Too much possession? How Wembley changed Arsenal’s season
Arsenal enjoyed 74% possession at West Brom, but still couldn't win
A pre-season draw to concede the Emirates Cup to the New York Red Bulls suggested that although it’s a new season for Arsenal, the same old problems remain.
In an attempt to get to the statistical core of those problems, I stumbled across a slightly bizarre relationship: last season the Gunners enjoyed more success when they actually saw less of the ball.
Accounting for differences in opposition strength and match location (home or away), a fall in the amount of possession increased Arsenal’s chances of winning any given match.
For a team that places high value on ball retention, this was incredibly counter-intuitive.
However, a little further digging revealed the true cause of the relationship to be the League Cup final, as demonstrated in these tables:
Not only was Arsenal’s win percentage significantly affected by their Wembley trauma, but their possession was too, increasing by a remarkable 6.7% on average in the league after that fateful Koscielny error.
The Gunners’ opponents, whether by accident or design, increasingly chose to sit back and surrender possession in those final eleven games of the season. If they could score a goal on the break – and they often did – then all the better.
It’s no coincidence that the three games after the League Cup final in which Arsenal enjoyed less than 60% of the ball – their average for the season – they collected seven points from a possible nine; against Spurs, Man United and Blackpool. It appears all three of those sides would have been better off allowing Arsenal to dictate the play; but then that’s not their styles.
The Gunners did not win any of the other eight post-Wembley matches.
The idea that teams chose to allow Arsenal more time on the ball is supported by the fact that total possession duels in Arsenal matches fell significantly in those final eleven games; teams were less keen to contest 50/50 balls during this period.
This begins to go against the notion that getting ‘stuck in’ is the best way to trouble Wenger’s side. Whilst the statistic doesn’t perfectly translate to hassling and harrying, five of Arsenal’s bottom eight-lowest duel frequency matches came after the League Cup final, whilst the top eight all occurred before.
This supports the suggestion that teams caught onto the best way of frustrating Arsenal; not by getting in their faces but by being well-organised defensively and not worrying about possession.
This has implications going into the new season. Will teams continue to take this seemingly defensive approach, or choose to take better control of matches against Arsenal? Continued success of the former would suggest that the League Cup hangover was not entirely mental, an arguably overrated phenomenon anyway.
It also leaves Arsene Wenger with a difficult dilemma. No team wants to concede possession, but would Arsenal be better off trying to invite teams onto them? One for the tacticians, I suspect, but the post-Wembley numbers don’t make great reading for Arsenal.
How should Arsenal approach the new season? Do they place too much value on possession? Or is higher possession indicative of the Gunners keener to have the ball when behind? Please give your thoughts below!
Statistics courtesy of @EPLIndex.
5 Added Minutes 
Interesting stats and thoughts but analysis can be extended. Right now I can’t see any basis for forming conclusions.
It could easily be the case that Arsenal were a lot more cautious – a mental issue – after the League Cup mistake and the subsequent exit from other cups. That would have led to a higher number of safe passing choices leading to more possession but less penetration. This would in turn give the opposition more confidence because they can sit back and defend with relative ease all the while looking for an opportunity to break. I think the number of duels would be related to the intensity of the game which would be lower if the team dominating the ball wasn’t doing enough with it.
Another issue could be the timing of the first goal. If the opponents scored one they would be even more inclined to sit back and invite pressure.
There are other angles that need to be seen but that would make this comment too long. Appreciate your effort but not sure I agree with the direction of your conclusion/explanation.
I think you’re possibly right. Quite often Arsenal took the lead in games in which they had ‘lower’ possession, and vice versa for higher possession. I can see where you’re coming from with duels but I’m not so sure, Arsenal still had plenty of high intensity games March onwards but this didn’t necessarily lend itself to more duels.
Think it’s kind of interesting if teams did decide to sit back against Arsenal post LC Final, because that isn’t, interestingly, how Birmingham won. Birmingham pressed Arsenal for the majority of the game (last 15 minutes excepted.) Defensive play, in some ways, yes, but not defend deep style.
Not surprised to see that Arsenal have high possession against Stoke. I think that this is why Stoke have such a great record against Arsenal by the way (3 out of last 4 won at Brittania and the one they lost was when Shawcross saw red). Because Tony Pulis knows that if you let them have the ball in the middle of the pitch they aren’t going anywhere, and its tough to break that sort of side down. Then of course they can counter attack, using the likes of Etherington on the wing.
Stoke rarely achieve more than 50% possession in any game, even at home. If they did play Barcelona, I can only imagine how little time they’d have on the ball!
Good article, but I don’t think that the stats really indicate that Arsenal’s possession based approach is wrong – just that teams have found a better way of neutralizing it. There’s no way to get round that really because you can’t dictate your opponent’s approach. Arsenal just have to get better in the final third.
I totally agree actually, mainly it just backs up the idea that Arsenal are great to watch but often don’t do enough with the ball. Helps identify one of the post-Wembley trends though, as opposed to pointing out issues to do with ‘bottle’.
Is there any stat/analysis to measure the value of ‘bottle’ or ‘mental strength’ would be interested to see it quantified…..
Ha likewise. It’s easy to find stats on points dropped from winning positions and the like, but most of the time it’s impossible to differentiate between the mental element and general team competency. There’s a lot of research on football psychology though, wonder if they’ll find something.