So often does the news feature pompous, frustrating, or otherwise regrettable individuals that sometimes we bloggers find ourselves, by no choice of our own, down there in the mud with them — slinging, squishing, and building little huts of ridicule to crawl inside.
It’s as refreshing as a premoistened towlette, then, to find someone we can point our dirty fingers to and say, “Ahh yes: This is someone we can easily admire and recommend.”
Nate Silver is one such person. The thirty-four year old baseball freak-turned-election year Nostradamus runs FiveThirtyEight.com, a blog that has lived at the top of every political junkie and junkette’s Favorites List since Silver began it in 2008. (Catonsvillains will note UMBC’s own Dr. Tom Schaller is a contributor.)
Five hundred and thirty eight is the number of electors in our country’s never-a-problem electoral college. Zero is the number of people more adept than Nate Silver at supposing the outcome of their votes.
FiveThirtyEight’s recipe for calling elections is simple but yields unbelievably rich results: Silver takes a wide assortment of polling data along with other factors relevant to the contest at hand, no matter if it’s at the district, state, or national level; he mixes all those variables together in a pan and lets them bake at 350 degrees in the comprehensive statistical models of his own design; and then he takes a slice, gives it a taste, and shares what he’s got.
Sometimes the result comes closer to devil’s food and the Democrat is more likely to win; sometimes it’s closer to red velvet and the Republican is the safer bet.
No matter what he dishes up, though, Nate Silver is a very, very good baker. His predictions on who can expect to win where in any given election have been startlingly accurate (though figuring out why is always more complicated).
After his nearly flawless call of the 2008 election, the New York Times was sufficiently impressed to hire this former full-time online poker player as their fair-haired boy in the numbers department. (Yes, that’s right: Silver used to be a professional gambler. It wasn’t until a few years ago, when Congress started moving to contain the online gambling industry with legislation, that he turned his eye to the complexities of how those do-gooders get elected in the first place.) He is one of the best examples of how someone can go from blogging on his living room sofa to getting paid by the New York Times to blog on his presumably much nicer living room sofa.
Silver recently chatted with the USDemocrazy team over enchiladas at an undisclosed Mexican restaurant in suburban Baltimore. He talked about a lot of things; he’s an affable guy — not one of these people who loves to hear himself talk but someone who just has a lot to talk about.
We asked him about his new book set to be released in September by Penguin under the simultaneously dramatic and anti-dramatic name, The Signal and the Noise: Why Most Predictions Fail — But Some Don’t. The title, he explained, comes from the tendency of pro predictors like himself to follow the wrong leads, identify patterns where there are none, see noise and mistake it for signal.
Slver shared
“In general, people overstate how well they can predict the future and how much knowledge they have about how they can apply from the past to the future.”
And in light of all the attention surrounding FiveThirtyEight and statistical projections, what about the future of predictions themselves? Are we relying on them too much?
“If people want more accurate predictions then they have some hard work to do, because that often means developing theories about the way things really work… It’s not just crunching numbers.”
We weren’t going to argue with him. After all, the numbers seem to be on his side.
