Six months ago David Brooks, Joe Scarborough, and company thought: "Mitt Romney is behind and has an uphill climb. But if we pretend that Romney is actually ahead, maybe he will get momentum and win." And so they launched the War on Nate Silver, seeking themselves and asking others to help them discredit the most prominent aggregator of state polls, Nate Silver.
Now, six months late, Gregg Easterbrook, suffering from trench foot and beri-beri, wanders out of the forest waving his katana, and seeks to join the effort, accusing Nate Silver of "spurious precision."
The remarkable thing, of course, is that on Election Day--when other poll aggregators were saying that Romney had only 1/200 odds of pulling his chestnuts out of the fire--Nate was saying that Romney's chances were about 1/6 because there was about a 1/3 chance that his model was wrong. In retrospect, Silver understated the precision of his model and the accuracy of his W-L calls. Nate Silver is the last person who makes his calls with "spurious precision"--and that Easterbrook thinks those are the grounds on which he should attack Silver speaks volumes…
Tuesday Morning Quarterback returns to mock the NFL mock drafts - ESPN: Absurd Specificity Watch: Americans seem to love hyperbolic claims of precision -- perhaps it makes us feel that science is more efficient than it really is. When Nate Silver of The New York Times forecasts, as he did on the morning of the 2012 presidential voting, that Barack Obama will win re-election with "314.6" electoral votes to "223.4" electoral votes for Mitt Romney, such numbers are received with gravitas -- as if the decimal places made them deep, rather than silly. In just two days, Obama's chance of re-election increased from "80.8 percent" to "83.7" percent . A claim of a "83.7" percent chance rather than "a good chance" is seen as turning the speaker into Mr. Spock, when actually ought to make readers giggle.
Nate Silver had a simulation model of the election. He fed it state poll numbers, drew a bunch of random variables, simulated the election, and did it over and over again with new draws from his random variables. The result on election morning was that 837 out of 1000 times Obama won. When he had done the same exercise two days earlier--with the polls as of Sunday morning--Obama won in his simulation model only 808 times out of 1000.
Reporting the results of these exercises--that election morning Obama won in the simulation 837 times out of 1000, 29 times more than two days earlier--makes Gregg Easterbrook "giggle". Why? Because MATH! And we shouldn't calculate sample averages because to do so is "silly" and creates "hyperbolic claims of precision"!
Gregg: the entire point of the War on Nate Silver was that David Brooks, Joe Scarborough, and company knew what they were doing were willing to burn their credibility in front of the statistics-literate community in order to try to push the election toward their desired Romney-wins outcome. And here you are burning your credibility in front of the statistics-literate community… for what? You are six months late. Why?