Gameweek preview · 2026/27
FPL Gameweek 4 predictions: captain picks, transfers & differentials
Every fixture simulated 500 times for 453 players — expected points with a floor and a ceiling. Updated Sat, 12 Sep 2026 02:50:04 GMT.
Gameweek 4 has been played. Below are the predictions we published before the deadline, unchanged, next to what really happened.
Gameweek 4 results: predicted vs actual points
Across 453 players, our predictions were off by 1.77 points per player on average (we over-predicted by 0.33 on average). No cherry-picking — this is every player we projected.
| Player | Predicted | Actual | Mins |
|---|---|---|---|
| Szoboszlai LIV | 6.7 | 3 | 90 |
| B.Fernandes MUN | 6.6 | 2 | 90 |
| Palmer CHE | 6.2 | 5 | 90 |
| Tavernier BOU | 6.0 | 8 | 82 |
| Gibbs-White NFO | 5.8 | 8 | 90 |
| Mbeumo MUN | 5.7 | 2 | 90 |
| Dewsbury-Hall EVE | 5.6 | 2 | 90 |
| Wirtz LIV | 5.5 | 3 | 90 |
| Gabriel ARS | 5.4 | 9 | 90 |
| Cherki MCI | 5.4 | 1 | 45 |
| Virgil LIV | 5.2 | 8 | 90 |
| Scott BOU | 4.9 | 4 | 90 |
| Neto CHE | 4.8 | 2 | 85 |
| Tarkowski EVE | 4.8 | 8 | 90 |
| Groß BHA | 4.8 | 17 | 90 |
See the full season record on is FPL AI accurate?
- GW4
- gameweek
- 6.7
- top projection
- 453
- players simulated
- 500
- sims per fixture
Captain picks — Gameweek 4
Szoboszlai (LIV vs FUL) leads on 6.70 xPts, 84% chance of 60+ minutes.
- Szoboszlai6.7xPts
LIV vs FUL · 84% to play 60'
floor 1.0ceiling 11.0 - B.Fernandes6.6xPts
MUN vs MCI · 82% to play 60'
floor 1.0ceiling 11.0 - Palmer6.2xPts
CHE vs HUL · 76% to play 60'
floor 1.0ceiling 10.0 - Tavernier6.0xPts
BOU vs BRE · 80% to play 60'
floor 0.0ceiling 10.0 - Gibbs-White5.8xPts
NFO @ AVL · 88% to play 60'
floor 1.0ceiling 9.0
Clean sheet odds
- Gabriel · ARS52%
- Calafiori · ARS52%
- White · ARS52%
- Raya · ARS52%
- Konsa · ARS52%
Differentials under 10% owned
- Tavernier BOU · MID · £6.0m · 3.6% owned6.03 xPts
- Gibbs-White NFO · MID · £7.9m · 9.4% owned5.75 xPts
- Dewsbury-Hall EVE · MID · £6.5m · 4.6% owned5.60 xPts
- Wirtz LIV · MID · £7.4m · 8.0% owned5.47 xPts
- Scott BOU · MID · £6.1m · 5.6% owned4.88 xPts
Best value picks — Gameweek 4 transfer targets
Most projected points per £1m among players likely to play 60+ minutes.
- Mykolenko DEF · EVE @ TOT · £4.5m4.65 xPts
- Bogle DEF · LEE vs NEW · £4.5m4.46 xPts
- Thomas DEF · COV vs BHA · £4.0m3.92 xPts
- Tavernier MID · BOU vs BRE · £6.0m6.03 xPts
- Szoboszlai MID · LIV vs FUL · £7.0m6.70 xPts
- Enciso MID · IPS @ CRY · £5.5m4.75 xPts
- Barry FWD · EVE @ TOT · £5.6m3.93 xPts
- Abraham FWD · AVL vs NFO · £5.5m3.78 xPts
- Awoniyi FWD · COV vs BHA · £5.5m3.18 xPts
FPL Gameweek 4 predicted points — top 25
| Player | Fixture | £m | xPts |
|---|---|---|---|
| Szoboszlai MID | LIV vs FUL | 7.0 | 6.70 |
| B.Fernandes MID | MUN vs MCI | 12.0 | 6.64 |
| Palmer MID | CHE vs HUL | 9.7 | 6.22 |
| Tavernier MID | BOU vs BRE | 6.0 | 6.03 |
| Gibbs-White MID | NFO @ AVL | 7.9 | 5.75 |
| Mbeumo MID | MUN vs MCI | 7.9 | 5.65 |
| Dewsbury-Hall MID | EVE @ TOT | 6.5 | 5.60 |
| Wirtz MID | LIV vs FUL | 7.4 | 5.47 |
| Gabriel DEF | ARS @ SUN | 8.0 | 5.43 |
| Cherki MID | MCI @ MUN | 7.8 | 5.41 |
| Virgil DEF | LIV vs FUL | 6.5 | 5.21 |
| Scott MID | BOU vs BRE | 6.1 | 4.88 |
| Neto MID | CHE vs HUL | 6.5 | 4.82 |
| Tarkowski DEF | EVE @ TOT | 6.0 | 4.79 |
| Groß MID | BHA @ COV | 5.6 | 4.78 |
| Saka MID | ARS @ SUN | 9.5 | 4.75 |
| Enciso MID | IPS @ CRY | 5.5 | 4.75 |
| Mykolenko DEF | EVE @ TOT | 4.5 | 4.65 |
| Semenyo MID | MCI @ MUN | 8.4 | 4.64 |
| Calafiori DEF | ARS @ SUN | 5.8 | 4.59 |
| McGinn MID | AVL vs NFO | 5.4 | 4.54 |
| Garner MID | EVE @ TOT | 6.0 | 4.49 |
| Bogle DEF | LEE vs NEW | 4.5 | 4.46 |
| Rice MID | ARS @ SUN | 7.4 | 4.45 |
| Haaland FWD | MCI @ MUN | 15.5 | 4.44 |
Want this for your own team?
Upload a screenshot of your squad and get who to sell, who to buy and why — based on your money in the bank and free transfers.
start 3-day trialHow these numbers are produced
Minutes are sampled per player, team goals drawn from a Poisson process from attack and defence strength, then goals, assists, saves, cards and bonus allocated under official FPL scoring. xPts is the mean of 500 simulations; floor and ceiling are the 10th and 90th percentiles. See the method or the FAQ.