Identifying 2013’s Most Unhittable Pitches

When I looked at this toward the end of August, a lot of baseball had happened. Enough baseball that I felt it worthwhile to take a look at this. Since that point, the rest of baseball happened, so now that we have season closure, it seems like it’s time for a second and final update. We can now officially answer the question: which individual pitch was the most unhittable during the 2013 regular season? Beyond that, which were the most unhittable pitches from relievers, and which were the most unhittable pitches from starters?

Of course, in updating the first post, I have to issue all the same caveats as in the first post. So in a sense I’m just writing the same thing again, with some different numbers and pictures. What this really is is a post containing contact-rate leaderboards. The pitches you’re going to see are the pitches that yielded the lowest rates of contact, getting therefore the highest rates of whiffs. That seems like a good way to explore unhittability, but as you understand, pitching is complex, and every pitch depends in some way on every other pitch.

I don’t know how one might test an unhittable pitch in a laboratory. Pitches work within a certain context. If you want to know about a certain pitch’s “true” unhittability, you have to strip it from the context. How unhittable is Matt Harvey’s fastball? Have Harvey throw only that fastball, hundreds or thousands of times, against generic competition. Then repeat with other pitchers throwing only fastballs. Whiff rates would plummet, because hitters would know what they were getting. You could get interesting information out of that experiment, but it’s an experiment that’ll never take place, and so much of being a good pitcher is keeping hitters guessing. Hitters are left guessing when they don’t know what’s coming. That’s a consequence of a pitcher throwing multiple pitches. If Koji Uehara follows a fastball with a splitter, and the splitter is swung on and missed, some of that credit goes to the splitter, but some of it also goes to the fastball.

Game theory gets in the way of conclusive results, of true unhittability results. Casey Fien‘s fastball yielded a contact rate similar to that of Aroldis Chapman’s fastball. You don’t need to be told that Chapman’s fastball is one of the best on the planet. You might need to be told that Fien is a pitcher in major-league baseball. But those fastballs are used differently, and the pitchers throw other pitches as well, the end result being that Fien’s heater was swung on and missed a lot even though it seems markedly more hittable in isolation than Chapman’s. In the numbers, there’s no such thing as a pitch in isolation. That’s something we understand, that’s something we have to deal with, and that’s something we’ll just choose to overlook today, from this point forward.

Even if this isn’t a post about “true-talent” unhittability, this is a post about pitches that were really hard to hit when they were thrown this season. There are still some things we can all learn from that, and it’s still a fun subject to explore. Are these pitches really as unhittable as they seem? Do they depend heavily on other stuff? That’s for you to consider and discuss. Read, consider, and discuss. Seems like a pretty good general policy.

All the data comes from the Baseball Prospectus PITCHf/x leaderboards and Brooks Baseball. We’ll begin with a combination of starters and relievers, but later I’ll break starters out separately. In August, I set a minimum of pitches that got at least 50 swing attempts. This time, I’m considering only pitches that got at least 100 swing attempts, which gave me a sample of 989. Just like 50, 100 is arbitrary, but it also feels to me like a good enough sample, so. Contact rate is calculated as non-whiffs over swings. The overall league-average contact rate was about 79%.

Just to throw it in there: Danny Farquhar’s curveball yielded a contact rate of 37.8%, getting 46 whiffs out of 74 swings. It didn’t meet the minimum threshold, but had I set the minimum at anything under 75, then Farquhar’s curve would be baseball’s most unhittable pitch, by a few percentage points. Instead, he just gets an honorable mention. Here now is the top ten of the most unhittable pitches of 2013, given at least 100 swings:

1. Pedro Strop slider, 43.8% contact
2. Cody Allen curveball, 44.0%
3. Craig Kimbrel curveball, 45.5%
4. Greg Holland slider, 47.0%
5t. Josh Outman slider, 47.1%
5t. A.J. Ramos changeup, 47.1%
7. Boone Logan slider, 47.3%
8. Patrick Corbin slider, 47.6%
9. Rex Brothers slider, 48.2%
10. Gonzalez Germen changeup, 48.5%

