February 25, 2011
Prospectus Hit and Run
Are You Experienced?
With Cliff Lee's surprising detour to Philadelphia and Andy Pettitte's unsurprising detour into retirement, the Yankees' major-league rotation is in a decidedly unfinished state. Happily, their farm system is chock full of quality pitching prospects. On the surface, the solution—fill the rotation's two open spots from within—appears obvious, but the Yankees are auditioning veteran retreads such as Freddy Garcia, Sergio Mitre, and Bartolo Colon for first dibs on those spots. Not long after pitchers and catchers reported to camp, general manager Brian Cashman dismissed the idea of either of the team's blue chippers on the mound, Manuel Banuelos and Dellin Betances, breaking camp with the big club, stating to the New York Daily News' Mark Feinsand that "(t)hey’re going to get their first taste of big-league camp, then they're going to get slotted into Trenton. They have no chance to make this team."
Cashman's words led Steven Goldman to wonder aloud about the dangers of rushing young pitchers to the majors. Specifically, he looked at a rather mixed bag of pitchers who debuted at age 21 or under since 1985, early success stories such as Zack Greinke, Dwight Gooden, Felix Hernandez, Clayton Kershaw, and CC Sabathia, as well as pitchers whose time in the limelight was more short-lived, among them Ed Correa, Hayden Penn, and Brent Knackert, guys for the Where Are They Now files.
Missing from Steve's analysis was any acknowledgement of the amount of minor-league experience those pitchers had before reaching the majors, particularly in the upper levels of the minors, where the higher caliber of play is more indicative of players who have made progress in converting their tools into skills, and from where pitchers most commonly jump to the majors. Gooden may have made the leap directly from A-ball to the Mets, but Hernandez tossed 57 1/3 innings at Double-A in the season before his debut, and another 88 at Triple-A before his late-season recall by the Mariners. Sabathia had 90 1/3 Double-A innings under his ample belt before breaking camp with the Indians the following spring. That's an especially salient issue when it comes to the Yankees' tyros: Banuelos, who turns 20 next month, has just 15 1/3 innings at Double-A to date, while the going-on-23-year-old Dellin Betances, whose progress up the ladder was forestalled by Tommy John surgery in 2009, has just 14 1/3 innings at Double-A.
Does the amount of minor-league experience really matter? With the data wrangling help of Colin Wyers, I decided to investigate if the amount of upper-level experience provided any clues as to a pitcher's future success, both in the short term or the longer term. I gathered data going all the way back to 1980 to allow myself the ability to examine whether such clues had changed over time.
To start with, 2,632 pitchers made their major-league debuts from 1980 to 2010. Using the standard July 1 age cutoff, those pitchers ranged from 19 (Gooden, Correia, Rick Ankiel, Jose Rijo, and seven others) to 39 (Japanese import Masumi Kuwata), with a mean age of 24. The pitchers posted a collective ERA of 4.65 in their debut seasons; during that time, the major-league average was 4.20. Over the course of their careers, those pitchers' collective ERA would come down to 4.28, while throwing an average of 340 innings and compiling an average of 4.4 WARP (some of these pitchers are of course still active). For their debut seasons, the average number of innings thrown was 33, the average WARP accumulated was 0.3.
Off the bat, while the correlation between a pitcher's debut season ERA and his career ERA was a solid 0.6, there was virtually no correlation between the number of upper-level innings (Double-A, Triple-A, or both) with debut-season ERA or career ERA. The maximum correlation for any of those inning combinations and debut ERA was with Triple-A innings, and that at just 0.011. For career ERA, the maximum correlation was with combined upper level innings, at 0.038. Correlations between those upper level innings and debut season or career WARP were similarly infinitesimal. Over that broad range of ages and debut years, there just aren't any useful generalizations to be made about the necessity of upper level experience.
