Not a bad situation to have as a pitcher. Wake-up expecting to spend your day as an Iowa Cub, end up pitching with a 12-0 lead as a Chicago Cub.
Mitch Atkins got his first inning of big league play, and our first taste of PITCHf/x data on the rookie right-hander. Atkins worked a bit in the new Yankee Stadium in an exhibition game, but the Gameday data was not released. So, three-plus months later, we finally get a look.
We only get a grand total of eight pitches, so nothing much can be concluded, but at least we know this much:
- Four-seam fastball, ranged from 89 to 91 mph.
- Slider, thrown 83 to 85, with decent lateral movement but little sink. One had enough back-spin to look a little like cutter
- Curveball, just one thrown, at 74 mph.
I think I read Atkins has struggled throwing a sinker, but we'll see if he's got one, along with a change-up he's willing to throw to a major league hitter, if he sticks around long enough.
In terms of PITCHf/x profile, Atkins throws a slutterish slider similar to Randy Wells' and a fastball that is similar to Wells' four-seamer in terms of speed and movement.
4 comments:
Hey Harry,
Jumping off of you last sentence there, have you done any work with similarity scores using pitch f/x or did you just eyeball your comparison? I've been thinking about making similarity scores my first big endeavor into pitch f/x (seeing if I can create an algorithm or something). Thanks
Steve (A Cards fan that doesn't mind sharing thoughts with a Cubs fan)
Hey Steve - Those were eyeballed, but, yes, that hints at something else I'm working on. Or thinking about, really.
Cool, I'll eagerly await anything you might do in that regard.
If I get a little free time sometime soon I may shoot you an email with what I'm planning on looking at, maybe there will be something valid in there that you hadn't thought of (unlikely, but ya never know)
Steve, I suspect you've put more thought into this than I have. Basically, the underlying keys, unless you have another way of looking at it, are getting pitches classified and park differences accounted for. On the other hand, I kinda have a feeling that could actually work both ways, as in classification and park correction aided by comparisons to known pitches. At some point, you need a baseline, and I suspect we have enough good data and classifications for that.
Post a Comment