Nate Silver has been a lightning rod for controversy over the final weeks of this campaign, with arguments from the right that his model is somehow biased in Obama’s favor. We’ll see tomorrow. But in terms of arguments about Silver’s statistical chops, I thought this piece by Colby Cosh for Maclean’s was pretty interesting. In short, before turning his attention to politics, Silver’s claim to fame was a model for projecting future player performance in baseball, and in hindsight, it wasn’t as good as is widely believed. Some potential solace for those of you hoping Silver’s model has the presidential election wrong.
(There’s also the fact that Silver’s electoral college model could be exactly right and Romney still wins — if I say your odds of tossing a six on a die roll are only 17 percent, and you toss the die and hit the six, it doesn’t mean my odds were wrong. And Silver has Romney’s chances at just slightly lower than 1-in-6.)
★ Monday, 5 November 2012