Mariano Rivera is the best reliever in baseball history, and the best paid. He earns $15 million each year, which would make him the ninth-highest paid starter in the game. Brad Lidge, at $12 million, is the only other reliever who would rate among the top 15 highest-paid starters. He earns as much as Bronson Arroyo, whose main virtues are that he pitches a lot and plays guitar.
Perhaps relievers are radically underpaid. It seems more likely that general managers have a good line on the value of short relief, and that they think it's worth a lot less than starting pitching. This makes sense, and should tell us something. Paychecks don't lie.
For this reason alone, I've never much understood why there was a controversy over whether the Yankees' Joba Chamberlain should be a starter or a reliever. After his 2007 debut, during which he gave up two runs in 19 games and struck out more than a third of the batters he faced, there was some sense in the notion that he was best left alone. There was more sense in the notion that it couldn't hurt to try him in the rotation, though. After all, he could always be moved back to the bullpen if it didn't take.
It did take, though. Chamberlain has been brilliant. Through his first 22 starts he has a 3.12 ERA. The only pitcher in baseball with at least as many career starts who has done better is Johan Santana. Chamberlain isn't a full-blown ace yet -- he can be oddly tentative and he hasn't mastered the changeup, both factors in a mildly unsightly 1.365 WHIP. And, as is the case with many young starters, his numbers are prettier than they might be because the team pulls him so early in most games -- his most recent outing was the first ever in which he pitched more than seven innings. Still, if those are the worst flaws in your game at 23, you're fine.
Starters are more valuable than relievers, and Chamberlain is showing every sign of developing into a wonderful starter; therefore, he should start. Simple.
Because it's so simple, it's mildly bizarre that more than a year after his brief run in relief ended there are still people who are passionately convinced that Chamberlain should be working out of the bullpen. Some of this is just contrived; some, like the insane ranting of New York City radio host Mike Francesca, is just comical. A lot of it, though, is sincere, which is what can make it so frustrating. Because when you run through the arguments, there's nothing there.
The best reason to move Chamberlain to relief would be to protect his health, and if he ever is moved, that will likely be why. Still, this is the least convincing case one can make. The Yankees are better positioned than anyone else in the world to know whether there's serious reason to doubt Chamberlain's ability to hold up to a starter's workload. That they're using him as a starter is the most credible evidence you could ever have that there isn't.
The next best reason to move him would be that it would give the Yankees their best chance of winning this year: They have some bullpen problems and six starters, of whom Chamberlain is theoretically the best suited for relief. This isn't entirely ridiculous, but while the Yankees have six starters, one of them is the suspect if improving Phil Hughes, another is Chien-Ming Wang, who earlier this year endured possibly the worst three-game stretch in baseball history, and a third is the rather injury-prone A.J. Burnett. Why you'd want to move a burgeoning ace out of the rotation when he'd have to be moved back soon enough because of ineptitude from, or injury to, one of these three sketchy pitchers is beyond me.
Still, at least this is an argument rooted in observable fact. The strangest argument people make is that Chamberlain would actually be more valuable as a reliever. This just isn't true.
There are a lot of ways to prove this with complicated math, but there really isn't any need. Chamberlain has a career ERA of 3.12 as a starter, and 1.53 as a reliever. Given that American League starters have run up a 4.64 ERA and relievers a 4.42 ERA this year, and that as a rule of thumb a random scrub pitcher -- or replacement player, per the term of art -- is about 20 percent worse than average, it's simple to figure out just how many more runs above replacement level Chamberlain would prevent in each role, at least assuming that he pitched the rest of this year as he has in the past. (You want to compare him to a replacement player, incidentally, because by setting the baseline at average you value the average players at zero. This is problematic, as Bronson Arroyo's agent could gleefully tell you.)
In 180 innings as a starter, Chamberlain would be worth about 50 runs above replacement. In 80 innings as a reliever, he'd be worth about 35. Fifty is more than 35.
Of course you can introduce a fudge factor here: Innings pitched by a reliever of Chamberlain's caliber are generally more valuable than those pitched by a starter, sometimes twice as much or more so. Double the value of Chamberlain's relief innings, and suddenly he is actually more valuable as a set-up man. The counter to that, though, is that he would absolutely not run up a 1.53 ERA as a reliever going forward. Spectators blinded by the brilliant memory of this apparition appearing from nowhere two years ago to heave bowling balls and vanishing benders at clueless batters might not believe it, but it's true.
1.53 is preposterously good. For perspective, Rivera's career ERA is 2.25. Just for argument's sake, though, say that if you put Chamberlain in the bullpen tomorrow he would be every bit as good as Rivera has been in his career. In that case he'd be worth 25 or so runs above replacement per 80 innings of relief. Double the value of those innings and he comes out as no more or less valuable than he would as a starter. This assumes, again, that he would pitch as well as the most magnificent reliever anyone has ever seen.
There are complicated decisions to be made in baseball. Neither this nor any similar argument involving any young pitcher is really one of them. If Chamberlain is hurt, or if the Yankees obtain access to a time machine and learn that he really is much better than Mariano Rivera, they should by all means listen to the outraged minority and consign to the bullpen a 23-year-old with every chance at being the best pitcher they've developed since Whitey Ford. Otherwise, they should leave him alone where everyone can enjoy him. He's doing fine.