You might recall that last year I wrote a trio of articles (One! … Two!! … Three!!!) that examined the minor league pitching statistics of two distinct populations of major league hurlers. One group was manifestly successful at the highest level, while the other group, while not as bad as a Slim Whitman concept album, still didn’t fare too well in the bigs. While far from conclusive, the findings of the study were more confounding and counterintuitive than anything else. It provided more questions than answers, which is usually what happens when you give a former English major the keys to Excel. Anyway, let’s summarize what came out of the wash:
Group Mi IP K/BB K/9 BB/9 HR/9 H/9 ERA A 12,657 2.00 7.50 3.74 0.48 8.05 3.38 B 22,363 2.17 7.58 3.49 0.57 8.31 3.55
Group A happened to be the population comprising modern-day greats like Roger Clemens, Pedro Martinez and Randy Johnson. Group B was peopled with lesser lights like Bryan Rekar, Kevin Jarvis and Jimmy Haynes.
Specifically, Group A included all current pitchers who’d spent the majority of their careers as starters and thrown, as of the end of 2002, at last 1,000 innings in the majors, all while posting a career park-adjusted ERA+ (the pitcher’s ERA relative to the league average) of at least 110 (meaning an ERA at least 10% better than the league average). Also in the mix were a few quality young arms who had pitched at least 500 innings and maintained a park-adjusted ERA+ of at least 120.
Group B, meanwhile, included all active pitchers who had, as of the end of the 2002 season, pitched at least 500 innings and posted a park-adjusted ERA+ of 95 or less (at least 5% worse than the league average). In both instances, I attempted to isolate those minor league innings that were developmental in nature–i.e., not an injury rehab assignment or late-career retread work.
It was quite surprising to find that Group B outperformed Group A in several key measures like K/BB ratio, K/9 and BB/9. In short, they struck out hitters at a higher clip, had better control and demonstrated more command. The only thing Group A did do better, albeit modestly, was keep the ball in the park and prevent hits (and the latter probably wasn’t entirely of their own making). As such, I’ve decided to revisit this matter with an eye toward home runs and hits allowed–the two measures that favored Group A in the original study. Additionally, this time I’ll remedy an oversight in the first study and bring age into the equation.
I’ve also whittled down the sample groups this time. Group A this time will include numbers from the 2003 season (which dropped Jarrod Washburn below the line of demarcation), and I’ve also made the 1,000 inning-limit a hard and fast one. With Group B, I’ve included only those pitchers who’ve spent the majority of their careers as starters, with the 500-inning and 95 ERA+ requirements still intact. Like last time, only developmental minor league innings have been included in the study. Let’s party…
Here’s a look at the individual pitchers of Group A (Age means age in the pitcher’s rookie season):
MLB ERA+ Mi IP Mi ERA Mi K/BB Mi K/9 Mi BB/9 Mi HR/9 Mi H/9 Age
Alvarez, W. 114 659 3.32 1.8 7.4 4.1 0.6 8.1 21
Appier, K. 121 401 3.12 2.7 7.4 2.7 0.2 8.6 22
Brown, K. 131 303 4.55 1.6 5.7 3.7 0.3 9.1 24
Clemens, R. 140 133 1.49 5.2 10.2 2.0 0.3 6.2 21
Colon, B. 121 394 2.42 2.4 9.3 3.9 0.4 6.5 24
Glavine, T. 121 537 3.22 1.8 7.1 4.0 0.7 7.6 21
Hentgen, P. 111 875 3.24 1.9 7.7 4.0 0.5 7.5 23
Hudson, T. 139 268 3.22 2.1 9.0 4.3 0.5 7.2 24
