In Game 7 of the ALCS, Astros pitcher Lance McCullers propelled the Astros to the World Series with four shutout innings in relief. Even more remarkable, 75% of his pitches in that outing were curveballs. He didn’t just suddenly start throwing a ton of curves, of course; he found regular season success with a nearly 50% curveball rate.

Storylines like this have led some to question whether it’s feasible for any pitcher with a good curve to throw it as much as a fastball. Traditional baseball wisdom holds that pitchers should work off their fastballs and set hitters up—probably a good idea for many. But, if a guy has a great curveball, why not throw it more often? Not every successful curveball would make the cut. Surely some curveballs—maybe those similar to McCullers’ curve–—could be viable primary pitches, at least in theory. This raises an intriguing question: who else could be a “curveballer”?

To answer that question, I decided to whittle down the list of curveball-throwers to those who possess characteristics that might translate well to McCullers-level curve usage.

Generally, the kind of pitcher we’re looking for is one who has a good curveball but doesn’t throw it a lot. I defined a “good” curveball using Pitch Info’s pitch value metric—wCU for curveballs—which measures how many runs a pitch prevents above average. It’s a rough heuristic, but worth a shot.

“Good” curveballs were those with a wCU value at least one standard deviation greater than the average for all pitchers who threw a curveball this season, minimum 50 innings. The average curveball rate was around 14%, so I set the ceiling to one standard deviation greater as well. Then I selected pitchers who threw their curves less than 25% of the time, but still generated at least five runs worth of value.

I came up with fourteen names:

 Player CU% wCU
 Stephen Strasburg 22.50 14.8
 James Paxton 21.60 5.4
 James Shields 21.10 6.0
 Santiago Casilla 20.80 5.6
 Ivan Nova 20.70 8.6
 Jimmy Nelson 20.60 9.7
 Robbie Ray 20.00 7.3
 Parker Bridwell 19.00 8.3
 Chris Stratton 18.50 5.8
 Chase Anderson 18.10 5.3
 Gregory Infante 17.60 5.6
 Clayton Kershaw 16.70 6.4
 Hyun-Jin Ryu 16.10 6.1
 Tanner Roark 15.60 8.6

First, I’m going to go ahead and cross Clayton Kershaw off that list because what he’s doing seems to be working pretty well. I’m also crossing off Gregory Infante and Santiago Casilla. They might benefit from using their curveballs more, too, but we’re looking for potential starters here.

Ideally, a potential curveballer wouldn’t have more than one other viable pitch. (If he did, then why bother making such a radical change?) If he does have a better pitch, it’s preferable that it’s not too much better than his curveball, or else he’d want to throw the other pitch a lot more. That clearly eliminates Strasburg since all his pitches are good. And with a fastball as electric as Paxton’s, that should be his primary offering. Not to mention his cutter and sinker. Same goes for Chase Anderson. Roark and Ryu both have two other good pitches, though they’re admittedly closer cases. So after that, the field is whittled down to six guys.

Next, let’s figure out what kinds of curveballs we want to see. If one of these pitchers wants to be like McCullers, it stands to reason that he should throw like McCullers. The task then is to figure out what McCullers—and others like him—have going for their curveballs that makes it possible to throw them so often. If we identify some of these curveballer-types and pinpoint commonalities between them, maybe we can figure out what it takes to be one of them.

To identify “true” curveball pitchers I looked at individual pitcher seasons over the last three years. Using that data, I attempted to come up with a list of guys who threw their curveballs often—as their primary pitch or close to it—and threw them effectively. After looking at the data, I settled on a threshold around 30% curveball usage. There aren’t too many surprises here:

 Seasons Player
 2017, 2016 Lance McCullers
 2017 Alex Meyer
 2017, 2016 Drew Pomeranz
 2017 Zack Godley
 2017 Jerad Eickhoff
 2016 Rich Hill
 2016, 2015 Tyler Duffey
 2016 Nate Karns
 2015 Jose Quintana

It isn’t a huge list, but keep in mind that each of these twelve pitchers threw their curves at least 30% of the time for an entire season, so there’s plenty of pitch data. Some of them only made the list once, and no one made all three years, but in each case this was more the result of pitch-mix changes than failure to sustain success.

Now that we have our sample of curveball pitchers, we can look for things they do well. Then, we’ll use this to come up with criteria to evaluate our could-be curveballers.

To start, a pitch that misses a lot of bats and gets a lot of ground balls has a good chance of being successful. So we could just look for pitchers who have good swinging strike rate and ground ball rate on their curves. To really get at what makes an often-thrown curveball successful, though, we can also examine underlying pitch characteristics that lead to whiffs and grounders.

Luckily, Eno Sarris has done some great work that tells us three things: that greater velocity is correlated with swinging strike rate, that horizontal movement is correlated with strikeouts, and that vertical drop is correlated with ground ball rate. So I’m going to look at movement, ground ball rate, velocity, and swinging strike rate, at least.

Second, it’s helpful to keep balls out of the air, so I also looked for low fly ball rates. Relatedly, if a curve is hit in the air it helps if it’s weakly hit, so I also included popup percentage. And I included average overall exit velocity for similar reasons.

Third, the last metric I looked at was spin rate. It is, of course, vastly oversimplifying to say that good spin equals good curveball, but high-RPM curves seem to be preferable.

