| Boyd's World-> Breadcrumbs Back to Omaha-> Defense-Independent Pitching Statistics in College | About the author, Boyd Nation |
Publication Date: December 3, 2002
Quiet Revolutions Are the Best
Sometime around 1998, a man named Voros McCracken, a regular over on the newsgroup rec.sport.baseball, started a simple little study, and in the process performed one of the better pieces of statistical science that I've ever seen. You can find a lot of the history by going to Google groups and searching for his name in rec.sport.baseball in the 1999-2000 time frame, but here's the gist of it. McCracken went into the study with some preconceptions but an open mind, and that turned out to make all the difference.
The preconception was that he thought, for the most part, that different pitchers had different abilities to affect how the ball was hit off of them -- that some pitchers tended to give up screaming liners, while other tended to give up lots of wimpy little ground balls. What he found was that, if you look at what happens to a pitcher from year to year, there's essentially no predictability in the percentage of balls in play that go for hits. In other words, a pitcher has great influence in home runs, strikeouts, and walks, but the rest of what happens is almost entirely determined by the defense behind him and by luck. The best place to start reading on the current state of the research is in this Baseball Primer article. This result contradicted not only his preconceptions but a great deal of conventional wisdom, and I've been amazed at how fast the stathead community has been won over to it; a lot of the credit for that goes to the rigor with which he's designed his studies.
One of the best uses of this research has come in trying to predict future behavior for pitchers. Pitchers are notoriously unreliable, but it looks like you can at least get a pretty good idea what they'll do on the three components included in DIPS (Defense-Independent Pitching Statistics) -- homers, strikeouts, and walks. I thought I'd take a look at some of the numbers from the 2002 NCAA season to see if there are any pitchers likely to be a surprise one way or the other.
There have been several different versions of single DIPS stats that he's produced so far, and, frankly, none of them have stuck around long enough to be considered canonical, so, based on the data I have available, I'm going with the following early version. It's based on some regression based on pro offensive stats, so it probably underestimates a bit for college, but it's certainly close enough for comparison between players. I'm not park-adjusting or competition-adjusting these in any way, so take them with the same grain of salt you'd take raw ERA with.
BFP * .83 + HR * 11.04 + BB * 1.98 - SO * 1.62
DERA = ----------------------------------------------
(BFP - HR - BB) * .244 + SO * .089
where BFP is estimated as
BFP = (IP * 3 - SO) * .966 + H + BB + SO
An interesting trend in doing this site over the last year or so has been that I seem to be hearing a lot more from the people within the game; I've probably heard from a few dozen players, coaches, and parents in that time. One side effect of this is that I really feel like censoring myself some times; I'm human, and I don't want to hurt anyone's feelings any more than anyone else. If I do that, though, I make the site more or less worthless; objective analysis is what I do best. In short, then, if someone like Tim Stauffer or his mom is reading this time, I'm sorry, but he did benefit a lot from great defense behind him and quite a bit of luck. That doesn't discount what he did, of course, but it does make him less likely to repeat it.
The Leaderboards
I considered all pitchers for whom I had statistics (around 265 schools worth) who had at least 10 starts last year. Here's the top 10 in DERA:
Team Pitcher DERA ERA Harvard Ben Crockett 2.70 2.79 Baylor Jared Theodorakos 2.77 5.18 San Jose State Matt Durkin 2.80 2.75 Old Dominion Justin Verlander 2.80 1.90 UC Riverside Chris Smith 2.82 2.91 Virginia Commonwealth Sean Marshall 2.84 4.45 Princeton Ross Ohlendorf 2.84 3.08 Davidson Tim Frend 2.85 3.56 Arkansas-Pine Bluff David Bayless 2.86 3.14 College of Charleston Matt Rackers 2.87 3.44
Some of these guys were almost exactly the same on both numbers, but some of them pitched considerably better than their raw ERA showed -- in particular, it wouldn't surprise me for Theodorakos to break out big this year if he gets some decent defense behind him, unless his excessive workload from 2002 kicks in. It's hard to pitch worse than your ERA shows and make this list, but Verlander managed -- he had a great year, but it wasn't 1.90 great. For comparison, here's the raw ERA leader list:
Team Pitcher DERA ERA Richmond Tim Stauffer 2.93 1.54 Wichita State John Tetuan 3.23 1.72 Northeastern Devin Monds 2.99 1.80 Houston Brad Sullivan 3.86 1.82 Old Dominion Justin Verlander 2.80 1.90 Rice Justin Crowder 3.00 2.00 Wisconsin-Milwaukee Geoff Lefeber 3.14 2.20 Oral Roberts Rene Recio 3.81 2.23 Maryland-Baltimore County Joe Wilson 3.23 2.26 Butler John Corcoran 3.30 2.29
Next, here are the unluckiest pitchers in the country -- those whose DERA was the most under their actual ERA. Looking at the team column doesn't really show you any teams that you think of when you think of great defense.
Team Pitcher DERA ERA Coppin State Stan Alston 4.88 14.22 Siena Brent Wengert 4.69 13.76 Alabama-Birmingham Andre Parks 3.62 10.18 Southern Illinois Bill Clayton 3.93 10.07 Lipscomb Matt Edwards 3.85 9.50 LaSalle Ruhland, Eric 4.82 10.28 Centenary Ande Collup 4.32 9.48 Bowling Green State Doug Flere 3.90 9.00 Marshall Rick Suter 4.40 9.48 Akron Frank Mendoza 2.94 7.78
This would probably be an eye-opening list if I were a scout, though. The one of these guys that I actually saw last year, Andre Parks, provides somewhat anecdotal evidence that there may be a hidden lode of talent lurking here on some really bad teams -- he was a fairly talented looking pitcher with some really atrocious defense behind him and a good bit of bad luck on the night I saw him.
Finally, the luckiest guys in the country. Notice that none of these DERA's are actually bad -- they're all well under the national average. Basically, these are all good pitchers who also benefitted from good defense and a bit of placement luck. It would appear that Houston may have been an extremely good defensive team last year.
Team Pitcher DERA ERA Houston Brad Sullivan 3.86 1.82 Oral Roberts Rene Recio 3.81 2.23 Wichita State John Tetuan 3.23 1.72 Richmond Tim Stauffer 2.93 1.54 Northeastern Devin Monds 2.99 1.80 Texas Justin Simmons 3.61 2.52 Illinois-Chicago Scott Anderson 3.68 2.60 Texas-Arlington Aaron Pullin 3.50 2.48 Butler John Corcoran 3.30 2.29 Houston Danny Zell 3.86 2.86
Just for completeness, I'll close with a link to a complete list of pitchers in the study so you can see how your favorite guy should expect to do this next year.
If you're interested in reprinting this or any other Boyd's World material for your publication or Web site, please read the reprint policy and contact me
| Boyd's World-> Breadcrumbs Back to Omaha-> Defense-Independent Pitching Statistics in College | About the author, Boyd Nation |