Feature

'Statistics can cheat you'

The Big Interview - Arsene Wenger

 This story first appeared in the December 2015 edition of the Arsenal Magazine. 

SUBSCRIBE TO THE ARSENAL MAGAZINE

Hector Bellerin is on the front of December's Arsenal Magazine

Hector Bellerin is on the front of December's Arsenal Magazine

Arsène Wenger has always been at the forefront of technical developments in the game, and one area in which football has made huge strides during his career is performance data.

At the club’s AGM in October he said: “In the last few years we have built a core of people around the team who can help us more. As manager, I get an unbelievable amount of detail and data on every single game on every single day. What was 18 years ago my eye, now I have to select the four or five pieces of information to be efficient.”

So with that in mind, the Official Magazine sat down with him to find out more about how data analysis has changed the manager’s job, to talk about its limitations and to look ahead to a future where numbers rule…

Arsène, you have said before that there’s a lot more data available to managers. What do you mean by this, and what areas of the game have been enhanced?

I just compare now to when I started, when I was on my own with my players. Today I have a team of about 20 people around me who take care of the players but also give me information: statistics, the analysis of the game, the quantification of the players’ work rate and their performances.

Modern managers have so much information available to them. Where the quality comes in is where they can choose three or four strands of information that are vital for them, and the modern manager is a guy who selects what is important and leaves what is less important.

Arsene Wenger

Arsene Wenger

Take what is vital – that’s where your eye comes in. So if the decision making is still a lonely job the manager isn’t a lonely man anymore. The manager’s qualities are tested to manage the support team around them as well as the players.

With all that information around you, how do you decide which is the important information?

In my experience, sometimes numbers highlight what I’ve seen after watching the game again, but sometimes numbers highlight what I have not seen, so I have to be humble enough to see that maybe there’s something there and I have to analyse it deeper.

Then I come to a conclusion – either it’s not important or I missed it and it’s important information.

As a manager who has grown up trusting your instinct and your eye, how difficult is it to accept that data can prove your instinct wrong?

It is your humility that helps you to think that your instinct can go wrong. Usually what you feel from the game has a truth as well and you have to respect that. The raw impression you get from the game has to have a weight, especially when you’ve managed 2,000 games.

But even when you have that experience you have to accept that you can be wrong and that you can have missed something. So you have to weigh the two things.

How can the data question what you have seen?

You have to come to your own conclusion and you have to question your impression as well. So a scout comes to you and says, “Boss there’s a player you should consider buying.” What data and stats do you focus on? Well, first of all I want to see the player – before I buy I want to see!

"Stats can cheat you because for example if a player has scored three goals you could think he’s a world-class player – but perhaps he’s just played against a poor centre back on the day"

Arsene Wenger

Because I know what is required at the top level, and if I like what I see I will get more information about the player. Most of the time it is his attitude – his competitive attitude – and of course his physical attributes. Can he physically compete at the top level? There is a selection at every level, to be a Championship player, to be a Premier League player and to be a top, top, top world-class player.

So when you go and scout the opposition, are you relying on data and stats more than you used to and are you conveying that to the players, or is it important not to bog them down too much with data?

No, we analyse our opponents and we try to make it as simple and efficient as possible. We show the players the opposition’s strong points and where they could be vulnerable as well.

We try to prove that through numbers, because the weight is greater if you can say, “Look, this team has conceded 70 per cent of their goals from the left.” The players will believe you much more. One of your players comes to you and admits they have a bit of an issue and wants to know where they’re going wrong.

How useful are stats to drill down and find out what the problem is?

We help them through video. He has to show you the situation where he doesn’t know what to do. You have to accept that the player has to make a decision on the pitch in a fraction of a second, and you have to accept that the game belongs to the players. You can educate them to make the right decisions but once they are on the pitch they are the ones who have to make the decisions.

Football is complicated because it’s 11 against 11 and you have a billion different situations possible. And each situation is never exactly the same every time so you have to be prepared to work with the analytics and make the right decisions. Stats can be misleading – if you have passing stats and every pass is sideways it doesn’t mean it’s a penetrative pass.

Are there other examples where you think stats can be misleading?

Stats can cheat you because for example if a player has scored three goals you could think he’s a world-class player – but perhaps he’s just played against a poor centre back on the day. So you have always to put that into perspective, analyse the game well and remember that dry numbers don’t fit into context.

But on the other hand if you repeat this kind of score for 100 games the numbers don’t cheat anymore. You cannot play 100 times against bad defenders! But one game isolated can get you to one conclusion through the numbers.

