FG on Fox: What Makes a Good Curveball?
There are all sorts of prescriptions for a good changeup. We talk about things like ‘velocity gap’ and ’tilt’ and ‘fade’ when we talk about how good a changeup will be. But what about the curveball? Can we spot a good one coming?
We can use our Arsenal Score work to identify good curveballs in general, just as we used to find bad changeups a couple weeks ago. Basically, by weighting grounders and whiffs on a pitch, we can assign it a score based on those easy-to-identify and important outcomes.
According to this metric, here are the best curves thrown at least 100 times last year.
| Pitcher | GB% | swSTR% | Arsenal Score |
|---|---|---|---|
| Brett Cecil | 56% | 29% | 5.34 |
| Mark Melancon | 55% | 25% | 4.19 |
| Nick Masset | 90% | 15% | 4.15 |
| Craig Kimbrel | 62% | 23% | 4.04 |
| Blaine Hardy | 91% | 14% | 3.75 |
| Vic Black | 78% | 17% | 3.72 |
| David Robertson | 60% | 22% | 3.69 |
| Carlos Martinez | 53% | 24% | 3.50 |
| Cody Allen | 53% | 20% | 2.35 |
With a phone full of pictures of pitchers' fingers, strange beers, and his two toddler sons, Eno Sarris can be found at the ballpark or a brewery most days. Read him here, writing about the A's or Giants at The Athletic, or about beer at October. Follow him on Twitter @enosarris if you can handle the sandwiches and inanity.
Even when there’s no big, new revelation, it’s still cool to see ideas that we know (or at least suspect) to be true demonstrated with shiny numbers. Good work. Go math!
Does spin rate matter or does that pretty much mirror break?
The spin rate coupled with the velocity are what determine the break on the curveball. A good curveball is location and variance from the fastball while being so subtle hitters can’t pick it up. Spin rate is very important.
This is true, but spin rate in PFx is calculated by velocity and break, so in my opinion it’s better to use velocity and break. If we had trackman data, where it’s observed, I would have used it.
Apologies – my comment was a bit vague. What I meant to ask was if spin rate provided new information on effectiveness (beyond what velocity and break provide)
I wasn’t aware the Pf/x spin was just backed out of velocity and break, I thought it was measured.
Thanks for the responses guys.
I don’t watch them as often as, say, Charlie Morton, but Kershaw and Wainwright have well-known excellent curveballs, but did not appear on this list. From what I’ve seen of them (I try to catch games Kershaw pitches as often as possible, as well as Kluber, but less so Wainwright), a lot of their curveball effectiveness comes from “freezing” hitters as opposed to generating whiffs and grounders. They do those two things, too, I’m sure, but the real value, I expect, is in setting up other pitches, and either getting ahead in the count, or freezing a hitter with two strikes.
Obviously, it’s impossible to truly accurately evaluate everything, but I do think that, at the very least, getting a called strike on a curveball in a two-strike count might be worth considering, since that’s the count a hitter is probably least likely to take it, and therefore when the pitch’s quality and deceptiveness can be best measured.