SuperDraft or Super Daft, Part 2: Who is good at drafting? / by Kevin Minkus

By Kevin Minkus (@kevinminkus)

Inspired by recent NFL draft analytics articles, I wrote an article developing an expected value curve for the MLS SuperDraft. Using that curve as a baseline for how well draftees in a given slot should do, we can compare that to how well they actually do, across the picks for a given coach or team. This then tells us which coaches and teams have done an especially good or an especially poor job evaluating NCAA prospects over the last few years, by looking at who exceeds and who underperforms expectations.

RankTeam# of Picks% Value Above
Expected
1LA2766
2NYRB2650
3DC2536
4NE3932
5CHI3422
6SKC3420
7HOU2410
8COL369
9CHV192
10TOR35-1
11PHI26-1
12ORL5-13
13SJ22-18
14SEA26-19
15FCD31-25
16CLB33-29
17VAN21-32
18RSL25-51
19NYC4-51
20POR17-59
21MTL13-66

To the left is how things look at the team level from 2007 to 2015. (I should note that I’m only going up to 2015 because the metric I’m using to measure value is the total number of minutes played by a player in his first two seasons).

The Galaxy top the list here, despite longtime former manager Bruce Arena’s noted disregard for the draft. LA has done really well in the first round, with Omar Gonzalez, Sean Franklin, Mikey Stephens, and Robbie Findley all becoming regular starters for at least one sustained point in their careers. They’ve also picked out some solid contributors from the middle of the draft, though. Bobby Burling, Ty Harden, and Brandon McDonald all have or have had solid MLS careers for guys picked outside the first round.

Besides Chris Korb, D.C. United doesn’t quite have the same solid pedigree as LA with picks in the middle, but they have consistently identified talent at the top of the draft. Their best picks include Perry Kitchen, Nick DeLeon, Steve Birnbaum, Chris Pontius, and Rodney Wallace. All were picked in the top-10, but, just ask Real Salt Lake - nailing those picks isn’t easy.

On the other end of things, Portland, in second to last, should not be surprising. A large part of this is that Caleb Porter hasn’t had much interest in developing the players he does select. Only three Timbers draftees since 2011, when Portland entered the league, and 2015, received any minutes in their first two seasons in the league (those are Darlington Nagbe, Taylor Peay, and George Fochive). More recent picks Ben Polk, Neco Brett, and Jeremy Ebobisse have not as of yet helped to buck that trend. Porter’s approach does raise the point, though, that we shouldn’t necessarily conflate poor drafting as its measured here with poor prospect evaluation. It might more simply be the case that a coach isn’t (for perhaps a number of reasons) interested in playing his draft picks at all, regardless of their quality.

RankManager# of Picks% Value Above Expected
1Carlos de los Cobos891
2Pablo Mastroeni1183
3Bruce Arena1870
4Ben Olsen1169
5Steve Nicol2853
6Curt Onalfo1545
7Oscar Pareja1044
8Dom Kinnear2334

We can do the same analysis at the head coach level. To the right are the top eight managers at drafting, filtered to those who made more than seven picks between 2007 and 2015.

Pablo Mastroeni stands out here. He and the Rapids front office deserve a lot of credit for picking out Axel Sjoberg, Jared Watts, and Dominique Badji at picks 14, 33, and 67, respectively. His 2016 and 2017 picks haven’t yet panned out, so that’s worth watching, but he did a pretty decent job in the draft his first two seasons in Colorado.

Dominic Kinnear’s best draft picks really are impressive, as is the fact that he’s drafted well through his entire career. He picked Geoff Cameron at 42 in 2008, grabbed Will Bruin at 11 in 2011, and Fatai Alashe at 4 in 2015. He also drafted Warren Creavalle, Corey Ashe, and Danny Cruz, players who are by no means stars, but carved out stretches of solid seasons in their careers, at picks 37, 26, and 41. His track record should give Quakes fans waiting on Jackson Yueill some measure of hope.

And below are the managers who are the worst at drafting, according to the method used.

RankManager# of Picks% Value Above Expected
30Caleb Porter13-86
29Gary Smith12-82
28Gregg Berhalter8-82
27Steve Morrow10-82
26Greg Vanney8-79
25John Hackworth10-53
24Carl Robinson13-52
23Fernando Clavijo9-40

Gregg Berhalter shows up due to some very poor drafting in his first two seasons at the helm. His only pick that outperformed expectations was Adam Bedell, the 45th pick in 2014. Berhalter seems to have gotten things right with Niko Hansen, but other than Hansen and maybe Rodrigo Saravia, his 2016 and 2017 record isn’t much better.

Carl Robinson’s only pick to outperform his expected value is Tim Parker. The 2014 draft was especially bad for Robinson. His two worst picks across all seasons are Christian Dean (#3 in 2014), who’s only now starting to really see the field, and Andre Lewis (#7 in 2014), who played 0 minutes in two seasons in Vancouver.

So there you have it. This type of analysis is certainly fun and entertaining to look through. As far as providing important insight into coaches - I think it does a bit less of that. Evaluating young players makes up a lot of an MLS coach’s job, but evaluating NCAA players for the draft specifically is only a small part of that. There might be some correlation between ability to analyze college players and ability to analyze young players in other environments, but determining to what extent that correlation exists would require its own involved study. Still, the draft occupies a large place in the American sports fan’s psyche, so keeping numbers like these in mind can definitely be useful when the MLS offseason rolls around.

*There’s a lot more to dive into here, so if you’re interested in playing around with the modeling file I’ve used, feel free! It’s on github here, as is the entire aggregated coaching data file.