Should any elite relievers be converted to starters?

Converting relievers to starters doesn’t exactly have the greatest track record. For every David Price, Chris Sale, and Adam Wainwright, there is a Joba Chamberlain, Neftali Feliz, and Daniel Bard.

Having a starter begin his major league career in the bullpen can be an effective way to acclimate him to the majors. Orioles Hall of Fame manager Earl Weaver was a big fan of this methodology.

The reason why Chris Sale spent his first two season in the bullpen is because scouts believed his delivery and body type were not suitable for the rigors of starting. Baseball Reference lists Sale at 6’6’’ and 180 pounds. He is a beanpole. He is by no means weak, but somebody that tall and slender pitching 200 innings a year seemed like a recipe for disaster. As for his delivery, take a look for yourself.

Yikes. That is some bad arm action that puts a lot of strain on the elbow. How he has stayed so healthy is amazing (and a testament to how difficult it is to predict health). Good genes might be a factor, but I think we would be remiss for not crediting the White Sox’s training staff.

Sale’s fastball averaged 96 MPH, he had a wipe out slider, and right-handed hitters didn’t kill him despite his low arm slot, which can allow opposite-handed hitters to see the ball better. It was quite a risk given the reasons I have already laid out, but it has worked out phenomenally well.

Joba Chamberlain was yanked back and forth between relieving and starting, which is never a good idea. Neftali Feliz was handled with more care and was actually quite effective in his seven starts. He had a 3.16 RA9, but that came with a .213 BABIP and 4.64 FIP. For Feliz, the culprit was Tommy John surgery, after which, he never started again.

Daniel Bard was the most interesting out of the pitchers I mentioned. He was converted to starting in 2012 after a stellar three-year run in the bullpen. It was a disaster. He had a 5.49 RA9 and struck out only 13.8 percent of the batters he faced. In what would turn out to be a foreshadowing of the rest of his career, Bard walked a whopping 14.6 percent of batters faced.

Bard developed Steve Blass disease. You don’t need to be a scout to hang a 20-grade on Bard’s control. For the rest of his professional career, he walked an unfathomably high 31.3 percent of batters faced.

Trying to convert a reliever with the potential to start is usually worth the risk. Even a mid-rotation starter is more valuable than an elite reliever, and his performance will likely be more consistent from year to year. Also, unlike Bard, if it does not work out, you can always put the pitcher back in the bullpen.

Let’s take a look at some of the elite relievers in the game today to see if it is worth trying any of them as a starter.

Aroldis Chapman

When he was on the Reds, there was talk of possibly trying to convert Chapman to a starter. For whatever reason, he was not interested in even trying. He signed a shockingly high contract this winter, so it is hard to say that it did not work out for him. That being said, he could have made twice that amount had he hit his ceiling as a starter.

Chapman has the greatest fastball ever. You read that right. He is a lefty throwing 100 MPH with some life and downward plane. He has the slider and changeup to make up a good repertoire of a starter. He would have to shave a few miles per hour off his pitches in order to be able to withstand the rigors of starting, but you could see something resembling a left-handed Noah Syndergaard if he reached his ceiling.

Let’s pretend for a moment that Chapman was open to starting. I would be all for it. The only caveat, as pointed out by ESPN’s Keith Law, is that he tends to throw exclusively fastballs when there are men on base. Throwing 100 MPH is much less effective when the hitter knows it’s coming. However, that is fixable; unfortunately, it’s not going to happen. Chapman doesn’t even like pitching outside of the ninth inning let alone starting every fifth day.

Kenley Geronimo Jansen

That is seriously his middle name.

The case against Jansen starting is pretty simple. He only has one pitch: his cutter. It is a freakin’ awesome pitch, but it is still just one pitch. He has struck out almost 40 percent of batters faced thanks to that pitch, which is extremely impressive. He had some control issues early in his career, but he has worked his walk rate down to an excellent 4.4 percent.

The interesting thing about Jansen’s cutter is that it is not gripped like a typical cutter, but rather is thrown like a fourseamer at 93-94 MPH. It just so happens to have a sharp, natural cut that moves away from right-handed hitters.

Kenley Jansen has the out-pitch and the command to start, but he is likely to have trouble turning over lineups without being able to supplement it with a second or third pitch. Perhaps his slider could be a second pitch, but what would be his third pitch? His sinker?

Regardless, for Jansen to work out in a starting role, he would have to learn to pitch very differently than he has for the past 6+ years. The Dodgers are clearly committed to having him close which is probably for the best.

[For those of you wanting to cite Bartolo Colón as a pitcher who succeeds despite throwing one pitch, an anecdotal piece of data does not help to prove a case. Also, though it is said he throws ~90 percent fastballs, it is actually a mixture consisting of sinkers and fourseamers.]

Andrew Miller

As you are probably well aware, Miller began his career as a starter. He was expected to be an Ace when the Tigers drafted him the first round in 2006. Dave Dombrowski leveraged that expectation to trade him and Cameron Maybin for a guy named Miguel Cabrera from the Marlins. You might have heard of him.

