Although the Yankees are pretty well situated in the late innings, their bullpen is likely to be one of their weaknesses this year. David Robertson will slide seamlessly into the closer's role, while Shawn Kelley and Matt Thornton should form a solid righty-lefty combo in the seventh and eight innings, but things get a little dicey after those front three. The team's other bullpen options -- Preston Claiborne, Adam Warren, Dellin Betances, Matt Daley, Jim Miller, Chris Leroux, Vidal Nuno, Cesar Cabral, and Fred Lewis -- all have their merits, but are almost entirely unproven at the big league level. Of course, any one of these pitchers can establish himself with a strong month or two, but for now at least, these guys are probably best suited for lower-leverage situations.
As the only established bullpen pieces, its not hard to envision Joe Girardi leaning heavily on Robertson, Kelley, and Thornton in high-leverage situations this year, using them on back-to-back days or even three days in a row -- something none of these pitchers have been asked to do in the past few years. Another factor to consider is that the Yankees aren't obligated to pay Robertson or Kelley beyond this year, giving them less reason to be concerned about any negative externalities extending beyond 2014. I wouldn't expect Girardi to work these guys to the bone a la Scott Proctor in 2006/2007, but he has little incentive not to squeeze a few extra bullets out of their arms. And the numbers suggest that would be a sound strategy.
Robertson has appeared in 205 games for the Yankees since 2011 and 53 of those appearances have come on one days rest.* Throwing on back-to-back days doesn't seem to phase D-Rob in the least. He throws just as hard and his strikeout and walk numbers don't seem to be affected either. If anything, Robertson seems to perform slightly better when coming back the next day.
*I use "one days rest" to describe pitching on back-to-back days, which is sometimes referred to as pitching on "zero days rest."
With 114 relief appearances over the last three years, Kelley doesn't have quite the track record of Robertson. Still, Kelley looks to be at his very best when pitching on back-to-back days. His average SIERA is nearly two runs lower and his strikeout and walk numbers are significantly better as well. Kelley's only pitched on back-to-back days 19 times, but even in such a small sample of games, a split like that is hard to ignore.
Like Kelley, Thornton is also at his best when pitching on back-to-back days, which he's done 45 times in the past three seasons. The lefty's SIERA drops by nearly a point and a half and his strikeout to walk ratio nearly triples. Thornton's split is particularly believable because it looks to be backed up by improved stuff. For whatever reason, Thornton's slider seems to be a little sharper when he's pitched the day before, dropping an inch or two further than when he's well rested. This added movement might explain why Thornton's ground ball rate jumps by over 10% when he pitches on back-to-back days.
Keep in mind that looking at reliever splits is always a risky proposition. Relief pitchers throw very few innings as it is, so sample sizes get dicey once you start slicing those innings up. Nonetheless, I think there's definitely something here. Robertson's split is probably negligible, but the splits for Kelley and Thornton are pretty sizable, and I think they may represent something more than just random noise.
At the very least, its safe to say that none of these three perform any worse when pitching on back-to-back days. So until someone else emerges a viable late-inning option, I'd hope to see Girardi signal for his top dogs just as often as is necessary, regardless of who pitched the night before. All three have proven they're more than capable of getting it done on little rest. When the game's on the line, the Yankees would probably be better served using their best pitchers for the second or third straight night than turning it over to the likes of Preston Claiborne or Cesar Cabral.