Demography of the Good Player, Part I: Amateur Origins

Recently, Jeff Sullivan wrote a piece here attempting to answer a question notable both for its simplicity and importance. The question: how many good players were good prospects?

As Sullivan notes, one typically finds the question pursued in reverse: of this or that group of prospects (top-10 prospects, top-100 prospects, etc), how did they fare in the major leagues (if they even made it that far)? There’s great utility in this sort of information — in particular where our understanding of prospect valuations is concerned. An appearance by a young player on one of these prospect lists tends to indicate, if not certain future value, at least present trade value. In other words: even those prospects who fail to record even one plate appearance or innings — even they are capable of possessing significant value.

Not every good player was a good prospect, though. In fact, as Sullivan found, about a third of good players weren’t good prospects — or, at least, about a third of them never appeared on Baseball America’s annual top-100 prospect list. They weren’t exclusively all non-prospects, of course, but a sufficient enough percentage of those top-100 prospects fail such that, for a rookie-eligible player to expressly not appear among that group immediately renders his chances of succeeding in the majors pretty low.

The purpose of this post (and two others to follow) is similar to Sullivan’s — insofar, that is, as it’s intended to provide some objective demographic data regarding those players who’ve become good major leaguers. Here, instead of examining which players did or didn’t appear among BA’s top-100 lists, however, what I’ve done is to look at data regarding the amateur origins of good players — in the case of this post by their final amateur status (college player, prep player, etc.). Tomorrow, I’ll also consider good players by their draft round (in the event that they were drafted) and also by college conference (for those players who attended college obviously).

I’ve used a similar methodology as Sullivan did in terms of defining certain terms. A “good” player is any one who produced 3.0 WAR or greater in a particular season. For pitcher WAR I’ve used a 50-50 split between the FIP and runs-allowed iterations of WAR. Where Sullivan used three years’ worth of data, I’ve used five, hoping that the larger sample might be of some benefit.

Also with a view to creating a larger sample, note that I’ve used player-seasons and not merely individual players. So, for example, Dustin Pedroia produced five “good” seasons between 2010 and -14. Therefore, he’s counted five times. I was originally concerned that the difference in results might be dramatic between player-seasons and mere players. In fact, the relationship is rather regular: among batters and pitchers, among college and prep players, the average good player produced nearly (but not quite) two good seasons.

Good Players by Their Final Amateur Status
The first demographic criterion we’ll consider in this three-part series is the last level at which a every good major-leaguer played as an amateur before joining affiliated baseball. For international players, that includes both the July 2nd-types from the Dominican and Venezuela and also high-profile free agents from Cuba and Japan (who obviously played as professional). There’s few enough of the latter group, however, that it doesn’t skew the numbers terribly.

Here’s all 397 good player-season between 2010 and -14 classified by the relevant player’s final amateur status:

Players signed out of high school and college are represented almost equally; as international free agents, about 10 points less than that. Players signed out of a junior college represent about 10% of good major-league batters. If you are looking at a good player, there’s about one-third of a chance he was drafted out of high school and another one-third chance he attended a four-year college.

Here’s the same thing for the 233 good pitcher-seasons between 2010 and -14:

The ratios are roughly the same. Prep players gain a few points; international and junior-college players lose a couple. I have zero constructive ideas to contribute on what the appropriate margin of error is for this sort of data. Let’s say that four-year and high-school players are represented roughly evenly, with the possibility of some advantage for the latter. If you’re looking at a good pitcher, it’s most likely — among the four possible options — it’s most likely that he was drafted out of high school. But it’s far from a certainty. There are a lot of former collegiate pitchers who’ve recorded good major-league seasons, too.

Now here’s all 630 player-seasons together:

Again, not surprisingly, the number are pretty close between those good players who signed originally out of high school and college. One note about this: a higher percentage of college players have been selected in the first round of the last 10 drafts than high-school players. Not by much, but it’s there. Prep players, though, have produced good seasons with slightly more frequency. That’s maybe not surprising. Because there’s less information available regarding their performance, high-school amateurs typically offer both a higher ceiling and lower floor. I’d guess that college draftees produce 0-3 WAR seasons with more frequency than prep draftees. I don’t know that, but that’s typically the appeal of the college player: more certainty.

Also of note: players signed out of junior colleges have accounted for nearly 10% of all good season over the last five years, but they represent only 2% of all players selected in the last 10 first rounds. Adding in all the international players that are signed each year for first round-type bonuses, and junior-college players have provided an excellent return on investment. More on that to come.