Where do Premier League sides attack from?
In an individual match, it’s easy to see that a side have a bias towards one particular flank. In Monday’s 2-1 win over Newcastle, Arsenal constantly attacked down the right, with Theo Walcott staying wide on that flank, and Alex-Oxlade Chamberlain moving inside from the left.
That was surprising when you looked at the opposition, because Newcastle were likely to be more secure down their left (where Jonas Gutierrez protects his full-back well) than down the right, where the ill-disciplined Hatem Ben Arfa was reluctantly fielded.
But the bias towards the right, while exaggerated in that game, is actually typical of Arsenal this season, and in the last couple of years. This season, they play 37% down the right, 30% down the left, and 33% through the middle of the pitch. Both their goals on Monday came after right-wing crosses, and Alan Pardew – a coach who likes statistics and analysing the opposition before matches – would have been aware of that threat. Stopping the danger, of course, is a different matter.
With that in mind, a look at the route of attack for each Premier League team over the course of the season, with stats from WhoScored, is interesting. If balance is more important than symmetry, sides have plenty of freedom to attack down one flank much more than the other, and there are plenty of variations between teams.
Here’s a graph showing the percentage of attacking play down the left, through the middle and down the right for each Premier League club:
Clearly, the graph is difficult to read with so many lines close to each other, and the obvious conclusion is that a lot of sides attack in the same way – when they get the ball in the centre, they spread the play to the flanks and drive forward from there.
However, there are some obvious exceptions to this pattern, and some sides stand out. Here are six sides whose approach is of interest, highlighted in different colours:
Orange: Wolves, who attack more down the left than any team.
Grey: Stoke, who attack down the right more than any team.
Black: Fulham, the only side to attack more down the middle than down either flank.
Blue: Everton, who attack down the centre the least of any team.
Pink: Arsenal, rising steadily to the right
Green: Blackburn, rising steadily to the left
To show these six sides more clearly, here they are without the other 14 sides, with the red dotted line denoting the average of all twenty sides:
All of these examples have their reasons. Wolves play through left-winger Matt Jarvis (despite a disappointing season compared to 2010/11), so they usually play down that flank. They don’t really have a permanent right-winger – no single player has started there more than six times this season.
Stoke’s right-sided is slightly more surprising, because Matthew Etherington has enjoyed a better campaign than Jermaine Pennant, and generally stays wide on the left. In fact, the right of the Stoke side is the one area of the team where places are up for grabs, so it’s surprising so much of their play is focused down that flank.
Fulham’s central tendencies have an obvious cause – the wide players generally come inside into the centre of the pitch, because few are natural wingers. Brian Ruiz, Clint Dempsey and Moussa Dembele are obvious examples, and even Damien Duff has been fielded more frequently on the right than the left, despite being left-footed. Fulham’s style in this respect hasn’t changed too much from their days under Roy Hodgson in 2009/10, despite changing coach twice since then.
Everton’s reluctance to play down the centre owes much to playing so many passes down the left. Despite being a left-back, Leighton Baines is one of their more creative, technically-gifted players, and playing through him is a key part of their approach. The return of Steven Pienaar also maximises this effect.
Arsenal’s right-sided bias was mentioned earlier, though it would be interesting to see how this has changed based upon which full-backs they’ve had available – their play down the flanks suffered significantly when they had to play centre-backs out of position there.
Blackburn’s left-sided bias can be attributed to Junior Hoilett, their best attacking performer this season. Like Wolves, Blackburn don’t have a permanent right-sided player, with no-one starting more than eight games this season on the flank.
Finally, it’s worth considering whether league position is related to attacking approach. This graph shows the top seven sides in green, the bottom five in red, and the eight other sides in orange (using the clear splits in the league table to form the three groupings):
It’s notable that the sides in green are all broadly in the centre of the spectrum.
In fact, the three sides closest to the average (35%-29%-36%) are Chelsea, Manchester United and Manchester City. It’s dangerous to form too many conclusions based upon a small sample size, but this might indicate that the best sides offer a threat from all three angles, and are less predictable than the likes of Wolves and Stoke.
