July 9, 2011
SABR 41 Live Blog, Day 3
2:20 am: At the Retrosheet meeting right now - this is like crack for baseball researchers. The big announcement is "deduced games," games where play-by-play accounts have been reconstructed from things like newspaper accounts. The release comes at an awkward time for me, as I'm on the road. I don't know how those deduced games will be handled by BP going forward, but I think that will give us a lot of new material to work with. -- CW
12:24 am: In the Statistical Analysis Committee meeting right now; Michael A. Humphreys is giving a presentation on his defensive metric, DRA. My laptop is about to die, so live-blogging will be light to nonexistant right now. I have Humphreys' book back at the hotel and will have a review available sometime next week, hopefully. -- CW
11:46 am: I guess my fundamental problem with Gennaro's rating system is I don't know what he's evaluating; I can't tell if he's trying to come up with a retrospective or a predictive valuation system.
Now we're getting into Game Score, what Gennaro calls a "nice, simple system." It's also vastly arbitrary. I know we like to mock things like OPSBI, but my impression is that what Gennaro is doing is just a much more complicated gloss on that same sort of "numbers stew" approach.
So let me go off on a tangent - I think a lot of people, including people in SABR, tend to conflate SABR and sabermetrics. SABR memers can often be sabermetricians and SABR was a great asset to sabermetrics in its infancy. But nowadays most of what occurs in SABR is not really sabermetrics - there's a lot of baseball study that isn't baseball analysis, of course. And most of what's happening in sabermetrics is happening well outside of SABR. As I mentioned on Day One, SABR is an organization at a crossroads; they are going to literally die off if they can't get younger people interested in the group. I get the sense that many SABR members think the challenge is getting young people interested in baseball research. Baseball research, however, is as vibrant as it's ever been. The question is how to get SABR relevant to baseball researchers. --CW
11:36 am: Vince Gennaro, SABR president, is talking about a pitcher rating system of his own devising. He's talking about the difference between what he calls "outcome" and "process" measures.
Now he's listing seven "success factors": command, stuff, velocity, batted ball tendency, durability, versatility and consistency. So far no statement on why he came up with these categories or how he related them to how a pitcher helps his team wins games. He says consistency is overlooked; it may be fairer to say it's overrated. If you look at it, an inconsistent pitcher is likely to be more beneficial than a consistent one, in terms of giving a team chances to win games (you can look at this crudely with things like quality starts, which Scott Boras got into in the the keynote).
He's talking now about valuing ground balls versus fly balls - I think we all know how I feel about this. He's talking about pitching deep into games, which is something we've incorporated into WARP. He's talking about credit for narrow platoon splits - why? Why should I possibly care about this? And then consistency (see above). --CW
11:05 am: David Smith, one of the greatest heroes of baseball reserach, is looking at changes in the size of pitching rotations over time. Massive shoot-up until 1910, and then it flattens off quite a bit since then. That's rather stunning to me.
Also: just saw video of Jeter's 3,000th hit on MLB.com. Living in the future is awesome. --CW
10:55 am: The umpiring study looked at all pitches, not just borderline pitches - in order to establish any umpiring bias you need to look at how a batter is reacting to pitches as well. If you look at the aging curve for hitters, they get more selective over time in terms of pitches they choose to take versus swing at. The study really doesn't allow us to separate batter selectivity from an umpiring bias. --CW
10:40 am: Now there's a presentation about how umpires behave based on batter experience. A lot of methodological notes up front, now we get a giant table that I can't hardly read. Well, I mean, I can read the numbers, I'm just trying to figure out what they're trying to tell us with this table.
Okay, now some explanation - they're saying veteran batters see fewer "false strikes," while veteran pitchers get more "false balls." Without having any sample sizes on this, I'm rather unpersuaded. --CW
10:26 am: Sean Forman of Baseball Reference fame is sitting two rows ahead of me. David Appelman of Fangraphs is sitting directy behind me. (Hi, David!) It feels to me like we should have some policy where one of us is mandated to stay at a remote location like NORAD as a precaution.
(Also, if some sort of dance fighting breaks out between me and the Fangraphs contingent, I'll make Larry Granillo take pictures for y'all.) --CW
10:18 am: Phil's theory is that HFA comes from a biological response to protecting one's home, and thus an increased testosterone production or something. I am, let us say, unpersuaded. Birnbaum does say that it's a theory, and he doesn't have the resources to prove it. And I don't know how to disprove it, either - unless you just want to get tissue samples from ballplayers at both home and road games, really not much that can be done to figure it out.
Mentions other theories - Craig Wright in "The Diamond Appraised" lists a number of causes, "home cooking" being one of them.
My personal feeling is that learning (or familiarity) has a larger share of it. HFA on things like triples, where familiarity plays a greater role, is larger than it is on other things. Phil of course knows this, though, so I may try and find him later for more of a discussion on this. We all pretty much agree that "Scorecasting" is wrong, though.
(And a shout-out at the end to BPer Mike Fast for help with the presentation.) --CW
10:08 am: Wi-fi is back on! Have an exciting day of presentations and meetings today (if you're here, come look for me in the Catalina room throughout the morning).
Listening to Phil Birnbaum talk about causes of home field advantage. This is somewhat old hat so far for me, as he's repeating a few things he's been saying on his blog (if you're not reading it, you should be). He's talking about the findings in "Scorecasting," where they found that home field advantage is almost all attributable to umpiring. This is, to put it bluntly, total bunk, and Birnbaum is getting into some other possible sources for HFA as well.
Also, if you want to follow along at home, his PowerPoint slides are available. --CW