Why Not Let Duensing Start Again?

Everybody knows that Brian Duensing is a great reliever and a terrible starter. He excelled out of the bullpen in 2009 and 2010, then moved to the rotation last year and flopped, putting up an ugly 5.23 ERA. This year the Twins have made it clear that Duensing will pitch out of the bullpen full-time. Everyone knows that’s where he does best, and everyone knows the Twins need more help there.

But everyone might be wrong.

One thing we like to do at Puckett’s Pond is present opposing viewpoints. So when something seems to be a foregone conclusion, like the fact that Duensing will be a reliever in 2012, it’s time for the Pond to delve a little deeper and see if things are actually the way they should be. And there might be good reason to reassess Duensing’s role on the team.

First, Duensing has not always been superior as a relief pitcher. Consider 2009; that year, Duensing joined the Twins in June and spent a couple months in the bullpen before the Twins’ desperate last-minute playoff run forced them to try Duensing in the rotation. He started eight games down the stretch, going 5-1 with a shiny 2.73 ERA. What we may not remember is that he had a 5.17 ERA as a relief pitcher that year. Duensing was actually not incredibly effective in 2009 until he started starting.

In 2010, Duensing was a good relief pitcher, and that cannot be debated. Despite his success the year before, Duensing again started out of the bullpen, since the Twins rotation was already fully-stocked with Scott Baker, Francisco Liriano, Carl Pavano, Nick Blackburn, and Kevin Slowey. In 40 relief appearances, Duensing had a 1.80 ERA, which proves he was very good. Then Blackburn started to struggle, and Duensing returned to the rotation. He made seven starts and went 7-2 with a 3.05 ERA with 1.25 WHIP. Sure, the ERA was higher in the rotation, but it was still very effective. An an effective pitcher in the rotation is more valuable than an effective reliever, because he can give the team more innings.

For most of the offseason, it appeared that the bullpen might need more help, and thus Duensing’s talents would be more needed in the ‘pen, whether he was actually better there or not. Now, though, Scott Baker is in danger of missing time to start the season. Anthony Swarzak seems to be next in line for that spot, and the Twins seem to be stretching him out in preparation; Swarzak pitched 3.2 innings on Monday. Duensing, though, has a longer history of success in the role, he might have more to add.

One common reason cited for keeping Duensing in the bullpen is that he struggles against right-handers, and the bullpen allows the Twins to match him up strategically against lefties. Righties have teed off to a line of .300/.357/.477 versus Duensing, compared to .203/.248/.263 for same-handed batters. Thus, there is a strong argument to keep Duensing as a LOOGY, a pitcher who only comes in the game to face left-handers. But it would make no more sense to use Duensing as a setup man than it would to use him as a starter, since he would have to face quite a few righties in either situation. Also, keep in mind that Duensing’s career splits are largely skewed by his dismal 2011 season – before then the numbers were not so stark. In 2009, for example, lefties hit .248 against him and righties hit .269.

Finally, Duensing’s pitching style might suit the rotation better than the bullpen. Ideally, a short reliever should not give up too many hits. A starter can get away with giving up hits, as long as he scatters them throughout five or six innings. But a reliever is often called upon to pitch with men on base in a close game, and if he allows a single it could tie the game or break it open. The best relievers are the ones who can come on and use his blazing fastball to strike out a batter or two. Duensing does not have a blazing fastball, and he does not strike out a lot of batters (5.9 per nine innings in his career). But he has proven to be an effective pitcher who can give the team quality starts during a pennant drive.

The Twins should at least consider this.