Since whatever revelations there are to be gleaned about this broad set of pitchers are mostly irrelevant to the Yankees' situation, I pared the set to include only pitchers who debuted at ages 23 or younger, leaving a set of 1,037 pitchers. Even that doesn't yield much more clarity. The average number of innings thrown in these youngsters' debut seasons was 41, the average WARP accumulated remained 0.3. This group compiled a 4.67 ERA in their debut season, with a collective career ERA of 4.26. Again, none of the correlations between upper-level innings and debut-year performance or career performance (as measured by ERA or WARP) reached even 0.1. Bupkis.
Thinking that the game's changes over this 31-season span might be concealing more revealing patterns, I split the group by decades: 1980-89 (276 pitchers), 1990-99 (331 pitchers), and 2000-10 (430 pitchers). The Eighties group (0.12) and the Aughties group (0.14) both reached correlations between upper-level innings and career ERA above 0.1, but that was undone by the Nineties group's -0.02, and even if it hadn't been, that's still nothing to write home about. The most interesting finding from this arrangement is that the number of upper level innings thrown by pitchers through their debut seasons appears to be decreasing over time:
The average pitcher debuting in the past 11 seasons has thrown 51 fewer innings in Double-A and Triple-A than the ones two decades ago. This may have something to do with a higher percentage of players coming from the college ranks than before. As for the career data, take it with a grain of salt, since a significant percentage of the Nineties and Aughties groups are still active. I suspect the fact that the latter's lower-than-average ERAs will probably be undone over time given the way pitchers generally stagger towards entropy.
Looking at the matter from a slightly different perspective, I backtracked and split the group of 23-and-under rookies into four more or less equally-sized subgroups based upon the number of upper level innings each had compiled (pitchers with the same innings totals were not split into different groups). While no clear pattern emerged, there were slight differences between the groups in terms of debut season performance, and larger differences in terms of career performances:
The group that reached the majors with the least upper-level experience was the most valuable according to debut season WARP (they edge the second and third groups 0.40 to 0.36, but let's keep our significant digits tidy), and the second most effective according to debut-season ERA. They tied for the lowest career ERA of the bunch, but averaged by far the fewest innings over the course of their careers, a pattern we might expect to see if these pitchers were rushed to the majors but were more susceptible to injuries or attrition once they got there. At the other end of the spectrum, the group that reached the majors with the most upper-level experience was the least effective in terms of both debut season and career performance, though they did stick around longer than the "rushed" group. This suggests a lower-ceiling bunch who were perhaps a bit more polished, or who may have been winnowed by injuries prior to reaching the majors.
Imposing such narratives on groups is a risky business, however. Drawing from the decade-based table, I discovered that the least experienced group prior to reaching the majors was also the most recent in terms of average year of debut, suggesting that a larger percentage of those pitchers may be active. Indeed, 24 percent of that group debuted in 2006 or later, compared to 19 percent for the two middle groups, and 10 percent for the most experienced group.
I trimmed the set of pitchers further to focus solely on the 21-and-under group, a cadre of just 258 pitchers. I did find that teams are doing a better job of getting value out of today's young 'uns, despite even less time in the upper minors:
At least in terms of their debut season, the Aughties group of 21-and-under pitchers has been more than twice as valuable as their predecessors, providing ERAs much closer to the major-league average. It's too early to tell what the long-term ramifications are, but this group has thus far performed particularly well in terms of ERAs relative to the league, a quarter of a run per nine lower than average. Whether that's due to better conditioning, stronger adherence to pitch counts, more college-level experience—I make no claims as to the cause of this surge, but it does appear as though the freshest young pitchers are faring better in recent years than they have in the past.
As to what any of that means to the Yankees with regards to their Killer B's, there's less to take from this quick-and-dirty exercise than hoped. Upper-level minor-league experience as determined by innings isn't a particularly helpful yardstick in projecting whether a pitcher will succeed either in the short term or the long term. Which shouldn't be too surprising given the complexity of pitcher development, and the importance of marrying scouting information to the stats. Nor should it be terribly surprising that the Yankees as well as most outside prospect experts feel that Banuelos and Betances could use more seasoning. That may mean more Garcia and Colon in the short term, but it doesn't preclude the possibility of the young whippersnappers reaching the Bronx later in the year.