Why doesn't football analytics focus more on how to make good footballers better?
From the Department of Not Wasting Money
This is a day late, though I don’t have a ‘schedule’ necessarily. But ideally I’d like to get these out Wednesdays, so if this negatively impacted your life, even in the slightest, then mea culpa.
Part of the reason is I have a column dropping tomorrow on the situation at Manchester United, in which Ed Woodward has made the counterproductive decision to hand Ole Gunnar Solskjaer a three-year contract before hiring a full-time technical director.
By appointing the manager without this still-fictional TD in place, you’ve essentially missed the entire point of having a technical director at all!
But this is not what I want to write about today.
Rather, in researching the article, I had a very interesting conversation with the founder of a football consultancy, that deals in part with executive placements in club front offices (among other things).
One of the things he touched on was the importance of finding ways to make already good players, even elite players, a little better.
I’ve harped on this for some time. We have every thinkable type of analytics work going on, in public, under the sun at the moment. Analysis of individual players, team analysis, managerial analysis, you name it. Then there are the tactics blogs, though I don’t read/see those as much as I once did. Thankfully, the tactics nerd debates that everyone insisted on having ever since Inverting the Pyramid and Zonal Marking have largely gone silent.
In much of this, though, there is very little discussion of whether it’s possible to use physical and data science to make already good players better.
You would think this would be a ripe topic for analysts wishing to be hired by a team. After all, is it not far better to make an 11-goal-a-season striker into a 14-goal-a-season striker?
Because unless you have a very generous merit pay system in place, it’s a lot more affordable to work on improving your existing players than to have to find a new one on the transfer market.
But the way most analysts talk about player form is that it’s a fait accompli. You get what you get, and what you get will inevitably age, move into decline, or get injured.
Yet I’m not convinced that anyone has really dug into this topic, and I think part of the reason is that we’re still not sure where team effects end and where individual talent begins.
I tried to make this point in a much-maligned Medium article I wrote shortly before I decided to launch this newsletter. In it, I tried (and I think, to be fair, failed) to get across my theory that while tactics matter, good teams at their core are aggregates of players who are very good at a discrete set of finely-tuned motor skills:
Nailing this kind of pass with something approaching regularity is the kind of elite technical skill that makes an okay team a little better. Knowing how to weight passes for maximum accuracy. Better ball control. Taking better and fewer touches. Being able to dribble past players more efficiently. The little things that add up to very, very good things. You add those skills up across eleven players and you have a top six club.
You might quibble strongly with this, but what is a footballer if not someone who is able to do footballing things very, very well and very consistently?
And if we accept this basic premise, why can we not accept that it might be theoretically possible to use data analytics in order to better identify these skills, but also how to improve them?
I realize I am getting dangerously close to analytics science fiction, the kind that I made fun of a few weeks ago when I picked on the Orlando Magic. But the kind of analysis I’m thinking about here doesn’t necessarily require complex tracking data to capture.
We could simply start from a very simple premise, as in, what precisely makes bad players bad (or at least what makes mediocre players mediocre), and what makes elite players good? I assure you, it’s more than ‘scores loads of goals and makes loads of assists’). What is the basic palate of repetitive actions that make up a professional footballer? And why are some players very good at them, and others very poor?
Maybe this work exists! Maybe there is a ton of it and I have managed to overlook it. But if you can find the basic, generalized, positionally agnostic language of football skill, you can find a way to teach some good players to become fluent in it, and perhaps, along the way, make them great.
And if you do that, and can prove it’s possible, then you have the potential to help a club save millions in transfer fees.