FPL GW5 review: City chaos, Brighton's roar and the model receipts

Gameweek in 30 seconds

• Man City won 5-3, Semenyo scored 17 and Haaland returned six. The attack remains premium; the defence currently resembles an unlocked conservatory.
• Brighton flattened Arsenal 3-0. Groß supplied 14 points, but Arsenal's Leeds–Forest–Everton run still argues against an emotional bonfire sale.
• Coventry's 1-0 win at Forest demolished the week's most fashionable stack. Muñoz, Gibbs-White and Neco Williams all played; the result, not the roles, failed.
• Barnes, Isak, Cherki and Barry rewarded published calls. João Pedro's zero-minute absence justified the caution; Porro's 18-minute cameo did not.
• Brobbey's hat-trick deserves applause, not an instant transfer. Sunderland's coming run is playable, but 17 points remain a party, not a pension plan.

Scoreboard and results

Gameweek 5 delivered 27 goals, seven clean sheets and an average of 44 points, although the official FPL feed still has the event marked as provisional. In other words, the football is finished but the spreadsheet has not yet removed its hi-vis vest. City own the headline after beating Sunderland 5-3, a match in which the champions won, conceded three and were outscored 4.32 to 2.72 on expected goals according to the Premier League's report. Semenyo's two goals and assist produced 17 FPL points; Brobbey matched him with a hat-trick for the losing side, the weekend's neatest proof that fantasy football and actual football remain on speaking terms but do not share a bank account.

Brighton supplied the proper statement result. A 3-0 win over Arsenal featured a goal and assist for Groß, with Kostoulas and the hosts' supporting cast turning a difficult-looking fixture into a seaside demolition. Aston Villa edged Spurs 3-2 as Manzambi scored and assisted on his first start. Coventry then completed the upset of the round at Forest: Jay Dasilva scored, goalkeeper Carl Rushworth assisted and a large pile of FPL clean-sheet projections quietly asked for the lights to be turned off.

Elsewhere, Barry won it for Everton against ten-man Ipswich, Barnes created both early Newcastle goals in a 2-1 victory over Hull, and Isak settled Bournemouth 0-1 Liverpool. Leeds and Palace produced the lone goalless draw. Brentford thumped Chelsea 3-0, while Cunha's 89th-minute equaliser rescued Manchester United at Fulham. United now have five points from five, which is less a launchpad than a suspiciously damp doormat.

Stock up, stock down

Stock up: Semenyo now has both a premium-team role and the sort of explosive route that makes midfield slots jealous. The 17 is not his new weekly allowance, but 90 minutes in City's attack matters. Groß was even more instructive: 90 minutes, a goal, an assist, a clean sheet and three bonus points. Brighton visit Sunderland and host Palace next, so his role deserves attention before Liverpool and City darken the corridor. Barnes also earned a promotion after two assists in 90 minutes. Newcastle's Coventry–Villa–Palace–Everton–Fulham run is the cleanest five-match runway on the board, and Lewis Hall adds a defensive route after scoring 13. Isak's 80 minutes, goal and maximum bonus were another welcome confirmation that Liverpool's new centre-forward is not merely decorative furniture.

Barry's six points look modest beside the weekend's fireworks, but the process was sound: a budget forward projected for 85 minutes played 90 and scored the winner. That is far more repeatable than buying Brobbey because he briefly turned the Etihad into his personal five-a-side pitch. Brobbey belongs on the watchlist. Sunderland face Brighton, Bournemouth, Leeds, Coventry and Chelsea; the fixtures invite further evidence, not a knee-slide towards the transfer button.

Stock down: the Forest concentration. Muñoz, Gibbs-White and Neco Williams all logged strong minutes, so this was chiefly an outcome miss, but a home defeat to Coventry is a fair reason to trim the badge premium. Forest's next five include Arsenal and City, with only Ipswich offering obvious relief. João Pedro's absence was better handled: the final coverage had already cut him to 33 expected minutes and warned against buying without bench cover. Porro is the more serious miss. After being reported fit, the final press model gave him 82.7 expected minutes; he played 18. Aina and O'Reilly supplied similar reminders that expected minutes are evidence, not an alibi.

Signal, noise and model accountability

Now for the receipts. Haaland led the captaincy model at 9.00 expected points and returned six: not a triumph, not a disaster, and still the correct conventional armband when judged before kick-off. Cherki beat 6.52 xPts with eight, Barnes beat 4.85 with nine, Isak beat 4.15 with eight and Barry landed almost exactly on his 5.41 forecast with six. Those were useful calls because the minutes survived contact with reality. The model also treated João Pedro as a major selection risk, and his zero was a warning doing its job rather than clairvoyance in a waistcoat.

The misses require less perfume. Forest were overestimated across attack and defence: Gibbs-White, Muñoz and Neco Williams combined for six points against Coventry despite three high forecasts. Their minutes were close, so the error sits with the match expectation and one-game outcome, not a hidden benching. Porro was different; 18 minutes against an 82.7 estimate is a selection miss. Aina's zero minutes against 80.2 expected and O'Reilly's five against 75.1 belong in the same file. When role evidence breaks, the response is to update availability and lineup inputs—not to blame finishing variance for a player who was watching most of the finishing.

Across all 659 archived players, the provisional mean absolute error is 1.51 points, with forecasts 0.56 points high on average. That broad pool includes hundreds of non-starters, so the more demanding cohort is the 289 players projected for at least 45 minutes: MAE rises to 2.63, with 37.7% finishing within two points. Among the top 25 xPts forecasts, MAE is 3.77 and the model over-predicted by 1.78 points on average. It is one wild Gameweek, not a referendum. Brobbey, Semenyo and Groß all landed deep in the positive tail; premium and Forest blanks pushed the other way. We record the cohort, sample and direction precisely, then refuse to turn one Saturday into a permanent personality test for the model.

The forward view is already more interesting than the autopsy. The current GW6 model has Bruno Fernandes first at 8.02 xPts at home to Spurs, with Haaland at 7.15 away to Liverpool and Saka at 7.15 against Leeds. Lewis Hall projects at 6.09 at Coventry, while Barnes sits at 5.16. Cherki remains attractive at 4.87, but 62.3 expected minutes make him a less comfortable buy than his eight-point return suggests. Treat these as a Sunday watchlist: an international break sits between the evidence and the 10 October deadline, which is ample time for hamstrings and headlines to form a small coalition.

Five actions for next Gameweek

1. Target Newcastle's run, but buy the role you need. Barnes offers attacking value, Lewis Hall offers defender upside and Wissa is the forward watch; Coventry, Villa, Palace, Everton and Fulham is a proper five-match platform.
2. Hold Arsenal assets through the Brighton bruising. Leeds at home, Forest away and Everton at home come next; Saka's current 7.15 GW6 projection and the defenders' schedule make a rage sale needlessly theatrical.
3. Keep City attackers, then reassess the pecking order. Liverpool away is awkward, but Ipswich, Villa, Brighton and Forest follow. Haaland remains elite; Semenyo has earned respect; Cherki needs a fresh post-break minutes check.
4. Do not chase Brobbey's 17 or bury Forest after one ambush. Sunderland's fixtures justify monitoring the striker, while Forest still have Ipswich in GW8. Demand another role-and-chance sample before making either verdict permanent.
5. Roll the transfer if your squad is healthy. The GW6 deadline is 10 October, and the break will deliver more useful availability evidence than any price-chasing sprint performed with one boot still in GW5.