BBTF's Newsblog Discussion :: wezen-ball: Top Talent-Producing Universities

Baseball Primer Newsblog
— The Best News Links from the Baseball Newsstand

Saturday, December 20, 2008

wezen-ball: Top Talent-Producing Universities

Feverishly scans for Jersey City State (AKA Snyder University) for totals…

While doing just that, I decided to look at what colleges have produced the best players. It’s easy to look at some lists and see that Arizona St. has produced the likes Barry Bonds and Reggie Jackson, or that USC has produced Tom Seaver, Mark McGwire, and Randy Johnson, but that seems a little too subjective to me. I decided to try something a little more objective to see what we get.

Using the database info available at Baseball Databank and the Win Shares info found on the Baseball Graphs Blog (limited through 2007 for now), I calculated how many players have come from each school and then summed up the total Win Shares of each of those players. In order to rank the schools in a meaningful way, I found the average Win Shares of each school’s alumni. So, for example, there were 26 MLB players who attended the University of Wisconsin - Madison, and those 26 players accumulated a total of 933 Win Shares throughout their career (with Harvey Kuenn earning the most with 223 WS). That gives UW an average Win Share value of 35.9.

Repoz Posted: December 20, 2008 at 03:20 PM | 17 comment(s) Login to Bookmark
Tags: history, sabermetrics

Reader Comments and Retorts

Go to end of page

Statements posted here are those of our readers and do not represent the BaseballThinkFactory. Names are provided by the poster and are not verified. We ask that posters follow our submission policy. Please report any inappropriate comments.

1. Misirlou's got a busy day, he's wearing a vest Posted: December 20, 2008 at 04:26 PM (#3034641)

Something ain't right here:

School # Major Leaguers Avg WS Notable Alumni
Columbia University 12 66.6 Lou Gehrig (489 WS), Eddie Collins (574 WS)

Add up Gehrig and Collins, divide by 12 and you get 88.6. Are there negative win shares now?

2. Scott Lange Posted: December 20, 2008 at 04:45 PM (#3034644)

Average win shares isn't a good way to rank the schools- it penalizes a school for turning out a bunch of average major league baseball players. That's how Ohio University (Mike Schmidt and 16 scrubs) beats USC (Tom Seaver, Fred Lynn, Mark McGwire, Randy Johnson, Spaceman Lee, and 80+ others). Its probably better to just add thm up and get a total. Of course, this critique is motivated largely by the fact that my alma mater didn't make either list- Georgia Tech (Nomar, Teixeira, Del Pratt, Varitek, Kevin Brown, and 35 others).

3. Misirlou's got a busy day, he's wearing a vest Posted: December 20, 2008 at 05:02 PM (#3034650)

To address my post above, of Columbia's 12 players, 6 played fewer than 15 career games, another played 47. The remaining 3 were probably below average but above replacement. Columbia gets above the 10 player threshold by virtue of 3 players with a combined 7 career games.

So yeah, average is a lousy way to assess them as well. Even worse than Ohio.

4. aleskel

Posted: December 20, 2008 at 05:06 PM (#3034652)

The remaining 3 were probably below average but above replacement

I'll thank you to not speak of Gene Larken that way!

5. Teheran's Uranium Enriched Missiles Posted: December 20, 2008 at 05:08 PM (#3034653)

my alma mater didn't make either list- Georgia Tech (Nomar, Teixeira, Del Pratt, Varitek, Kevin Brown, and 35 others).

Patience. All these insults and sufferings will be noted, and Wieters will punish these sinners all. Patience!

Weird to see the USC, and see no Mark Prior on it.

6. Jolly Old St. Neck Wound, Moral Idiot

Posted: December 20, 2008 at 05:19 PM (#3034659)

Okay, I just learned that unlike with several other states, "Cal Poly" and "Cal Tech" aren't the same thing, and that Ozzie Smith and Mike Krukow didn't attend Cal Tech. That kinda had me wondering there for a minute.

7. bfan Posted: December 20, 2008 at 05:45 PM (#3034670)

Isn't it clear that he did it this way to produce the optimum result for his college? Total WS would be interesting and much more informative.

8. Sandlapper Spike Posted: December 20, 2008 at 07:12 PM (#3034718)

I wonder what he did with guys who played for more than one school.

Also, I'm guessing that Wisconsin's player win share average will not be changing any time in the near future...

9. Walt Davis Posted: December 20, 2008 at 08:13 PM (#3034734)

Go Fighting Quakers!!

