Fábio Carvalho's finish in the 90th minute on Friday night read, on the scoreboard, as the third goal in a comfortable home win for Brentford. On a longer ledger it did something else. It pushed Chelsea goals conceded in the Premier League across the 2026 calendar year to 43, a figure no other club in the competition has matched, per the count circulated by Squawka.
Crystal Palace sit second on that list with 42. Newcastle United and Tottenham Hotspur have each allowed 39. The gap at the top is one goal, which is the kind of margin that can flip on a single deflection — but Chelsea have led it for a while, and the trend within the year is the part that should concern them.
43, 42 and 39: the company Chelsea are keeping
| Club | PL goals conceded in 2026 |
|---|---|
| Chelsea | 43 |
| Crystal Palace | 42 |
| Newcastle United | 39 |
| Tottenham Hotspur | 39 |
Calendar-year totals cover Premier League matches played between January 2026 and 18 September 2026.
Split the 43 and the picture sharpens. Twelve of them have come in the opening five fixtures of 2026/27. The remaining 31 were conceded in the stretch of last season played from January onwards — which, set against Chelsea's final 2025/26 total of 52 in 38 games, means roughly 21 were shipped across the August-to-December portion of that campaign.
The defence, in other words, has been getting worse for three quarters of a year rather than three weeks.
It is also not a table Chelsea usually appear in. They finished 2025/26 tenth for goals conceded, one behind Crystal Palace and well clear of Burnley's 75 and Wolves' 68 at the foot of that particular ranking. Arsenal conceded 27 across the same 38 games; Manchester City 35. Chelsea were mid-pack, with 58 scored and a 14-10-14 record. The calendar-year number is what happens when a middling defence stops being middling in the second half of the sample.
Twelve in five: Chelsea goals conceded since August
| Date | Match | Result | Conceded |
|---|---|---|---|
| 24 Aug | Fulham 2-3 Chelsea | W | 2 |
| 30 Aug | Chelsea 4-3 Brighton | W | 3 |
| 6 Sep | Arsenal 2-1 Chelsea | L | 2 |
| 12 Sep | Chelsea 2-2 Hull City | D | 2 |
| 18 Sep | Brentford 3-0 Chelsea | L | 3 |
Five matches, twelve goals, no clean sheets. Chelsea's last shutout in the league came in January, a 2-0 win over — of all opponents — Brentford. That is eight months without one.
The rate matters more than the raw total. Chelsea kept nine clean sheets in 38 league games last season, a 24% rate. This season the figure is zero. Their average of 2.40 goals conceded per game is up from 1.37 across the whole of 2025/26.
Set against Chelsea's own recent openings, the start is their worst defensively in more than a decade. Twelve conceded after five matches matches the 2015/16 side and sits far adrift of the 2018/19 and 2010/11 teams, who had allowed four and one respectively at the same stage. Points-wise the damage has been contained: seven, one fewer than last season's opening five and three fewer than 2024/25.
That containment is the attack's work. Chelsea have scored 10 goals from 52 shots and 26 on target, with Morgan Rogers and João Pedro on three each and Cole Palmer on two. Two of the three positive results were a 3-2 away win at Fulham and a 4-3 home win over Brighton. At Brentford, with no goals at the other end, there was nothing to cover the cracks.
The number travelled well beyond Stamford Bridge over the weekend, and not only among Chelsea supporters. For a fanbase whose own defensive year has been rough, being beaten to this particular table takes some doing.
The expected-goals gap, and an unsettled goalkeeping picture
There is a case that Chelsea are being punished harder than the underlying play deserves. One tracking dataset covering four of the five league matches puts their expected goals against at 1.28 per match against 2.25 actually conceded — a gap of nearly a goal a game, from roughly 13.5 shots allowed per match.
Small sample, obvious caveat. But it is a new kind of gap for this squad. Across all 38 games of 2025/26, Chelsea conceded 52 against an xGA of 52.49 — essentially dead on. Defences that track their expected numbers for a full season and then overshoot by a goal a game tend to be losing something specific: individual errors, set-piece organisation, or saves.
The supporting detail points in a few directions at once:
- Goalkeeping is not settled. Martínez has taken the majority of the starts, with Robert Sánchez also featuring. Five matches into a season is early for a rotation at that position.
- Discipline is heavy. Nine yellow cards and 43 fouls in five games, following a 2025/26 campaign that produced 98 yellows and eight reds — the worst card record in the division last season.
- The volume work is there. 126 clearances, 49 interceptions and 20 blocks in five matches suggests a defence that is busy rather than absent.
Bournemouth, Everton, Tottenham, United: the October test ahead
The fixture list offers a chance to reset rather than an escape. Chelsea's next scheduled league match is at home to Bournemouth on 10 October, followed by Everton away, then Tottenham and Manchester United at Stamford Bridge across the back half of the month. Three of those four are at home.
Of the 33 matches still to play, 17 are at Stamford Bridge and 16 away. The opponents Chelsea host have so far averaged 1.29 points per game on the road this season; the sides they visit have averaged 1.24 at home. On that measure the remaining schedule is close to neutral — no obvious run of relief, no obvious ambush either.
Chelsea sit tenth with seven points and a goal difference of minus two, six points behind Brighton in third and eight behind Manchester City, who have won all five. The attacking numbers are comfortably good enough for that gap to close. The calendar-year total of 43 is the reason it hasn't.
