i applaud your efforts but i have to wonder if this question should only be applied to the first 1-3 rounds of the draft. after round 3 the signing bonuses are much smaller and the likelihood of finding a gem becomes more and more of a crapshoot. in the first 3 rounds i would think that the question is much more relevant. furthermore, i think that the philosophy of college vs. high school would yield larger result gaps. maybe you can narrow down your results to answer this question.

Good job.

I wonder how the numbers are affected by where they are drafted. I know there would be no causation but there might be a correlation.

Maybe like in poker. Sometimes to call to draw to the flush and sometimes you don't. The odds will tell you. Same with pitchers. maybe a pitcher with a scouted score of 72 (or whatever a good scouted score is) factored with chance of success will tell you if you are getting value.

I personally don't like High School pitchers. To much chance of injury. let them get hurt at Cal State Fullerton is what I say.

... i wish i wasn't so busy, so i could follow up better. i think draft position is probably somewhat important, but i imagine that if you control for that then the college pitchers will look a lot better. there probably aren't many high school pitchers going in the 12th round. so, if a lot of the college pitchers are going in the late rounds, and their top potential is as a middle-reliever, that could skew the numbers, making college pitchers look worse as a group than they really are.

i'd also like to normalize for IP in the majors, but that probably isn't possible without tons of data work. still, i think it is relevant to parse whether these guys are getting September call-ups for the Royals and just throwing a handful of innings, or whether they're pitching full seasons for contending teams. since WHIP is the measure of quality, presumably a few dudes that come up and throws 5 innings of 0.85 WHIP in the majors could skew the numbers. the same could happen at the top of the distribution.

i also would like to see some sort of breakdown of standard deviations. the mean values only tell us so much. high school pitchers could still be more high risk/reward types even if the mean values are exactly the same as those of the college pitchers. the standard deviations could clue us in on that.

but it's still a great diary.

That college pitchers tend to outperform high school pitchers is one of Bill James' tenets:

Players taken in the June draft coming out of college (or with at least two years of college) perform dramatically better than players drafted out of high school.

Since pitchers tend to peak later than hitters—in their late twenties—it makes a ton of sense to draft pitchers who'll be peaking in their cost-controlled seasons, and who you've shown are more likely to contribute in the show.

That being said, I think it's a good idea to draft a few more HS pitchers. With the new CBA, you have 5 years to develop them before they need protection—having some talented arms that can be broken in as relief pitchers before converting to starters (like Wainwright) is a solid way to run a system.

Once things settle down a bit, I'll look into this too. It'd be a fun off-season project to put together a community-wide collaborative draft strategy report.

I've always thought that the answer was somewhere in the middle.  There is the occasional arm that is to good to pass up out of the high school ranks.  However, a wiser choice is for the majority of your picks to be college players (not just pitchers).  This data seems to reinforce this, as you're gambling with a lot of money by taking highly touted HS arms.  

Outstanding diary, btw.

is that most of the best pitchers never went to college.  They were drafted out of high school and went straight to the professional ranks.  Therefore, comparing HS to college pitchers isn't comparing apples to apples.  This isn't a criticism; rather, it's a reflection of the available talent pool.  

If HS pitchers, generally, perform better -- they should.  They're the better pitchers, by and large.  

If college pitchers, generally, perform better (or even as well) -- doesn't that speak to the fact that pitching is so difficult to predict?  You'd be taking lesser talent and having it produce as well or better.

So, to your question, which should we draft?  The HS pitchers should be the more talented b/c the best of them never make it to college.  College pitchers are probably less risky and closer to the majors but rarely have the upside of the HS pitchers.  The major infusions of talent come from the HS ranks.

I have been intrigued by this question for a long time.  It is very difficult to really prove anything, but lots of fun to talk about.  I don't really agree with the argument that the best pitchers sign out of high school.  My son is a D1 pitcher and I have seen several other prospects who were very highly regarded who essentially took themselves out of the draft by raising "signability" issues.

So, just for fun on a lousy day I decided to take a closer, although admittedly unscientific look, at the draft class of 1997.  I chose 1997 solely because it was ten years ago and it seems like anybody who is going to make a mark would have done so by age 28.

Interestingly enough there were sixty pitchers taken in round 1-3, including the supplemental round.  The data are as follows:

Summary       
    HS    College
Active    7    6
Done    7    13
DNP    16    11
Total    30    30

Active  MLB       
IP       
Mean    471    706
Min    55    175
Max    1428    1714
Median    242    417
ERA +       
Mean    99    101
Min    77    87
Max    118    127
Median    101    99

Retired  MLB       
IP       
Mean    108    216
Min    11    2
Max    228    544
Median    84    188
ERA +       
Mean    72    76
Min    63    30
Max    97    95
Median    72    81

From the summary information you can see that college pitchers (63%) were more likely than HS pitchers (47%) to at least get a cup of coffee at the big league level.  FWIW, I counted JUCO players as college level.  There were only three and only one made it to the show with insignificant results.  I don't think it is too surprising to see that college pitchers are more likely to get to the big leagues.  

It was interesting to me that a higher number of high schoolers (23%) than college pitchers (20%)are still active in the big leagues.  Probably because they are younger than the college guys, but the college guys should only be 31 or 32 years old.  Not really a significant difference anyway.

The active pitchers look pretty comparable to me with the college guys having a slight advantage - again probably due to the age difference.  One could easily assume that the career accomplishments of each group will be pretty similar.  So it comes down to if you like the college guys from this draft, most notably:

Jeff Weaver
Randy Wolf
Scott Linebrink

or the better HS guys:

John Garland
Aaron Cook
Jeremy Affeldt
Rick Ankiel

I have no idea how to deal with the Ankiel factor, but he still was a very good pitcher with a career ERA+ of 118 which would put him in the top 150 of all time if he could have done that over a career.

The only really significant difference in the two popluations is that the college guys who made it for a while and faded away seem to have made a better mark than the HSers.  To me, the fact that college pitchers might offer you 188 innings at an ERA+ of 76 versus only 84 career innings at an ERA+ of 72 is statistically interesting, but should have no bearing on draft strategy.

I would like to take a look at all the current big leaguers with ERA+ over 100 and see where they came from.  My guess is the answer will be Latin America, Asia, and high school.  I will try to put that together in the coming weeks unless someone else wants to step up.

I have been thinking about this HS vs. College thing for a while now and it occurred to me I was thinking about it all wrong.  While it is a great thing to draft in a way that maximizes your potential for developing major league talent, I have been thinking about it in terms of quantity rather than quality.   I thought why not just look at the best pitchers and see where they lurk.  After all, if I'm a GM I get a lot more excited about drafting an Albert Pujols than a Skip Schumacher - nothing personal against Skip.

So I started trolling around for data as I am not exactly a sophisticate in this type of analysis.  Thankfully I didn't have to go too far to find a comprehensive list of the top pitchers as ranked by ERA+.   This may or may not be the ultimate determinant of a pitchers value, but I was pretty sure it was a valid enough measure for trying to rank pitchers.  As an ERA+ of 100 represents major league average, I decided to limit my study to pitchers with an ERA+ of 110 and above.  Of  all the pitchers active through 2007 with at least 1000 IP, there are 38 with an ERA+ on the good side of 110.

These 38 pitchers come from one of three places:  they are amateur free agents (AFAs) signed outside the U.S. or they are either high school or college pitchers draft in the MLB amateur draft.  I don't know the total number of prospects signed each year from each source, but of the top 38 pitchers the breakdown is as follows:

*    8 AFAs (21%)
*    9 College picks (24%)
*    21 High School picks (55%)

So much for my earlier post that, "I don't really agree with the argument that the best pitchers sign out of high school."  Either I was wrong or the minor leagues are a much better place to develop than college.  I don't think the disparity could be that great, so it is a pretty safe bet I was wrong.  

Some argument could be made that while the HS picks have been more successful some of the very best pitchers have come from both the AFA signees and the college picks.  After all, Pedro Martinez, Johan Santana, and Carlos Zambrano are three of the top eight.  Three college players, Roger Clemens, Brandon Webb, and Randy Johnson are also in the top eight.  That leaves only Roy Oswalt and Greg Maddux to represent the high schoolers.    On top of that the average ERA+ for the AFA group is 124, the college group is at 125, and the high school group trails at 119.  A better way to look at is who would you want to be in your starting rotation.  I would love to have any of these five during their peaks.  Ranked by ERA+ each group would be represented as follows:

AFA               College      HS
Pedro Martinez      Brandon Webb      Roy Oswalt
Johan Santana      Roger Clemens      Greg Maddux
Carlos Zambrano      Randy Johnson      Roy Halladay
Andy Pettitte!!      Tim Hudson      Mark Buehrle
Bartolo Colon      Mike Mussina      Derek Lowe

A couple of notes on the pitchers above:  I don't know Andy Pettite's history, but my source showed him as an amateur free agent.  Doesn't sound right, but as Joaquin would say, "youneverknow".  Mark Buehrle doesn't have a college listed, but I thought for sure he went to Jefferson College for at least a year.  Maybe he was a draft and follow.

One additional bit of interest, especially to Cardinal fans, is which teams actually identified these guys.  I don't think many will be surprised that the Cardinals join the Devil (or not) Rays, Giants, Mariners, Phillies, Reds, Rockies, Royals, and Twins as the only MLB teams not to have signed or drafted a single one of these 38 pitchers.  My guess is these are all teams that don't have overseas operations and draft mostly college pitchers.   The winner in drafting the most of them is, of course, the Cubs with four.  I think that shows the value of drafting high....The Astros, Blue Jays and Braves come in second with three apiece.  Interestingly enough, the Astros were fortunate enough to have been the only team that signed two AFAs who made the list.  Unfortunately for them, they have the distinction of not having seen a single MLB inning from either Johan Santana or Freddy Garcia.

Finally, it apparently isn't all that hard to identify these guys as 15 of the 30 drafted players on the list were taken in one of the first two rounds.  Since eight were AFAs, only 15 players drafted after the second round have put up significantly above average major leagues careers.  The latest pitcher drafted was high schooler Kenny Rogers by the White Sox in the 39th round.  No college player drafted after the 8th round made the list while nine of the HS guys were drafted in the 8th round or later.

I could use some help drawing conclusions here.  It seems like it is not a particularly good idea to draft a college pitcher high unless you are pretty sure he will make it to the bigs pretty quickly.  Six of the eight college pitchers on the list were drafted in the first two rounds while Jake Peavy, John Smoltz, and Roy Oswalt were drafted in the 15th, 22nd, and 23rd rounds respectively.  

by the way

seems all this boils down to the individual.  clemens probably had a choice out of high school (wild guess), but chose college.  i suspect his career would have been largely the same either way.  the later maturation of pitchers seems on the surface to simply suggest concentrating on HS position and college pitchers is the best way to go.  for college players, the landscape is changing a bit this year with the new transfer rules.  i think these rules will help because it gives the college coaches and the p[layers more of a long term outlook.  i think this will enable college baseball programs to move a bit closer to the talent development factories they are for basketball and football.

on the guys who are least statistically above average.  More specifically, this is a followup to cdb's request that I look at the pitchers with over 1000 IP and ERA+ ratings from 101-109.  These are the guys Cardinal fans are more likely to have some familiarity with:  the Matt Morrises, Jeff Suppans, and Woody Williamses of the world.  Surprisingly enough there are only 21 starters on this list.  I guess that means if you are only barely above average it is hard to stay around long enough to rack up 1000 IP.  

I don't have a way to identify the population down below 1000 IP, so I will stick with what these data show me.  Anyway, if I am drafting and looking for guys who are significantly above average I am for the most part looking for starters.  Most relievers just kind of happen.  I think it is fair to define a starter who is significantly above average as one who can rack up 1000 IP. That's only 5-7 years of starting.

The numbers are a bit different as the college pitchers fare better with the same percentage of the pool as HS pitchers - 43%.  That leaves only 14% for the amateur free agents (i.e. Latins from outside U.S.)  Combined with the numbers from the previous post on pitchers with ERA+ of 110 and above, it means that of all the active pitchers with ERA+ above 100 the breakdown is as follows:

AFA           11/59     19%
College       18/59     30%
High School   30/59     51%

Combined with the total draftees numbers provided by cbd indicating that historically a lower percentage of HS pitchers (42% and declining) have been drafted, it is striking that the numbers are skewed so far in favor of the HSers.  I guess you could say that a HS pitcher is close to twice as likely as a college pitcher to become an above average MLB starter.

Again, it doesn't seem all that difficult to identify who these guys are as 13/21 or 62% of this group were either AFAs or drafted in the first two rounds.  Woody Williams has the distinction of being the lowest drafted college pitcher to make the cut as he was drafted in the 28th round.  As most of you are probably aware, Woody was drafted as a shortstop.  That means that NOT A SINGLE ONE of the 59 currently active above average starters with over 1000 IP was a college pitcher drafted after the eighth round.  I hope Luhnow and company are aware of this.  Eight of the HS pitchers on the list, or 14% of the total were drafted after the 8th round.

Thanks to Matt Morris with an ERA+ of 109 the Cardinals were not entirely shut out.  He is the only active big league starter with 1000 IP and an ERA+ above average to have been selected by the Cardinals in a period that roughly covers at least a dozen drafts.   Again, the overall leader in selecting top pitchers is the Cubs with six.  

If you wonder what an ERA+ of 100 buys you there were three of those:  Livan Hernandez, Aaron Sele, and Steve Trachsel.  Some names that were under 100:

Joel Pineiro     97
Matt Clement     96
Jason Marquis    96
Kip Wells        96
Kyle Lohse       95
Jeff Weaver      93

This is a really great discussion.