Wednesday, October 31, 2012

Good post by Ezra Klein on Nate Silver's less impressive critics. I think there are a variety of criticisms one can legitimately make of Silver's methods. Maybe he's overfitting, and maybe he sometimes misapplies things like the convention bounce corrections from September. But a lot of the criticisms that traditional media analysts like the people at Politico make of him really amount to an inability to understand probability and its applications to life.

The craziest Nate Silver opponent out there is Dean Chambers at right-wing site, who is rejiggering all the polls with party ID numbers that fit his estimation of how many Republicans there are in the electorate, and getting results where Romney wins the election with 321 votes in the Electoral College. He criticizes Silver not only as biased, but as "a man of very small stature, a thin and effeminate man with a soft-sounding voice." This leaves me wondering if conservatives have different mathematician stereotypes than I do. Do they expect them to be large burly men with deep voices?

A clever business opportunity for anyone interested in such things would be to grab the domain (it appears to be still available) and use it to sell non-FDA-approved medical products to the sorts of people who read Mr. Chambers' website. I'm sure there's quite a market out there for medical advice that isn't skewed by the biases of government regulators and academic scientists.

No comments: