Well, that hasn’t gone to plan. Unless coronavirus spares their blushes by putting an end to the season, Manchester City are looking at finishing with 20 points fewer than 2018–19. After a ludicrous 198 point haul over two seasons, City are back to Pep Guardiola’s first season in pure results terms. Mike Goodman wrote about City’s struggles for StatsBomb recently and it seems like, while they have issues to work out, they’re still pretty good. But this is Guardiola we’re talking about here. “Pretty good” is not the acceptable standard. He strives for perfection. The noise for a while is that summer 2020 will lead to a bigger overhaul of City’s squad than usual. But where do they need to change things? Who is worth sticking with or twisting? Let’s take a look.
Goalkeepers
Ederson is having a season that could be described as “okay”. 
City’s number one has faced 226 non-penalty shots on target in the Premier League since he arrived in the North West. He has conceded 65 non-penalty goals, while StatsBomb’s post-shot expected goals model estimates that the average ‘keeper would have let in 66.16. He is a fairly average shot-stopper with exceptional passing range. Guardiola almost certainly isn’t going to compromise on the latter to find someone better at making saves, so Ederson’s place looks safe, but it remains noteworthy. Behind him, Claudio Bravo’s contract expires this summer, and as someone originally signed to be a starter, he’s probably on high wages for a number two. The talk of Zack Steffen returning to the club is likely fueled by cost-saving measures. After that, Scott Carson still exists to fill a homegrown spot, but unless he is a member of your immediate family, you don’t need to care about that.
Defenders
Everyone knows Aymeric Laporte is great. Everyone also knows John Stones and Nicolás Otamendi are prone to errors. But then again, the latter pair played a lot of football together before Laporte’s arrival, when City already had an excellent defence. They had the benefit of a midfield much more effective at preventing counters back then, but recent form feels like a fairer indication of their abilities. We’ll get to that later. It seems very likely that Laporte will retain his status as City’s first-choice centre back when he eventually gets fit. Adding a new signing to partner him might be the biggest no-brainer prediction anyone will make about the transfer market. John Stones’ age and homegrown status probably keep him around for a little while longer, so Otamendi looks the most likely casualty if another club can pay his wages. Fernandinho can see out the last year of his contract as fourth choice. In theory, City have two different and exciting options at left back. Benjamin Mendy perfectly fits the mould of a conventional attacking fullback, combining an excellent ability to move past people with really good technical delivery from wide areas. Oleksandr Zinchenko, meanwhile, fits as a more distinctly Guardiola fullback, comfortable enough in possession to fill in as a midfield player. The graphic below shows how differently they progress the ball. Zinchenko moves the ball into the final third a lot, more akin to a central midfielder at another club. Mendy does less of this, but once it’s in the final third, he’s key in then getting the ball into the box, more akin to how most teams use attacking left backs. 
It’s a little difficult to show in the data, but as anyone can see, their problems are in the defence. In past years, City’s ability to prevent teams moving the ball through midfield helped Mendy and Zinchenko from being badly exposed. That’s not happening now, and they look less than invincible. The best option would be to fix the midfield, but let’s assume that’s rather difficult to do. City then need an excellent two-way left back who is likely to cost a lot of money, with one of Zinchenko or Mendy making way. Right back was supposed to be an area City fixed last summer. João Cancelo hasn’t flown in his first season, but this seems forgivable. The Portugal international spent last season at Max Allegri’s much deeper defending Juventus team, while the year before he was at Luciano Spalletti’s uninspiring Inter side. This isn’t just a different league but a different style of football for him, and he should be granted patience. 
Kyle Walker is about to turn 30 and it already feels like he’s a step off his best, but he’s homegrown, so congratulations, Kyle.
Defensive Midfielders
Now here’s the meat of the issue. With Fernandinho ageing and no longer able to offer the energy he once did as a defensive midfielder, Rodri came in as his permanent replacement. But Rodri and Fernandinho are, well, different. 
Rodri is the kind of calm passer Guardiola himself played as, and later utilised Sergio Busquets as at Barcelona. But Fernandinho is a box-to-box midfielder moved into a deeper role, who used his mobility to break up opposition attacks and offered a slightly more direct passing option. Guardiola isn’t stupid, and he’s adjusted his midfield accordingly. City often play a 4-4-2ish formation that becomes a 4-2-4 in possession at times, with İlkay Gündoğan sitting alongside Rodri. The passmap from the defeat to Tottenham clearly shows the shape (we’ll talk about that attacking four in the next section). 
The problem is that Gündoğan and Rodri don’t really compliment each other. Both tend to be fairly passive in possession and lack the mobility to be aggressive without it. Gündoğan’s many injury problems certainly seem to have taken their toll, and at age 29, he’s only likely to physically decline further. 
It seems as though Guardiola is fine with this partnership, which is mystifying. If for some reason he were to ask my opinion, I’d suggest investing in a physically dominant box-to-box midfielder, not unlike Arturo Vidal when he signed for the Catalan’s Bayern side.
Attackers
A lot of players have been condensed into this section due to how fluid Guardiola’s selections can be —plus, there’s just a lot of talent to choose from. Kevin De Bruyne, Bernardo Silva, Riyad Mahrez, Raheem Sterling, Sergio Agüero, Gabriel Jesus: they’re all great, and they’re all likely to be great again next season. Someone who isn’t going to be great for Man City next season is David Silva. It seems like the club really are just going to follow through and let Phil Foden take his (admittedly fewer) minutes next season, which is admirable. There will be a short-term cost to this, but City have so many good attacking options that they can cope if he struggles next year. 
Leroy Sané has been injured throughout the season, though it seemed like he had one foot out the door anyway. But he did offer something totally absent now. Sané is almost unique at the top level of modern football as a left-footed winger who exclusively plays from the left, which lets him stretch the play differently than City’s other attacking options. With Sané playing wide on the left, Zinchenko can more easily come into midfield without the team losing structure. He’s not just tactically interesting, though. He’s quite good at kicking the ball with that left foot of his, too. When looking at his left-footed shots from the left side of the pitch over two seasons, he makes a mockery of xG. No one else does that. 
City were linked with Mikel Oyarzabal last summer as a replacement. This makes sense, since the Real Sociedad man is also a rare left-footer who plays on the left, but with nothing like the same verve. 
The Sané question is a fairly minor one, though. City’s substantive issue is the ease with which teams are breaking their press. Their xG per shot conceded is the second-worst in the league, strange for a side that’s elite at basically everything else. The assumption is that the midfield is more porous without Fernandinho, opening up the defenders to increased vulnerability. Fernandinhos don’t grow on trees, and a collective rethink of the defence and midfield will be needed to get to 2017–19 levels. It sounds like a tall order, but Pep Guardiola has done more from worse positions in the past.