Book excerpt: Why it’s still a bad idea to draft high school pitchers in the first round

For years, Daniel Kahneman’s iconic work of behavioral science “Thinking Fast and Slow” has been required reading in front offices across Major League Baseball. In this book, Law applies Kahneman’s ideas about decision making to the game itself. Combining behavioral science and interviews with executives, managers, and players, Law analyzes baseball’s biggest decision making successes and failures, looking at how gambles and calculated risks of all sizes and scales have shaped the sport, and how the game’s ongoing data revolution is rewriting decades of accepted decision making.

Excerpted from the book “The Inside Game: Bad Calls, Strange Moves, and What Baseball Behavior Teaches Us About Ourselves” by Keith Law. Copyright © 2020 by Keith Law. From William Morrow, an imprint of HarperCollins Publishers. Reprinted by permission. Order a copy from Bookshop.org.

If you ask any longtime scout or baseball executive what the riskiest category of player is in the MLB draft, the majority of answers will probably come down to high school pitchers. The industry regularly selects between two and ten such players — teenaged pitchers who often throw hard, but are still growing physically and emotionally, and whose careers thus have a panoply of ways to go off the rails — in its first round, despite the long-held belief that such players are high-risk bets, because the belief is that the potential rewards justify the risks. Some front offices have eschewed the category almost entirely, including the Oakland A’s of the era documented in “Moneyball” and the Blue Jays while I worked there, but such strategies have been shortlived; in 2005, just two years after that book was published, the A’s had two second-round picks and used them both to select high school pitchers, neither of whom ever saw the majors. (The A’s still haven’t used a first-round pick on a high school arm since taking Jeremy Bonderman with an extra first-rounder they received in 2001.)

The truth of the matter seems pretty clear: high school pitchers selected in the first round have a higher failure rate than other categories (high school position players, college pitchers, or college position players), and do not offer higher upside. That’s not to say that it is always wrong to take a high school pitcher in the first round, but that such players should be pushed down on draft boards to reflect the greater risk of them failing to reach the majors or produce first-round value once they get there. If you have two choices for your next draft pick in front of you, one a high school pitcher and one any other type of player, and you believe their value is about equal, you should take the other guy.

This information isn’t necessarily new, although recent data seems to bear it out again. Yet teams continue to — in the vernacular of the industry — “pound” high school pitching in the first round. The data say that teams should be moving away from these players, but until the 2019 draft, which came close to historic lows for this category, there was no evidence that the industry as a whole was backing off high school pitchers.

Focusing on the individual case in front of you while ignoring the data you have on the more general category to which that case belongs is known as base-rate neglect. Base-rate neglect is the name for the phenomenon where you have a mountain of evidence saying one thing, but you choose to ignore it in favor of the specific case in front of you. You favor the information that stands out because it’s more recent or it’s more memorable or it’s just the first thing you found, and in the process you forget or fail to check on the larger sample of data over a longer period of time. The former could be misleading, while the latter should be more predictive because of the greater sample size. You know high school pitchers flame out more often than other players do, but dangit, this here kid Joey Bagodonuts — you just know he’s different, right? He’s the exception that proves the rule, or something. Your scouts are just sure he’s the one kid who’s different.

And sometimes there are indeed exceptions. The Royals took Zack Greinke with the sixth overall pick in 2002, and said at the time that they thought he was as advanced as a college pitcher even though he was just eighteen. They were right about that, and they were right about his upside; as of this writing, he’s been the most valuable player drafted in 2002, and has enough of a margin over the second-most valuable player (also a high school pitcher, Cole Hamels) that he seems likely to retain that title. He really is an exception, though, and one Greinke don’t stop no show. Taking a high school pitcher with your first-round pick remains a high-risk, medium-reward proposition, and we should see fewer such players taken in the first round.

Major League Baseball’s draft differs from the drafts in North America’s other two largest sports leagues (the NFL and the NBA) by allowing teams to select players who have finished high school but not yet matriculated at any college. To be eligible for the draft, you must have graduated from high school (or be about to do so), or have completed three years of eligibility at a four-year school, or have completed one or both years at a two-year school, or have turned twenty-one by the draft, or be about to turn twenty-one within forty-five days of the draft. That’s a lot of rules, but for most players, the upshot is that you can sign right out of high school, or you can go to a four-year college and come back out after your junior year.

This presents major-league teams, from their individual scouts on up to the general manager, with the challenge of evaluating high school players who are typically seventeen or eighteen years old against college players who are twenty-one and occasionally twenty-two years old, a comparison that isn’t quite apples to oranges, but more like red apples to green. They’re similar, but you probably have different feelings about Fuji apples than you do about Granny Smith apples. (And don’t get me started on Golden Delicious apples, which are not delicious, and turn to mush when you bake them.) It’s not easy to decide whether the seventeen-year-old wunderkind you’re watching is going to be a more valuable player in the long term than the twenty-one-year-old player who’s not as exciting or athletic but who promises to get to the majors a year sooner, if not more.

With pitchers, the comparison is further complicated by perceptions that high school players are more likely to get hurt, because they’re still growing and developing physically during the years they might be in college, and that going to four-year schools tends to thin the herd — whether just because of natural attrition or because there are still college coaches who’ll overwork their pitchers in pursuit of soon-to-be-forgotten wins.

