Over the last week, numerous people have emailed me and asked for my prediction regarding the election. My predictions are of limited value, of course, given my notorious inability to correctly predict either 2008 nomination in 2003, but since I dislike cowardly commentators who refuse to go out on a limb for fear of damaging their precious reputations, I consider myself bound to honor these requests and make a prediction about who will win the presidency tomorrow.
The media experts are pretty much all predicting an Obama win of varying degrees. FiveThirtyEight rates it at a 96.3 percent probability. Ever since the September economic meltdown occurred and McCain blew his handling of the bail-out so badly, I also assumed that Obama would win going away, but as the election has drawn closer and I’ve paid more attention to the details, I’ve noticed a few significant cracks in their basic assumptions. I’m not just being contrary here, I’m basing my prediction on three pre-established factors that don’t involve my instinctive distrust of expert consensus.
The first is my five-point rule. In general, if a Democrat leads by less than five points in the polls, he loses. I was planning to run the numbers today, but the Baseball Savant already did so I don’t have to. He points out:
Look at the states that would flip to McCain if the 5% adjustment holds true:
Colorado: Obama +4.3
North Carolina: Obama +0.5
Nevada: Obama +2.6
Ohio: Obama +2.1
Virginia: Obama +4.4
I haven’t checked the Savant’s math, but he’s a competent guy and assuming that he added the states up correctly, it looks as if McCain will not only win, but will win by a larger margin in the electoral college than George W. Bush did. Also note that during the primary voting, Obama’s exit polling was overestimated by SEVEN percent on average.
Second, I’ve been very skeptical of the polls this time around because they’re based on new and improved models that have not been hitherto used and did not work well in the primaries. Thanks to Alan Greenspan and the Black Swan Boyz, we’ve all had it driven very well home that untested new models seldom work anywhere nearly as well as advertised. These new polling models are predicated on the idea that due to the sheer wonderfulness of Barack Obama, the turnout patterns will be very different than they’ve historically been. The problem is that turnout among new voters, young voters, and Hispanic voters has actually been lagging among those voting early. It’s certainly true that voter registrations have been extraordinarily high, but it must be kept in mind that the ease with which registration can now be accomplished logically suggests that the percentage of registered voters who fail to vote will also be at an all-time high.
Third, the anxious actions of the Obama campaign belie the seemingly serene confidence of the Obama candidate. Obama’s success in building a political career has largely been based on illusion, and I think the air of calm assurance that he’s been projecting is a false one. By all rights, he should be absolutely slaughtering McCain; Hillary would probably be 20 points ahead at this point. But he isn’t, and since he hasn’t sealed the deal by now with such powerful Democratic winds at his back, it suggests that the deal is not sealable by him.
I’m not voting for either man. My considered opinion is that America would be worse off in the short term and better off in the long term with an Obama victory. But based on my observations, I am forced to conclude that despite the way things superficially appear, John McCain will win the election. The two key states to watch are New Hampshire and Virginia, with a particular emphasis on Virginia. If either of them go for McCain, you can safely conclude that it’s over in the GOP’s favor.
John Scalzi has his own take on things electoral, upon which I’ve commented in some detail in case you’re interested.
UPDATE – Bob Krumm puts significantly more work into his analysis and reaches very similar conclusions. All hail Logic: the lazy intellectual’s shortcut. It’s entirely possible that we could both be wrong, of course, but look at it this way. If your life was on the line, would you bet on the precision of the media’s new polling models?