So how do you beat Manchester City?

Sam Allardyce revealed that he had a plan for beating Manchester City.

“We all know that Manchester City play a certain way. They come away from home expecting to win and go and attack you. We’ve seen a lot this season about how those spaces they leave can be exploited by other teams.”
Source: telegraph.co.uk

So what did West Ham United do to exploit Manchester City’s weaknesses? Let’s look at a few game images, and what jumps out from the stats.

Their aim was to win the ball 40-60 meters from the goal of the visitors, even though they started pressing the City defenders in possession very early on. The press on the defenders happened one-on-one, Valencia or Downing ran straight to the defender in possession, and stopped a yard from him. This way the defender couldn’t knock the ball past him and dribble out of pressure, he had to pass it. When the ball left his feet, another West Ham player – Sakho, Downing, Amalfitano or Valencia – did the same with the player who was about to receive the ball. The aim of these individual pressures was to rush the defender into playing a subpar ball into midfield.

West Ham United tackles. Their interceptions show even more that they defended in the midfield. However we wouldn't expect a lot of interceptions in the final third unless we are dealing with an extremely aggressive team.

West Ham United tackles. Their interceptions show even more that they defended in the midfield. However we wouldn’t expect a lot of interceptions in the final third unless we are dealing with an extremely aggressive team. Source: FourFourTwo Statszone

When the ball was passed to the midfielders, West Ham simply overloaded them.

West Ham's press.

West Ham’s press when the ball comes to a midfielder.

West Ham aimed for preventing City from playing one-on-one in midfield. It worked tremendously well, Touré had just four dribbes the whole game, but none in the first half. Fernando had only one dribble in ninety minutes.

Prevent one-on ones at all costs. Be narrow, give support. The full-backs follow the wingers loosely. This way they have more time to make a decision in case the opponent decides to overload them.

Prevent one-on ones at all costs. Be narrow, give support. The full-backs follow the wingers loosely. This way they have more time to make a decision in case the opponent decides to overload them. Silva and Navas doesn’t have a realistic chance of getting the ball the moment this picture was taken, otherwise the full-backs would be closer.

Track the overload attempts of the opponent's full-backs, prevent one-on-ones at all costs.

Track the overload attempts of the opponent’s full-backs, prevent one-on-ones at all costs.

WH defend2

Notice how West Ham overload the half of the pitch where the ball is/coming from.

Interestingly Touré didn’t attempt any dribbles when Dzeko was on the pitch, but attempted four after Jovetic came on. The reason behind this lies in the positions the two forwards like to take up.

Dzeko vs. Jovetic

Dzeko's passes. Notice how deep he got into possession, and how little influence he had around the box.

Dzeko‘s passes. Notice how deep he got into possession, and how little influence he had around the box. Source: Squawka

Jovetic's passes. More influence around the left corner of the box, there is more space left for the midfielders like Touré.

Jovetic’s passes. More influence around the left corner of the box, there is more space left for the central midfielders to exploit. Source: Squawka 

Jovetic is positioned diagonally to Touré. This way Touré has space to dribble,  and Jovetic is not marked.

Jovetic is positioned diagonally to Touré. This way Touré has space to dribble, and Jovetic is not marked. Noble is caught between two minds, Jovtic can be found with a one-two (Touré-Agüero-Jovetic triangle).

 West Ham’s directness

I am using PADA to measure how direct West Ham United were in attack. PADA (Passes in Attacking third per Defensive Actions) is the number of passes a team makes in the attacking third divided by the number of defensive actions (interceptions+tackles) they made all over the pitch. The idea is that teams that take a direct route do not keep possession in front of the goal for long. The lower the value, the more direct the team is.

You can get a zero PADA by dribbling towards the goal each time you win the ball, or loosing the ball before you get to the attacking third, but neither of those will win you games. Even a PADA of 1 is extremely direct, this would be the equivalent of booting the ball forward after a tackle, and shooting from wherever you receive it.

West Ham had a PADA through the game of 1.83, so for each ball won they had less than two passes in the final third. This was the first time I used PADA, so I don’t know how this compares to the rest of their games, but it surely seems very direct on an intuitive level.

