Revisiting Carlos Martinez's rotation conversion

At the beginning of the year, I questioned a claim made by Howard Megdal on Effectively Wild: that Cardinals reliever-turned-starter Carlos Martinez would actually improve his strikeout rate moving from the bullpen to the rotation. It's a bold take, simply because not many pitchers do so. As I stated in my article back in January, there's been plenty of research that indicates that moving from the 'pen to the rotation usually includes a velocity drop and diminished effectiveness.

But, Howard's an expert, and so I wanted to figure out what it could look like if Martinez did improve. Did very many pitchers actually get better when moving to the rotation? Given that Carlos was likely to lose velocity in the transition (almost everybody does), were there success stories that we could point to and say there was the potential for big-time success?

In that original article, I found that 26.5% of relievers moving into the rotation saw their strikeout rates increase in the year of their move ... a number that surprised me at the time, and lent more weight to Howard's opinion. After all, if a quarter of pitchers actually strike out more hitters after moving into the rotation, and experts on Martinez seem to think he could benefit from such a move thanks to his pitch mix, then it sure sounded like it could be possible.

Now, almost a full year later, I wanted to revisit Martinez's conversion. As we know now, Carlos Martinez was, in fact, a damn fine starter in 2015. He even went to the All-Star Game this season and posted a 3.26 RA9 (runs against per nine innings) and a 3.21 FIP.

And, of course, Howard was right. Martinez did in fact raise his strikeout rate in 2015. He managed a respectable 21.8% K-rate as a (mostly) reliever in 2014, but he bumped it up to 24.4% as a (mostly) starter in 2015. That delta of +2.6% would've slotted in between Dallas Keuchel and Dan Haren on my top-15 K% gains list from the end of last season.

So ... how'd he do it? Did his velocity stay high? While Martinez does still have elite velocity, his velocity did drop, like all good relief-to-starting conversions do. But as you can see from the Brooks Baseball chart below, his velocity drop wasn't too bad at all. To wit:

His fastballs (four-seamer and sinker) did in fact lose a bit of velocity, especially at the beginning of the year. According to Brooks' numbers, Martinez averaged about 1.5 mph less on his four-seamer and about 1.0 mph less on his sinker. His slider lost about half a mile per hour as well, and his change (which we'll talk about in a bit) had the same drop as his heater ... which is probably good from a pitch separation standpoint.

Perhaps the most interesting thing to note from this chart is how Martinez's velocity kept on ticking up all season long -- in the monthly aggregate -- until he sat out the back-end of the season due to his shoulder injury. (We'll talk about that later on too.)

With his velocity (slightly) down, what about his pitch mix? Jeff Sullivan smartly broke down the alteration of his changeup mid-season over at FanGraphs, where it appears he modified his grip on the pitch to a circle-change grip. The shift in his change -- note here that folks always thought it was destined to be an average-to-plus pitch -- allowed him to throw it much more often than he had as a reliever, bumping it's usage from about 9% in 2014 to about 17% in 2015.

As one might expect, Martinez leaned heavily on the pitch against left-handed hitters, which makes sense given that the right-handed changeup is a lefty-killing pitch. He used it about 28% of the time and got a very healthy 23.5% whiff rate on the pitch. So though lefties still beat him up a bit -- his wOBA against from left-handers was .329 compared to right-handers at .276 -- he was able to do well enough to keep him in games and strike down the righties he did face.

(It's also worth noting that Baseball Prospectus's Deserved Run Average (DRA) metric doesn't believe in Martinez's performance quite the same way other statistics do, giving him a 4.31 DRA, or almost a run and a quarter above his actual ERA. That seems to indicate that his true performance was aided and abetted by contextual factors and his "deserved" runs allowed might have told a very different story.)

Also in that previous article, I noted Tyson Ross as a potential comparison for Martinez despite their very different pitching styles. Where Martinez throws a lot of four-seamers and changeups, Ross is a classic "sinker-slider" pitcher, who doesn't have a regular pitch below 87 mph and has effectively thrown his changeup off a bridge. Both pitchers, however, are able to leverage their sinkers into high ground ball rates while still managing to get Ks with their off-speed or breaking offerings. And both guys, of course, were very good last season.

The conversion from the bullpen to the rotation was a very successful one overall for Martinez. He was able to keep his strikeout rate high -- raise it even -- while inducing scores of grounders and without letting his walk rate get too bad. The only true negative, in my book, is the shoulder injury that pulled him out of commission late in the season and for the playoffs. Perhaps that's something that we should've given a little more consideration as part of the conversion. After all, for a pitcher to throw as hard as Martinez, but then also to lose only a little velocity and add 80 innings of regular season work ... that's quite a bit.

But any way you slice it, Martinez's conversion was a success, and it shows how just because the odds are against someone (or the statistics say some event is improbable) doesn't mean that the event is outside the realm of possibility. It's an example of a way to interpret what Russell Carleton wrote about in the BP Annual a couple years back: N=1. Instead of trying to fit our paradigms for a system / metric / set of events to every single player, it's wise to take a step back once in a while and look at the individual events or players and see if there's something there that applies specifically to that one case. It's what Howard Megdal did in his analysis of Martinez, and it's played out in the real world, despite the shifty odds against relievers adding strikeouts as starters.

All Carlos Martinez needed to do to succeed was maintain much of what made him a solid reliever while improving a changeup that many thought was already on the up. He did that, and now we have another example of why some of these conversions can work. We don't know what the future holds for Carlos. Will his shoulder hold up? Will he be able to continue development of his sinker and change for the days when he starts to lose velocity, a la Felix Hernandez?

The odds are against Martinez having a Hernandez-like trajectory. But the odds aren't everything.

* * *

Bryan Grosnick is the Lead Writer at Beyond the Box Score and a columnist for Baseball Prospectus - Boston. He also throws four pitches, but they're all eephuses.