10. lar @ wezen-ball Posted: December 20, 2008 at 10:11 PM (#3034772)

@#1: thanks for pointing that out. It's been corrected. Columbia's average is actually 102.2 WS. I originally transcribed the numbers incorrectly (I was having troubles formatting the table, and I screwed something up there).

@#2: Scott, those thoughts crossed my mind too, and I think I'm going to revisit the numbers to see what I find. I never really had an agenda when I started looking at this data, so I only went through a pretty simple methodology. I think two things that can be done to fix this are: use a floor of how many Win Shares a player must earn before being counted in the average, and/or adjusting the average Win Shares value by some factor to account for a larger number of players (like the USC vs Ohio Univ. comparison you brought up). Any thoughts on that?

I'm not sure I agree Total Win Shares is the best metric because that can also misrepresent a school (eg, having 1 HOFer and 50 replacement-level players probably isn't a better school than one that has 5 HOFer and 10 All-Stars - not that there's necessarily any two schools like that).

I'm open to any suggestions...

11. Scott Lange Posted: December 20, 2008 at 10:37 PM (#3034780)

having 1 HOFer and 50 replacement-level players probably isn't a better school than one that has 5 HOFer and 10 All-Stars

I don't think adding up total winshares would indicate that it was. Just for a quick and dirty look, consider USC. If you take out Johnson, Lynn, McGwire, and Seaver, the average of the other 87 players is 36 win shares. That means it takes ten of those guys to equal one HOF-caliber player, and those guys aren't replacement level- they include lots of middle-of-the-road players that pull up the average. 5 HoF by themselves would beat 50 replacement level guys in total win shares, even before you throw in the 10 All-Stars.

12. Bitter Calculus Instructor Posted: December 20, 2008 at 10:42 PM (#3034782)

[6] As a Cal Poly student, I found that quite amusing.

[10] The thing is though that tons of replacement level players don't produce all that many win shares. And there should be a reward for getting guys to the majors. I think total win shares is pretty good actually.

13. Misirlou's got a busy day, he's wearing a vest Posted: December 20, 2008 at 11:03 PM (#3034791)

I'm not sure I agree Total Win Shares is the best metric because that can also misrepresent a school (eg, having 1 HOFer and 50 replacement-level players probably isn't a better school than one that has 5 HOFer and 10 All-Stars - not that there's necessarily any two schools like that).

yeah, but having 2 inner circle guys, 3 barely adequate regulars, and 7 guys with no real career shouldn't put you in a commanding lead.

I mean, go back to USC. The second tier guys from there like Brett Boone, Don Buford, Jeff Cirrilo, Roy Smalley, and Barry Zito absolutely destroy anyone from Columbia not named Gehrig or Collins. Brett Boone played more games than all 10 other Columbia players combined.

It would be a ##### to calculate, but something on the order of (the #1 players WS) + 2(#2 player's WS) + 3(...)/sum of the multipliers would be better. That would reward both quality and quantity. Kind of like James did in the baseball families section of the latest historical Abstract. That way, the 1 HOFer + 50 replacements wouldn't get anywhere near the top.

14. Teheran's Uranium Enriched Missiles Posted: December 20, 2008 at 11:16 PM (#3034795)

Or just take the average of the top 10% to indicate quality, and total to indicate quantity.

15. Misirlou's got a busy day, he's wearing a vest Posted: December 20, 2008 at 11:28 PM (#3034800)

Or just take the average of the top 10% to indicate quality

Well, then Columbia's average jumps to 574.

And my equation in 13 is obviously not right. Maybe divide by total number of players? Don't average?

16. RMc is the loyal supporter of the MLB event Posted: December 21, 2008 at 01:06 PM (#3035153)

I seem to remember James saying something about "rent" of a roster spot; if you're, say, the starting SS, you get charged a rent of 10 WS (or whatever) because anything less and you're hurting the team. Subtract each player's rent from their career WS, and you'd eliminate the scrubs. Or something.

17. lar @ wezen-ball Posted: December 22, 2008 at 03:07 AM (#3035523)

I took another look at the numbers, this time trying to account for quantity and quality. The list ends up looking a lot like the lists ranked by Total Win Shares and Total MLB Alumni, but there are some differences...

http://wezen-ball.blogspot.com/2008/12/top-talent-producing-universities_21.html

You must be Registered and Logged In to post comments.

<< Back to main

Saturday, December 20, 2008

wezen-ball: Top Talent-Producing Universities

Reader Comments and Retorts

Support BBTF

Bookmarks

Hot Topics