High School Draftees and Elite Seasons
Over the weekend, I stumbled upon this article that quoted some thought-provoking numbers from agent Scott Boras. In the article, he voices his support for increased collegiate baseball scholarships and lays out a few statistics that his agency has unearthed:
- >> 79% of collegiate first-round draft picks reach the major leagues for at least one day.
- >> 62% of high school first-round draft picks reach the major leagues for at least one day.
That is a significant 17% difference, though not necessarily surprising. High school draftees come with increased risk. Generally, their skill sets remain more unrefined than their collegiate counterparts. Thus, predicting the future talent for those players becomes much more difficult. This uncertainty causes teams to miss on a greater portion of high school players.
So, what continues to draw teams into drafting a high school player over a collegiate player every year?
Answer: The allure of acquiring potentially elite talent early in their development cycle.
Above is the aging curve for a baseball player. Teams desperately wish to pair young prospects with professional coaches as quickly as possible because the increase in ability is so dramatic during the early years. Teams believe they can have more of an impact during the formative years rather than later in the career when the curve begins to level off.
Theoretically, if Player A (a high school player) and Player B (a collegiate player) were both guaranteed to reach the big leagues at some point in their career, most organizations would gravitate toward Player A because he is lower on the development curve. He has more time to improve. Or more simply, he possesses greater upside.
This made me wonder if teams understand the greater risk of drafting high school players, yet continue to draft them over collegiate players in the early rounds because the eventual reward is greater. Perhaps fewer high school draftees make the big leagues, but if they do, they provide more value to their respective teams than the collegiate draftees. Perhaps the elite players in Major League Baseball are more likely to be drafted out of high school than college.
After all, why else would teams continue to draft high school players if the likelihood of their panning out was clearly lower? The reward has to be higher to entice teams to take a higher risk. That’s how it works in gambling and investing. Perhaps it works the same way in the MLB Draft.
For the sake of this analysis, the definition of an elite season from a player is +4.0 WAR or greater. There have been 235 players who have compiled a single season worth at least +4.0 WAR since the year 2000. Those 235 players have, at one time in their career, been considered elite players in Major League Baseball. We want to see if a higher percentage of those elite players were drafted out of high school rather than college.
As we can clearly see, 32.8% (77 of 235) of elite players were drafted out of high school. Compare that to the 45.9% (108 of 235) that attended some sort of college, and we are left with the conclusion that high school players are not worth the risk. The increased reward of an increased likelihood of drafting an elite player is simply not corroborated by the numbers. Not only do collegiate players reach the big leagues at a higher rate, but collegiate players also make up the vast majority of the elite players in the MLB since the 2000 season.
Despite the evidence that drafting high school players decreases the potential of reaching the big leagues and becoming an elite player, teams will undoubtedly continue to load up on the young high school players. On the aging curve, that makes complete sense. When analyzing overall success at the major league level, however, the numbers suggest that teams would be more wise to opt for the collegiate players.
That is not to say your favorite team will not hit the jackpot on a young, toolsy high school player from time to time. Of course, they will. Nothing is ever black and white in the world of scouting and player development. What the data does suggest, though, is that if your favorite team has the choice between Player A (a high school player) and Player B (a collegiate player) and both are considered first-round talents or second-round talents, your favorite team should select the collegiate player. They have a higher probability of drafting an elite player that way.
J.P. Breen is a graduate student at the University of Chicago. For analysis on the Brewers and fantasy baseball, you can follow him on Twitter (@JP_Breen).
Is that curve based on average WAR for all players that age in MLB? If so, it suggests a 20 year old plays on roughly the same level as a player who is 31-32. I can’t believe that’s true…
That might be true, actually. Remember there’s a lot of selection bias as far as which 31-32 year old players stay in the big leagues. The ones that REALLY drop off end up out of baseball, and the ones that stick around are very very good, making the aging curve look gentler than it actually is.
