Newly American Prospect Julian Green Projects to be a Top 100 Player

Last week, Bayern Munich wunderkind Julian Green committed to switching his national team allegiance from Germany to the United States. The switch was approved by FIFA on Monday, and on Wednesday it was unsurprisingly announced that Green has been called up by Jurgen Klinsmann for next week’s Mexico friendly. The reaction to this news has been interesting, as some fans trumpet the arrival of a perceived savior of US Soccer, while smart writers work to temper short-term expectations, dive into the process of wooing Green, or bring undertones of how nationality has and should be defined in these circumstances up to the surface.

Naturally, I looked for an analytic that might shed light on Green’s current and potential level of play. Unfortunately, as can be expected with an 18 year old, his statistics are sparse. ESPNFC had only his 3 Champions League minutes, Transfermarkt listed 28 matches’ worth of data (1870 minutes), and the best I could find was SoccerWay with 50 matches, covering 3,967 minutes.

Green’s stat-line in that largest resource seems impressive, with Green scoring 25 times, meaning he has 0.57 goals per 90 minutes, and this year alone with Bayern Munich II, he’s on an astronomic 0.81 pace. But how to adjust for playing a few years above his age bracket, while several tiers below the Bundesliga? The only USMNT comparative I could find was Terrence Boyd’s SoccerWay page. Boyd played 2,041 minutes for Hertha Berlin II up through 2011, while he was 20, scoring 0.62 goals per 90. Scale back to Green’s age and there’s only a miniscule sample of 2 goals over 429 minutes (0.41 G/90).

Thankfully, GoalImpact swooped in with their scoring of Green’s young career and how he projects to mature:

GreenFor those that haven’t heard of GoalImpact before, it’s a metric that basically takes the goal differential while each player is on the pitch during every match it can find, and aggressively adjusts that ± based on opposition strength, an aging curve, and other factors. So, we are now beyond Green’s individual scoring rate, and on to how much better his teams have performed when he has been facing older opposition, and where his likely talent peak lies. It’s a very interesting process, which I have grossly simplified, but you can read more about it here. The mind-bogglingly large data mine driving this thing has been used to track the Carter progressions of an absurdly high proportion of professional footballers worldwide.

Through Twitter exchanges, GoalImpact offered some useful context for his projection of Julian Green as an eventual worldwide top 100 player. The current rating and forecast come from data on 50 matches with 3,847 minutes. While GoalImpact defends his projection, he categorized it as “unsecure,” and would not give a confidence interval for it due to the possibility of injuries, etc. that can hamper young players’ development. He also identified Shawn Parker (another dual national who has yet to decide between Germany and the US) and Jack McBean as pretty good, but not top 100 prospects, for the US.

GoalImpact does have a good track record of young player prognostication. Recently backtesting of 2007 prospects identified by the metric and those players’ present accomplishments is of particular interest here. Impressive that it identified the likes of Gareth Bale & Gonzalo Higuain via only 2007 data, but even more so that it tagged relative unknowns at the time like Axel Witsel, Neven Subotic, Jonny Evans, and the Bender twins, Lars and Sven. To be fair, the list also includes Jozy Altidore and others who haven’t lived up to their early GoalImpact billing. However, the mass of 2007 prospects identified by the metric have become abundantly more valuable over the last seven years.

Nothing is guaranteed, but given Green’s Bayern pedigree and good ratings from both traditional and quantitative scouting, there’s a very respectable chance that a few years from now he will be the best player wearing whatever outfit Nike will have fashioned for the US National Team. His combination of club pedigree, and accomplishments at a young age separate him from such vaunted names as Tab Ramos (First USMNT cap at 22), Thomas Dooley (31), Earnie Stewart (21), and Roy Wegerle (28). Eventually passing any of those figures in the hierarchy of US Soccer heroes is a daunting goal, but Green’s ceiling goes well beyond that. Never before has a truly promising young player valued by a national team as powerful as Germany’s chosen the USA instead.

