THE BOOK--Playing The Percentages In Baseball

One thing I’d love to see is if pitchers throw harder/get more break when facing guys like Rodriguez and Pujols, and cruise against, say, Adam Everett and Tony Pena.

For the opposite perspective from mine, here’s a question from Will Carroll’s chat today at Baseball Prospectus:
http://www.baseballprospectus.com/chat/chat.php?chatId=450

Joe (Washington, DC): Will: Do you know of any teams that are doing anything with the MLB pitchfx data?

Will Carroll: All of them are looking at it, but I don’t think it’s changing what they’re doing in most cases. Almost all teams were using advanced video systems, some more than others, and honestly, as good a PitchFX is, the eyes these batters have is better. I wish I could tell you the story about a player you’d probably say “he sucks” about seeing the most minute thing on video, something I could barely see in slo-mo, and being able to note it, make a decision, and hit the ball in a fraction of a second. These guys are all freaks.

I don’t know if anyone who comments here knows, but are major league baseball teams actually giving more than a cursory glance at the PITCHf/x data or employing anyone who would know how to make good use of it?  I’ve seen only a few very isolated indications that maybe a team is using the data and some very strong indications that most teams are not using the data.  I won’t go into the details of those because they were private conversations, but my overall impression, partly for lack of indications otherwise, is that MLB clubs are not using this data to any great extent, certainly not to anywhere close to the extent it is being analyzed publicly.

Secondly, it seems to me that while video is suited for some applications, the power of PITCHf/x data is that it can be manipulated as recordsets.  With video you have to look at one pitch at a time, and it’s awfully hard to realize in that context that a hitter is hitting .230 against the curveball down and in and .280 against the curveball away.  With PITCHf/x data in a database, you can look at every curveball thrown in a whole season, or whatever combination of factors you want, and those sorts of understandings can be developed very quickly.

I thought it was interesting that Brian Bannister said he preferred the PITCHf/x data to video.  I feel that way, too, but I’m surprised that more people don’t see it as more than just a neat little toy.

What Will is talking about is not what we’re talking about frankly.  His point regards one of mechanics.

PITCHf/x is about pitch and hit selection, sequencing and game theory.  The “game within the game” for lack of a better word.

Basically, we’re having two conversations here.

I’ve had conversations with analysts in three different front offices, and to some extent all three indicated that they were at least *thinking* about how they would use pf/x data.  What concerned me was that I gathered that as they begin to find it more and more useful, they will want to make the pf/x data private, and therefore unavailable to public researchers. 

I told them that I sure hope that doesn’t happen!

Okay.  I’m feeling a little less pessimistic today, with Dan Fox joining the Pittsburgh front office.  I know Dan has done a lot beyond just his PITCHf/x work, but I’d like to think that’s an important part of what he’ll be doing for the Pirates.  In any case, I’m excited for Dan.  (And maybe a little bit sorry that we have lost the voices of two PITCHf/x giants from the public arena in the last month.)

Mike, if you play your cards right, you and John Walsh might be next…

Remember, we are all part of the oldest profession in the world.  We’re just negotiating a price.