Scouting by time machine - France
My belief is that data scouting works. You can narrow a massive field to a very small field of potential players by identifying players with the attributes you want. You can identify these attributes by looking at how players perform these tasks which are reflected in their performance statistics.
But how many statistics do you need to look at to see if a player is actually any good?
Well, I have made an exciting discovery.
I reckon one statistic alone has been enough to identify pretty much the cream of Europe's defensive talent.
As regular readers will know I think Ligue 2 is the best value league around, for many reasons, so we will base this experiment on that.
What I want to do is use the data from previous years to look at young players who showed up well in a category and follow their career from that point.
I'm choosing the Wyscout category of "progressive passes" which I think means accurate long passes and normally shows up good passers from deep positions.
The PL top 5 this year are:
Looks fairly accurate to me.
If I restrict it (as I will be doing for France) to players who are 22 or younger:
I don't think anybody would look at that list and think it was ridiculous. All these are young players, getting regular minutes who can pass the ball well from deep positions on the pitch.
The great thing now is that data has been around for a few years. My Wyscout data goes back to 2015/16 so what would happen if I looked at the good passing players from Ligue 2 in that season who were then <22. Would they have all disappeared into obscurity?
Nope.
Lenglet of Barcelona, Gbamin and Niakhate of Mainz, Ait-Bennasser of Monaco. Cyprien of Nice. The most obscure is Cisse of Olympiakos but he is extremely highly rated and likely to move to a big 5 league club this summer. If we take (along with a pinch of salt) the value off Wyscout (adding in 10m fo Cisse) then that is around £124m of players picked up by a single "meaningless stat"
OK, you say, what if that was a one-off? What happened in the next season?
Top of the list is a player I'm constantly promoting Jean-Kevin Duverne AVAILABLE FOR FREE THIS SUMMER, he had a bad injury but is now back and looking good. Niakhate and Cisse are both in the list, showing the repeatability of the data. Then we have lots of new arrivals.
Ferland Mendy, of Lyon, heavily linked to Real Madrid. Briancon of Nimes who has anchored the midfield of the 7th placed team in Ligue 1. Florian Miguel too, who is now also at Nimes.
Junior Sambia defensive midfielder for 5th placed Montpellier. Fulgini of Angers now a French U21 international with 30 Ligue 1 games this season. Dylan Bronn? Player of the season for Gent and linked with Lyon.
Every single player over two seasons of data has gone on to success at a much higher level.
So it works for defenders? What about attackers. Let's go even lower and choose an even more basic statistic of "goals and assists".
The French 3rd division in 2015/16 sorted by goals and assists of players 23 and under:
Well, hello.
JP Mateta, who has just scored 14 goals in the Bundesliga in his first season.
Nicholas Pepe, 22 goals with Lille and wanted by every major club.
Wissa who is now with Lorient near the top of Ligue 2.
Bouanga who is now at Nimes with 10 G+A this season in Ligue 1
Kamara of Fulham
Sada Thioub also of Nimes.
Two things stand out.
1. The sheer depth of talent in France
2. That Nimes and Mainz are two clubs who have gone heavily into this market and are reaping the benefits finishing much higher than clubs with much higher budgets.
Smart clubs should be working with smart scouting companies to find a way to find the best value players around.
But how many statistics do you need to look at to see if a player is actually any good?
Well, I have made an exciting discovery.
I reckon one statistic alone has been enough to identify pretty much the cream of Europe's defensive talent.
As regular readers will know I think Ligue 2 is the best value league around, for many reasons, so we will base this experiment on that.
What I want to do is use the data from previous years to look at young players who showed up well in a category and follow their career from that point.
I'm choosing the Wyscout category of "progressive passes" which I think means accurate long passes and normally shows up good passers from deep positions.
The PL top 5 this year are:
Looks fairly accurate to me.
If I restrict it (as I will be doing for France) to players who are 22 or younger:
I don't think anybody would look at that list and think it was ridiculous. All these are young players, getting regular minutes who can pass the ball well from deep positions on the pitch.
The great thing now is that data has been around for a few years. My Wyscout data goes back to 2015/16 so what would happen if I looked at the good passing players from Ligue 2 in that season who were then <22. Would they have all disappeared into obscurity?
Nope.
Lenglet of Barcelona, Gbamin and Niakhate of Mainz, Ait-Bennasser of Monaco. Cyprien of Nice. The most obscure is Cisse of Olympiakos but he is extremely highly rated and likely to move to a big 5 league club this summer. If we take (along with a pinch of salt) the value off Wyscout (adding in 10m fo Cisse) then that is around £124m of players picked up by a single "meaningless stat"
OK, you say, what if that was a one-off? What happened in the next season?
Top of the list is a player I'm constantly promoting Jean-Kevin Duverne AVAILABLE FOR FREE THIS SUMMER, he had a bad injury but is now back and looking good. Niakhate and Cisse are both in the list, showing the repeatability of the data. Then we have lots of new arrivals.
Ferland Mendy, of Lyon, heavily linked to Real Madrid. Briancon of Nimes who has anchored the midfield of the 7th placed team in Ligue 1. Florian Miguel too, who is now also at Nimes.
Junior Sambia defensive midfielder for 5th placed Montpellier. Fulgini of Angers now a French U21 international with 30 Ligue 1 games this season. Dylan Bronn? Player of the season for Gent and linked with Lyon.
Every single player over two seasons of data has gone on to success at a much higher level.
So it works for defenders? What about attackers. Let's go even lower and choose an even more basic statistic of "goals and assists".
The French 3rd division in 2015/16 sorted by goals and assists of players 23 and under:
Well, hello.
JP Mateta, who has just scored 14 goals in the Bundesliga in his first season.
Nicholas Pepe, 22 goals with Lille and wanted by every major club.
Wissa who is now with Lorient near the top of Ligue 2.
Bouanga who is now at Nimes with 10 G+A this season in Ligue 1
Kamara of Fulham
Sada Thioub also of Nimes.
Two things stand out.
1. The sheer depth of talent in France
2. That Nimes and Mainz are two clubs who have gone heavily into this market and are reaping the benefits finishing much higher than clubs with much higher budgets.
Smart clubs should be working with smart scouting companies to find a way to find the best value players around.
Comments
Post a Comment