Johnson, R. 143 418 3.51 1.4 9.6 7.0 0.5 6.8 25
Leiter, A. 115 692 4.15 1.3 6.9 5.5 0.3 7.9 22
Maddux, G. 143 491 2.86 2.1 5.7 2.7 0.3 7.9 21
Martinez, P. 174 518 3.01 2.4 8.6 3.6 0.4 7.1 21
Millwood, K. 114 503 3.97 1.9 8.3 4.5 0.9 7.7 22
Mussina, M. 129 178 2.43 3.9 8.2 2.1 0.6 7.6 22
Pettitte, A. 117 584 2.51 2.8 7.2 2.6 0.3 7.8 22
Radke, B. 111 570 3.22 3.3 7.0 2.1 0.4 8.3 22
Schilling, C. 129 725 3.34 2.3 7.1 3.2 0.5 8.5 23
Smoltz, J. 124 377 4.13 1.6 6.0 3.8 0.7 8.4 21
Wells, D. 110 519 3.80 2.1 7.5 3.6 0.5 8.7 25
And Group B:
MLB ERA+ Mi IP Mi ERA Mi K/BB Mi K/9 Mi BB/9 Mi HR/9 Mi H/9 Age
Anderson, J. 81 755 3.55 1.7 6.8 4.0 0.4 8.9 24
Baldwin, J. 92 756 3.37 2.2 8.6 3.9 0.8 7.8 25
Bere, J. 86 422 2.65 2.0 8.6 4.2 0.3 7.2 22
Clement, M. 95 726 3.57 2.1 9.0 4.3 0.4 7.9 24
Dempster, R. 85 484 3.85 2.4 7.8 3.3 1.0 8.7 21
Estes, S. 91 442 4.36 1.7 9.0 5.2 0.5 8.3 23
Hamilton, J. 94 327 3.44 2.0 6.2 3.2 0.3 9.5 23
Haynes, J. 85 893 3.13 2.8 8.8 3.1 0.7 8.1 23
Hitchcock, S. 91 649 2.95 3.3 9.2 2.8 0.4 7.3 23
Jarvis, K. 75 653 3.45 2.6 6.9 2.7 0.6 8.3 25
Johnson, J. 91 704 3.78 2.6 7.6 3.0 0.6 8.8 24
Lima, J. 88 777 4.14 3.1 7.4 2.4 1.0 8.8 22
Mahomes, P. 83 851 3.13 1.8 8.9 5.0 0.5 7.1 22
Meadows, B. 83 533 4.02 2.7 5.6 2.1 0.8 9.4 22
Oliver, D. 94 324 2.50 1.9 9.1 4.7 0.2 6.1 23
Parque, J. 89 120 3.38 2.0 8.4 4.3 0.8 6.5 22
Pavano, C. 93 566 3.01 3.4 8.1 2.4 0.6 7.7 22
Rekar, B. 88 583 3.91 3.0 7.1 2.4 0.8 9.3 24
Rusch, G. 86 654 3.47 3.4 7.7 2.3 0.6 8.5 22
Wilson, P. 90 331 3.24 3.6 8.8 2.5 0.4 7.7 23
Wright, Jam. 94 471 3.55 1.8 6.3 3.5 0.3 9.2 21
Wright, Jar. 84 342 2.9 1.8 8.9 4.9 0.5 6.5 21
Now let’s see how they stack up on an aggregate basis:
Group Mi IP K/BB K/9 BB/9 HR/9 H/9 ERA A 9,145 2.00 7.51 3.76 0.47 7.84 3.32 B 12,363 2.35 7.93 3.38 0.58 8.17 3.45
As you can see, Group B looks even better this time around. They’ve increased their advantage in K/BB, K/9 and BB/9. Group A, meanwhile, made modest relative gains in HR/9 and H/9. Now let’s bring in another couple of elements. First we’ll look at rookie age, and by that I mean the age at which each pitcher exhausted his major league rookie status. Also, Group A’s consistent advantage in ERA and H/9 made me wonder how they’d fare on a balls-in-play scale. That is, how do their homer rates stand up relative to balls in play (BIP), and how will the two groups fare with regard to opponents’ average on balls in play? The cumulative results:
Group Rookie Age HR-BIP% BABIP (avg. on balls in play) A 22.42 1.87 0.299 B 22.78 2.17 0.293
As you can see, only modest advantages for Group A in all three new categories. They hit the majors a few months earlier, give up about one fewer homer every 300 balls in play and actually surrender hits on balls in play at a worse rate.
Granted, the sample population I’m dealing with is quite small, but the results here, even after delving a little further, are still puzzling. Do a few months in age and a homer allowed here and there constitute the difference between success and failure for a pitcher at the highest level? Of course not. A more reasonable question might be: Do we know as much as we think we know in terms of projecting minor league pitchers? That’s what I’m left wondering. I’ll continue to chew on this in the coming months.
Thank you for reading
This is a free article. If you enjoyed it, consider subscribing to Baseball Prospectus. Subscriptions support ongoing public baseball research and analysis in an increasingly proprietary environment.
Subscribe now