In last Friday’s Prospectus Hit and Run, I took a look at the Yankees’ “decision” to shift Joba Chamberlain to the bullpen to start the season, making some ballpark estimates as to his workloads in the rotation and in relief and how they fit in with the team’s plans:

So, the recent Yanks [2005-2007] have averaged about 123 starts per year from their top five, leaving 39 starts to be absorbed by the rest of the staff. Even if they were to play towards the high side of that time span, 136 starts, those extra 26 are more than one extra pitcher can be expected to absorb. Of that surplus, how many can we expect Chamberlain to make?

An aggressive estimate might put him at about 15 starts, leaving the Yanks to come up with another dozen or two from among the Jeff Karstens/Darrell Rasner/Kei Igawa/TBNL herd. Suppose Chamberlain pitches on a 180-inning pace as a starter for three months of the year, and takes the guise of a heavy-use 90-inning reliever for the other three months. In his starter phase, he could total 15 starts averaging six innings per start (90 innings) and as a reliever, he’d average an inning every other day (45 innings). That would put Joba at 135 innings, not far off from his Rule of 30 target. Of course, that wouldn’t leave a ton of headroom for October, if the Yankees get there. One can’t determine whether the team might want him to be part of their front four, or to return to the set-up role without knowing the rest of the staff’s strengths and weaknesses at that point in time. Either way has its merits so long as he can usurp a greater percentage of innings than the regular season, but both have their drawbacks as well.

From there I took a look at Chamberlain’s performance in relief last year using BP’s Reliever Expected Wins Added stat (WXRL), which measures the incremental impact, in wins, of a reliever’s performance based on the changes in score margin and game state (runners on base and number of outs) when he arrives and departs. Joba was 1.85 wins above replacement level in his 24 innings last year; on a per-inning basis, only Cleveland’s Rafael Betancourt had more of an impact, and all Betancourt did was set a record among the subgroup of nonclosers that I’m calling MISERs (Maximum-Impact Setup Relievers, crudely defined as pitchers with at least 1.0 WXRL and fewer than six saves in a season). The leaderboard for 2007:

Pitcher             Team    IP     WXRL    WX/9
Rafael Betancourt    CLE   79.1   6.845   0.777
Joba Chamberlain     NYA   24.0   1.848   0.693
Hideki Okajima       BOS   69.0   4.429   0.578
Mike Gonzalez        ATL   17.0   1.079   0.571
Heath Bell           SDN   93.2   5.656   0.543
Brandon Lyon         ARI   74.0   4.357   0.530
Zack Greinke         KCA   53.1   3.114   0.525
Carlos Marmol        CHN   69.1   3.694   0.480
Rafael Perez         CLE   60.2   3.142   0.466
Pat Neshek           MIN   70.1   3.572   0.457
J.C. Romero          2TM   56.1   2.822   0.451
Bobby Seay           DET   46.1   2.303   0.448
Akinori Otsuka       TEX   32.1   1.566   0.436
Tony Pena            ARI   85.1   4.103   0.433
Ehren Wassermann     CHA   23.0   1.075   0.421
Henry Owens          FLO   23.0   1.051   0.411
Chad Qualls          HOU   82.2   3.608   0.393
Derrick Turnbow      MIL   68.0   2.752   0.364
Matt Herges          COL   48.2   1.927   0.356
Justin Speier        ANA   50.0   1.950   0.351
Scott Downs          TOR   58.0   2.260   0.351

By comparison, Mariano Rivera’s WXRL/9 was 0.467, which ranked 20th in the majors. Still, as good as Chamberlain’s performance was, 24 innings is a pretty small sample size to base many conclusions on, so when I retroactively computed leaderboards for the seasons 2001 to 2006, I raised the bar to 40 innings to allow the occasional LOOGY or midseason callup to sneak through. What the rankings show is that the Yanks have been fairly deficient in this department recently. Here are the qualifying Yankees dating back to 1996:

Pitcher           Team   Year    IP     WXRL    WX/9   Rk
Mariano Rivera     NYA   1996  107.7   6.876   0.575    1
Mike Stanton       NYA   1997   66.7   3.541   0.478    2
Ramiro Mendoza     NYA   1999   84.0   2.385   0.256   27
Mike Stanton       NYA   1999   58.3   1.480   0.228   34
Jason Grimsley     NYA   1999   75.0   1.608   0.193   41
Jeff Nelson        NYA   2000   69.7   1.534   0.198   35
Mike Stanton       NYA   2001   80.3   3.333   0.373   10
Jay Witasick       NYA   2001   40.3   1.037   0.231   32
Ramiro Mendoza     NYA   2002   91.7   1.621   0.159   44
Chris Hammond      NYA   2003   63.0   1.954   0.279   21
Tom Gordon         NYA   2004   89.7   6.438   0.646    1
Paul Quantrill     NYA   2004   95.3   1.546   0.146   51
Tom Gordon         NYA   2005   80.7   3.269   0.365   11
Scott Proctor      NYA   2006  102.3   1.833   0.161   49
Luis Vizcaino      NYA   2007   75.3   2.081   0.249   35
Kyle Farnsworth    NYA   2007   60.0   1.142   0.171   51

Since Mo ascended to the closer role, only Mike Stanton and Tom Gordon have cracked the majors’ top 20 in a given season. For a team spending as much money on its bullpen as the Yankees have, that’s no relief. And it’s yet another reason why sending Chamberlain to the bullpen for at least part of the year makes sense.