hits 47% Season Accuracy
Published 1 September 2026 · FootyPulse AI Analysis
The AI managed five correct predictions from ten Gameweek 2 fixtures, a 50% hit rate that slightly dipped our season accuracy to 47% (8 from 17 overall). While the model nailed its biggest calls—Manchester City and Arsenal away—it struggled with home favourites across the board.
The scoreboard
Gameweek 2 delivered 5 correct calls from 10 fixtures, bringing our season tally to 8 from 17 (47% accuracy). A mixed weekend saw confident picks land cleanly, but several mid-probability home favourites failed to deliver.
Biggest upset
Liverpool's failure to beat Nottingham Forest at Anfield stands as the week's major miss. The model backed Liverpool with 58% confidence, but Forest fought back to secure a 2-2 draw—a result that defied both the pre-match probability and home advantage.
Best correct call
Chelsea's 4-3 thriller against Brighton represents our finest work of the gameweek. The model gave Chelsea just 38% win probability—making them the underdog in their own stadium—yet Enzo Maresca's side delivered seven goals and three points. This was the least-confident correct prediction we registered, highlighting the value in probability-based framing over binary certainty.
Every fixture
- Crystal Palace 1-4 Manchester City: Predicted away win (60%) ✅ Correct
- Liverpool 2-2 Nottingham Forest: Predicted home win (58%) ❌ Incorrect
- Coventry City 0-1 Hull City: Predicted home win (48%) ❌ Incorrect
- Bournemouth 1-1 Everton: Predicted home win (36%) ❌ Incorrect
- Tottenham Hotspur 0-2 Newcastle United: Predicted home win (48%) ❌ Incorrect
- Chelsea 4-3 Brighton & Hove Albion: Predicted home win (38%) ✅ Correct
- Leeds United 1-1 Brentford: Predicted home win (48%) ❌ Incorrect
- Sunderland 1-0 Fulham: Predicted home win (48%) ✅ Correct
- Manchester United 5-2 Ipswich Town: Predicted home win (48%) ✅ Correct
- Aston Villa 0-1 Arsenal: Predicted away win (58%) ✅ Correct
Five draws and away upsets punished a model framework built on identifying marginal edges. The cluster of 48% home-win probabilities tells the story: tight matches where small factors tipped outcomes against the numbers. See the pre-match preview for how we framed these probabilities before kick-off.
Gameweek 3 offers a reset. The data adapts, the model recalibrates, and we keep tracking every probability against every result.
Half-time re-predictions this gameweek
At half-time the AI re-reads each match and re-predicts. Here is how those calls did against the pre-match calls, on the same set of Gameweek 2 matches.
Pre-match
5/10
50% correct
At half-time
6/10
60% correct
Where half-time changed the call
Bournemouth 1–1 Everton
Pre-match: Bournemouth ✗ · Half-time: Draw (1–1) ✓
Tottenham Hotspur 0–2 Newcastle United
Pre-match: Tottenham Hotspur ✗ · Half-time: Draw (1–1) ✗
Leeds United 1–1 Brentford
Pre-match: Leeds United ✗ · Half-time: Brentford (1–2) ✗
This is a single gameweek — far too small a sample to read anything into, in either direction. See the running pre-match vs half-time comparison on our accuracy page for the season picture.
Share this verdict:
Prediction figures in this article reflect the data available at the time of publication and may have shifted since. See the live Gameweek 2 fixtures for the current numbers on each match.
See how the AI called it before kickoff:
Read the Gameweek 2 preview →