Predictions and Probabilities

Yesterday here at PA Kathleen Geier noted that Nate Silver was moving from a brutal rollout week for his new ESPN-affiliated FiveThirtyEight site to some more familiar ground: election forecasting. But as in earlier cycles, Nate’s insistence on framing his forecasts for particular contests–in this case, the 2014 Senate races–in terms of probabilities rather than “leans” and “likelies” and “toss-ups” or other such oracular terminology simultaneously exposes him to charges of too much and too little precision. Looking at his forecasts one way, there’s only one true “toss-up” contest at present: North Carolina, where Kay Hagen has a 50/50 chance of re-election. But in his own analysis of his numbers, Nate suggests a much wider range of possibilities:

There are 10 races that each party has at least a 25 percent chance of winning, according to our ratings. If Republicans were to win all of them, they would gain a net of 11 seats from Democrats, which would give them a 56-44 majority in the new Senate. If Democrats were to sweep, they would lose a net of just one seat and hold a 54-46 majority.

So our forecast might be thought of as a Republican gain of six seats — plus or minus five. The balance has shifted slightly toward the GOP. But it wouldn’t take much for it to revert to the Democrats, nor for this year to develop into a Republican rout along the lines of 2010.

Comparing 538’s forecast to its most credible rival, the Cook Political Report, is instructive. Cook’s Jennifer Duffy lists Arkansas as a toss-up race; Nate shows a 70/30 probability that Mark Pryor will lose. Similarly Cook shows the two vulnerable Republican seats, Kentucky and Georgia, as toss-ups. Nate gives Democrats a 25% chance of winning Kentucky and a 30% chance of winning Georgia. But his numbers would change rapidly with a few more likely-voter surveys in any of these states showing Democrats running even or ahead; Duffy tends to project races as very close until evidence emerges that they are not so close.

But there’s not a great deal of divergence in the factors used by 538 and Cook–polls, electoral history, money, national trends–and it’s very likely their forecasts will converge as we get closer to November.

Ed Kilgore

Ed Kilgore, a Monthly contributing editor, is a columnist for the Daily Intelligencer, New York magazine’s politics blog, and the managing editor for the Democratic Strategist.