City 5, Sunderland 3: You Can Score on This Team. You Just Can't Outscore It
City beat Sunderland 5-3 in an eight-goal thriller, and Brian Brobbey scored a hat-trick and still lost. No surprise that Sunderland out-created us either. By Sofascore's numbers it was 4.32 xG to 2.72, with Sunderland putting 9 shots on target to our 6. They played amazing and they're on fire, even with Enzo Fernández's free kick putting us ahead in the 9th minute. I don't read this scoreline as a defense falling apart, though. I read it as Maresca's system doing exactly what it was built to do, and the trade is the whole story: you can score on this City, but you can't outscore them.
Here's the read. I already had a middling view of how this team defends, and if you look back at Maresca's Chelsea it's the same movie: high line, aggressive stepping, plenty of goals let in, with worse players than he has now. The way to score on that is exactly what Sunderland did. Enzo Le Fée played balls over the top, Thomas Meunier and Brobbey ran onto them, and when City's offside trap got beaten before the 33rd-minute equalizer, Meunier was clean through and squared it for a Brobbey tap-in. VAR looked at it, it stood, 2-2. The catch for Sunderland is that playing that way leaves you wide open against what I think is the best attack in the league, which is how they still lost by two. Give them real credit, too. They lost 0-2 to Arsenal eight days earlier and 5-3 to us, played well in both, and got zero points out of it.
You can score on this City. You just can't outscore them.
Start with Enzo, because the free kick set the tone. Right-footed, from a difficult angle, whipped around the wall into the near post, and it was his first goal in a City shirt. That's the kind of finish that makes a £125m fee feel a little less insane. He was everywhere, winning the ball back next to Elliot Anderson in the high press, and by the second half he was playing more like a second striker than a midfielder while Cherki dropped deeper to get on the ball. I said before kickoff I wanted to see if Enzo would sit deeper to give Cherki the freedom to roam. It went the other way and it worked. My Fanalysis rating had him at 79, best on the team.
Then Semenyo, who had a hand in 3 of the 5. He set up Cherki's goal in the 29th minute, after Ndiaye won it in midfield, Gvardiol found Haaland, and Semenyo prodded it on for Cherki to finish calmly. Then he scored in the 43rd and the 57th, two goals off a combined 0.14 xG. On the first, Cherki cut inside and crossed with his right, it bounced around the box, Guéhi got a touch, and Semenyo smashed it. On the second, Cherki played him through the middle and Haaland decided to leave it for him at the back post. Semenyo is not going to beat you one-on-one every time, and his first touch isn't always clean, but he's more physical than whoever is marking him and once the ball gets to him he ends the move with a goal or an assist. He also covered 10.2 km and completed all 29 of his passes, and Sofascore had him at a 9.3 rating.
The other winger I wanted to see was Ndiaye, and I got what I hoped for. With Foden serving a three-match ban after the red card at Old Trafford, Ndiaye got the start on the right, pinched inside next to Cherki, and got into good positions all afternoon. The final ball took a while to show up and his touch let him down on a Haaland lay-off around the 52nd minute, but he was lively, and Haaland missed a sitter off one of his passes right before halftime. Fanalysis had Ndiaye at 78, second among the starters behind Enzo. Foden gives you familiarity and security building up. Ndiaye is relentless and hungry and wants the ball, and it's hard to compete with someone who wants it that badly when you've had every luxury of winning. Foden also misses Liverpool on October 11. Ndiaye has the spot right now, and he might close the door behind him.
The part I don't love is the back end. I said before kickoff that another clean sheet would add to my confidence in this defense. We got three goals against instead. Donnarumma missed two balls he came for around the 68th and 70th minutes, and either one could have made it 4-4, and Dias had to clear a Ballard header off the line after Ballard beat Donnarumma to a set piece. Claiming crosses was my worry about him last season, and FotMob had him in the 19th percentile among goalkeepers for high claims. This season it's up to the 39th, so it's better, but this game looked like a slide back toward last season's mean. He wasn't bad in goal, to be fair: Sofascore has him at 4 saves and +0.15 goals prevented. It was the claiming. Fanalysis gave him a 62, the lowest of anyone who started. The line ahead of him looked stretched as well. Brobbey kept dragging Dias forward by man-marking him, which left Guéhi out of position, and Nunes and Gvardiol were both pushed high on either side. I'd keep the high line anyway. A good center forward is going to pull Dias out of that flat four and open up the middle, and Brobbey spent the afternoon proving it, but the answer is to fix the panic, not to sit deeper.
Keep the high line. Fix the panic.
Haaland's goal in the 81st minute finished it. Doku slid the ball through to Gvardiol, Gvardiol crossed, Haaland tapped it in, and then we waited forever for VAR to confirm an offside call that wasn't close. Where is the automated offside? It's blatant and it takes forever. That goal completed a set for Haaland, too: he's now scored against every team he's faced in the Premier League, and Sunderland were the last one standing. Doku played the final 15 minutes and Fanalysis gave him a 78, which is a decent return on that little time. If you ask me who starts on the wings when everyone's fit, I say Semenyo and Doku, but I think it depends on the game. Foden and Ndiaye both have a case, and that's a good problem to have. City are 3 points clear of Arsenal at the top of the table with Liverpool up next. I'll take chaos when the chaos looks like this.
