I'd say FiveThirtyEight was pretty good

Posted by Ryan Jerz on Tuesday November 11, 2008.

Nate Silver took the polling world by storm during this past election with FiveThirtyEight. In a nutshell, he applied similar statistical analysis and predictive analysis to that which made him a well-respected baseball analyst to polls surrounding the presidential election.

The effect his site had was put perfectly by John Gruber:

A lot of people asked me over the last few weeks why I was so certain Obama would win, or why I wasn’t a nervous wreck (especially for the brief period in September when McCain took a small lead in most polls). The answer is Nate Silver. Or, more specifically, Nate Silver’s math.

The election was never in doubt for me. No poll showing a shrinking lead was going to make me start thinking one way or another about it. Until Silver said it was going to happen, I was positive it wasn’t going to happen. Even with the relative closeness of the popular vote, Silver had it pegged that Obama would dominate the electoral vote and the win would never be in doubt.

With this kind of success comes many imitators, so I’m predicting that there will be a shift in the next election and subsequent ones toward a more predictive model as opposed to the horserace daily poll grabbing the headlines. Nate Silver did all of us a favor by kicking ass. He won’t put pollsters out of business, but he should make them re-evaluate how they operate.

I just hope somebody takes the time to put together a local copycat that works. I saw a couple of people I know who ran for office here looking worried. Of the three that were scared, only one had a close-ish race, and the he still won. The other two crushed their opponents, and were needlessly freaked out.

Ryan JerzRyan Jerz is an all-around good guy who wants people to eventually refer to him as "that dude who climbs mountains."

Archive | About