Handicapping the handicappers

From Dynamical Bias in the Coin Toss, Persi Diaconis, Susan Holmes, and Richard Montgomery. Via Coding the Wheel.

Dylan Byers at Politico tries to bite a piece out of Nate Silver's trousers:

Prediction is the name of Silver's game, the basis for his celebrity. So should Mitt Romney win on Nov. 6, it's difficult to see how people can continue to put faith in the predictions of someone who has never given that candidate anything higher than a 41 percent chance of winning (way back on June 2) and — one week from the election — gives him a one-in-four chance, even as the polls have him almost neck-and-neck with the incumbent.

And "more than a few political pundits and reporters, including some of his own colleagues, believe Silver is highly overrated." The "some of his own colleagues" that he goes on to quote are David Brooks, famous mathematician, and the "reporters" are former congressperson Joe Scarborough, which adds up to somewhat less than "more than a few". All two of them are Republicans, too. I guess if he quoted more the column would have gone over length. Anyway, they are not very well informed about probability.

I think I know enough about how Brooks thinks: if the poll says the candidates are at 50% for A and 48% for B, that's pretty close, right? So A probably has a 50% chance of winning or thereabouts, because you have to do some magic with the number first, so it won't come out exactly the same—but it'll be pretty close in the same way. That's just common sense!*

The implication is that there's something like dark magic involved in Silver's calculations, or they wouldn't be giving him something so far off the common sense mark. Either he doesn't know what he's doing, or he's manipulating them in Obama's favor out of secret liberal bias, as an underhanded way of influencing the voters.

And is that why Danny Sheridan, football handicapper from Mobile, Alabama, gives the election to Obama 2:1? No, actually.

Obama, he said, was favored by 6:5 to win Colorado, by 9:5 to win New Hampshire, by 2:1 to win Iowa, by 3:1 to win Nevada, by 3:1 to win Ohio and by 7:5 to win Wisconsin.
Romney, he said, was favored by 2:1 to win Florida, 10:1 to win North Carolina and 7:5 to win Virginia.

"I may not vote for Obama, but I still think he's going to win the electoral college," Sheridan said. Sheridan said it was "even money" -- a toss-up -- as to which candidate wins the popular vote.

Romney's 10:1 odds in North Carolina don't mean he's going to get 90% of the vote there, either. It means that if you ran the election in eleven more or less identical universes the Romney would win in all but one of them, at around whatever shares of the vote are expected (Silver says the vote will be relatively close at 51.3:48:3, but his estimate of the odds is 82:18).

Danny Sheridan is entitled to disagree with Silver if he wants, because he knows what probability is. The self-denominated pandits should really make an effort to find out, and stop embarrassing themselves.

*Or an intuition, born out of math anxiety; the main thing is to minimize the number of numbers he has to contemplate.

From Image-Archeology.