How Problematic is Manchester City’s Leaky Defence?
Manchester City are top of the Premier League with a perfect record of five wins from five, but they haven’t been entirely convincing at the back. Is there any cause for concern?
“I prefer to win 1-0 than 5-3.”
Enzo Maresca’s comments after watching his Manchester City side’s rollercoaster victory over Sunderland on Sunday afternoon sparked an interesting philosophical debate. Rather than lean towards the are-you-not-entertained ethos of an Ange Postecoglou, Maresca intimated he would like to see rather more control from his team, and a little less excitement than they provided in this win.
The Italian may well prefer to watch a game with lots of goals, but he was likely stating this preference on this particular occasion because the chaos of City’s win, in which they needed to retake the lead three times in the first half alone – something that has only happened six times in Premier League history to any team – was not exactly the mark of champions. As the scoreline suggests, City allowed their opponents far too many opportunities to score.
Sunderland are a very good team but not one that routinely creates a lot of chances. On this occasion, away to title hopefuls, they had chances worth 3.57 expected goals, which is by far their highest since their return to the Premier League last year.
In fact, in 43 games since the start of last season, they’ve only managed more than 2.0 xG twice, with their previous best being 2.44 vs Newcastle in March. They’ve had under 1.0 xG in more than half (22) of their matches.
But they carved City open time and again. Brian Brobbey scored a hat-trick of what were essentially tap-ins, allowed into areas where he could not miss, time and time again. And at 4-3, Sunderland came within a Rúben Dias goal-line clearance of equalising again.
While match-going fans got their money’s worth and neutrals were thoroughly entertained, even though City won and kept up the only remaining 100% record in the Premier League, it’s understandable that Maresca came away a little dismayed at his side’s defensive display.
“It’s much better to win games,” he said. “In this moment, it’s better to have three points more than less, no doubt. We scored five. We could have scored probably more goals.
“So that means that we created enough, but at the same time we need to avoid conceding goals like today. We had four clean sheets in a row and today we conceded three goals.”
It is a fact that City had kept four successive clean sheets before the Sunderland game, but it was slightly disingenuous of Maresca to try and reframe their form that way. That run included the midweek win over Championship side Norwich in the EFL Cup and a 1-0 win at home to newly-promoted Coventry, who caused City all manner of problems and could quite feasibly have scored three goals, let alone one.
Credit where it’s due: City were mightily impressive defensively in their 1-0 win at Manchester United, keeping their rivals at arm’s length despite playing most of the derby with 10 men. But across that four-game run of clean sheets, they conceded chances worth 2.82 xG. So, in other words, they could reasonably have been expected to concede three goals.
They do have Gianluigi Donnarumma in goal, so perhaps they just felt confident that they could rely on one of the best shot-stoppers and most commanding goalkeepers of his generation.
But the evidence from the Sunderland game suggests that is not a reliable method of defence. Sunderland’s 3.57 xG is the third highest ever by a team that went on to lose a Premier League game since xG data is available (2013-14). On another day, Sunderland would have won that match, just like United will feel aggrieved not to have won the derby, and Coventry were left rueing missed chances against City.
On the opening weekend of the season, too, City required 84th- and 91st-minute goals to come from behind and rescue a win at home to Bournemouth.
In other words, their ‘perfect’ record of five wins from five has been rather less convincing than it might appear from a quick glance at the league table. They are clearly feeling the loss of Rodri from the base of their midfield.
They have, though, still won each game, which is something no other team can say. They are able to rely on the goalscoring of perennial golden-boot winner and early leader in the top-scorers’ stakes, Erling Haaland (five goals).
Rayan Cherki (three goals, two assists) and Antoine Semenyo (two goals, three assists) have started the season very well, too, while Enzo Fernández (one goal) provides further attacking threat from the double pivot. Iliman Ndiaye looks to have settled well following his summer move from Everton, and Phil Foden (two assists) is another exceptional attacker when he’s not kicking Bruno Fernandes in the you-know-wheres.
But is their arguably unrivalled attack enough? Can they rely on being bailed out often enough to win the title? “Attack wins you games, defence wins you titles,” as goes the famous Sir Alex Ferguson quote. If that’s true, then City might have something to worry about.
Indeed, only Brighton (who have played two halves of football against 10 men) have scored more goals (16) in the Premier League this season than City (13), but six teams have conceded fewer than them (5). Those teams are Leeds, Everton, Brentford, Hull and, perhaps crucially, likely title rivals Arsenal and Liverpool.
man city xg premier league 2026-27
man city xg premier league 2026-27
“There’s no way Manchester City win the league,” Sky Sports pundit Paul Merson said after the Sunderland game. “Going forward, [they] were good but at the back they were all over the place. [That] catches up with you.”
In defence, City certainly didn’t look like a team who would win the title on Sunday, but might their attack be sufficient to see them through?
In the 31 completed 20-team Premier League seasons, the team with the best attack has won the league 21 times. The team with the best defence has won the title 15 times. On the basis of those numbers, it appears as though attack trumps defence.
But on many of those occasions, the title winner has had both the best attack and the best defence, which isn’t – we’re assuming, given their start to the season – something City are going to manage.
In those 31 seasons, there have been 12 instances of the team with the best attack and not the best defence winning the title, compared to just six times that the title winner has had the best defence and not the best attack.
That backs up the previous finding: that Fergie wasn’t quite right; goals tend to win titles ahead of sturdy defences.
Clearly, though, neither is any kind of guarantee. And if you’re capable of scoring at a rate of 2.6 goals per game – as City have done so far – for the whole season, you’re always going to be in with a decent chance.
That said, history suggests we shouldn’t read too much into City’s perfect start. They are the 15th team to begin a Premier League season with five wins from five, but only four of the previous 14 have gone on to win the title.
And two of those came in the first four times it happened, meaning that only two of the last 10 teams to win their first five games have won the league. Most recently, for example, Liverpool won their first five games last season, and they only just snuck into fifth place.
City’s form looks more sustainable than Liverpool’s did 12 months ago, and a slump to fifth appears pretty unlikely, but if they carry on like this, there may be reason not to expect them to remain quite so fortunate.
The expected points table is arguably a better indicator of performance than the actual table, as it uses expected goals data to show how every game ‘should’ have finished based on the quality of chances at both ends of the pitch. And in our expected points table, City are second to Arsenal.
Actual position vs expected position - Premier League expected points table after MD5 26-27
Actual position vs expected position - Premier League expected points table after MD5 26-27
xG data becomes more reliable over a greater number of games than the five we’ve had so far, so if City are to continue as they have started, Arsenal may well reel them in.
Then again, given they have Haaland, Cherki, Semenyo, Fernández, Foden, the returning Jérémy Doku, and goalscoring defenders like Nico O’Reilly and Josko Gvardiol, there’s at least the possibility that City will continue to have too much in attack for any opponent they come up against.
If they are going to become champions again, we may well be in for an entertaining season.
Premier League Opta Stats
Premier League Opta Stats