President Obama and Governor Romney will debate again tonight, and I probably won’t watch it. I’ve never cared much about debates, and I find them frustrating on many levels.
But that doesn’t mean that they don’t matter. Romney has done much better in the polls since the first one and seems to have gotten about a 4 point bounce.
That bounce has certainly tightened the race, but Obama remains the favorite, according to the most reliable source that I know: Nate Silver’s blog FiveThirtyEight. Silver and his team use a complex statistical model that inputs all poll results — some days dozens of them — and gives a few key outside factors, including some economic data, a certain amount of weight. Those outside factors diminish, however, as we get closer to the election.
Right now, the FiveThirtyEight model gives Obama a 63.8% chance of winning on Nov. 6. That’s down dramatically from the 87.1% chance before the first debate. FWIW, InTrade has Obama with a 60.4% chance of re-election.
It looks like a very close election. And that’s why I like Silver’s work — he doesn’t cherry-pick poll results and looks closely at the historical record:
Heading into the second presidential debate, the FiveThirtyEight forecast still showed Mr. Obama as a modest favorite, with about a 2-in-3 chance of winning the election and just over a 1 percent lead in the popular vote.
But historically, the second presidential debate has moved the numbers by about 2.5 percentage points in one direction or another.
If that gain were in Mr. Obama’s favor, he would re-establish enough of a lead that there would be little doubt about who was ahead.
Another shift toward Mr. Romney, however, and he would probably lead in most national and enough swing-state polls to show him on a path to 270 electoral votes.
So here are some maps based on FiveThirtyEight calculations. I have filled them out and then grabbed them from 270toWin.com.
If we assume that the FiveThirtyEight model holds, here’s what the final result would look like on Nov. 6th:
The most common single result in FiveThirtyEight’s simulations, routinely occurring more than 10% of the time, shows a more favorable map for Obama with more than 330 electoral votes. That would look something like this, Virginia, Colorado, and Florida being picked up compared to the above map.
The above map isn’t quite a best-case scenario for the President, but it’s awfully close to that.
But if we see a movement toward Romney of just a point or two, and if that movement occurs more or less uniformly, there’s a slight chance of an electoral tie:
FiveThirtyEight predicts just a 1% chance of an electoral tie. Let’s hope that doesn’t happen.
A sharper move toward Romney in the polls seems unlikely to me given the polarization in the country and the simple fact that most polls show a tiny sliver of the electorate undecided or in play. But a hard move toward Romney in the final three weeks could give us something like this:
Keep in mind that in all of these scenarios, the President-elect will likely win by a very small margin in the popular vote.