Education of a pitcher

At this time last year, when Kent Bonham and I announced that we’d be collecting play-by-play data for all 300 or so programs in NCAA Division I, much of the interest in our project focused on one thing: batted ball data. Until last year, most college baseball data-collection efforts ended with standard, overall stats—after all, we’re talking about thousands of players, most of whom aren’t of much interest to fans of professional baseball. But more data is almost certainly better, especially when you’re looking at colleges.

After a year of collecting batted-ball data for Division I players, now seems like a good time to take a first crack at determining what that information tells us.

What we have to work with

Due to the limitations of the data sources, there isn’t the same breakdown of batted-ball data for college players that we’ve grown accustomed to for the pros. The play-by-play descriptions offered by college teams generally say whether outs are ground balls, line drives, fly balls, or popups, but hits generally aren’t specified.

Although the college season is relatively short (most teams top out around 50 to 60 games per year), that brevity isn’t too much of a problem when analyzing college pitchers, as I’m going to do in a moment. Those roughly 60 games are more spread out than they are in the pros, so top college pitchers can log more innings than they would in a stretch of 60 minor league games. In 2007, several college pitchers exceeded 100 innings, and Wes Roemer, a Fullerton product and D-Backs draftee, racked up more than 140 innings.

College pitchers in the pros

Any time you want to compare two leagues, it’s helpful to have a population that has played in both. We have batted-ball data for only one college campaign (namely, 2007), but most 2007 draftees at least got their feet wet in the pros last year. There’s our sample.

I probably missed somebody, but my effort to find college pitchers from 2007 who also pitched in the minors last year netted nearly 300 names. Although many of those pitchers amassed a decent-sized sample of innings in college, fewer pitched very much in the pros; only a handful topped 60 innings in the minors. (Note that my college IP totals might differ a bit from “official” totals because my numbers are extracted from play-by-play descriptions, and I don’t have play-by-play for every game.)

For this research, I included players from all minor leagues (almost all players pitched in either rookie or short-season leagues). And I limited my focus to those who pitched at least 20 innings in both college and the pros. That’s not a lot of production, and it’s certainly not ideal for this kind of study, but it still gives us a minimum of about 40 batted-ball outs at each level.

We were left with 213 pitchers who pitched in both college and the pros last year. Although complete batted-ball data (for both hits and outs) is available for minor leaguers, I took only rates that we also have for college pitchers—that is, GB/FB rates for outs only.

Comparing environments

Overall, these 213 pitchers got slightly more groundball outs in college than they did in the minors. Grounders represented about 58% of outs in the minors, compared to 61% in college. (Somewhat surprisingly, these pitchers had about the same number of total batted-ball outs at both levels; that is, their strikeout rates held steady.)

The correlation between GB/FB rates at the college and pro levels is 0.42. That’s high enough to say there’s something to it, but probably not enough that it should make a huge impact on any team’s draft board. However, a lot of the noise is due to the limited number of innings in our sample. When we look only at the 100 or so pitchers who totaled 35 or more innings both in college and in the minors, the correlation jumps to 0.56. (Any threshold higher than 35 innings wipes out many of the remaining players.)

The minor league standard

Without a reference point, those correlations aren’t very meaningful. To get one, I ran the same basic algorithm on minor league players who pitched in both 2006 and 2007. That’s a much more satisfying sample, both in the number of pitchers and in the number of innings.

At the 20 IP threshhold, there are more than 1,600 minor league pitchers who qualified in both 2006 and 2007. The correlation in GB/FB rates for that population is 0.61—quite a bit better than the corresponding number for college to the minors (0.42). Probably a little of that improvement is due to the fact that the average inning total is higher in the minor league pool, but more is due to the consistency between the environments. At the 35 IP threshhold, there is less of a difference between our minor-to-minor and college-to-minor comparisons: The correlation for 2006/2007 minor league pitchers’ groundball rates is 0.66 (vs. 0.56 for college-to-minors).

The extremes

Since some of these guys were discussed quite a bit last year at draft time, I thought it would be interesting to look at some names. Here are the 10 pitchers who had the highest college groundball rates and how they fared in the pros (these lists are limited to pitchers above the 35 IP threshhold):

First      Last         collIP  milbIP  collegeGF   milbGF
Michael    Tarsi        84      56      2.07        2.96
Jordan     Powell       54.7    45      2.15        1.13
Charles    Fick         45      37.6    2.16        2.56
Paige      Dumont       44      39.3    2.17        3.05
Anthony    Slama        49      37.3    2.18        1.35
Clayton    Mortensen    115.3   56.4    2.19        3.13
John       Mariotti     87.3    35.3    2.35        6.08  (wow!)
Erik       Crichton     53.7    36      2.41        1.28
Marc       Rzepczynski  70.7    49.3    2.58        2.71
ChristopherProvince     43.3    36      2.91        2.59

And the pitchers who got the most flyball outs in their college season:

First   Last       collIP   milbIP  collegeGF milbGF
David   Hurst      51.3     39      0.56      1.05
Evan    Reed       41.3     41.3    0.61      1.05
Bradley Meyers     105.3    38.6    0.62      1.60
Bruce   Billings   85.7     81.7    0.70      0.86
Arik    Hempy      63.7     35.3    0.71      1.00
Ryan    Falcon     101.3    46      0.80      0.39
Adrian  Alaniz     101.3    59.3    0.85      1.02
Harris  Honeycutt  93.3     38      0.85      1.06
Joshua  Johnson    63.7     62.7    0.85      1.50
Omar    Arif       52.7     48.3    0.86      1.10

Next steps

It seems to be generally accepted these days that, all else equal, being a groundball pitcher is a good thing; certainly it keeps home runs down, especially in hitter-friendly parks. In addition to using larger samples, then, it seems logical to wonder whether college pitchers with high groundball rates have more success in the pros than do similar pitchers with substantially lower groundball rates.