THE BOOK--Playing The Percentages In Baseball

That is super cool.

Ok, so to avoid the spam filter, I need a longer comment.  This is great, and it’s really too bad we’ll never get this data for the past (unlike PBP data, which some motivated folks could get from broadcasts...)

You know, maybe it’s not so bad.  We talk about the 30 frames per second.  And on a single frame look, you can be off by 10% in the speed of a pitch (was that pitch thrown in 10 frames or 11 frames?).  If you have say 100 fastballs thrown by video, then we’d be off by 1% in the fastball speed.

As we know, a fastball is thrown pretty much at 1SD = 1% mph (or something like that). 

Therefore, if we want to know how fast Charlie Hough or the Neikro boys threw their knucklers, couldn’t we get 100 pitches of theirs on video, count the frames, and estimate it that way?

Couldn’t we do the same thing with Feller’s fastball?

Yep.  More importantly, once Dickey gets to good counts, the fast knuckler is all he throws (well it’s like 80%+ on 0-2, 1-2, and 0-1 counts), maximizing his ground balls.  But on 0-0 counts, the slow knuckler is his main pitch. 

Of note, the velo-GB trend is true for nearly every pitch, not just knuckleballs.  Faster = More GBs.

The faster the pitch, the more likely the hitter is to swing late.  The later you swing, the more likely you are to hit the ball on the ground.

Both of these results are known, but I have found they are also borne out in the HITf/x data, which is something I want to publish on at some point.

The first effect is actually weaker than I might have imagined, but the second effect is extremely strong, in accordance with Matt Lentzner’s theory:

http://www.hardballtimes.com/main/article/why-flies-go-one-way-and-grounders-go-the-other/

A 30-something Dickey can throw an 80mph knuckler. When I was 20, I could throw an 80mph fastball. I don’t know if that’s awesome or embarrassing. (Probably both.)