Football Intelligence – the game’s direction of travel

Some people are turned off by statistics in football. Frequently I get responses when I tweet basic stats saying ‘I use my eyes’ – or words along those lines – disregarding any form of data or numbers. That’s fine – we all appreciate football in different ways.

Well, I unashamedly like data. I don’t ‘do’ stats  particularly well; I don’t have a formal background in statistics and I’ve never been involved in that type of scholarly research, and so I leave it to the experts and mostly keep it simple, focusing on ‘events’ – tackles, passes, assists, shots, goals. But there’s a growing community dedicated to ‘advanced’ statistic and, at major football clubs, teams of statisticians are starting to use ‘football intelligence’ data.

I had been recommended this presentation by various people over the last couple of months, and today I finally got around to watching it. It is Damien Comolli – formerly Sporting Director at Spurs (2005 – 2008) – presenting a Performance.LAB Innovation Seminar about Squad Management.

I don’t want to spoil it, as it’s a must watch. But it’s fairly Spurs-centric, and Nabil Bentaleb is used as an example throughout. In addition, if this doesn’t make you want us to sign Christoph Kramer, nothing will.

At the same event, Paul Power gave a presentation about Game Intelligence. He states towards the beginning of the presentation that they used 260 million points of data to prepare for that presentation alone. That is remarkable in itself. But look at how they’re using it, and it will blow your mind.

It amused me somewhat that Comolli talks about how Luka Modrić – a player who doesn’t stand out when you look at event data alone – helped drive them to work on some of these models. To essentially prove, statistically, that he is the brilliant player that we know he is from simply using our eyes.

Where this will lead next is fascinating. This sort of data is already being used to recruit players, and to analyse existing players’ performance. Indeed at Spurs we have Paul Mitchell, recruited at the back end of 2014 from Southampton as Head of Recruitment and Analysis. He who is famous for his ‘black box’ and use of data. He uses data and visuals together, presumably in ways similar to what is seen in Comolli’s seminar.

The next step will be to drive coaching forward – to apply the numbers to coaching methods and to help players improve based on very specific data. To prepare players and teams for specific opponents by targeting their weaknesses and stopping them from using their strengths.

Of course you can’t just have the data. You need someone to analyse it, you need someone to present it, and you need to make judgements and have conversations – to remember that different coaches want different players to do different things. It’d be no good to give Mauricio Pochettino a player who is brilliant at controlling possession and tempo by taking lots of touches because he actually wants his players to move the ball quickly.

And so conversations need to be had, logic needs to be applied, and also personalities need to be analysed. But it’s fascinating nonetheless.

I expect some form of backlash in terms of how players will adapt to being analysed. For example, players all get their numbers back immediately after a match, and there have been stories about players playing lots of safe, square passes between one another late in games to drive their pass completion up. Advanced analytics and football intelligence will know this, and these sorts of safe passes will not allow the player to stand out – because only passes that matter will add value.

You can, of course, still continue watching football without digging any deeper – the beauty of the game need not change. But rest assured that beneath the surface is a team aiming to ensure that clubs get every inch of value from players in an ever-changing environment.