It is interesting to compare West Ham’s PADA when they were leading compared to when they were drawing. They had a PADA of 1.56 when the game was still goalless, but they became less direct after they took the lead, with a PADA of 1.89. 

I can think of two reasons for these values. One is the statitical fluke, after all the first goal was scored after twenty minutes, not  a big sample, the value could be off. On the other hand maybe the Hammers kept more posession once in the lead in order to run down the clock.

West Ham had a PADA of 2.07 when the game was 1-0, their highest during the game. This suggests that Manchester City didn’t press the ball well enough.

Where did City loose this game? Lack of pressing or lack of midfield steel?

Manchester City are one of the most aggressively pressing teams in the league when they get behind, it was strange to see them allow West Ham such high PADA. After West Ham scored the second and City got one back the PADA of the Hammers dropped to 1.25, but strangely not due to the excellet pressing of their opponent.

The press

The intensity of a team’s pressing can be measured with the PPDA metric. It stands for Passes allowed Per Defensive Actions. In this article I divided the passes in a team’s defensive third with the sum of the opponent’s tackles and interceptions in the same area. The idea is that the less uncontested passes the opponent can make at the back, the more intensive your press is. You can read an excellent piece on this subject on Statsbomb.

It was astonishing to look at the PPDA numbers of Manchester City. Their overall value was extremely high, they only had 2 tackles in the attacking third, which resulted in a PPDA of 17! That is very high for a team that is aiming for the title. The two tackles happened when they were 1-0 down, so during that time they had a very aggressive 8.5 PPDA, but after they conceded the second goal they stopped pressing well.

The steel

When City were 2-0 and 2-1 down they allowed West Ham to make 1.86 passes per minute compared to 2.74 when they were 1-0 up, and 2.29 when it was still goalless. The sudden drop at the end suggests together with the low PADA and the non-existen pressing that West Ham were getting tired, and didn’t open up their shape to keep the ball in the attacking third. The 2.74 passes per minute West Ham had leads me to beleive that Manchester City had problems breaking up the play in midfield. If we compare the stats of three central midfielders in the game, this becomes visible.

Dribbles Successful Tackles Unsuccessful Tackles Interceptions Pass Completion
Alex Song 1 3 6 1 77,00%
Yaya Touré 4 2 2 0 88,00%
Fernando 1 2 2 1 98,00%

It is striking how few balls Fernando and Touré won. The four successful tackles and one interception between the two of them is not enough to dominate a midfield battle. For me it seems like their poor defensive contribution delayed City from getting back into the game earlier. On the other hand they deserve kudos for their distribution, Fernando had 17 successful foward passes out of 18, and Touré’s 52/61 stats look commanding as well.

 

Creating chances with through balls vs. crosses

If you divide the number of crosses West Ham atempted (19) by the number of shots they had (11) you get how many crosses they had to attempt in order to get one shot on goal on average. (1.72) If you do the same with through balls and shots you get 0.63, so basically out of 2 through balls West Ham could get one shot on goal on average.

I didn’t expect West Ham to be so efficient with through balls, so I did a little bit of research. Two of the through balls happened after free-kick combinations in the box, this is something to be expected from a Sam Allerdyce team. Two were played at the end of counter-attacks, which further shows how direct West Ham were. Two happened after Manchester City could not put pressure on one of the midfielders – the old three against two overload that so many teams do against City – so he had all the time to pick out a pass. Of all the through balls two were played behind Clichy, one of them led to the first Hammers goal.

“I think they leave Mangala exposed. He’s only just joined the Premier League. They don’t protect him and he’s only just got here, so the more and more you can get down the side of him, with the space that Gaël Clichy leaves [the better].
Source: telegraph.co.uk

This lack of pressure was due to Fernando and Touré being outnumbered in midfield. This becomes a problem against sides who have quick forwards who can give depth to the team, while the midfielders have enough game intelligence to know when to keep the ball, and when to pick out that slicing pass that cuts the defense open.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s