| 15°C Dublin

'It's pretty incredible' - Arteta highlights statistics to explain Arsenal’s bad start


Hammer blow: Chelsea striker Timo Werner collides with West Ham goalkeeper Lukasz Fabianski. Photo: Reuters

Hammer blow: Chelsea striker Timo Werner collides with West Ham goalkeeper Lukasz Fabianski. Photo: Reuters

Hammer blow: Chelsea striker Timo Werner collides with West Ham goalkeeper Lukasz Fabianski. Photo: Reuters

Most observers would surely agree with Mikel Arteta when he describes Arsenal's disastrous start to this season as "pretty incredible". It is certainly incredible that a club of their stature can fall so far down the table, and that all of the positivity from the summer can be lost so quickly in the winter.

Arteta was not referring to the club's alarming slide, though. The Arsenal manager was instead discussing the underlying numbers in his team's performances, the metrics that he believes show their recent results are not reflective of the reality of those matches.

He and the club will never divulge the inner workings of their analysts, who operate in an area that has become a battleground for top sides (Liverpool, for example, employ a team of elite physicists in their data unit). As a field of study it is competitive and secretive, although Arteta did shine some light on the conclusions they have reached when studying recent defeats.

According to the club's own metrics, Arsenal's performance against Everton on Saturday, when they lost 2-1, meant they had a 67pc chance of winning. In their recent defeat against Tottenham Hotspur, their figures showed just a 7pc chance of losing. Against Burnley, when they lost 1-0, it was a 3pc chance of losing.


"When you look at the perspective about how we are losing football matches and how we are where we are, it is pretty incredible," said Arteta, who faces his former side Manchester City in the League Cup quarter-final tonight. His overarching message was that the underlying numbers from their data team show that Arsenal have been unlucky to lose these games.

Unfortunately for Arteta, it is a reality of football culture in the UK that any reference to underlying statistics is likely to be met with scepticism. And Arteta is not helped by the fact that his internal data is built on models that are not accessible to the public.

He says the club's figures suggest they were unlucky to lose, but teams will never reveal how these models work. So Arteta is asking fans and observers to trust in his metrics rather than more commonly used statistics, or indeed their own eyes.

The claim about the north London derby was particularly hard to comprehend.

Yes, Arsenal were dominant for much of the match, but Tottenham seemed to allow them to control the ball, knowing they were unlikely to be broken down and that they could then strike on the counter. All this is dangerous territory, for Arteta and for those of us on the outside.

The Halfway Line Newsletter

Get the lowdown on the Irish football scene with our soccer correspondent Daniel McDonnell and expert team of writers with our free weekly newsletter.

This field is required

We cannot say the figures are definitely wrong, because we have no access to them. But we cannot say Arteta is definitely right, because we do not know the finer details of how Arsenal analyse their matches.

What is certainly true is that the numbers are far more complicated than basic measurements such as possession - "I am zero interested in possession," Arteta said - and shots on target. Top teams have all sorts of different ways of measuring key metrics like chance creation and defensive stability, going far beyond figures such as "expected goals".

As an example, earlier this year Liverpool data scientist William Spearman (who has a PhD in particle physics from Harvard) delivered a presentation on the club's complex "pitch control" model. Football is not rocket science, but it is getting there. And rocket scientists are driving it in that direction.


So, if Arsenal's internal analysis shows they are performing better than the results suggest, why are those results not improving? In Arteta's eyes, it largely comes down to efficiency in front of goal.

Earlier this year, the strikers were burying a high proportion of their chances. Now, they are not.

"Our finishing quality is what is letting us down at the moment because the rest is what it should be to win many football matches," said Arteta. "It worked there before and we were winning because our finishing quality was through the roof."

It is interesting to note that Arteta's referencing of these statistics was unprompted. He was answering a question about the atmosphere in the dressing-room when he brought up his percentages.

Asked what the figures are based on, he said: "Every possible area that can happen in a match is considered. And it works. But then football is not like basketball where you shoot 50 times and the opponent shoots once so you win every time.

"It is pretty tricky but I am telling you, last year we were winning games with a 25 per cent chance of winning, but we were so efficient when we got into the (key) positions. But if you are constantly in the 60s and 70s (percentages) long term, you are going to win many more matches."

His conclusion is that the underlying numbers show that the short-term pain will eventually give way to success. Whether the supporters agree, whether they believe in his metrics, is up to them.

© Daily Telegraph, London

  • Arsenal v Manchester City, Live, Sky Sports, 8.0

Telegraph Media Group Limited [2022]