Note: A couple of teams total 99 or 101% because of rounding!
Interesting stuff… a thought: perhaps Stoke attack down the right more as the defending teams understand their threat down the left and as such leave more space on the right (and as such Stoke end up attacking on the right as players have more space that side).
Focusing on Arsenal, it’s funny because when you look back at 2004-06, their attacks were so heavily focused on the left (Cole/Pires/Henry), with Ljungberg coming in from the right. Would also be interesting to see how they conceded their goals in that era.
Nice article. It would be interesting to see how this breaks down if you sorted teams via payroll… you’d probably see something similar to sorting sides based on league position. My guess is that you see the top teams offering a more broad-based attacking strategy because they can afford to have talented players a multiple positions. The poorer teams have to focus where they allocate their resources and then the strategy follows from that.
well this are relative values, not absolute. So if you attack 5 times over right, left and trough the middle each, you are not more dangerous through the middle than a team that attacks 8 times through the middle but 16 times over each side.
It’s quite normal that the numbers about attacks through the middle are not that high than over the flanks, because it’s a normal tactic to protect the centre of the pitch and the area in front of the back four. But I think it’s right that a good team should be able to attack over both sides and through the middle.
Would be interesting to see how counter attacking and possession based styles differ in this stats.
That is a very interesting way to look at teams. It will be so interesting to further compare between leagues. For example, PL is considered to traditionally base a lot of the attacking play on crosses and wing play, and in Italy, as you regularly point out, the approach is much narrower. It will be nice if you have data on the top 5 teams from Spain, England, Italy, Germany and maybe France and to show similar plots…
You asked?
http://t.co/zE0cPlCY
that’s great
First I thought: that’s strange, Premier League and Seria A differ from the other Leagues on the same way. But the Seria A stat is no surprise, with the lack of wingers (even if the centre is usually packed with defenders). Premier League was the bigger surprise, but that may result from the fact that most leagues in Europe try to protect the back four with two holding midfielders (like 4-2-3-1) and in England more teams play 4-4-2 with more box-to-box midfielders and an open midfield. But that’s only an assumption.
Great job, wengerball!
On the subejct, it’s to expect those kind of numbers since the majority of EPL teams are based around the 4-4-2.
European teams are much more varied formation wise, but 4-3-3 and 4-2-3-1 are the predominant ones, and obviously they favour wing play.
Wow, nice!
PL being the second narrowest from the top 5 leagues is very surprising to me!
Love the second vis, https://twitter.com/#!/wengerball/status/180313159779889152/photo/1 ! Are Barcelona attacking the least from the left and being the most unbalanced? If so it only confirms ZM’s piece on the importance of Alves.
I think this is interesting for managers during pre game analysis. Which is the “strong” flank of the opponent and how will that change with different line ups. On the other side the manager has to know the strong flank of his team, not only in the attck but also during defending. Which flank have opponents been more active or more successful. With that knowledge a manager should be able to adjust his team to the strenghs and weaknesses of the opponent (the specifics of the next game).
Next step would be (imho) to take a look at when teams play the ball out wide or switch the flanks (passing stats, typical build up), which players are involved, where and how they receive the ball.
this is what Bielsa does all the time, watching the entire 10 last matches of the team they have to play against…(thats one of the reasons some pleople call him “el loco”)
All managers should do something like that. If they don’t watch the games themselves, they pay some video analyst to do it.
I understand why United doesn’t reinforce much in the center, basically there is not so much attacking trough center from opposition, full back and wide players defensive ability is more matter I guess.
So, to sum up: top teams vary their attacking play, and don’t focus too much on one particular area. However, it should be noted that this may not be significant, considering the fact that Manchester United, Manchester City, and Chelsea all have the resources to be able to vary their play more, and not have one focal point within their team.
Could Fulham’s tendency to attack down the centre have anything to do with their pitch size? I could be way off here but it always appears to be one of the smallest in the league.
