Thursday, April 05, 2007

2007 Yankees to win 110 games, says math prof's simulation

According to math professor Bruce Bukiet, the Yankees will win 110 games this year.

Bukiet apparently uses a simulation to project the outcomes of games, and ran it on the 2007 season.

He also suggests a plan to reinvent the "win probability" approach:

Were the model to be commercialized, it could be updated on a play-by-play basis, which fans could monitor to see how every play changes the outcome of a game. “I think some fans would think that’s cool,” Bukiet said.
It's not possible to evaluate Bukiet's system from the article, but I don't think there's been any team in baseball history who would have been *expected* to win 110 games. The very few teams to win that many games mostly did so by luck. For what it's worth, here's my presentation where I found the best teams since 1961 only had 102-game talent (the 1969 Orioles and the 1998 Braves).

Also, TradeSports has the Yankees with only a 39-40% chance to win as many as 98 games, never mind 110.

(Thanks to John Matthew for the pointer.)

Labels: ,


At Friday, April 06, 2007 2:25:00 PM, Anonymous Anonymous said...

I like your presentation -- I'd seen it before, but not looked at it in a while. (Although I totally disagree with Jaffe's application of the method to managers.)

Question: did you ever try to identify team qualities associated with good/bad luck? For example, are fast teams any "luckier" than slow teams? Looking at the very lucky teams, 5 of the 6 were good-to-great basestealing teams (1960 Pirates the exception). Overall, they went 1016 SBs in 1343 SBAs, or 76%. SB% is a very crude measure of team speed, of course, but I thought that was interesting.

At Friday, April 06, 2007 2:37:00 PM, Blogger Phil Birnbaum said...

Thanks, Guy! No, never did anything to do with speed. I wonder if what some of what I measured as luck might be good fielding, assuming speed and defense are correlated somewhat ... the pitching luck didn't consider the defense behind it.

What is it about Jaffe's application that you disagree with?

At Monday, April 09, 2007 9:54:00 AM, Anonymous Anonymous said...

On Jaffe's manager analysis: I think we can be sure that three of the 5 measures -- pythag, RA +/- RC, and RS +/- RC -- are truly luck. Jaffe shows that long-serving managers do well on these metrics, but you can also show that winning teams will on average "outperform" on these measures, by about the same amount. And winning managers tend to keep their jobs. Also, I'm pretty sure someone did a study comparing 1st and 2nd-half of seasons, showing zero correlation for teams over/under-performing their pythag.

The other two measures -- players over/underperforming their true talent -- seem logically like they could be impacted by managers. But there are at least two problems here as well: 1) to the extent a player "overperforms" compared to surrounding seasons, it could as easily reflect good GM decisions on obtaining and releasing players; and 2) I think you acknowledge that these two metrics are somewhat crude. Still, if it can be shown that some managers "get the best" out of their players, that would be very interesting.

Jaffe does some interesting work, but I find his approach frustrating at times. If the data confirms his gut instinct, he presents the data; point out methodological flaws in his analysis, and he tends to say "look, I'm not a statistician, you can't hold me to those standards." My view is, in that case, leave the numbers out and just tell us you're giving us your subjective opinion. (On the other hand, he's very open about his methods and responds politely to criticisms -- seems like a nice guy.)


Post a Comment

<< Home