Pitchf/x Page Fixes
Just some quick notes on the 2010 pitchf/x data:
– There appears to be some new pitch type categorizations that were breaking some pitchf/x pages. There is EP (Eephus Pitch), SC (Screwball), KC (Knuckle Curve) and FO (Forkball).
– SC and FO I have temporarily lumped into FT (Two-Seam Fastball). I’ve been explained that screwballs (and possibly forkballs) are not really two seam fastballs, but there have been a total of 19 screwballs thrown this year, 18 of them by Daniel Ray Herrera and then the other one was thrown by Dallas Braden. There have been 12 Forkballs thrown, 8 by Livan Hernandez and 4 by Kenshin Kawakami. Overall, these two pitches didn’t seem to warrant their own category, but I can be convinced otherwise.
– Eephus pitches (EP) I have lumped into the UN (Unknown) category. Vicente Padilla has thrown all 16 of them.
David Appelman is the creator of FanGraphs.
David, I’m curious why you chose to lump SC and FO with FT. In terms of velocity, break and usage I would think they’d be more similar to CB or CU. Or is that assumption wrong?
A screwball breaks the opposite way from a curve or cutter.
And I’ve noticed Aaron Harang has 28.6% of his pitches listed as unclassifiable. He has not changed his repertoire unless you consider his slider’s severe degradation. Bad data? Bad classification?
Harang pitched last night and the BIS data runs 1 or 2 days behind in terms of pitch type classification. Those will all turn into “known” pitches tomorrow or the next day.
Please note that the data on the player pages is BIS data and is in no way related to pitchf/x data. Pitchf/x data is only displayed in the special pitchf/x section.
Yeah we chart pitches next-day, hard deadline of 6:30 p.m. eastern. Not sure how long it takes Fangraphs to process that and update their databases after that.
Oh, and one other thing to keep in mind is that on the player pages, the BIS pitch type %’s are displayed as a percentage of known pitches and leave out unknown pitches entirely when calculating the %’s. So even if you see 25% of unknown pitches, the sample size in the known bucket tends to be large enough where the numbers would still be fairly accurate. (This may not completely be the case at this point in the season, but definitely by mid-season).
Yeah, caught that they weren’t part of the classified percentages when crunching some numbers earlier. Good to have it confirmed.
Regarding the unknown % from the BIS data, it was there the other day, prior to yesterday’s start as well. Perhaps I noticed them immediately after his prior start? Am I correct in understanding that there will always be a 1-2 day period after each start where the pitches are listed as unclassified? If so, perhaps it would make sense to delay their display until the classifications come in?
Still interesting to see Harang at 18.9% two-seamers in the pitchf/x data considering he’s basically never thrown them before, just 2.0% last year. They do seem to be coming directly our of his 4-seamers, not surprisingly.
I guess I never noticed the separate pitchf/x section of the player pages. Too many improvements to keep up! Thanks the clarification.
Well, with the way the % are being calculated it doesn’t make a difference whether they show up there or not.
For instance, if I decided to not include any of the unknown pitches in the data the only thing it would change is the % of unknown pitches because they are excluded entirely from the pitch type percentages. All the data would remain exactly as is.
So, you can either know that a pitcher still has some pitches that need to be charted (in Harangs case 20-some %) or unknown pitches could always be labeled as 0% (or the column removed at entirely).
It’s my understanding that the XX classification is actually an uncharted pitch and not an unknown pitch, maybe that just needs relabeling. It should not be mistaken for some bizarre pitch a pitcher is throwing that can’t be figured out.
My understanding has always been that a forkball is very similar to a splitter, and a screwball is a changeup with a lot of horizontal break (moves down and in on same-handed hitters)
I guess eephus pitches could be considered very slow changeups…
For the sake of accuracy, and my love of uncommon pitch types, I think all 4 warrant their own category.
Herrera’s screwball is a slower changeup with extra drop. I’d put it either with the changeups (probably) or with the curveballs (probably not). A 65-mph pitch shouldn’t be lumped in with fastballs. Same thing for Braden’s screwgie, if in fact he’s thrown it in the majors.
The forkball also should either go with the changeups or the splitters. It’s just a slower-spinning changeup. Some splitters are that also–slower spinning changeups–and some are thrown with more normal spin rate. So I’d probably put it with the changeups, but I could see the logic behind putting it with the splitters, too.
Mike, thanks. I’ll put the screwballs in with the changeups starting tomorrow.
It’s more similar to a change-up, and that certainly works (agree that there just aren’t enough to need a type), but I wouldn’t really say Herrera is just throwing a variant. Don’t know for sure, obviously, but a screwball can definitely be a distinct pitch…
Mike and TCQ have the screwball right. When it was taught to me (as a lefty), I pretty much re-learned my stuff to have a cutter, 2-seamer, and screwball right “next to” each other on the ball.
Start with a pure two-seam grip. Ball is “vertical” in the hand with the two seams straight in line with the index and middle fingers. That’s the fastball, with some down-and-away tail against righties.
Rotate the ball slightly clockwise so that the middle finger crosses the left seam. Same arm and hand positioning. Same release, except pressure the ball with the middle finger, almost as if it’s trying to snap the thumb. As a lefty, this pitch will break left-to-right, in on a righty. Cutter.
From the original two-seam grip, rotate the ball an equal amount counterclockwise as was done clockwise for the cutter. The 2-seam tail, I found, increases with this grip change alone. But it’s an off-center position. The ball needs added support to maintain control. Swallow the ball slightly deeper into the hand, with the thumb-to-index “C” fully in contact with ball; and allow the loosely curled ring finger to provide some tension and support on the lower back corner-part of the ball. The necessity of support, which by definition adds surface area contact in the hand, causes the desirable byproduct of change-up velocity from that initial drag. Release is pretty much the same, except for the little bit of wrist rotation that, more than anything else, frees the ball from the hand. Screwball.
Guile was pretty much all I had. And having, essentially, one pitch that operated at 3 speeds with 3 flight patterns coming from one release point allowed me to just throw without worrying about tipping or doing anything different. Comfort on the screwball was always a little less than ideal, but it sure was more effective than anything else I could pull off without tearing up my arm (dear lord, I still have nightmares about the sidearm eephus-like slurve I used on lefties…the pain…)
Now if only somebody would start using a foshball again…Keep the spirit of Jason Bere alive.
TCQ, I agree. It functions like a super-slow changeup for Herrera, but presumably he’s twisting his wrist on release, which is not a changeup feature, of course.
Sometimes I think Pedro’s changeup looks a little like a screwgie. I’d love to see what some of the older screwball pitchers would have looked like on PITCHf/x. Mecir just missed out. Braden I’ve never seen evidence that he threw one in the majors, although I haven’t looked recently.
Well, a screwball is kind of just an exaggerated (very exaggerated, of course) changeup in wrist action, so someone like Pedro could easily have thrown something that would function as a cross – which could be a really nasty pitch, I’d bet.
The “eephus” that Padilla is throwing is really just a big slow curve. I guess MLBAM’s neural net is matching it to the El Duque eephus because of the slow speed, but even a 60-mph slow curve isn’t quite as slow as the 50-mph El Duque eephus pitch.