Finally, I took our set of curveballers and found the average value for each of the above metrics:

 Horizontal   Movement (in.) Vertical Movement (in.) GB% FB% HR/(FB+LD) Velocity  SwStr% Popup% Spin Rate Avg. Exit Velocity
 4.19 -6.00 58.31 14.9 6.17 81.31 15.30 3.81 2456 86.48

Now that I’ve found some pertinent characteristics of our curveballers, I can compare our six candidates against them. I indicated in green when a candidate value was better than the curveballer average and in red when it was worse.

James Shields

 Horizontal   Movement (in.) Vertical Movement (in.) GB% FB% HR/(FB+LD) Velocity  SwStr% Popup% Spin Rate Avg. Exit Velocity
 3.66 -3.18 42.25 25.35 5.88 78.6 16.29 9.86 2499 84.4

Shields might not have the movement to be a curveballer. His curveball is slower (compared to the sample), but he manages to get whiffs anyway. He gets good outcomes on balls hit in the air, but that’s hard to rely on in the current home run environment, especially when the other skills aren’t there. I’m not sure the whiff rate would be sustainable if he threw the pitch a third of the time without more horizontal movement and velocity. He does throw a slow curve as well, so there’s a chance he could survive by mixing both pitches in. It might be worth looking into for Shields.

Ivan Nova

 Horizontal   Movement (in.) Vertical Movement (in.) GB% FB% HR/(FB+LD) Velocity  SwStr% Popup% Spin Rate Avg. Exit Velocity
1.11 -3.21 38.6 21.05 10.34 81.03 17.12 10.53 2353 83.3

Nova’s curve was pretty good this season, but it just didn’t look enough like our sample pitchers’ to make me think he could sustain it at high percentages. His ground ball rate was only 38%. While his whiff rate was good, it’s likely not enough to compensate. He threw the pitch fairly hard, but it didn’t have much depth or ride. His strength appears to be getting popups, but it comes at the price of lots of balls hit in the air and a higher home run rate.

Parker Bridwell

 Horizontal   Movement (in.) Vertical Movement (in.) GB% FB% HR/(FB+LD) Velocity  SwStr% Popup% Spin Rate Avg. Exit Velocity
4.67 -3.09 47.27 18.18 13.64 79.84 17.69 12.73 2805 84.20

Bridwell had a decent ground ball rate (just not near our sample) and a good whiff rate. His velocity is about average, but with that much horizontal movement he could continue to miss bats. Like Nova, he has a bit of a fly ball problem, even if he does induce popups. If he could keep the ball on the ground a bit more, he might have a chance.

Jimmy Nelson

 Horizontal   Movement (in.) Vertical Movement (in.) GB% FB% HR/(FB+LD) Velocity  SwStr% Popup% Spin Rate Avg. Exit Velocity
5.81 3.25 53.03 13.64 3.33 84.89 15.05 1.52 2555 84.20

Nelson throws his curve hard and has good horizontal movement, so he has a decent swinging strike rate that he might be able to sustain. While his vertical movement isn’t great, he gets a pretty good number of ground balls. His fly ball rate is favorable, and he keeps the ball in the yard, so Nelson has a chance at being a successful curveballer.

Robbie Ray

 Horizontal   Movement (in.) Vertical Movement (in.) GB% FB% HR/(FB+LD) Velocity  SwStr% Popup% Spin Rate Avg. Exit Velocity
0.27 -2.36 53.62 18.84 10.34 81.81 18.93 4.35 2038 86.20

Robbie Ray’s curveball is largely what turned him from a high-strikeout-bad-result pitcher into what he is now. His whiff rates are fantastic, and he throws his curve fast. He has a bit of a homer problem, but the new humidor at Chase might fix that. Still, he’s had tremendous success since he started throwing this curveball midseason 2016, so he should probably keep doing what works for now, especially since he doesn’t get a lot of movement.

Chris Stratton

 Horizontal   Movement (in.) Vertical Movement (in.) GB% FB% HR/(FB+LD) Velocity  SwStr% Popup% Spin Rate Avg. Exit Velocity
8.31 -5.42 60.87 13.04 0.00 77.7 16.29 8.7 3105 81.4

There isn’t a lot of data for Chris Stratton, but the data we have is quite encouraging.
His ground ball rate is over 60%, compared with a small FB%. And he didn’t give up a home run on the pitch all of last year, despite throwing it over 200 times. His velocity isn’t great, but he’s gotten plenty of swings-and-misses regardless. That’s due, in part at least, to the whopping eight inches of horizontal movement (compared to a pitch with no spin, of course). Only six other pitchers have more run than that on their curveballs. What’s more, he doesn’t have to sacrifice much drop—a little less than our sample, but still well over league average. He also has the least exit velocity among all our candidates and sample pitchers. For these reasons, he looks like a good bet to ramp up his curveball usage—he gets good results, and he has the movement to back them up.

We’ve identified some potential curveballers by evaluating them against other recent curveball throwers. From this cursory examination, it appears that Stratton is our best candidate. Jimmy Nelson is another viable option. Shields and Bridwell are fringier possibilities. With these results in mind, I’ll next examine these pitchers in more depth and reevaluate my conclusions. I’ll look more closely at pitch mix, mechanics, and other factors to see whether any of them could, in fact, become the next great curveballer.

Image via screengrab, “Lance McCullers Breaking Ball Mechanics,” Stats via Fangraphs, Brooks Baseball, and Baseball Savant.

Leave a Reply

Your email address will not be published. Required fields are marked *