Happy Labor Day! Regularly Scheduled Articles Will Resume on Tuesday, September 2.
May 19, 2005
Doctoring The Numbers
The Draft, Part Two
Before we continue our study, Jim Callis of Baseball America chimes in with this interesting fact about the draft's early days:
"While teams did prefer to get their hands on guys out of high school and mold them themselves in the early days of the draft, the draft rules also led to the high number of high school picks in the first rounds. If a player had been taken out of high school, then he wasn't eligible for the June regular draft (which most closely parallels the single draft of today). He would go into the secondary phase of the January or June draft. In 1971, which you cite, Pete Broberg, Burt Hooton and Rob Ellis (all college guys who shot to the majors) were the top three prospects and went 1-2-3 in the delayed phase of the June secondary draft. Steve Rogers and Steve Busby also were part of that draft. Steve Garvey and Dave Kingman are two examples of previous years.
"After 1971, baseball changed the rules, and all college guys not taken within the past 13 months were eligible for the regular phase. Teams did prefer the young guys, and it wasn't until 1981 that colleges had a majority in the first round."The existence of multiple drafts through 1986 greatly complicates the analysis of draft reviews from that period, and is all the more reason why draft research from that era, such as the work done by Bill James, is essentially obsolete. (That's not to say James' conclusions are obsolete…we'll find out soon enough.)
In this series' first article, we established the general value of each draft pick in their first 15 years, ranging from more than 45 WARP for the #1 overall pick to about 4 WARP for a second- or third-round pick.
That's the average value of all picks. What happens if we separate out the picks based on whether the player was drafted out of college or high school?
Of the 1526 players in our study, 749 were drafted out of high school and 715 were drafted from college. (The remaining 62 were junior college players…we'll address them later.)
The first place to start would be to compare the percentage of players from high school and from college who eventually reached the majors. Once again, players were broken down into five-pick segments--picks 11-15, 16-20, 21-25, etc. were grouped together--except that #1 and #2 overall picks were separated from picks 3-5. Here's a chart comparing the major league matriculation rate of both segments of players.
As you can see, the edge in favor of college players reaching the major leagues is substantial and sustained. At one point the lines touch, but they never cross; in every subset of draft picks, a higher percentage of college players reached the majors than high-school players. I calculated a weighted average for all players in the study from picks 1 through 100, and found that 41% of high-school players reached the majors, compared to 59% of college picks.
What if we break the study down into two eight-year chunks, and look at the data from 1984-91 separately from the data from 1992-99? Here's the same chart, but broken down into four different groups.
The chart looks pretty messy, but the important thing to take from this is that the blue lines--representing college players--are a lot higher than the red lines, which represent high-school picks. As you would expect, the lines representing player groups from the earlier era (1984-91) are generally higher than their counterparts from the more recent era, as players from 1992-99 have had less time to reach the majors and some eventual major leaguers have yet to debut.
Even so, the baby-blue line--the more recent cohort of college players--is consistently higher than the burgundy line representing the older cohort of high schoolers. At only two points (11-15 and 66-70) are high-school players from 1984-91 more successful at reaching the majors than college players from 1992-99. This, even though the high schoolers have all had a full opportunity to reach the majors (the youngest of them is about 32 years old.)
In chart form, the weighted averages of these groups are:
High School, 1984-1991: 41%
The gap between college picks and high school picks reaching the majors, which was 19% in the earlier era, has shrunk all the way to…18%.
Which leads to Draft Rule #3:
Draft Rule #3: College players are roughly 50% more likely to reach the major leagues than high-school players of equal draft caliber. This advantage has not changed over time.This is not a revelation, mind you. Even those studies, like the one done by Jim Callis, which argue that high-school picks are of essentially equal value to college picks, have shown that college players are more likely to reach the majors. The argument has been that the extra advantage in college players is primarily in getting more marginal players up to the majors for a brief cup of coffee. In terms of quality players and stars--in other words, in terms of real value extracted from the draft--the groups are not substantially different.
So let's compare these two groups again, only this time we'll look at 15-year WARP, the same metric we used to determine the value of draft picks in the first place. For this study, let me introduce a row of numbers you're going to see more of in future articles:
Y0 Y1 Y2 Y3 Y4 Y5 Y6 Y7 Y8 Y9 Y10 Y11 Y12 Y13 Y14 Y15 WARP 0.00 0.73 3.10 2.57 3.27 4.72 5.00 5.14 5.78 5.76 5.24 4.05 3.53 2.63 0.55 52.07This set of data represents the year-by-year WARP data for all #1 picks out of high school. Some explanations are in order:
The final WARP figure is the sum of all the totals from Y0 through Y15, which is our estimate for the overall worth of a draft slot. Certainly, most players will reach free agency before Y15, but I wanted to design the study in such a way that high-school players would not be unfairly penalized by virtue of the fact that they require longer to develop. If I had looked at only a draft pick's first 10 years, say, most high school players are just 28 at that point, and unless they established themselves in the majors very quickly, they were probably still short of free agency. In an ideal world, we would figure out the point at which every single player reached free agency, or was released, or picked in a Rule 5 draft…this isn't an ideal world, and there are only so many hours in the day. This data should suffice for most of our conclusions.
