Small sample. Only 403 minutes this season, so the rates below swing heavily on a handful of moments. Read them as description, not as form.
Non-penalty xG per 90
0.454 non-penalty goals from 14 shots
league 0.14
Expected assists per 90
0.422 assists, 14 key passes
league 0.11
Finishing
+2.20goals above what the league converts the same chances into
Build-up per 90
0.908% of Manchester City’s build-up value
league 0.30
Mathis Cherki has not reached the 450-minute floor this season, so they are not ranked against the league yet. They shoot from 17.3m on average. They carry 12% of Manchester City’s attacking involvement, the 2nd largest share in the squad.
Against their position
87 midfielders
Average match rating8.0299th
Shooting
per 90
Goals
0.8999th
Non-penalty goals
0.8999th
xG
0.4595th
Non-penalty xG
0.4597th
Shots
3.1394th
Shots on target
1.0086th
Shot accuracy
–not ranked
npxG per shot
0.1490th
Shot distance
17.3m76th
Set-piece share
4%24th
Goals against xG
+2.098th
Creating
per 90
Assists
0.4594th
Expected assists
0.4294th
Key passes
3.1397th
npxG + xA
0.8798th
xG chain
1.3199th
xG build-up
0.9097th
Assists against xA
+0.168th
On the ball
per 90
Touches
66.0083rd
Passes
52.6783rd
Pass completion
92%93rd
Take-ons attempted
2.6775th
Take-on success
–not ranked
Dispossessed
1.6713th
Defending
per 90
Tackles
0.002nd
Tackles won
–not ranked
Interceptions
0.0013th
Clearances
0.0010th
Aerials won
0.009th
Aerial success
–not ranked
Dribbled past
0.0094th
Discipline
per 90
Fouls committed
1.3334th
Caught offside
0.0063rd
Errors
0season count
Yellow cards
1season count
Red cards
0season count
Ranked against 87 midfielders with half a match a round or more, the same comparison the compare page makes. Where less is better, such as being dispossessed, the order is turned round, so a high percentile is always the good end.
This week
Gameweek 8
Next fixture
Gameweek 8
away toAston Villa
Sat 24 Oct, 11:30 UTC
Aston Villa concede 15.8 shots a match and are 1.02× the league norm against central attack (31 shots so far). A favourable draw.
Projected shots
3.39
3.13 per 90 × 1.08
Projected on target
1.29
1.12 per 90 × 1.16
Season rate times what this opponent concedes, over a full ninety minutes. Check availability.
Shots
6 appearances
Over 0.55 of 683%± 15
Over 1.54 of 667%± 19
Over 2.53 of 650%± 20
Over 3.51 of 617%± 15
The band is one standard error, the tick the estimate. Blanks count as misses, 1 of 6.
Corners and set piecesSupplies 12% of Manchester City’s dead-ball danger, from 5 deliveries that became shots.
When it mattered1 of 4 goals came with the game level or Manchester City behind.
On the pitch75% of available minutes.
What would actually have to be replaced. Losing the penalty taker is not the same as losing a proportional slice of Mathis Cherki’soutput.
Chances received
who supplies them
Josko Gvardiol0.8 xG2 shots, 2g
Phil Foden0.5 xG1 shot, 1g
Antoine Semenyo0.3 xG3 shots, 1g
Ayyoub Bouaddi0.1 xG2 shots, 0g
Enzo Fernández0.1 xG1 shot, 0g
Chances created
who they supply
Erling Haaland0.6 xG3 shots, 0g
Phil Foden0.5 xG2 shots, 0g
Josko Gvardiol0.4 xG2 shots, 1g
Marc Guehi0.1 xG1 shot, 1g
Iliman Ndiaye0.1 xG2 shots, 0g
Season record
2026/27
Involvement in build-up
excludes their own shots and final passes
Mathis Cherki carries 7.8%of Manchester City’s build-up value.
Build-up per 900.899
98% of midfielders
Possession value per 901.306
100% of midfielders
Thin sample. Only 403 minutes, so this describes a cameo. Build-up is the value of possessions a player was involved in, with their own shots and final passes removed, so it credits the work before the chance rather than the chance.
Match by match
Match log
6 appearances, oldest first
over 1.5
2g
at Palace3 shots 3 on target
v Coventry3 shots 0 on target
at Man Utd1 shot 0 on target
1g
v Sunderland2 shots 1 on target
1g
at Liverpool5 shots 1 on target
1 blankno shot not placed
Height is shots, solid is on target, and the dashed rule is the 1.5 line. The feed cannot date an appearance with no shot in it, so one sits at the end, counted but unplaced.
How to read the ties. Each rate ties back to the league average: solid where the sample is settled, broken where it is still being shrunk toward the league, an open circle where there is no cohort to rank against. 403 minutes leaves every tie here broken.
