Using analytics to match team style to recruitment

England vs Mexico 2010 World Cup warm up.

Leighton Baines lines up at left back with a place in the world cup looking certain after a brilliant season as an attacking left back for Everton. His link up play with Pienaar has seen him create a huge volume of open play chances with 8 assists.

In the game he lines up with Gerrard nominally covering the left side of midfield.

Very soon it becomes obvious something is wrong, he is making all his normal runs, getting into good attacking positions but the ball isn't coming back to him. When he goes beyond his midfielder he isn't covered. Several times he is caught upfield, several times he faces both the Mexican winger and right back without any defensive cover.

England win but Baines is dropped from the World Cup squad with rumours of homesickness given as the excuse. Denied by Baines.

Baines was the same player he was at Everton, but the lack of a covering midfielder, passes that didn't come, and free kicks he was down the pecking order for, nullified his game.

In the following season he directly assisted 11 goals, from full back. The season after that he lead the whole of Europe for chances created in the first half of the season, and not just because of his set pieces, his open play saw him 6th in the league for passes in the final third. All whilst playing in a tight defence.

At Everton Baines played in a side that was, until the emergence of Seamus Coleman, very biased to attacking down the left hand side. With Hibbert or Neville at right back Baines was encouraged to venture forwards as often as possible knowing there would be sufficient numbers back and Pienaar would cover the left hand side.

So does the player make the system or does the system make the player?

How many players are in squads where a change of style have rendered their skillset useless?

We've seen plenty of players struggle in one league and thrive in another. We tend to write them off as not being "up to" the English game, but is it not suited to the game or suited to the style of the particular team they are at?

After all they seem more than capable of beating English club sides when playing against them in European competition. And the likes of De Bruyne who were written off came back within 12 months to be one of the best players. Or Serge Gnarby who couldn't make the matchday squad for West Brom but was a German international playing for Werder Bremen and Bayern Munich within a year.

So is there any way that analytics can help identify the likes of Baines or Gnarby?

The main issue, as always, is if they don't play you can't get numbers on them.

You can look at older data for more experienced players, but you need to consider the impact of inactivity and the age curve.

For youngsters coming through the situation is even more precarious. I sometimes wonder if players turning pro at 17 would be better (as a few Manchester City and Chelsea ones have done recently) shopping around for a club that is committed to a style that suits them and offers a first team pathway, rather than just signing for their parent club.


Think like a business

Clubs invest hugely in their playing staff and academy systems. But I don't really get the impression, outside of a few clubs, that they are thinking of player trading like a business.

Emotion takes over, which is kind of the point of it, nobody celebrates their club being a money making machine with nothing to show for it. There are humans involved too and us numbers people mustn't underestimate that sentiment (that final contract for your club hero) plays a big part in keeping a group happy.

But you wouldn't put your money into a film production company who couldn't decide whether they were going to make a horror, a comedy, or a cartoon and kept hiring actors on long, unbreakable contracts at your expense.

Data is increasingly important in clubs buying decisions. If you don't have the data on a player you have no use for in your style of play, you won't get maximum resale value. Therefore even if you won't directly benefit from his form it is always in the interest of the selling club to ensure peripheral players are playing in a team that suits their style.

Which is why I think the next breakthrough in recruitment is going to be via stylistic matching.

Stylistic matching

Stylistic matching, for buying clubs, means scouting teams that play in a similar style to your desired style.

To find similar styles I'd look at formation, volume of passing, passing length, tendency to cross, involvement of positions in build up play, shot location and frequency, pressing data.

I am aware some analysts have already done work in this area so please flag up yours if you are reading this.

So rather than scouting all players look for players who already play in the system you do (or want to).

This narrows the pool of talent but increases you chance of finding a perfect match.

There is no point in buying the next Baines if you don't make use of his talents. 

And if you do have the next Baines in your squad then at least maximise his value for transfer by carefully considering the style of the club you are loaning him to.