The data behind the Manchester divide
By Jacob Whitehead and Mark Carey 4h ago 26
Manchester United might have only lost by two goals to Manchester City, but the gulf between these sides is seismic.
A 2-0 defeat may be less outwardly embarrassing than United’s 5-0 home defeat to Liverpool two weeks ago, but City’s derby victory was just as comprehensive.
Saturday saw United hit a series of statistical lows, several suffixed by the phrase “since Opta started collecting that data”.
A paucity of shots, a sidelined midfield and jailed inside their own half: this is the data and analytics that show just how much City outplayed Ole Gunnar Solskjaer’s side.
Limited chances
United’s expected goals tally (xG) of 0.73 was not especially low — a rebound missed by Mason Greenwood contributing to much of that — but the tiny quantity of chances demonstrates how difficult they found it to put City under any pressure.
They took five shots, the second-fewest of the Solskjaer era, and David de Gea had to save more shots from United players (two) than Ederson (one).
These shot maps clearly illustrate the disparity. Aside from Greenwood’s rebound, United (on the left) were reduced to long-range efforts and a couple of set-piece headers.
Pinned back
Watching the Manchester derby, it appeared the only meaningful action was taking place around De Gea’s goalmouth.
This is backed up by a metric called “field tilt”, a simple way of showing the territorial dominance between teams. It measures the balance of possession a team has in a game, considering only touches or passes in the attacking third.
United’s share of attacking-third possession was 22 per cent, their lowest tally since the beginning of last season. It effectively means City attacked United’s goal for nearly four times as long as United attacked City.
This imbalance was especially pronounced in the first half. United’s half-time possession was only 32 per cent — the results of their average position map were bleak.
Except for Victor Lindelof (No 2), who is sitting so deep that his average position is almost inside his own penalty area, United’s players have been compressed into a flat wall by City’s relentless pressure.
Cristiano Ronaldo (No 7) is the furthest United player forward, sitting just inside his own half, as if receiving a kick-off.
The verticality of the wing-backs is particularly damning. Overloaded with defensive duties, and without any opportunity for attacking endeavour, the average positions of Luke Shaw (No 23) and Aaron Wan-Bissaka (No 29) are barely more adventurous than the wide centre-backs (Eric Bailly, No 3 and Harry Maguire, No 5).
The Cancelo effect
This was not just a poor game for United — it was also an exceptional performance by City, who were indefatigable with and without the ball.
(Nominal) left-back Joao Cancelo had an outstanding afternoon. If you were to put on a blindfold, start rewatching the match and peek out at a random moment, it would likely show Cancelo attacking United’s right flank.
Look at the touch maps for Cancelo and Wan-Bissaka.
Aside from the sheer volume of Cancelo’s involvements, look at how far up the pitch they occur. He only has two touches in the segment closest to his own goal (compared to 11 from Wan-Bissaka), with the vast majority of the Portugal international’s touches occurring ahead of the halfway line.
These two maps are mirrors of each other. Cancelo’s advance provokes Wan-Bissaka’s retreat, like dance partners joined in a grim embrace.
Cancelo’s individual attacking output was comparable to United’s collective effort. According to Opta, Cancelo managed 45 passes in the final third — for comparison, United’s team total was 73. Ronaldo, their star attacker, had 27 touches. Cancelo had 125, 98 more.
Misfiring midfield
So far, we’ve seen how United’s opportunities in possession were limited, often beginning deep inside their own half.
Correspondingly, looking more deeply at touches shows their inability to impose themselves on the game and build any basis for attacking output.
Opta records Lindelof as having the most touches of any United player (67). Every one of City’s outfield players had more.
Donny van de Beek, cheered on to the pitch by the United fans, had only three touches in 15 minutes. He completed one of two passes, with the misdirected effort almost leading to a City goal.
However, one of the reasons Van de Beek was brought on was to help bring about quick pass-and-move sequences into a midfield that achieved little throughout the game. Compare these two pass maps.
One (above) shows how many times Scott McTominay and Fred managed to link with each other, the second (below) shows the equivalent for Rodri and Ilkay Gundogan.
Outmatched in attack by admittedly superior ball-players, the United midfield also failed to disrupt City’s flowing attack. McTominay and Fred completed only one of eight attempted tackles in the first half, while according to Opta data, City completed 753 passes at Old Trafford. This is more than any team has completed against Manchester United in the Premier League since 2003-04, when Opta records began.
Lack of intensity
Worryingly for Solskjaer, this was not just a case of being outplayed, but also being outrun and outfought.
Passes per defensive action (PPDA) measures how many passes a team allows the opposition to make before attempting to win the ball back with a defensive action, such as a tackle, interception, or clearance, within the final 60 per cent of the pitch. A low figure indicates a team that is active in its pressing, a high figure suggests a side more willing to sit back.
United’s tally was 24.2, their second-highest figure of the season, behind the 5-0 defeat to Liverpool, indicating low intensity. That those two figures have occurred in matches against their biggest rivals could be concerning.
City’s PPDA was 11.5, their seventh-lowest number of the season. They barely had to exceed first gear.
However, they did force seven high turnovers in the game. High turnovers measure when a team win the ball back in the final third, and then complete at least two touches.
That figure is the highest of City’s season, indicating the difficulty United had getting out of their own half, and their profligacy in possession close to their own goal.
Solskjaer’s strain
The headline numbers of Saturday’s derby, 2-0, may not be particularly embarrassing. But the underlying numbers certainly are, revealing a game in which United were marginalised, forced back and statuesque.
Since the start of last season, Manchester United have lost more home matches (nine) than Crystal Palace (eight).
Yesterday was the 15th Premier League game that they’ve lost by two or more goals under Solskjaer, compared to five under David Moyes, five under Louis van Gaal, and eight under Jose Mourinho.
If football can be reduced to numbers, then it might not be long before Solskjaer’s number is up.