Miller was terrible on the Marlins, so they gave up on him and traded him to the Red Sox for a fraction of what they paid for him. Red Sox pitching coach Curt Young could not fix him, so they demoted him to the bullpen. The rest is history. From 2012 to 2014, Miller posted a 2.90 RA9 and 35.4 percent strikeout rate. For his career as a reliever, he has a 3.00 RA9 and 37 percent strikeout rate.

The key to his success is that his slider is a destroyer of worlds. Could he leverage that back into a starting role? At this stage, probably not. Besides the fact that he will turn 32 in May, he doesn’t have the command to start, nor does he have the third pitch. Miller has only thrown two types of pitches since 2013: a fourseamer and a slider.

Mark Melancon

To be honest, I’m not even high on Melancon as a reliever. His fastball sits at only 92 MPH, and his 22.8 percent strikeout rate is roughly league average. I don’t see that combination aging well, despite the low walk rates, especially for somebody who will turn 32 by Opening Day. Recently at FanGraphs, Eno Sarris looked at pitchers similar to Melancon, and it’s not great.

Like K. Geronimo, Melancon succeeds by relying heavily on his cutter, albeit in a different way. He does not throw it nearly as much, and he uses it to induce weak contact, not to miss bats. He has a career 56.1 percent groundball rate. He doesn’t sound too different from Mariano Rivera, huh?

Despite his ability to induce weak contact, Melancon does not outperform his FIP by much. He has had a 2.11 RA9 since joining the Pirates versus a 2.25 FIP. He might have mediocre strikeout rates, but he limits the free passes and keeps the ball in the park.

Melancon’s curveball is also quite effective in inducing weak contact and is a good second pitch. Like Miller, he lacks the third pitch and is too old to start. I also can’t imagine him being able to throw his cutter over 60 percent of the time in a starting role. The only starting pitcher who threw a cutter more than 30 percent of the time in 2016 was Josh Tomlin. Only twelve starting pitchers threw a cutter more than 20 percent of the time.

Dellin Betances

Betances’s 3.82 RA9 looks like he had an off year, but he really didn’t. He still had a 1.78 FIP and 1.57 DRA. He pitched in a very hitter-friendly park and suffered from a .353 BABIP. He still struck out an outstanding 42.1 percent of batters faced and had a career high 53.9 percent groundball rate. A 98 MPH fastball and an 85 MPH curveball can do wonders when it comes to making hitters swing and miss. The curveball doesn’t have much depth to it, but it plays up thanks to his height.

The Yankees of course tried him as a starter in the minors, and he kept getting hurt likely because of his mechanics, including having Tommy John surgery. Similar to Miller, Betances having just two pitches combined with his poor mechanics, shaky command, and injury history are why he was converted to a reliever in the first place, and that is probably where he should stay.

Zach Britton

Britton had a great 2016, but it was overrated by people who put too much weight on runs allowed. Still, a 60 cFIP and 2.17 DRA are excellent and put him among the best relievers in baseball in 2016. Even with his crazy high 80 percent groundball rate in front of the stellar Orioles’ infield defense, a .230 BABIP is unsustainable. Yes, he does have a career .251 BABIP as a reliever, but at only 591 balls in play over three seasons of relieving. A pitcher needs 2,000 balls in play for his BABIP to represent his true talent and because relievers pitch so seldom, their true talent is likely to change significantly between when they start pitching and when they reach that 2,000 BIP threshold.

Britton followed the Andrew Miller model in starting out terribly in the rotation and then becoming lights out in the bullpen. He had a 5.41 RA9 and 4.20 FIP as a starter, and a 1.72 RA9 and 2.40 FIP as a reliever. His 30.2 percent strikeout rate over the past two seasons is very good, though not quite at the level of the likes of Chapman, Betances, and Miller.

Similarly to Jansen, Britton succeeds by using one pitch: his sinker. He throws it over 90 percent of the time. At 97 MPH with sink, it is a devastating pitch, causing swings and misses as well as the weak contact already mentioned. It is a big 5 MPH faster than it was when he was starting. The velocity increase is likely due to being able to pitch at max effort while relieving, but what I don’t understand is how Britton is also able to add even more sink to the ball.

Again, having only one pitch means a player can’t be an effective starter. Empirically, we already know that Britton was terrible as a starter, and if his sinker were back to 92 MPH in that role, it would be a lot less effective.

Craig Kimbrel

Kimbrel is already on the decline as a reliever, so I don’t know how he could be converted to starting. One could very well argue that he is no longer an elite reliever, despite his high strikeout rates. His velocity and movement have not changed at all in recent years, but in 2016 his groundball rate and control plummeted. Only Tony Cingrani and Kyle Barraclough had worse walk rates than Kimbrel’s 13.6 percent. The good news for the Red Sox is that if his decline continues in 2017, they can buy out is 2018 for $1 million.

As a declining reliever with only two pitches who could be gone after 2017, it makes no sense for the Red Sox to try and convert him to the rotation. Their rotation doesn’t need much help, anyway.

I went into this exercise really hoping to find at least one elite reliever who would be worth converting to a starter. All I got was a hypothetical ‘yes’ for Aroldis Chapman, though it will never happen at this stage (despite the Yankees desperate need for starting pitching). Unsurprisingly, teams always have really good reasons for keeping relievers in the bullpen.