A Brief Superdraft Analytic Preview

 In 2021, the MLS Superdraft (now reduced to three rounds) is barely a blip on the sports landscape. With the majority of players in MLS coming from outside the college ranks, does the draft serve a purpose?

The immediate success of players like Henry Kessler and Daryl Dike indicates that value can be found in the draft. Obviously, the bulk of that value is in the earliest rounds, but some diamonds in the rough can still be found. 

What's in a Pick?

What kind of value can be expected in a draft pick? The draft has changed a lot over the years, but looking back at the last 5 drafts there are clear trends that can be seen. I'm using ASA's Raw Goals Added to assess individual player contributions. Because this data does not include goalkeepers, I have left them and their seasons for this analysis.

First, I want to look at the seasons players have played over what's available. The draft is in many ways feast or famine; around 75% of players drafted never appear in an MLS game. The odds are good in the first 10 picks; about 91% play at least one season of MLS, and 66% have participated in every season available. By the 50th pick we reach equilibrium; players from 1-50 have a roughly equal likelihood of never playing than playing at all.

Obviously, that means that many of the picks before 50 are not playing to balance out the number from picks 1-10 that do. So what is the true equilibrium point, at which players have a roughly 50-50 chance of playing at all? In the next chart, we see that around pick 18 is where we hit that equilibrium. Picks from 11-18 are likely to play, but are not completing all of their seasons; that means either they are leaving MLS early or spending time elsewhere before starting their MLS careers. Interestingly, picks from 20-40 have a very similar total chance of playing regardless of where they are picked.

Now, how good are the players that are coming out of the draft? Small differences in pick selection are less due to differences in player quality and more differences in team situation. For this reason I've looked at a 5 pick average rather than the individual pick. This helps smooth out the graph and show a more realistic expectation per pick. Even making an appearance after pick 60 is not the norm so I have left it out of the analysis.

In the graph above, the rolling 5 pick average GA/90 is in blue. The line in red is the production adjusted for the likelihood of participating in a season they are eligible for. 1st picks, for example, have completed 12 of 12 potential seasons. The 30th picks, meanwhile, have completed just one of an available 9.

Immediately noticeable is the fact that production hovers between the 0.1 and 0.2 mark. MLS replacement level is roughly 0.173, without accounting for different contributions per position. The likelihood of playing drops off severely after the 10th pick.

Intuitively this makes sense. Replacement level MLS players are not expensive and are a prerequisite to competing at an acceptable level in MLS. If players cannot produce at that level, they simply won't play. The spikes in the graph in later picks are mostly affected by the fact that this is just one or two players making an impact.

A further question would possibly be why the back half of the first 10 picks produces at a higher level than the first half. It may be the case that teams in this position are selecting for specific needs, and therefore those players are more likely to enter a favorable situation.

A quick note: The ratio between the performance of defender, midfielders, and forwards picked in the draft is fairly similar to the ratio in MLS as a whole, suggesting that there is not a significantly greater pool of talent in a certain position from the draft. This also should suggest that teams should approach the draft with an eye towards filling needs rather than simply identifying the best player.

Who is picking well?

It's pretty easy to think of coaches and teams that use the draft well. Bruce Arena led the New England Revolution deep into the 2020 playoffs with drafted players in key spots. Meanwhile, Sporting KC's drafted players have played just 3 of 31 available seasons in the past 5 years.

To determine what a player might be expected to produce, I looked at the average raw G+/90 for players selected at that position. I then multiplied that by the amount of 90s the average player at that position received. This gives a rough idea of what a player might be expected to contribute per season. Then, I looked at the actual production of players selected by teams.

This chart shows which teams are getting the most value above expected per drafted player:

In the last 5 years Philadelphia and Atlanta have outperformed expectations in the draft, while RSL and the Galaxy bring up the rear in expected production. Inter Miami and Nashville, the most recent expansions, have not had much of an impact through the draft yet. As time goes on the difference between beating or underperforming expectations becomes more clear. 

On the surface, draft performance is not hugely important in building a squad. Very few drafted players become even average performers, and fewer still MLS stars. The net returns are usually small. That said, good team construction relies on marginal gains, especially in a league with as much parity as MLS. If your team struggles to create value from the draft, it makes sense for them to trade picks to accumulate value in other areas.

What can we expect from this year?

If patterns hold from the last few years, the first 10 picks have a good likelihood of providing  value for their teams if they draft well. Atlanta and Portland in particular are in strong drafting spots, and have a history of good performances from draft picks.

Houston has the pick with the most value historically, but they also have a history of drafting poorly. Tab Ramos and Martin Vazquez both have a reputation with youth and this could be a major opportunity for them to show they can find good value in young players.

Philadelphia's maneuvering to trade picks for other considerations may seem odd considering their history, but as the team moves towards identifying talent in their academy setup it makes sense. The ability to spot young talent in the draft should be transferable to identifying young talent in your academy players, and if you can acquire assets from other teams you still come out ahead.