CELINA, Ohio — As Mitt Romney's campaign works to exude confidence in the home stretch, one wonky, liberal, self-made New York Times election forecaster has emerged as a particularly tricky obstacle.
Nate Silver's FiveThirtyEight blog runs dozens of polls through a model that weights and averages them, and then spits out a daily forecast determining the likely outcome of the race. Silver has consistently identified Mitt Romney as a serious underdog, and his projection did not swing as dramatically as did the conventional wisdom after the first presidential debate in Denver: his latest projection gives the Republican just a 26.4 percent chance of winning.
The model's stubborn insistence on a likely Obama victory has earned Silver the wrath of conservatives, who accuse him of shoddy methodology at best, and sinister Obama-boosting at worst.
But out on the campaign trail, Romney's team is at pains to show it's taking the FiveThirtyEight projections in stride. Senior strategist Stuart Stevens is reading Silver's recently published book, The Signal and the Noise and says he's "enjoyed" it. Aides laugh and roll their eyes when reporters tease them with mentions of the model. And one senior adviser, when asked about the numbers, tried to avoid Silver-bashing — even as he pushed back against his forecast.
"In the primary we'd go from having close to zero chance, to winning a state, to 80 percent [likelihood] within some short period of time," said the adviser, predicting that FiveThirtyEight could well give them a better chance of victory as the swing state polls tighten in the final days of the race.
The adviser added that Silver's blog is not widely discussed within the campaign.
The relationship between Silver and Romney world hasn't always been so congenial: Earlier this year, an adviser griped to BuzzFeed that the openly liberal blogger was biased, complaining that he had studied the Obama campaign's internal polls in 2008 without disclosing it.
For his part, Silver told BuzzFeed he's "a bit dismayed that the FiveThirtyEight forecast has become a story rather than what the polls say. It seems unhealthy, on a number of levels."
He said the Romney adviser could be right that the state polls will tighten — along with his forecast — but that the more likely positive outcome for Romney is that the state polls are just systemically biased, and will be proven wrong on election day.
"Nobody is, or ought to be, 'calling' the election," Silver said. "We do have Obama favored, since he's ahead in the polling averages in states that would suffice for him to win the Electoral College... But the odds are, also, that Nov. 6 will be a long night."