4 replies to this topic

[aped1611]

Hey guys I was wondering if you could provide some feedback on this study I recently conducted on the draft using historical WAR data.


There has been plenty of work published regarding the value of draft picks, most notably the work done by Victor Wang. Wang's previous studies (Part 1 and Part 2) on the draft were a major help for me in conducting this study.

Using the first 100 picks from the 1992-1999 drafts I came up with a sample of 388 players who reached the major leagues. Due to such a small sample size I disregarded Junior College and Community College draftees. I used Wins Above Replacement as my benchmark for evaluating players which differs from Wang's used of WAB (thanks to Sean Smith of baseballprojection.com for the WAR database). I took the average of each player's WAR over their first six seasons. I also used CHONE's 2010 projected WAR for certain players who did not yet have six full seasons.

Much like Wang's study I broke the data down into first round picks (1-30), second round (31-70), and third round selection (71-100). I have broken down the data various ways below.

First Round

College hitters-- 1.336 WAR/year
High School hitters-- 1.204 WAR/year
College pitchers-- .679 WAR/year
High School pitchers-- .782 WAR/year

As previously noted college hitters hold the edge in terms of production, although high school batters are not far behind. One thing I found surprising is that my study shows that high school pitchers have actually outperformed college pitchers. This differs from all other previous draft study. This could be a result of a sample bias or a flaw in my methodology.

Second Round

College hitters-- .773 WAR/year
High School hitters-- .672 WAR/year
College pitchers-- .087 WAR/year
High School pitchers-- .084 WAR/year

Here we see a noticeable drop of in pitcher's WAR/year for both college and high school players. College hitters are still the most productive in this group.

Third Round

College hitters-- .115 WAR/year
High School hitters-- .424 WAR/year
College hitters-- (.015) WAR/year
High School pitchers-- .033 WAR/year

In this round we see high school batters hold the edge. College pitchers actually had a negative WAR in this range.

We can see clearly that hitters are the safest picks, particularly college batters within the first two rounds. Victor Wang noted that he felt the best strategy would be to draft hitters early and then stock up on pitchers. My analysis leads me to believe that drafting pitchers in the first round is a pretty good strategy. Pitchers drafted in the first round are far more likely to be productive during their first six seasons.

I broke the first round selections down further based on picks:

1-10-- 1.417 WAR/year

11-20-- 1.115 WAR/year

21-30-- .353 WAR/year

There is a drastic drop in performance between picks 11-20 and 21-30. The difference between the top 10 and 11-20 is not nearly as large.

Here is the WAR/year broken down by position. I only used the positions listed in the Baseball-Reference draft database although most players surely switched positions. I also did not include two-way players Brooks Kieschnick and Rick Ankiel. For ease I combined first baseman and third baseman into a group of corner infielders and second baseman and shortstop into middle infielders. In parenthesis are the number of players in each position.

outfielders (64)-- .977 WAR/year

middle infielders (48)-- .561 WAR/year

corner infielders (41)-- 1.046 WAR/year

catchers (29)-- .829 WAR/year

right-handed pitchers (151)-- .314 WAR/year

left-handed pitchers (55)-- .404 WAR/year

I think the most interesting note from this is that lefties outperformed righties. This could obviously be due to the fact that there are far fewer lefties than righties which makes them that much more valuable to a team.

[Rinoblast05]

Nice work. My biggest question would deal with standard deviation. More specifically, where were the outliers in each category in terms of WAR/year? If, for instance, most pitchers were within a short range of the average WAR/year from the first round (IE you have three pitchers at .700, .600, and .800) and in the 3rd round were farther apart (in a boom-or-bust fashion, with -1.2, -1.3, 1.2, 1.1) it would still support Wang's theory on drafting them later under the guise of high-risk, high-reward.

Then again, while the statistical analysis gives us a glimpse at draft value, it doesn't consider organizational need.

[philly sox fan]

As previously noted college hitters hold the edge in terms of production, although high school batters are not far behind. One thing I found surprising is that my study shows that high school pitchers have actually outperformed college pitchers. This differs from all other previous draft study. This could be a result of a sample bias or a flaw in my methodology.

I think the biggest thing to keep in mind is the size of the edge. It's 0.1 WAR/yr. Over 6 years that's barely over 0.5 WAR. These measures aren't that precise. The drafts of the 90s were much more balanced so depending on the years that the other studies used that could explain the difference as well.

We can see clearly that hitters are the safest picks, particularly college batters within the first two rounds. Victor Wang noted that he felt the best strategy would be to draft hitters early and then stock up on pitchers. My analysis leads me to believe that drafting pitchers in the first round is a pretty good strategy. Pitchers drafted in the first round are far more likely to be productive during their first six seasons.

The distribution might be as important as the average here too. Again though (and just of the top of my head) the 90s were pretty good for early rd pitchers - especially HS pitchers. There'd be a huge difference between 1st rd HS pitchers of the 80s and 90s for example.

There is a drastic drop in performance between picks 11-20 and 21-30. The difference between the top 10 and 11-20 is not nearly as large.

I found that exact same thing more or less although I think I had a break around #17 or so. Teams seem to be pretty good at determining the best 15-20 prospects, but not necessarily getting them in the right order. So there may be a huge difference between #17 and #22 in favor of the earlier pick, but it's a flat expectation between #3 and #13 with no surprise when #13 is actually the better player.

Here is the WAR/year broken down by position. I only used the positions listed in the Baseball-Reference draft database although most players surely switched positions. I also did not include two-way players Brooks Kieschnick and Rick Ankiel. For ease I combined first baseman and third baseman into a group of corner infielders and second baseman and shortstop into middle infielders. In parenthesis are the number of players in each position.

outfielders (64)-- .977 WAR/year

middle infielders (48)-- .561 WAR/year

corner infielders (41)-- 1.046 WAR/year

catchers (29)-- .829 WAR/year

right-handed pitchers (151)-- .314 WAR/year

left-handed pitchers (55)-- .404 WAR/year

I think the most interesting note from this is that lefties outperformed righties. This could obviously be due to the fact that there are far fewer lefties than righties which makes them that much more valuable to a team.

The position player differences aren't too large except for the MIs being so low which is interesting. I can't recall what my research said about that in the early rds.

Not sure what to make of the pitcher difference.

[aped1611]

Nice work. My biggest question would deal with standard deviation. More specifically, where were the outliers in each category in terms of WAR/year? If, for instance, most pitchers were within a short range of the average WAR/year from the first round (IE you have three pitchers at .700, .600, and .800) and in the 3rd round were farther apart (in a boom-or-bust fashion, with -1.2, -1.3, 1.2, 1.1) it would still support Wang's theory on drafting them later under the guise of high-risk, high-reward.

Then again, while the statistical analysis gives us a glimpse at draft value, it doesn't consider organizational need.

Thanks for the comments. I did check the SDs and the for the first round HS pitchers had a SD of 1.261 and College pitchers had 1.049. The SD were much lower in the second and third rounds as one would expect.

I'm not sure how this relates to Wang's hypothesis of taking pitchers in the later rounds. I still think the drop in average performance from the first round to the second in enough to warrant taking pitchers early in the first round.

[Eric Van]

Again though (and just of the top of my head) the 90s were pretty good for early rd pitchers - especially HS pitchers. There'd be a huge difference between 1st rd HS pitchers of the 80s and 90s for example.

The little study I was doing of top 15 (in MLB) pitching prospects suggests pretty strongly that any shift in draft value for pitchers is entirely the product of keeping them healthier. Which would make a lot more sense than any improvement in scouting.