J.P. Howell: The Man, the Myth, the...Starter?

During Eric's interview with Carson Cistulli on Thursday, Carson brought up a point that I was pondering myself about a month ago: how good of a starter would JP Howell be if given the chance again?  Here's what he had to say on the matter:

"The real question, for me, is how this affects J.P. Howell. I've been the giantest Howell apologist for some time now. It seems like he could make a fantastic starter. From 2005 to 2007, Howell made 33 appearances (all starts) and posted xFIPs of 4.41, 4.08, and 4.12, respectively. His tRA*s (that is, regressed tRA, available at StatCorner) are less friendly over that time period (5.16, 5.21, and 5.24, respectively), so maybe I should just can it. Still, if he's down with it -- and I don't know, maybe he's not -- then I think it's worth a shot."

We all know that starters are much valuable than relievers, with even the best relievers only racking up around 2+ WAR over the course of a season (there is an argument to be made that WAR undervalues high-leverage innings, but the point still holds).  This is why players like Joba Chamberlain and David Price are transitioned back into the starting rotation, even after proving to be dominant coming out of the bullpen; they're simply too valuable to not have them in the rotation.  JP Howell has been dominant out of the pen - no doubt about that - but there hasn't been any sort of murmurings to move him back into the starting rotation.  How come?  Should he remain in the pen, or are we being silly to not consider him for the starting rotation?

For those that forget, JP Howell was initially a starter when first acquired by the Rays in exchange for Joey Gathright.  He started 15 games over the course of the 2006 and 2007 seasons with the Rays, compiled ERAs of 5.10 and 7.59 respectively, and was moved to the pen at the start of the 2008 season as the long man.  Since then, well...I don't need to tell you how he's done since then.  Here's a look at all his important stats (remember, the numbers from 2005-2007 are from the rotation, and 2008-2009 are in the bullpen):

Despite the bloated ERAs, it looks like Howell may have been a fairly talented starter.  His K/BB ratio is very similar to Matt Garza's K/BB from this past year (and you notice, didn't change any upon moving to the bullpen), and his BABIP and LOB% suggest that he was the victim of bad luck.  His xFIPs (a predictive statistic) also seem to imply that he's a serviceable major league starter; however, tRA begs to differ, probably because tRA accounts for batted ball data as well:

The first thing that jumps out at me (and I mean jumps out at me) is the steep drop-off in line-drive percentage in 2008.  Howell gave up a pretty high percentage of line-drives as a starter, including a ridiculously high 25.5% in 2006.  Upon moving into the bullpen, though, Howell started giving up significantly less line-drives and creating more groundballs, both of which are huge helps.  Looking at all of these numbers, we can come to one of two conclusions: either we believe that Howell was an effective starter but merely an unlucky one, or we believe that there is some underlying reason why Howell let up large numbers of line-drives and was ineffective as a starter.  Which one is it?

To be honest, this argument could probably be argued effectively in both directions.  Over the past two years, Howell has proven that he has dominant stuff and that he's an expert at deception and mixing pitches.  Maybe his luck improved when he moved to the bullpen, and he received the typical boost in performance that starters typically gain upon pitching out of the pen (approximately .75 FIP, if I remember correctly).  At the same time, though, I can't help but notice this fact:

Before moving to the bullpen, Howell's fastball averaged around 82 MPH, giving him only 5 MPH separation from his change-up.  This lack of velocity made both his fastball and change-up ineffective, which changed when Howell moved to the bullpen.  In the bullpen, Howell's fastball gained velocity and separation from his change-up, and as a result, they became some of his most effective pitches.  Heck, as FreeZo discussed in this piece, Howell's fastball froze batters for called strikes 37.5% of the time this past season.  Even with good mixing and deception, would that same fastball freeze batters if it only went 82-83 MPH?  I tend to doubt it.  It's a small decrease in velocity, but it may be enough to make a meaningful difference.

With these reasons in mind, I believe the Rays are making the correct choice leaving Howell in the bullpen.  If moved back into the rotation, he probably wouldn't be as bad as he was in 2007, but also not nearly as good as he's been in the bullpen.  Best estimate: a 4.25 FIP, which certainly isn't horrible, but also isn't awe-inspiring.  Of course, he could always be better than that, but considering the starting pitching depth the Rays have in the franchise, their lack of dominant bullpen depth, and their position on the win curve, it'd make no sense for them to move a valuable, cheap, dominant arm out of the bullpen.  If Howell were still on the Royals, on the other hand, it might make more sense for them to leave him in the rotation, since they have little starting depth and they could afford to take the time to see how Howell fared.  Also, what's the point of having a dominant closer if you can't get him a lead (*cough*Soria*cough*)?

So Carson, I definitely agree with you in that I think J.P. could still be a valuable starter and it's fun to ponder about intellectually (I spent many lunch hours daydreaming about it myself), but it isn't going to happen anytime soon.  At least not on the Rays.