THE BOOK--Playing The Percentages In Baseball

Actually, is the speed of the pitch when it crosses the plate the most important?  I used to think so when pitch data on gameday started to come out, but now I’m not so sure. 

Maybe it should be at the beginning, since that represents the speed the ball is going when the hitter needs to identify that it’s a fastball.  Or maybe it’s the time the ball takes to get to the plate, so it’s somewhere in between the two?  It’s probably an interesting question for people who actually do the PitchFX stuff-- which version of speed is most relevant?  Obviously this is only possible if the ball slows down at a faster rate for different pitchers.

Well, what counts is when the batter makes the decision.  So, rather than “pitch speed”, I want to know “reaction distance”: WHERE is the ball 0.25 seconds prior to crossing the plate.

My larger point though was that some people report pitch speeds at 50 ft, or 40ft, or 30ft.  But, we can all get behind reporting the pitch speed at 0 ft.

Yeah, I definitely got your larger point.  I think it’s probably an interesting question to consider, but clearly people want to see 100 MPH so it’s moot. 

I didn’t know about the 0.25 second reaction time.  Is that a common trait among major leaguers-- the ability to recognize a pitch and react within a quarter of a second?  Wow, it’s amazing how far I must have been from being a major leaguer!

I’m not sure what that chart is that you show in the post. I checked it for a few different pitchers and it’s the same for everybody. You can’t see flightpath for individual pitchers (like the graphs Harry Pavlidis does). Maybe I’m missing something.

Cool site overall though.

Neat!

I like the R implementation. I use R for NLME and some other stats but just really started getting into it, maybe I will write this guy. =)

And, the interesting thing about the 250ms reaction time is that’s actually quicker than most people can tell you the color (red or green) of an square, presented at fixation, with a simple keypress (one button or the other). Even after 100s of practice trials.

Yeah those visualizations are really interesting, but there are couple little issues.  The second set of figures is titled “Location and velocity by pitch type,” but I think it shows break and velocity.  Not that break and velocity by pitch type isn’t interesting, it is.  And then at the top of Beckett’s page that Tango linked it says “PitchFX data exists for 571 pitches thrown in 2008.” That is only like 6 games worth.  Between 2007 and 2008 I have over 4500 pitches for Beckett.  I don’t know if they are working off an incomplete data set or maybe it is not practical to make those images off lots of pitches.

Those reservations aside, it is an incredible resource that complements the existing pitchf/x web tools of Brooks and Kalk.

In some of the first work done on pitch f/x data, Dan Fox showed how the difference between the initial ball speed and the ball speed when it crosses the plate is highly variable due to elevation and weather (+-1 mph).