Thursday, October 14, 2010

Bill James and the Simpsons

Over at SI.com Joe Posnanski analyzes Sunday night's Simpsons baseball episode, which cleverly took on the supposed sabermetrics-traditionalist divide.

Even in making fun of statistical analysis, the episode (inadvertently?) got at a larger truth: stealing bases is actually pretty dumb much of the time (not always, just much and possibly most of the time).

10 comments:

El Aguila said...

I hate sabermetrics. Geeks have taken all of the fun out of playing and watching sports. You should read _Opening Day_ by Jonathan Eig. In it he has an intriguing description of how just the threat of stealing when he was on base, made Jackie Robinson the most valuable player for the Dodgers in his first year in the majors. Fantasy sports also sucks by the way. Unless, I could only choose my entire team, why would I put myself in a position where I had to root for anyone but the Cowboys, Dodgers, Lakers, etc.

dcat said...

I don't understand this argument about "taking all the fun out of watching sports." I don't understand how replacing bad statistics with good ones impacts one's ability to enjoy the game. On a simple basis, OBP is a better statistic to understand hitter effectiveness than batting average. Wins are a terrible way to understand how good a pitcher is. We already have and use statistics. Why not have and use better ones? If baseball were not a stats driven sport, I'd take your point. It's just that too many people look at bad stats or stats that do not tell us all that much.

Statistical analysis has proven especially effective in helping teams level the playing field and exploit inefficiencies in the marketplace. this is where Moneyball is so great -- it helps us to understand why those A's teams were able to compete for so many years despite having half the payroll of the Yankees.

I've read Eig's book. I've read a couple of his books, actually (I like the bio of Gehrig). But Jackie Robinson is an outlier and in any case, the argument against stealing bases is that it too often gives away outs, the most valuable commodity in baseball. Robinson averaged just under 20 steals and 3 caught stealing a year, so any sabermetrics person would not have an issue with him stealing base. The best measurements indicate that anything less than about a 70% steal percentage is costing your team runs. Jackie Robinson stealing is good. A stolen base for the Red Sox helped them win in 2004. But on the whole, between stealing and not stealing (and don't get me started on bunting) I'd generally prefer not to.

I like fantasy but I also have two rules -- there are teams I will not own players from (Colts, Jets in football, eg; Yankees in baseball) and I don't play someone on a day they are playing the Pats or Sox.

dcat

El Aguila said...

Figuring out ERA and batting averages were easy and enjoyable when I was a kid. But with today's emphasis on more complex stats, I don't want to have to carry around a calculator in order to enjoy a ballgame. And arguing over statistics as opposed to the actual ballgame is frankly, well in the words of John Bender, "So it's sorta social, demented and sad, but social. Right?"

I don't need a statistical argument to tell me who to root for. If you have to revert to statistics to defend a player, then the player probably isn't deserving.

For example, given baseball is a uniquely stat driven sport, but the Giants suck and the Yankees can go fuck themselves. That's it, life is just that easy and so should watching sports.

dcat said...

I think that most of us who care about stats do so for our own enjoyment. But there are times when stats folks conflict with others -- strategy and scouting might be one.

Awards are another. Felix Hernandez was the best pitcher in the AL this year. Period. But people who think that wins matter will give the award to CC Sabathia simple because of wins, as if wins are not a team stat rather than a pitching stat (quick, which center fielder led the league in wins this year?). Seattle was terrible and they simply did not score runs. But Hernandez was the best pitcher by any real measurement. And yes, statistics like FIP and ERA+ and WHIP help us to understand why.

I don't need stats to understand who to root for. That's pretty clear. But those stats help me to understand the game better, especially when I'm not watching (in other words, this discussion isn't happening at the expense of a game, as no games are being played now.)

The Yankees do, indeed, suck. We're all Rangers fans this month.

dcat

Paul said...

Can any casual fan really tell me what VORP, BAPIP, and the others are during a game without distracting me from actually watching the game? Let the managers (lineup, defensive positing, pitching match-ups) and advanced scouts [finding fat shlubs who can walk and hit balls into the gap (McCann)] take care of this and let the casual fan enjoy the aesthetics of the game.
And yes, the Yankees suck. Dumbass.

Haven't read all of it yet, but this looks interesting:
http://joeposnanski.si.com/2010/10/13/the-baseball-playoffs/?eref=sihp

dcat said...

A good announcer could, I think, explain these things. Listen to your typical broadcast and there is a lot of inanity. I would imagine that there would be time to explain that VORP simply places players' values relative to a "replacement player" and that BABIP basically is a way to try to assess the role of luck.

Here is how I would do the latter: I'd talk about how baseball does have an element of luck -- think of all the times a guy has drilled a line drive directly at someone, and then how many times a guy has squirted a dribbler through. That is a function of a lot of things, but as much as anything it is a function of luck. Over the course of a season you'd expect that to even out. And in most cases it does. I wouldn't use a term like "revert to the mean," but that's what you expect to have happen. A guy with a BABIP of .380 has been exceedingly lucky. A guy with a BABIP of .218 has likely been incredibly unlucky -- especially if he is not otherwise hitting .380/.218.

Look, I know it's a hard sell. But cut a brother some slack on a Friday afternoon, turd for brains.

dcat

Paul said...

I started reading firejoemorgan a few years ago and I appreciated the statistical analysis and learned a lot from it, but, damn, sometimes I just want to have a game on and not have to do too much thinking. You're right, good announcers can make it simple to understand, but the actual breakdown takes too much time to process. Joe Buck is NOT that announcer, though; he's proof of the limits of genetics. Mccarver is smart but his attitude turns me off. (You're a big man, Deion!) How a great announcer like John Miller puts up with Joe Morgan without losing his mind is beyond me. Joe Simpson is excellent. Did you happen to catch The Tenth Inning by Ken Burns? I'm half way through episode one and it's fantastic. Moron.

dcat said...

Paul --
FJM was genius. I miss it daily.
I listen to a lot of games on radio because I have XM. In all honesty I don't care of they incorporate a lot of sabr type stuff. I just care that they not be hostile to it. I'm certainly not the perfect stats guy -- I think something like wion shares is pretty imperfect inasmuch as it attempts to attribute shares of a win to guys in ways that seems as flawed as rbi's, which while flawed are at least reflective of a team-related sport. I understand why rbi is imperfect, but I also think trying to isolate guys from the contingency of a fundamentally team sport is itself flawed. Why yes, a guy gets rbi based on guys before him being on base in the same way that a quarterback throws touchdowns to another guy on the field. That does not make the metric of the touchdown necessarily flawed, or at least not fatally so.
As I write this: Rangers 5, Yankees 1. Glorious.

dcat

dcat said...

And then a while after I wrote that the Rangers pissed down their legs. Ugh.

dcat

Chetmeister said...

God, I hate baseball. Thank god we've got a couple of good months for football still.