A few years ago I wrote a guest blog at OptaPro proposing the Premier League’s stars of the future, based on a simple check of some old models. And guess what – after a year or two, it seemed like I hadn’t done very well! So I decided to figure out a better way of spotting young prospects. Here is the gist.
Young players can’t be expected to perform on the pitch in the same way as established regulars. They’re still learning the game, they may be playing an unfamiliar position, and the psychological pressures of senior football can be intense. Every season, we see huge contrasts in how debutantes perform. Some play too safe on the ball, while others can’t seem to hold onto it. Some want to dribble all the time, while others barely ever get forward. As a result of these extremes, young players’ decisions might not be as telling in the crucial situations of a match.
What’s more important, I’ve found, is to see how much they can get involved. Part of this comes from the player’s own initiative: how good is his positioning, can he make himself available to receive the ball, is he willing to challenge an opponent, etc. That’s not the only factor, though. During every minute of play, there’s also a jury of ten other players giving constant feedback on a young player’s performance – feedback that comes in the form of bringing the player into the game: passing him the ball, putting him in situations where he has to defend, giving him a chance to take set pieces, etc.
In addition, it’s critical to realize that young players at the same position may develop different profiles. A centerback could be a deep playmaker or an aerial specialist who prefers to clear the danger. A defensive midfielder could be a passing metronome or a storming destroyer. A winger could be an expert crosser or a take-on specialist. When we assess potential, we have to allow for these different forms of success.
With all of this in mind, I put aside my mathematical models of the overall game to focus on the individual efforts of young players. In the end, I came up with a completely different approach that used benchmarks and Boolean analysis to identify top prospects. I selected the benchmarks through a statistical process, and it was remarkable to me how consistent they were: you could use roughly the same benchmarks to evaluate players as young as 16, even in academy and reserve teams.
Of course, the benchmarks were slightly different for players expected to succeed in the top European leagues versus the second or third tier. And some players were versatile enough to combine several profiles. Niklas Süle was an outstanding example, having been flagged in all of my centerback profiles starting with his data from Hoffenheim in 2013-14, when he was just 18.
Recently I looked back at my data to see which players had been flagged in the Championship in 2017-18. Here is the complete list of outfield players, with a few popping up at multiple positions. We can produce similar lists for any season in dozens of leagues around the world.
RCB: Fikayo Tomori
LB: Joshua Earl, Josh Clarke, Ola Aina, Matt Targett, Rúben Vinagre, Antonee Robinson, Lloyd Kelly
RB: Ola Aina, Wes Harding, Josh Clarke
DM: Joe Williams
LW/LM: Ollie Watkins, Bersant Celina, Josh Murphy, Diogo Jota, Barrie McKay, Oliver McBurnie, Ryan Sessegnon, Harry Wilson, Ryan Hedges
AM: James Maddison, Josh Harrop
RW/RM: Josh Brownhill, Grant Ward, Mamadou Thiam, Ben Brereton, Barrie McKay, Martin Samuelsen, Jon Toral
SS/CF: Diogo Jota, Ashley Fletcher, Keinan Davis, Jack Grealish, David Brooks, Ollie Watkins