Thus major-league front offices are confronted by an epistemological question. The conventional wisdom in baseball has long held that high school pitchers are higher risk than other categories, but that the higher reward justifies the risk. Clayton Kershaw was the best pitcher in baseball for a five-year stretch, and he was a high school pitcher (taken 7th overall in 2006 by the Los Angeles Dodgers). Greinke has been incredibly successful since he was drafted, as has Hamels, as well as Matt Cain, also taken in that same first round, 25th overall. (That draft yielded six high school pitchers who were drafted in the first four rounds and produced at least 20 Wins Above Replacement in their careers: the three I just mentioned, as well as Scott Kazmir, Josh Johnson, and Jon Lester.) Partisans of prep pitching prospects point to these players as proof of their proposition, and can invoke a baseball-centric FOMO argument: if you don’t take high school pitching prospects, you’ll never get a Kershaw or a Greinke or a Roy Halladay.

Baseball-Reference has every draft in MLB history with players linked to their career stats, which makes this a question we can easily answer just by choosing our parameters. Does taking a high school pitcher in the first round make sense? That is, is the base rate for such a draft pick on par with those spent on other types of players?

Of course it doesn’t — otherwise we wouldn’t be talking about it. Yes, there are stars who come from the high school pitching ranks, but such hits are infrequent enough to say that, in the absence of other information, spending a first-round pick on a high school arm is a poor strategy.

I looked at every first round from 1985 through 2012, which is the most recent first round to see any high school players accumulate at least 10 WAR in the majors, and compared high school pitchers’ production to those in other categories. For example, how frequently did players taken in the first round reach the 10 total career WAR threshold?

Pitchers have more ways to fall short of expectations or to bust completely than hitters do, thanks to higher injury rates, so it's unsurprising that both categories of pitchers fare worse than even the less successful category of hitters ... but first-round high school pitchers are so much less likely to pan out that it's hard to believe they're still chosen there as often as they are.

What if we look just at the uppermost echelon, the top ten overall picks in the draft, where one would assume only the absolute best high school arms are taken, those who should have a lower failure rate while still carrying the same potential upside?

If you pick in the top ten and you choose a high school pitcher, you're the guy at the craps table betting on eight the hard way because it sounds cool to say it. You might hit big on your bet, but you're accepting a much higher risk that you get little to no return than you would by taking any other category of player.

As you might expect, some of the hits in that group of ten high school pitchers taken in the top ten picks of their drafts are enormous. Two are probably going to the Hall of Fame: Clayton Kershaw and Zack Greinke, inarguably great picks for their respective clubs (the Los Angeles Dodgers and the Kansas City Royals) where they were selected. Two others, Josh Beckett and Madison Bumgarner, racked up more than 30 career WAR each, and were both major contributors to World Series titles won by the teams that drafted them, with Beckett earning a second championship ring after a trade to Boston.

A fifth, Kerry Wood, is one of the great might-have-been stories in recent baseball history; as a rookie, he recorded what was at the time just the second 20-strikeout game in MLB history, then won the NL Rookie of the Year award. He blew out his elbow and missed all of 1999, but returned to post three seasons as an above-average starter, culminating in a 2003 season where he led the NL in strikeouts and was part of a Cubs team that came within five outs of reaching the World Series.

Unfortunately, a long history of overuse that dated back to high school — he started both ends of a doubleheader just days after he was drafted — and ran through 2003, where manager Dusty Baker had him throw 120 or more pitches six times in seven starts to end the season, ended Wood's career as an effective starter after age twenty-seven. He's both a successful pick, taken 4th overall in 1995 and producing 27.5 career WAR within nine years of his draft, and a cautionary tale of the risks of taking high school pitchers (or any pitchers, really) with high picks.

The other 37 high school pitchers taken in the top ten picks who didn't reach 10 career WAR, however, include a lot of names you've probably never heard before, or perhaps heard once on draft day and never heard again: Matt Hobgood, Chris Gruler, Clint Everts, Colt Griffin, Mike Stodolka, Matt Wheatland, Mark Phillips, Joe Torres, Josh Girdley, Bobby Bradley, and Kirk Presley combined for exactly zero major-league appearances.

Taking high school pitchers against these odds amounts to a bit of magical thinking: Yes, the failure rate for high school arms taken high in the draft is high, and the opportunity cost of those picks is also pretty high (you could have taken a position player who was more likely to return sufficient value for the pick), but we think this guy right here is the exception. This amounts, at least at some level, to saying that you believe that your group of evaluators, from the general manager to the scouting director on down, can figure out which players will be the exceptions to the base rate. That's quite possible — perhaps some teams or individuals have proprietary tools or ideas that allow them to sift through high school pitchers and find prospects with higher probability of success. Yet in the sample I examined, covering 28 first rounds, three teams drafted three high school pitchers each who met the 10 WAR threshold, and no team drafted four. Of those teams, only the Toronto Blue Jays, who took Steve Karsay in 1990, Chris Carpenter in 1993, and Roy Halladay in 1995, never whiffed on such an evaluation: They took five high school pitchers in the sample, those three plus two who didn't sign. If there's a way to distinguish these high school pitchers from one another, no team has found the secret.

I asked multiple front-office executives why they continued to take high school pitchers with first-round picks, or why they thought other teams did so. One simply called them "stubborn" but asked me "don't convince them otherwise." Another executive who has taken high school pitchers in the first round in the past said, "I think people still think they'll draft the exception. The logic has always been if you don't take HS arms then that means that you'll pass on guys like Kershaw, Bumgarner, etc. I've definitely learned my lesson."

(Photo of Clayton Kershaw in the 2007 All-Star Futures Game: Jed Jacobsohn / Getty Images)