J$P Extra! Likely voters, Bin Laden, Dick Morris, and The Amazing Kreskin. What do they have to do with the election? They tell us the winner. Updated!

Voting day begins in less than 48 hours. We have never been able to resist the prediction game, so bear with us. And always keep in mind it's just one guy's opinion.

In our view, there are several faulty assumptions coloring perceptions of where the race stands at this point. One of these regards the last-minute appearance of Usama Bin Laden. Conventional analysis asserts that this turns the thoughts of the electorate to the war on terror, benefitting the incumbent and causing a swing to Bush in the final days. But such a trend has not appeared in the tracking polls--in fact, in the Fox News poll, Bush has lost his five point lead and now stands tied with Kerry. And the Democrat has been relentless in hammering his dubious Tora Bora claim, effectively turning Usama’s latest cameo appearance to his advantage. Since this old allegation has been allowed to fester without strong rebuttal from the incumbent, Bin Laden’s October surprise could well represent no advantage (or worse) for Bush.

The corollary to this, of course, are all the polls that show Bush holding a strong advantage over Kerry on issues of national security. But the polls themselves are based on some questionable assumptions. Almost all the coverage, including on Fox, is emphasizing the “likely voter” numbers. Opinions differ on what a likely voter is, and every statistician has his own pet theories about where to draw that line. But in an election where turnout is expected to reach record levels, and millions of new voters have been registered, how exactly can likely voters be reliably predicted? We don’t believe they can.

If the new registrants were roughly split equally between the parties, this issue would assume less importance. But surveys show that Democrats have an edge among these potential first-time voters. Both parties have mounted impressive get-out-the-vote campaigns, but even if both were equally successful, that edge should favor Kerry.

The media fascination with likely voter results, rather than the numbers for all registered voters (which usually tilt toward the Democrat), is little more than tea-leaves reading, leavened with a dose of statistical technospeak. It’s guesswork--educated guesswork, but still guesswork. In a year when turnout may be unpredictably large, only poll numbers for all registered voters can be considered reliable in any predictive sense.

So when the latest numbers show the race essentially (or, in the case of Fox’s poll, literally) tied among likely voters, that means that Sen Kerry is actually a few points ahead. And we predict that is how it will end: Sen Kerry about a point-and-a-half over President Bush. And that margin should translate to an electoral college victory.

Other factors, like systematic voter fraud and Big Media hit pieces, also play into our forecast. Still, a lot can happen in 48 hours, so we’re not ready to bet the savings. This is our version of educated guesswork (OK, maybe not so educated). There is a significant piece of evidence that runs contrary to our theory--The Amazing Kreskin is on record predicting a Bush win. (And we don’t like to disagree with Amazing.) On the other hand, the clincher may well be Dick Morris: he predicts Bush, and based on the Morris record of prognostication, that means champagne in Boston.

Update: Oops.