Well, the graph goes all the way down to 14 years old, so I’m not exactly sure what the y-axis is supposed to be (I’m not sure the last time a 14 year-old was in the majors). Maybe it’s not actually reflecting real data and the fact that the duds drop out of baseball.
In that case, shouldn’t it count for all the 15 year olds who don’t pan out?
Disagree. Need to know the formula used to assume that.
What if they count the 32 year olds who are out of the game as 0 WAR players? Then it’s making the aging curve look more “harsh”.
Agree, needs to be further elaborated.
Click the link.
Not that clear (from the BP article):
Baseball’s aging curve is fairly well known by now. Most hitters peak at or around the age of 27, and their performance usually proceeds along a parabolic curve, rapidly improving in their late teens and early 20s, then a more gradual improvement in their mid 20s, before a gradual decline in their late 20s that accelerates in their 30s. The following chart simulates the aging curve for players in the aggregate, plotting talent as a function of age:
(Note that the chart is an approximation not based on actual data, which is hard to come by for 14-year-old major leaguers.)
Here’s a thought. How many total players in MLB attended some sort of college, and how many total were drafted out of high school? Given that college players are much more likely to make the big leagues at all, I would think that the percentage of former college players is higher, which means that any given elite player is more likely to have gone to college.
If it turns out that high school draftees make up just 25 percent of MLB players, but produce almost 33 percent of the elite seasons, that could indicate that high school players have higher upside.
+1
My thoughts exactly. This is a glaring mistake unless I’m missing something.
Goes beyond glaring mistake. It pretty much invalidates the whole conclusion.
That was my first reaction too; that is to say, I would like to have seen the chart take into account the percentage of HS players and college players drafted in the upper rounds.
Also, teams generally have to make comparisons between specific players that are available at various slots. So, for example, if the team has a high first round pick, it may have a college and high school pitcher rated as the two best available pitchers. Suppose one is like Josh Beckett (2d pick, 1999), out of high school, and the other is like Paul Maholm (8th pick, 2003), out of college, One might be viewed as a potential top of the rotation pitcher, and the other as a low ceiling/high floor pitcher. My point is that the team isn’t faced with a generic college vs. high school pitcher choice, but rather the specific scouting reports on each player. I really don’t think you can generalize into saying teams should pick college players over high school players—there are too many individual factors.
Also, the chart in this article says absolutely nothing about the likelihood of draft choices producing elite talent. In addition to representation of hs guys in MLB, it fails to consider the proportion of the draft that attended some college vs. straight out of high school. I have no statistical basis for this, but I’m guessing that substantially more than 50% of drafted players who sign contracts came out of either 4-year programs or juco.
It would be really labor-intensive to figure out the exact proportion of “elite” performances to guys drafted in each category, but a representative sample of 3 years or so should give a decent idea of how many guys are drafted (and would be even easier to compile if you consider elite guys coming after the first 10-12 rounds or so to be an anomaly and cut the analysis off there).
For the second part of the analysis (regarding elite seasons), shouldn’t you adjust the numbers based on the relative MLB populations of college and HS players? What I mean is, even though there were more “elite” seasons posted by college players (108 to 77), that difference doesn’t have as much meaning if the number of college players in the majors is far higher than the number of HS players in the majors.
If there are more college players in the majors than HS players, then it should not be surprising that the majority of 4.0+ WAR seasons belong to the college players. College players could also have far more negative WAR seasons than HS players too.
I think it would be more informative to say that X% of MLB college players have produced 4.0+ WAR seasons and Y% of MLB HS players have produced 4.0+ WAR seasons. That would actually tell us if the higher-ceiling theory is true.
Or even the number of college players taken in the draft as opposed to the number of high school players. College players make up a larger percentage of both players drafted, so they make up a higher percentage of players who make the majors, and a higher percentage of players with +4.0 WAR seasons.
Interesting article.
Wouldn’t it be helpful to include the % of players drafted out of HS and college (say in the first 20 rounds or so) and then look at the % of those specific players that became elite?
