Politco

Defending Nate Silver and Math

As the November 6 election approached, a lot of commentators trained their fire at Nate Silver, a numbers guy who writes the FiveThirtyEight blog for The New York Times. Critics said Silver was crazy for projecting an 80 percent chance Barack Obama would win re-election.

Faux conservative commentator Stephen Colbert hosted Silver on his show and pooh-poohed his prediction that the election wasn't a "coin toss." Dylan Byers, media blogger at Politico, said Silver could become a "one-term celebrity."

Post-election, Paul Raeburn has blogged that critics who sniped at Silver seriously misunderstood what he does. Silver isn't a pollster. He is a statistician who uses polls and other data to model projected outcomes.

Raeburn says Silver's work should be compared to weather forecasters. They offer predictions based on probability, such as a 75 percent chance of rain. It's not the same as saying it will rain, just that there is a high likelihood it will rain.

For many people, from Karl Rove to New York Times columnist David Brooks to the New York Times public editor, it was illogical to predict such an overwhelming prospect of an Obama victory when most polls showed the popular vote as neck and neck.

However, Silver was modeling the probability of what would occur, not prophesying it. A Romney victory November 6 would not have gone against what he modeled; it merely would have demonstrated that Romney beat the odds.