Friday, November 9, 2012

Nate Silver Is Not a Witch

(Last post about the election this NaBloPoMo, I promise.)

So one of the meta-stories about the election has been the vindication of Nate Silver, the statistics/politics/gambling nerd behind the vaunted FiveThirtyEight blog. Long story short: some people thought his model was way off, but it was extremely accurate. It feels like the last day and a half have been all about how supernaturally spot-on Silver's predictions are, with the great single-serving site even changing its answer from "No" to "Probably" in the wake of his success. Everyone is convinced that Nate Silver is pretty much perfect. He can predict who the president will be and even which way every single battleground state will go.

But I wonder how people will react when Nate Silver is wrong on the outcome of a presidential election. It's not a question of if he will, but when. Ironically, his entire process is meant to tell us that, and even how often he'll be wrong. And yet I'm sure that once he predicts that Hillary wins over Huntsman (or whoever over whomever) in 2016 and gets it wrong, even if he told us he was only 54% sure, people will start discussing how his skills have atrophied and asking what's wrong with his model. I hope I'm wrong, but I wouldn't bet on it.

No comments:

Post a Comment