What if 70% of draftees were college kids? Then I’d expect that there would more overall elite players that were drafted out of college as opposed to HS.
Classic example of getting the wrong answer because you asked the wrong question. Percentage of players with elite seasons is utterly irrelevant. What you need to figure out, when comparing relative value on a draft pick, is what is the likelihood of THAT player becoming elite. What percentage of the draft picks at a given point in the draft become elite, then compare with where they came from.
Unusually poor logic for this site.
“The reward has to be higher to entice teams to take a higher risk. That’s how it works in gambling and investing.”
Not exactly. In gambling, in fact the longer the odds (higher the risk) the WORSE the reward is on an expected value basis.
And in investing, there is startlingly little evidence of a risk premium in most assets.
I would also note that a baseball team might logically prefer to increase variance even if the expected value remained the same, or even decreased- trying to hit that elite player jackpot.
What Ian and bhollida said.
There also might be a sample issue here. You’re looking at results post-2000. College players are likely closer to the majors when drafted than high school players. If you were a high school player drafted in 2008, it’s highly unlikely you’ve had the time to develop enough to challenge that 4-win threshold, versus the player drafted in 2008 out of college who might have had a shot.
This article is a mess.
As a Jays fan this is a very interesting analysis, but I agree with most of the foregoing and more specific detail is needed.
Is there a difference between position players and pitchers? What about talent selected in the later rounds of the draft? In the previous CBA, teams could blow the bank on hard to sign HS talent. Is there evidence to suggest that that practice was an unwise gamble?
Most importantly, what about players who sustained 4+ WAR over more than one season. I realize I’m referring to a very small cadre of players, but perhaps the allure of drafting a player from one group over the other is based on the (hypothetically) increased likelihood of sustained stardom.
How much is this affected by pitcher injuries. Many pitchers don’t pan out because of an inability to handle the workload. A large part of the difference is that pitchers who threw at least two years in college have proven to be durable enough to at least make it to the majors. Something high-school kids have not done. I think you should seperate this study into position players and pitchers and see if that changes the numbers drastically.
I’m really impressed with the comments on this article. I would love to see this question asked with the above adjustments made.
I agree with the comments, though I’m not that impressed. The logical flaw in this is so glaring that I saw it immediately, and I’m not even that smart. I think some analysis of cost might be helpful, too. Don’t college players usually cost more to sign?
College players usually cost less to sign, as they have less leverage than High Schoolers.
I’d be interested if you did this analysis again in maybe five years. If you look at the group of young stars, or potential stars, today, a large number of them are high school players. Off the top of my head (all of these are younger than 25): Hosmer, Moustakas, Heyward, Lawrie, Trout, Harper, Upton, Stanton, Morrison, Bruce, McCutchen, Kershaw, Bumgarner, Freeman, Rasmus, M Moore, Cahill, Hellickson. Harper is the only one of those who hasn’t played in the majors yet, but I’m comfortable putting him in that group. This doesn’t include recent drafts that have been high school heavy. Whether it is better scouting, or just a greater emphasis on scouting in general, or a short term blip, a lot of elite young players recently have come from the high school ranks.
Can you find the average WAR of both the high school and college elite players? If the high school average was higher it would show why teams draft them. Even though they have a less likely chance to make it, they provide more WAR than their counterparts making them more valuable.
arghhhhh, the graph! the graph!
Part of my judgement on this question is prejudiced by the fact that I am more familiar with the prospects of the team I follow, the Giants.
Recent Giants top prospects list are loaded with 23-24-year-old players who have not played above A-ball, and 22-year-olds who haven’t played above rookie leagues. I presume this is because most of them come from colleges.
By the time they get to the majors, if they ever do, most of these players will be close to the peak age of 27 already. Before their 6 years of team control are over, they will be past their peak.
If these players were 3 years younger (no college), they might be in the majors at 24 and their 6 years would be 24-30 rather than 27-33.
This appears to be a significantly better value to me, though there are, of course, many other factors to consider.