No, it’s just the way we’ve been playing this year. Jol tried to get us into a 4-3-3/4-2-3-1 that saw an awful lot of play funneling straight into the heart of the defence.
In previous years we’ve had more progressive full-back play but until relatively recently Riise hasn’t been attacking as regularly and Steve Kelly on the other side is an occasional contributor rather than a regular joiner-inner.
Two things have changed that make me think this might start to change:
Dembele’s been moved back to a deeper role where he finds more space. Typically he was trying to do too much in very congested areas and it was causing our attacks to slow down. Now there’s less pressure on him and he’s spreading the play more, taking defenders out with his amazing dribbling and attracting 2/3 people at once. This makes space for others.
Perhaps more importantly, the sale of Bobby Zamora. He was a very good focal point for attacks but we almost always went to him with a range of passes, the idea being that he’d tee up the midfielders joining him with late runs. Duff’s cutting inside was deadly for a while but teams know about it now and stop it. Dempsey has more variety and seems to be making more runs towards the six yard box rather than Zamora/the D.
We were really, really frustrating to watch as the team transitioned to Jol’s new style of play, with attacks petering out in the middle of the pitch because it was so crowded and so few options. Now we’re finding space and making better use of it, with Murphy making a huge contribution and Ruiz demonstrating some wonderful vision (watch out for him next season).
We do badly need some width and pace but it’s not clear where it’ll come from. Upgrades at full-back and one scorching winger would make a huge difference here, but in the meantime a narrow attack with Ruiz, Dembele, Dempsey, Johnson and Pogrebnyak will probably be fine, if still attacking defences’ more crowded areas. Ho hum.
Fascinating article and something i was caught up in last week analysing why Sergio Aguero has scored 75% of his 16 Premier League goals at home. One of the main factors was that Man City favour attacking down the left flank at home, whilst the right flank when on the road. You can read the full analysis here http://footballfollower.com/football-blog/sergio-aguero-plays-home-man-city/
This is very interesting stuff. More number-driven analysis please!
No, more conceptual driven analysis please! Then use the numbers to prove the theory.
http://scienceblogs.com/evolgen/2008/07/the_end_of_theory.php
What do you want to say with this link? Are you backing Chris Anderson’s article or the opposition?
It think the fact that today so many stats and numbers are available, scientists could use this massiva amount of data to do science. Of course they can, but that is not a prove in an scientific way it’s only statistics.
Like: Arne Friedrich never scored a goal for Germany, is no prove that he never will score. But of course the opponent knows that this defender is not a big threat. (In fact Friedrich scored his only goal at the World Cup.)
In theory, praxis and theory are the same. In praxis they are not.
Thats why ZM’s analysis is better than say Soccer Analyics, where they just number crunch without any conceptual framework which leads them to ludicrous conclusions, sometimes I fear some of the moneyball influenced writing is Charles Reep all over again.
I appreciate you for all your work you do. I remember you did a piece on why EPL matches between the top teams have been so open. I was wondering if you could do a similar piece on why Italian teams have not been able to preserve their lead in the CL. I mean they are the best at parking the bus but for some reason they have become the worst at it.
It would be interesting to contrast the premier league, where crosses are king…
http://arsenalcolumn.co.uk/2011/05/18/crossing-is-footballs-greatest-divide/
…with other leagues like perhaps Serie A where the preponderance of the 4-3-1-2 channels a lot of attacks through the middle.
Also, it would be illuminating to contrast different time periods, since as Jonathan Wilson pointed out, in the modern era full-backs are the only players with much space to play with, then it follows that we would see an overall increase in attacks down the wings.
I took a quick glance at the stats of TOP3 in the BPL and La Liga to make a comparison, and though the differences are marginal at times, there’s something interesting there while looking at the fullbacks on preferred attacking flanks.
La Liga TOP3:
Real Madrid use the left flank the most. Marcelo starting usually on the left, arguably the best FB in the side going forward.
Barcelona use the right flank the most. Dani Alves starts usually on the right, same thing.
