There’s Nothing Out There Like the Odrisamer Despaigne Changeup by Jeff Sullivan February 24, 2015 Last Friday, I took a look at the exceptional nature of Carlos Carrasco‘s split-change. I had trouble finding any kind of decent comparison for it, and when I did a little further analysis in the comments, the best I could come up with was Masahiro Tanaka’s splitter, and Tanaka’s splitter is supposed to be one of the best in the world. So, that was neat, and fun, and people are all about Carrasco hype because of the statistics he just posted down the stretch last summer. This post is about Odrisamer Despaigne. Despaigne isn’t nearly as statistically appealing as Carrasco is, and he’s not even assured a starting spot in the majors this coming season. But, just as Carrasco has an unusual split-change, Despaigne has an unusual changeup. A very, very, very unusual changeup. Maybe that’s not a total surprise — last July, Eno talked about his weird grip. And Eno, see, knows a lot about grips. Here’s an image: On the other hand, here’s a .gif, and in this .gif, it looks like Despaigne is using a different grip. Do you see him using a different grip? I’m not totally certain. But it’s an awesome .gif anyway, so here it is: Let’s get to the point. Despaigne makes use of multiple arm angles, and multiple pitches. He’s a complicated guy, who pitches in a classic Cuban style, so he has a lot going on. What grabbed my attention was his changeup, which he threw about once per nine pitches. This table is the result of the same process I’ve been using for a couple weeks, selecting liberally from the Baseball Prospectus PITCHf/x leaderboards. I looked at starting pitchers in 2014, and this time I didn’t bother to separate righties and lefties. I compared changeups and splitters to Despaigne’s changeup in terms of velocity, horizontal movement, and vertical movement, and shown are the top five comps. I consider a comp score up to 1 a good comp, and a comp score up to 2 a half-decent comp. Pitcher Pitch Velocity Horizontal Vertical Comp Rating Odrisamer Despaigne CH 74.7 3.9 6.4 – Jeremy Hellickson CH 80.9 5.0 5.9 3.1 Bruce Chen CH 76.3 8.7 6.3 3.1 Jered Weaver CH 78.2 5.5 8.8 3.3 Kyle Hendricks CH 79.4 2.8 4.2 3.3 Jeff Locke CH 80.2 6.2 6.9 3.5 The closest comp: 3.1. Which isn’t a particularly good comp. And you can see where the differences are — Despaigne’s changeup is a lot slower than Hellickson’s. It’s a bit slower than Chen’s, with a lot less run. On these traits alone, Despaigne’s changeup is unique among 2014 starting pitchers. But, just as with Carrasco, there’s another layer here. If you have an interesting changeup, and you throw it 100% of the time, it’s not an interesting changeup, because it’s not a changeup from anything. The whole point is that a changeup looks like a fastball, except it isn’t a fastball, and if the hitter things it’s a fastball, he’s screwed. Changeups, ideally, are to be compared to heaters, and the magic is in the separation. So I looked at the separation. I compared changeups (and splitters) to four-seamers, and I compared changeups (and splitters) to two-seamers. I calculated gaps in velocity, horizontal movement, and vertical movement, and then I generated comp scores for the differences, relative to Despaigne’s differences. Up first, let’s look at changeups (and…well, you get it) vs. four-seamers, for 2014 starters: Pitcher Velocity Gap Horizontal Gap Vertical Gap Comp Rating Odrisamer Despaigne -16.7 2.5 0.2 – Marco Estrada -11.1 3.1 -2.6 4.8 Trevor May -8.8 1.9 -0.5 5.0 Scott Kazmir -14.2 1.8 -6.1 5.0 Wily Peralta -11.5 0.6 -2.0 5.0 Marco Gonzales -11.2 2.7 -3.5 5.0 The closest comp is 4.8 standard deviations away. That’s an enormous difference, and you can observe where it comes from — Despaigne’s changeup was almost 17 ticks slower than his four-seam fastball. For Estrada, 11 ticks. Kazmir shows up with a 14mph separation, but then look a couple columns over — Kazmir’s changeup sinks a lot more than his four-seamer does. Despaigne’s shows no difference in vertical movement. It’s just weird. It’s clearly unusual. Maybe