For now, we should all be excited to see what the kid can do in friendlies, particularly the meeting with Mexico next week. Many want to talk about whether Green can or should make the World Cup team. As Alexi Lalas has been saying, there are legitimate chemistry issues to consider, because the last thing this team needs is a repeat of the 1998 fiasco with David Regis replacing Jeff Agoos at the last minute.To his credit, Green (and his father) are saying the right things, and he seems to be getting along with his new teammates.

Plus, a player of Green’s prospective skills seems well-suited for a bench role on this squad. It would be shocking to see him pass Landon Donovan, Clint Dempsey, Graham Zusi, and Alejandro Bedoya in the next two months, but Klinsmann probably wants at least one more midfielder who can bring something to the attack, and none of his other options in Brad Davis, Brek Shea, Sacha Kljestan, Joe Carona, Mix Diskerud, and Jose Torres seem to have booked their ticket to Brazil yet. Klinsmann won’t need to bring Green along for cap tie purposes, his one-time switch already means he is permanently a US player, so such an assignment would have to be merit-based. To win one of the last roster spots in Brazil, Green should need to prove his worth on the field, and win over the other players in the locker room, both in short order. That is a lot to ask, but Julian Green has the opportunity to deliver, and his chances of succeeding have to be respected.

8 thoughts on “Newly American Prospect Julian Green Projects to be a Top 100 Player

  1. But Subotic’s plus/minus is only so good because he ices the puck and can’t even defend a decent dump-and-chase team.

    Oh, I confused my sports.

  2. Loved the article, very interesting read. However, how does Green’s projection stack up to Altidore’s? When compared at the same age. It seems Altidore’s peak GI was 138 while Green’s seems to be 139ish. Also is it harder to use this model on certain positions and their “values” to their respective teams? Say it would be easier to predict strikers since the goal output is a easy way to tell if they are truly talented or not? Or do goals maybe inflate certain views on strikers? That is only a example of course.

    • Thanks for your comment, Nicholas. The prediction of the future development of very young players is difficult and hence all predictions involve uncertainty. In the case of Altidore, the original projection was maybe too optimistic. In many other cases, it was accurate or too pessimistic. If Julian Green will be once as good as predicted, or even better, or worse, we don’t know. But chances are that he will be excellent. Altidore couldn’t fulfill all hopes, but he isn’t a complete failure too.

      The algorithm doesn’t analyze which player scored the goals. The only criteria is the goal *difference* when a player is on the filed as compared to when he is not. A defender on the field benefits from a goal by a striker of his team as much as the striker himself. Therefore, unlike many other algorithms, it does not overvalue strikers and offensive midfielders. It works just fine on all positions.

      • Ah I see, so then does it overvalue players who are on better teams? I guess most great players play on good teams. However, some potentially good players could come from a losing team.

        • Check out the hyperlink I put in there that goes into more detail on Goal Impact’s calculations. Claims to adjust for teammate strength, alongside opposition strength, age, home field advantage, etc.

          The biggest questions for me regarding GI are:
          1) whether it overrates offensive or defensive (dis)advantages of a player who isn’t asked to do much on that side of the ball. For me, this is most problematic for keepers. Strikers and central defenders to an extent as well, though this is much more dependent upon what their managers are(n’t) asking of them. This doesn’t seem terribly relevant in Green’s case, as Bayern generally asks outfield players to work hard whether they have the ball or not.
          2) the role of luck in scorelines. While an individual player can certainly help and hurt his team on either sides of the ball, goals often have little to do with a particular player on the field, and I have a suspicion that this can be a big problem when looking at smaller samples. This is a substantial hurdle in rating and projecting players as young as Green, but GI’s history of prognostication is nonetheless impressive.

  3. As a Master of Public Administration student (Grad Dec 14), statistics, analytics, methods & research are always interesting to apply to various fields of study. The data analyzed is certainly one way to try and predict the future, based on past performance. I’m a Nate Silver fan, his track record, especially during the last election cycle was spot on. Suffice it to say, as i’ve argued in class from time to time, stats cannot measure heart <3!

  4. Pingback: A couple things for Tuesday « Over There

  5. Pingback: Educated Guesses, Analytics, and Tactics of United States World Cup Roster Decisions | Stat Hunting

Comments are closed.