October 19, 2010
Checking the Numbers
I'm Relieved to Start
A couple of weeks back, I wrote an article called ‘The Turnaround Kids’ which identified the players with the most substantial improvements from 2009 in TAv, WARP, SNWP, and WXRL. Everyone loves a good turnaround, and it was interesting to see how Aubrey Huff’s 6.1 WARP was not only high, but representative of a vast improvement from a year in which he fell below replacement level. Similarly, Brad Lidge might not have put together the greatest season in 2010, but he was solid, and his numbers look masterful when compared to the dreck he put on display a year ago. In all likelihood, Josh Hamilton experienced the greatest turnaround, improving from a disappointing and injury-plagued campaign that found him at the replacement level into the odds-on favorite to win the American League’s Most Valuable Player award.
In discussing the turnarounds, some comments mentioned other players who did not appear on the list, but who maybe should have earned a spot. Brett Gardner was one, though his exclusion stemmed from my plate appearances filter; there needed to be some restrictive measure in place to ensure that the turnaround was more than the byproduct of a change in playing time. Colby Lewis was mentioned as another possible candidate, though he did not play in the major leagues last season. A comparison for his 2010 numbers would involve statistics produced in small samples prior to his trip to Japan, or some hypothetical construct. Neither would be an appropriate comparative baseline.
The other name that surfaced was C.J. Wilson, who was the Rangers’ best starter this season. Cliff Lee is certainly more talented, but he didn’t arrive in Arlington until the final months of the season, and Wilson was consistently effective for the entire season. SIERA might be bearish on his ability to remain as successful moving forward, but he still put up great numbers, producing 4.9 WARP in 33 starts and 204 innings. A year ago he logged 73
Wilson’s improvement is interesting, because it wasn’t as if he went from bad starter to good starter—he went from good reliever to very good starter. In other words, Wilson didn’t turn anything around as much as he flourished in a newer and tougher role. Today my goal is to find the pitchers who made the Wilson Transition successfully, and for those schadenfreude advocates out there, the pitchers who failed in their attempts. This is different from evaluating swingmen, who make the transition in the same season, as Wilson changed roles from one year to the next.
The first step is to actually define starters and relievers. This is not as cut-and-dried as labeling a starter as a pitcher with zero non-starts, or a reliever as a pitcher with zero starts. Roy Oswalt pitched in relief on the final day of the season, so a query as restrictive as zero non-starts equals a starter would exclude him. Similarly, if a player makes five or six relief appearances and then settles into the starter’s role, proceeding to make 20 or more starts, he should qualify as a starter for this exercise.
With that in mind, I defined relievers as making no more than five starts in a season while pitching more than 40 innings. I defined starters as making 20 or more starts in a season with a difference of no greater than 10 between games pitched and games started. From there, I found which relievers meeting this criteria went on to fulfill the starting requirements in the very next season, and sorted by SNWP, a very good measure for evaluating starting pitchers. Here are the top transitions:
Derek Lowe very clearly produced the best pitched season for those making the transition out of a group consisting merely of 55 pitchers since 1960. On average, the relievers produced 0.87 WXRL the year before becoming a starter, and a .523 SNWP while in the rotation the following year. Neither of those numbers necessarily screams success, but they provide some value nonetheless. In that 2002 campaign, Lowe put up a 2.58 ERA in 219
In the cases of R.A. Dickey and Shawn Chacon, both pitchers performed poorly out of the bullpen and were thrust into the starting rotation by virtue of necessity the following season. Chacon produced a 7.11 ERA as the Rockies closer in 2004, putting together what would go down as, at the time, the worst season ever for a reliever given as many save opportunities. The aforementioned Lidge would make that season look like an Dennis Eckersley special when he set the all-time low in WXRL last season. In 2005, Chacon split time between the Rockies and Yankees and had one of those my-ERA-is-low-but-peripherals-stink seasons, with a 3.44 ERA but ugly HR, BB, and SO rates. Dickey was still fine-tuning his knuckler in the Twins bullpen last year and turned into the Mets’ top starter this past season. Could he be the new Tim Wakefield?
This top-10 list also consists of Ryan Dempster and John Smoltz, who were starters that successfully became relievers before going back into the rotation. As is clear from Smoltz’s WXRL, he was dominant no matter where he went. What about the opposite list? Which 10 pitchers made the transition and produced the worst SNWPs?
For the most part, even these bottom feeders performed relatively well. A .467 SNWP isn’t great by any means, but a team wouldn’t see their hopes of contention drastically diminished if someone with that mark occupied the fifth spot in a rotation. Several of these players went back to the bullpen after their attempt at starting didn’t work out, including Scott Schoeneweis, Paul Quantrill, and Danny Graves.
Wilson produced an SNWP of .560 this season, which is very solid, but not spectacular relative to the pitchers in the list above. One would think, however, that his success will be longer lived than that of Chacon or Matt Morris. Have any other pitchers made noteworthy shifts from the bullpen to the rotation?