City barely had the ball. When they kept it, they went straight through.

Manchester City spent 26.8 minutes on the ball at the Etihad, their sixth-lowest in 43 league matches across two seasons, and played 93 passes out of their own third against a season average of 139. Sunderland’s front three pressed closer to City’s goal than any side has managed this season, and used fewer pressing runs doing it. That won the ball high and produced 1.42 expected goals. It also gave City the cleanest line-breaking route they have had all year. Both of those are the same decision.

Twenty-seven minutes on the ball

Before anything else: City were not allowed to settle. They had less of the ball than in all but five of the 43 league matches in this sample, played fewer passes out of their own third than in all but six, and kept restarting from scratch.

MINUTES ON THE BALL

White tick = City’s 2025/26 average of 35.4 minutes. Only the Manchester derby has kept them lower this season.

PASSES PLAYED OUT OF THEIR OWN THIRD

White tick = 2025/26 average of 125. City are a side built on circulating the ball in their own half until a line opens. Ninety-three touches of circulation is not enough to find one.

THE READ

The volume numbers all point the same way, and they are not small effects: this was one of the handful of matches in two seasons where City could not get the ball down and keep it. Whatever else the afternoon became, the pressing plan did the first thing a pressing plan is asked to do.

Not a chase. A grid.

From tracking, across 995 frames with City settled on the ball in their own third: how much space each City line was given, measured as the mean distance to the nearest Sunderland player.

SPACE ALLOWED, BY CITY LINE

Lower is tighter. Sunderland were the tightest side City have played this season on the holding midfielders and the full-backs, the two outlets City use to turn a back-line build into progress.

WHERE THE PRESS LIVED

Measured over the same frames, in City’s attacking direction: 0 is City’s own goal line, 100 is Sunderland’s.

The two panels say the same thing from opposite ends. Sunderland’s front three stood closer to City’s goal than any other side has managed this season, with roughly nine players inside City’s half, and the man on the ball had under seven metres of room. They did it on half the pressing runs of the average City opponent this season.

That is not a chasing press. It is a marking grid: hold the shape, take the receivers away, and let the ball-carrier find out there is nobody to pass to.

The plan, two minutes in

City play out from Donnarumma. Every City midfielder has a red shirt on him, both wingers have dropped onto the full-backs, and Brobbey alone covers the two centre-backs. Fifteen passes and forty-three seconds later City have got no further than the halfway line, and the move ends with the ball lost back in their own half.

The marking map

Over the same frames: the Sunderland player who was nearest to each City man most often, and how much room that City player had. This is where the players actually were, not a team sheet.

Man-oriented almost all the way through: Brobbey shuttling across both centre-backs, Le Fée on Anderson, Xhaka on Cherki, Sadiki on Fernández, both wingers dropped onto the full-backs, and Ballard stepping out of the back four onto Haaland.

Room was left at one end of the list only: Ndiaye and Semenyo on the flanks, and Guéhi, the left centre-back, who Brobbey shaded onto barely a third of the time while spending 69% of it on Dias. Leave the far side and one centre-back free, take everything else away. That reads like a deliberate trade, and it is the same trade section 06 comes back to.

The first half is where it bit

Split the same build-ups either side of the break and the halves are not the same match.

Heights are on the same 0-100 scale as section 02, measured in City’s attacking direction: 0 is City’s own goal line, so a lower regain number means Sunderland won the ball closer to City’s goal.

In the first half Sunderland won the ball four times inside City’s own third and the median regain landed 16 points of pitch closer to City’s goal. After the break: zero.

The tracking says the press did not loosen. If anything Sunderland stood tighter and higher after the interval. What changed is that City stopped accepting the invitation: they spent 231 fewer seconds settled on the ball in their own third and went over the press instead of through it.

Straight after the restart

Fifty seconds into the second half. City work it out to the right, get barely a third of the way up the pitch, find nothing and come all the way back. The ball is lost inside their own penalty area. This build-up never leaves their own third.

What the press actually produced

A press is worth what it creates. Every shot in the match, sorted by where the possession that led to it began.

SUNDERLAND SHOTS, BY WHERE THE MOVE STARTED

Bars are expected goals; hover for the shot count.

MANCHESTER CITY SHOTS, BY WHERE THE MOVE STARTED

Same scale, so the two panels can be read against each other.

Winning the ball high gave Sunderland three shots worth 1.42 expected goals, including the equaliser on eleven minutes. That is a press paying for itself: roughly a goal and a half of chance from three moves. City, building out of their own third against it, managed three shots worth 0.94.

Eleven minutes: the press pays

Gvardiol is forced into a clearance under pressure and Methalie wins the ball back inside City’s own third. Sunderland recycle it, come again, and thirty-seven seconds later Brobbey makes it 1-1, at 0.55 expected goals, the best chance of the first half.

Through it, not around it

Here is the honest other half. Sunderland cut how often City built, but not how far each surviving build-up travelled. Per build-up, City converted at almost exactly their own rate.

The reason is in the next table. A man-oriented grid has one structural cost: beat your marker and there is no second layer behind him. Counting line-breaks, completed open-play passes under 30 m out of the own third that take at least two opponents out of the play.

City broke a line in 15 of their 38 build-ups: 39.5%, a higher share than in any other match they have played this season, and two-thirds of every City arrival in the final third came through one. Their progression rate out of their own third was their fifth-highest of the 43 matches in this sample.

So both things are true at once. The grid held the volume down and won the ball high; each time it was played through, City were immediately into open space. That is the trade the plan makes, not a mistake in it, but it is the number that decides whether this press is worth running against a side of City’s passing quality.

Where the match actually went

Eight goals, and only one came out of the phase this report is about.

Sunderland created more than City and lost by two. Four of City’s five goals began with a set piece, a rebound, or Sunderland giving the ball away around their own half, not with City playing out. The one City goal that came from a genuine build-up out of their own third arrived on 80 minutes at 4-3, with the game already stretched.

Worth separating cleanly, because the two problems are unrelated. The pressing plan did its job. The afternoon was decided in the restarts, the rebounds and the giveaways.

The read

FOUR THINGS THIS MATCH SAYS

1. The press is an identity, not a one-off. Sunderland’s front three has sat inside a three-point band on the same scale in all five league matches this season, with eight or more players in the opposition half every week. City got the most aggressive version of it.

2. It did the first job. City’s 26.8 minutes on the ball, their 93 passes out of their own third and their 36% long-ball completion are all bottom-nine figures across two seasons.

3. It paid: 1.42 expected goals from three moves that began with a high regain, the equaliser among them.

4. It also cost. City’s line-break rate against it was the highest they have recorded this season and their progression rate out of their own third the second-highest, behind only the opening day. That is the lever to pull next.

This was built from the event and tracking data of one match, benchmarked against both sides’ own seasons, in an afternoon. MyGamePlan builds these for clubs every week, on their own matches and on the next opponent’s.