High School Arms and the #1 Pick

The MLB draft kicks off tonight, and for the third year in a row, the Houston Astros will have the top selection. For the third year in a row, no one really knows what they’re going to do. Two years ago, they went for an under-slot high school shortstop, taking Carlos Correa ahead of Byron Buxton and Mark Appel, though Correa certainly looks like a terrific prospect in his own right. Last year, they went with the consensus, taking Appel and paying closer to the slot value of the pick.

This year, however, there’s a wrinkle; the consensus top prospect is high school left-hander Brady Aiken. No high school pitcher has been take with the #1 overall pick since 1991, when the Yankees picked Brien Taylor. He was a bust, as was the only high school arm who had been taken at 1-1 before him; David Clyde in 1973. In 49 years, teams have picked a high school arm just twice, and both of them failed.

However, as I noted in a piece for the Wall Street Journal today, as long as teams keep using those two failures as a warning to avoid high school arms at the top of the draft, the warning will become a self-fulfilling prophecy. Until someone takes another 18 year old pitcher with the top selection, no one will be able to disprove the truism that high school arms are too risky for the top spot.

But it isn’t like there aren’t examples of years where, with the benefit of hindsight, taking a high school pitcher at 1-1 wouldn’t have worked out. Perhaps we would regard the risks surrounding high school arms at #1 overall differently if the Pirates had taken Zack Greinke in 2002, or Josh Beckett went to the Rays in 1999, or Dwight Gooden was selected by the Cubs in 1982. Should we really suggest to the Astros that they shouldn’t take Brady Aiken with the top pick because no team has been willing to take that risk in the last two decades? Based on a sample of two failures?

What if we expand the pool beyond just the top pick? How much better have college pitchers performed than high school arms if we look at just the top 10 picks?

Thanks to Baseball-Reference’s draft database, we can do just that. So here the some career WAR numbers for pitchers selected in various buckets of the top 10 of the draft since 1965.

Top 10 Pick	High School	College
Number Selected	103	114
Average WAR	8.7	9.0
Median WAR	2.3	3.5
—–	—–	—–
Top 5 Pick	High School	College
Number Selected	47	65
Average WAR	10.1	10.9
Median WAR	2.2	4.7
—–	—–	—–
Top 3 Pick	High School	College
Number Selected	21	37
Average WAR	8.0	11.9
Median WAR	2.2	10.1

Keep in mind that these numbers are just total WAR (Baseball-Reference version) accumulated by players selected in those spots divided by total number of picks, so you don’t want to interpret those numbers as the expected value a team will get from a given selection. After all, guys like Gerrit Cole and Madison Bumgarner are included in the sample, as well as guys who are still in the minors but have very bright futures ahead of themselves. This is absolutely the most basic way of looking at production per pick, and plenty of more thorough studies have been done previously, but they basically draw the same conclusion as the super simple method above; top college arms get to the big leagues at a much higher rate than high school pitchers — thus the higher med, but the overall differences in total value for players who do make it isn’t that much higher.

Basically, college arms have a much higher floor, and you’re much less likely to get a nothing return on a pitcher who is closer to the Major Leagues than on one that requires several more years of development; this is why the median WAR is much higher for college pitchers than it is for high school arms. But look at much smaller differences in average WAR totals; there’s hardly any difference at all for the top 10 and top 5 buckets. HS arms taken in the top 3 haven’t fared particularly well relative to ones from college, so the truism that high-end college pitchers are generally better picks than high school pitchers holds up, but it’s still not so extreme that the “never draft a high school arm at 1-1” theory should hold up.

Talent is cyclical, and the best college arm in one draft is not always going to be better than the best high school arm. According to most of the guys who cover draft prospects for a living, this is one of those years where the upside of the best high schooler may outweigh the extra risk that comes with taking a guy further from the majors.

A decade ago, the sabermetric community roundly mocked the idea of using top draft picks to select high school pitchers. Today, there’s a decent chance that perhaps the organization that is most aggressive with their use of data is going to draft a high school pitcher with the #1 overall pick. Whether Brady Aiken is going to be more Josh Beckett or Brien Taylor is basically unknowable, but if we’re going to invoke the names of Taylor and Clyde, we should acknowledge that plenty of high school arms have turned out just fine, and would have been quality #1 picks had teams not been scared off by the risks.

I certainly don’t know enough to say whether taking Aiken at 1-1 is a good strategy. History tells us that it probably comes with more risk than taking a college arm, but not every draft class is the same, and perhaps Carlos Rodon is a higher risk college arm than most, or maybe Aiken is a lower risk high school arm, and maybe even both of those things are true. We just don’t really know. I think, in an ideal world, a team with the #1 pick would get to take an impact hitter, but those seem to be in short supply this year.

Brady Aiken’s career is going to turn out however it is going to turn out, and it won’t really matter much if he goes #1, #2, or even #10. Let’s not doom him to failure just because David Clyde and Brien Taylor didn’t make it. And let’s be thankful that the sabermetric community has moved beyond the idea of “high school pitcher = bad pick.”