ManchesterCity have conceded a goal from their opponents’ first shot on target in each of the past four games.
“Today it’s Edi (Ederson) or Rodri (to blame), the next time it will be a full-back, the next time a central defender, another day whatever.”
City dominated the game against Forest and missed a hatful of clear-cut chances, only to defend badly once and be exposed for it.
Read Also
City are giving their opponents too many one-on-one chances with the Brazilian but questions have been asked about his management of those situations.
One example came against Aston Villa the weekend before last, when Bernardo Silva gave the ball away to allow Ollie Watkins a clear sight of the goal.
Against Tottenham in January, when Guardiola bemoaned the goals given away, Ederson played a ball to Rodri under pressure and gave Dejan Kulusevski a relatively simple task.
The numbers do not provide any solace: he is underperforming in terms of shot-stopping, having conceded about four goals more than the data suggests he should have, allowing 20 goals this season from an xGOT (expected goals on target) value of 15.8.
xGOT, essentially, measures how likely, on average, a goalkeeper is to save that shot, based on the combination of chance quality (xG), the placement of the shot and the angle from which it is taken.
Average xG per shot faced
| Season | xG/shot against |
|---|---|
| 2022-23 | 0.1 |
| 2021-22 | 0.1 |
| 2020-21 | 0.08 |
| 2019-20 | 0.13 |
The graphic below shows the placement of shots on target faced by Ederson in the league this season, with goals in red and bigger circles representing higher-quality shots.
Drilling down into the first shots on target he has faced in every game this season, he has conceded 10 from an xGOT value of 8.3, another underperformance.
In terms of those recent examples against Tottenham and Villa, The Athletic’s goalkeeper expert Matt Pyzdrowski is better placed than many to have a say on the ins and outs of goalkeeping.
“He had been getting much better in the past two seasons, much more in control of his decision-making, but he’s regressed a little bit into his old technique of just rushing out and trying to smother the attacker instead of being more composed and patient.
Pyzdrowski says there are perfect examples of these from the past few games alone, starting with Kulusevski’s goal at the Etihad after Ederson’s poor pass to Rodri.
“He would have been better suited staying set closer to his goal line, and if he did he probably would have made the save.
This is the perfect example where, had he been calmer and more in control, it would have given him a better chance of making the save.”
Kane’s record-breaking goal appears similar on the face of it, but Pyzdrowski spares Ederson on this occasion.
“To a fan, I could understand why they would think that’s a ball he should save — because it’s not pounded into the net — but the easiest way to explain it is this: in baseball, if you’re expecting a fastball but they change it up, that’s very difficult.
Pyzdrowski also argues that goalkeeping xG stats can be a little unfair because things such as the position of defenders, and the pace and cleanliness of a shot — including if it is deflected — are not taken into account by all models.
“That’s not to take any blame off of Ederson — someone like Liverpool keeper Alisson is at the top of those statistics because he is so excellent — but when you give up so few chances in the game it’s hard to have a really good xG metric.
I always found it more difficult when I was playing in a team where we weren’t getting a lot of opportunities against us and then you’re expected to make a save.
Conversations about Ederson’s future would be premature — he brings so much more to the table than merely shot-stopping and would not be easy to replace — and even if he is underperforming, his problems do not exist in isolation, clearly, as City keep giving him one-against-ones to save in the first place.
“That’s football, it happens sometimes,” Guardiola said after the Forest game, and that is one conclusion to draw: these 10 goals are something of a statistical quirk, or can be put down to things like good finishing, bad luck or penalties.
The goal they conceded on Saturday against Forest highlights how well City had defended overall, as much as anything.
Goals from first shot on target
| Opponents | Minute | |
|---|---|---|
| Chris Wood | Nottingham Forest | 84 |
| Bukayo Saka (pen) | Arsenal | 42 |
| Ollie Watkins | Aston Villa | 61 |
| Harry Kane | Tottenham | 15 |
| Demarai Gray | Everton | 64 |
| Pascal Struijk | Leeds | 73 |
| Andreas Pereira (pen) | Fulham | 28 |
| Leon Bailey | Aston Villa | 74 |
| John Stones (OG) | Crystal Palace | 21 |
It was the only chance Forest created all game.
Many of these goals conceded have come late in the second half, which is certainly disappointing but shows that teams find it very difficult to create chances for long periods.
But that goal came from some very poor defending and City really have given away cheap goals of late.
In the 2019-20 season, when City struggled to contain opposition counter-attacks, the average xG value per shot faced was 0.13, which is easy to quantify given it was the highest (ie, worst) of any team in any season for which data was available.
City got the figure down to 0.08 the following season, a fairly significant change that showed they were not giving their opponents quality chances on the whole.
The problems of the past four games are pretty easy to identify in that the defending has been poor or City have made a mistake in the build-up and Ederson has, on occasion, not been ideally placed to come to the rescue.
 or Rodri (to blame), the next time it will be a full-back, the next time a central defender, another day whatever. ”City dominated the game against Forest and missed a hatful of clear-cut chances, only to defend badly once and be exposed for it.){kind=link}