Let me relist the data above with a second line, the 15-year WARP data for all college players taken #1 overall:
Y0 Y1 Y2 Y3 Y4 Y5 Y6 Y7 Y8 Y9 Y10 Y11 Y12 Y13 Y14 Y15 WARP HS 0.00 0.73 3.10 2.57 3.27 4.72 5.00 5.14 5.78 5.76 5.24 4.05 3.53 2.63 0.55 52.07 Col 0.00 0.94 1.86 3.10 3.84 2.93 2.65 4.23 3.71 3.86 3.75 2.88 3.48 1.30 2.38 0.78 41.69Here, we see that while the college players taken #1 overall reach the majors faster, and are more valuable than their high-school counterparts in the first four years after the draft, they tail off quicker (understandable, as they're about three years older on average), and overall they fall far short of the value of high-school players. Their average 15-year WARP is a full 10 wins less than the average WARP of high-school picks.
So with the #1 overall pick, at least, high-school picks seem to be worth more valuable than college players. Of course, this is an extremely small sample size, which makes the conclusion suspect. The evidence that high-school players are more valuable really comes down to three data points--Griffey, Rodriguez and Jones--as two of the other three didn't even reach the majors, and Shawn Abner had 3.3 WARP for his career.
Still, exactly one-half of the #1 overall picks out of high school went on to become superstars. That's a pretty good ratio. If we look back further in time, four high-school players were picked #1 overall from 1974 to 1983: Harold Baines, Al Chambers, Darryl Strawberry and Shawon Dunston. So out of 10 players, three may end up in the Hall of Fame, one fell just short, one had a Hall of Fame start to his career, and one spent a remarkable 18 years in the major leagues without ever taking a called strike.
So I'm comfortable making this rule:
Draft Rule #4: In a year where there is a clear superstar talent available in the high school ranks, it is a perfectly acceptable draft strategy to select that player with the #1 overall pick.I qualify this rule because there are some years where, going into the draft, no high-school player stands out as a can't-miss superstar. That's happened twice in the last five years; it was pretty clear on draft day that Adrian Gonzalez and Matt Bush were #1 picks by default, and not out of any clear conviction that they were future Hall of Famers. The two players in the last five years who did fit that profile, Joe Mauer and Delmon Young, are coming along rather nicely.
Why am I separating the #1 overall pick from the rest of the draft? That will become clear when we look at the rest of the data. The follow chart once again uses five-player groupings, with the #1 overall pick removed from the study, and compares the 15-year WARP value between high school and college picks.
The edge for college players on this chart may not be quite as dramatic as it was in this article's first chart, but it's equally as significant. With the exception of two small spikes at 31-35 and 46-40, there is no cohort of high-school players that outplays its college counterpart. The gap is particularly striking in the first half of the first round; high-school players picked from #2 to #10 average just 16.70 WARP, while college players drafted in the same spots average 29.16 WARP--nearly double the career value.
If we break it down round by round, here's what we get:
Round HS Col % Diff 1st (#1-30) 14.31 20.32 42.0% 2nd (#31-70) 4.37 6.59 50.6% 3rd (#71-100) 1.88 5.34 183.8% Overall 6.68 10.33 54.5%College players taken in the first round have, on average, 42% more overall value than their high-school counterparts…and the gap (in relative terms) only widens from there. Third-round picks out of college have almost triple the value of picks out of high school. (Remember Draft Rule #2, which states that there isn't much difference between second and third-round picks? Apparently, that really only applies to collegiate players.)
What's particularly interesting is that earlier, we saw that college players were more likely to reach the majors than high-school players, by a margin of 59% to 40%, which works out to a 48.6% increase in the likelihood of reaching the majors. But college players have 54.5% more WARP than high school players. In other words, if you do the math, not only are college players more likely to reach the majors, but the ones who do reach the majors have more career value than the high-school players who reach the majors. The advantage college picks enjoy in terms of reaching the majors is not simply because of the added bulk of a bunch of marginal draft picks.
Which leads us to Draft Rule #5, which is pretty much the most important draft rule of them all:
Draft Rule #5: In the first three rounds, not only are college players about 50% more likely to reach the major leagues than high-school players drafted in the same slot, they produce approximately 55% more value over the course of their careers. This advantage is persistent at every point after the #1 pick.Twenty years ago, Bill James wrote that "the rate of return on players drafted out of college is essentially twice that of high school players." If you simply change "twice" to "55% more," that statement still holds true today.
Next time, we'll look at whether the advantage enjoyed by college players has diminished in the more recent set of draft picks.