Equally do stats ever surprise you when you see one of your own players…

Most of the time the stats are most surprising are work rate, because you might think a guy who is decisive on one or two occasions has worked very hard but then find out he hasn’t worked as hard as you thought.

Or I can give you concrete examples of the fact that Mesut Ozil’s work rate is very high, but people don’t always think that. Because people can sometimes have a style – fluent, not as aggressive – that can cheat the eye.

Are there other players who are underrated for certain abilities because the stats don’t really show the whole picture?

Yes. Work rate for example is not always the absolute necessity because for a player like Francis Coquelin, who works in a tactical position, what’s important for him is that he arrives at the right moment and intercepts the ball, and then when he intercepts the ball that he is decisive.The decisiveness in some positions is more important than the number of kilometres you cover in a game.

Are there more concrete examples in the current squad?

What is important as well is the quality of the positioning, of the distribution, of the final balls – you can’t count only assists but a player like Santi Cazorla gets you out of tight situations and sometimes gives the ball before the assist, which can be even more difficult than the assist itself.

Arsenal analysis

Arsenal analysis

So when you know the players, you know their qualities, you just try as manager to find the right balance so the whole teams functions well. Players like Ramsey, Cazorla and Coquelin are important to get the ball into situations where you can be dangerous. If we go back a few years we have the example of Alex Hleb, who used to give ‘pre-assists’ – something that’s now recognised by OPTA as the ball before the assist. Alex Hleb had that quality.

He wasn’t too interested in scoring goals but he was capable of getting out of very tight situations under pressure, and if you can do that then it opens up the whole game. If you’re one against two that means somewhere else you outnumber the opposition.

Looking at the stats for Academy players, how significant are they at such an early age?

They are very significant at the age of 17 to 20 because until the age of 16, 17 you know that the players are talented. After that the speed of understanding the game and the speed of relating to other players in your team becomes important, because quality alone will not be enough to make a difference anymore.

Some players suddenly have a problem because maybe they had a technical difference at 15 or 16 but that’s no longer enough – you need to know how to quickly get rid of your opponent and how to get into pockets where you are available. Or players who have made the difference physically at a young age suddenly face men – their physicality is not enough anymore and they have to find other answers as well.

So as a manager you’ve bought into stats, as have other managers, but what about players? When they’re younger and perhaps more impulsive, are they reluctant to look at stats and data?

No, they want to know. They’re curious. We live in a society where information has become very important and people want to be informed – they want to know how well they do and not how badly they do, but sometimes to do well you have to accept that you don’t do as well as you think! Sometimes for the manager the numbers are a good way to convince players, “Look, it’s not only my impression, it is confirmed HERE.”

So it makes you stronger sometimes to convince people who don’t want to see the truth. Overall I must say that at the top level players always have a fair and objective analysis of their own performances. Otherwise you can’t improve.

So, after more than 2,000 games in charge, can you think of any specific games where you’ve used data to get a positive result or make the right decision?

Unfortunately at my age I remember more the bad decisions than the good ones! What you never forget is a big defeat so you always question what you did – but if you win the game you think it was a great decision.

What I remember is every wrong decision because I question myself a lot – but maybe I survive in this job because I do that. Today you always look for more information before you make a decision because the margin for error is smaller, the difference between the teams is smaller, so every single decision can make a difference.

Do you think football lags behind other sports when it comes to data analysis?

No, generally I don’t think so. I think football today is ahead of other sports because we have more money and it’s easier for us to improve the way we analyse things. In some ways maybe we are behind – considering for example the diet, or the chemical analysis of our performance. Maybe in cycling they are a bit ahead on that front, but overall I think football has made a huge improvement in the last five years.

So what’s next for analytics in football? Where’s it going? Consider the 20 years that have gone and then project yourself forward 20 years.

I believe that the power of the pure technical people will diminish.

Why?

Because they will be surrounded by teams who analyse absolutely everything and make decisions for them, so the weight of the eye and the experience of the manager will diminish. So I think over the next 20, 30 years and beyond the power of the pure technical person will sink a little bit and they will be surrounded by more machines who will make decisions for them.

You can foresee an era that will provoke a huge transformation of the way we make sports. And I’m sure that change is always quicker than you think and it has hugely changed between the moment I arrived here and today, but that acceleration will become faster in the next 20 years.

Wenger and stats

Wenger and stats

See Full List

Fixtures & Results

Premier League
Ticket Info