there are better comps, comparing changeups vs. two-seamers? Pitcher Velocity Gap Horizontal Gap Vertical Gap Comp Rating Odrisamer Despaigne -16.5 -3.9 2.1 – Wily Peralta -11.7 -2.0 -0.1 5.0 Clay Buchholz -9.8 -4.9 0.4 5.2 Jeremy Hellickson -9.9 -3.4 -0.8 5.4 Jeff Locke -11.2 -2.4 -1.0 5.5 Carlos Martinez -7.4 -3.3 2.7 5.7 Haha, no. These comps are even worse, relatively speaking. In reality, they’re all just really far away, so far away it doesn’t make sense to care about decimals or anything. Peralta shows up 5 standard deviations away, and while the movements aren’t outlandishly different, Peralta’s changeup is five miles per hour closer to his two-seamer than Despaigne’s. That’s a big difference. That’s why it’s actually fair to say Despaigne might be unique. Out of curiosity, I expanded the window to cover the entire PITCHf/x era. This way, I could compare Despaigne against starters from 2008 – 2014. Here are changeups vs. four-seamers: Pitcher Velocity Gap Horizontal Gap Vertical Gap Comp Rating Odrisamer Despaigne -16.7 2.5 0.2 – Sean O’Sullivan -13.9 0.8 -1.6 3.4 Franklin Morales -13.6 2.6 -2.9 3.4 Mike Mussina -16.0 1.1 -4.6 3.7 Dallas Braden -15.2 3.6 -4.1 3.8 Sean Gallagher -10.9 1.6 -0.9 4.2 Getting a little closer, but that’s still quite a distance. You know what changeups usually do? Sink more than fastballs, especially four-seam fastballs. Not so, for Despaigne. That’s the other weird thing, aside from the velocity separation. Finally, here are changeups vs. two-seamers: Pitcher Velocity Gap Horizontal Gap Vertical Gap Comp Rating Odrisamer Despaigne -16.5 -3.9 2.1 – Mike Mussina -16.7 -4.3 -1.1 2.2 Brandon McCarthy -12.9 -2.3 1.0 3.7 Mitchell Boggs -12.0 -4.9 0.1 4.2 Cha Seung Baek -11.6 -2.3 0.7 4.5 Wily Peralta -11.2 -1.8 0.8 5.0 Hey, look! Something in the 2s! Indeed, Mussina is almost a half-decent comp for Despaigne in this way. But then, this captures just Mussina’s 2008, which was the last season of his career. He didn’t throw very many changeups — under 6% — and according to pre-2008 data, Mussina’s changeup had previously been much faster. That year, there was a 16mph difference between Mussina’s fastball and his change. Over seven years for which we have some data, the separation averaged 13mph. So our one somewhat half-decent comp is a 39-year-old version of a should-be Hall-of-Famer who didn’t pitch like that earlier in his career. I’m calling it: Despaigne’s changeup is unlike any other thrown by a starting pitcher on recent record. Part of that is the movement, and a bigger part of that is the huge separation in velocity between Despaigne’s changeup and his fastball. We’ve observed a gap of more than 16mph. The next-closest starter is probably Dallas Braden, with a gap of a little over 14mph. And Braden’s changeup had more relative sink. Despaigne’s changeup is extremely weird, and here’s a sequence you might have playing over and over in your head: I mean, what else would you expect, right? The changeup traits are silly. What is a hitter supposed to do? But! Despaigne, mostly, used the change against lefties. It had one of the lower whiff rates for changeups among starters, and Despaigne came away with a big platoon split, walking almost as many lefties as he struck out. Lefties saw a cutter more than they saw a changeup. Maybe this has something to do with it (via Baseball Savant): Those changeups are scattered. Despaigne isn’t a command pitcher. He hasn’t been graced with a great location skill. The traits of Despaigne’s changeup are exceptional. The changeup itself is not great. Having a unique changeup/fastball separation doesn’t necessarily have to be a good thing, and it’s not going to be a good thing when you don’t locate too well. Almost everything begins with location. And in a way, almost everything ends with location. Despaigne’s changeup is bizarre, and Despaigne’s changeup is only moderately useful. Weird and great don’t automatically go hand-in-hand. But weird’s weird, and weird’s notable. Even if weird isn’t good for the player, weird’s good for the writer. So I’m satisfied, at least.