Valencia use the left-side the most. Alba / Mathieu tandem being preferred there, no difference to Madrid or Barcelona in terms of FB’s ability to get forward / preferred attacking side.
BPL TOP3:
Manchester United use the right flank the most. I think it’s fair to say that Evra offers more going forward than Smalling or Jones.
Manchester City use the left flank the most, though edging it just by a single percentage point above the right flank. Richards, perhaps the most effective FB in the league going forward this season, plays on the opposite side.
Tottenham use the left flank the most. Though I appreciate Assou-Ekotto’s passing ability, Walker on the right side is much more in the mold of their La Liga compatriots. Fast, decent dribbler, good stamina etc.
While the top3 La Liga teams have very distinct preference to attack down the side where their more attacking fullbacks play, the phenomena is almost the opposite among the top3 teams in the BPL. I really don’t know what to make of this, but I think it’s still interesting.
May be they try to play at one flank so that space open up on other flank where there’s attacking player is playing but it limits the chance of playing ur best FB attacker to certain period only when space opens up ..
Maybe, with Man City, Clichy is more involved in the passing game, while Richards conserves himself only to maraud forward on occasion.
Quote
“It’s dangerous to form too many conclusions based upon a small sample size, but this might indicate that the best sides offer a threat from all three angles, and are less predictable Than the likes of Wolves and Stoke.”
Don’t see any danger here. The more routes you have to goal, the more goals you score and the more games you win!
I am also reminded of Herbert Chapman who never trusted the cross as a means of creating a scoring opportunity. I think he called it a “mindless policy” becasue the odds were 9 to 10 in favour of the defenders. Instead of focussing on goals scored, it might be better to campare the above with a plot of “balls delivered into the box” to see if crossing really is all its cracked up to be in the creating chances stakes. I suspect this would show that an awful lot of crosses actually go to waste.
“..becasue the odds were 9 to 10 in favour of the defenders.”
I think that’s right (depending on the quality of the crosses and the ability of the strikers), but playing wide does not always mean plaing crosses.
“playing wide does not always mean playing crosses”
Agreed, but in the premier league all too often it does. Escpecially when you have Big centreforwards to aim for.
An absence of crosses does not mean a team is not playing Route 1 football though.
The punt down the middle from a goal kick to the centreforward is classic Route 1, but its still a pass through the middle.
How are chances normally adjudged. I’ve seen some excellent crosses in the box that would have been easy chances if a forward had made a different run (or a run at all in Carroll’s case!). Likewise things called chances that are nothing more than crosses hitting a forward on the head with little chance of really scoring.
Who is the team that has almost identical numbers to Stoke? Looks like same left and middle percentages, and right is one point lower.
Route 1 football
Is there any way to do the same kind of comparison but for all the top teams in Europe, and not only EPL teams???
Barcelona ===>Middle usually through Messi/Xavi/Iniesta combo
ZM, now that’s good stats. I criticized your previous post regarding EPL goals versus European progress…but this one you actually used the data you have and explained the observation well. You are not making a theory or anything like that, but it was some good observation, and you made it better by using the right (and unbiased) graphical representation. It was a really good read. Thanks.
How about putting Wimbledon from the 80’s early 90’s in that graph
Barcelona play this -> Messi, Xavi, Iniesta, Xavi, Messi
Judging from the Manchester United – Athletic game, it appears that Manchester United prefer to attack from no sides.
Interesting article. As someone who has often been ’stuck out’ on the right flank, it might be worth adding that right footed centre midfielders simply have a natural tendency to pass the ball to the left flank. In Arsenal’s case, for example, since lots of build up play goes through RVP, it is possibly instinctively easier for such a leftie to move the play towards the right. Obviously an incredibly unsophisticated supplementary point, but if you look at Spurs’ play, Modric’s passes always seem to go left, yes because of Bale and BAE, but also because righties find passing the ball to the left easier and faster, since they don’t have to open their body up
Am I the only one who thought the “orange” lines were yellow?
They were “Wolves gold”!