Obviously, there aren’t any fastballs, because fastballs usually aren’t a swing-and-miss weapon, and they’re thrown in all counts, very often, and sometimes they’re thrown trying to get contact. You don’t find any fastballs until you get into the 60% contact range. We’ve got here six sliders, two curves, and two changeups, one of the changeups coming from whoever Gonzalez Germen is. Also, one starter and nine relievers, because relievers more often try to get strikeouts since they don’t have to pace themselves for several innings at a time. This is a list of putaway pitches, and they’re all mighty fine putaway pitches.

Because I get paid first and foremost for my .gifs, here are .gifs of the top six. Would’ve been .gifs of the top five, but Outman and Ramos came out identical and I didn’t want to pick.

A.J. RAMOS CHANGEUP

• 102 swings
• 54 whiffs

JOSH OUTMAN SLIDER

• 136 swings
• 72 whiffs

GREG HOLLAND SLIDER

• 213 swings
• 113 whiffs

CRAIG KIMBREL CURVEBALL

• 110 swings
• 60 whiffs

CODY ALLEN CURVEBALL

• 125 swings
• 70 whiffs

PEDRO STROP SLIDER

• 153 swings
• 86 whiffs

I said earlier that I would break starters out separately, so now let’s look at another top-ten list, excluding relievers and also relief appearances in the cases of guys who filled multiple roles. The most unhittable pitches of 2013 thrown by starting pitchers:

1. Patrick Corbin slider, 47.6% contact
2. Tyson Ross slider, 49.6%
3. Jeff Samardzija splitter, 50.5%
4. Jarrod Parker changeup, 51.2%
5. Derek Holland slider, 52.2%
6. Stephen Strasburg changeup, 52.7%
7. Kris Medlen changeup, 53.0%
8. Yu Darvish curveball, 53.3%
9. Matt Garza slider, 54.2%
10. Ivan Nova curveball, 54.4%

Four sliders, two curveballs, three changeups, and a splitter, which moves a lot like a changeup. You see a couple sliders from lefty starters, where ordinarily the conventional wisdom is that those lefty starters need effective changeups. We have a lot of evidence that Corbin’s slider is something special against hitters on both sides, and it’s going to be interesting to see how the league responds to him in 2014, if it responds differently at all. I suppose he also pitched plenty in 2012. It’ll also be interesting to see if Tyson Ross maintains his unhittability in a starting role, since he blossomed almost out of nowhere.

Some .gifs now of the top five:

DEREK HOLLAND SLIDER

• 416 swings
• 199 whiffs

JARROD PARKER CHANGEUP

• 324 swings
• 158 whiffs

JEFF SAMARDZIJA SPLITTER

• 293 swings
• 145 whiffs

TYSON ROSS SLIDER

• 272 swings
• 137 whiffs

PATRICK CORBIN SLIDER

• 370 swings
• 194 whiffs

Two years ago, Corbin’s slider yielded a contact rate of 52.6%, so it’s pretty demonstrably elite. This year, lefties hit it 39% of the time, while righties hit it 52% of the time. In fairness, Corbin did show a wide platoon split, with a righty xFIP higher by almost two full runs, but the more important point here is that Corbin’s slider works against both sides, so he can start and start well without possessing an amazing changeup or splitter or curve. In this way, Corbin is somewhat unusual.

In closing, the highest pitch-type contact rate belonged to Aaron Harang’s changeup. Batters attempted 102 swings, and they missed three times. Here is one of those misses, which was actually a foul tip:

97.1% contact. A .327 batting average with a .146 ISO. Interestingly, over the rest of the PITCHf/x era, Harang’s changeup allowed a contact rate of 84.7%, but with worse in-play results. This post has now concluded with a brief discussion of Aaron Harang, just like you expected.