UPDATE (January 4th): I see it and I believe it. Yeah, the Register poll was right, and I was completely wrong; Obama got his 80,000 first-timers (more like 100,000 actually), in what has got to be one of the most remarkable GOTV accomplishments of all time. So kudos to the Obama campaign (as well as Selzer and the DMR), and read below if you'd like to enjoy my complete and utter wrongitude. See you all in New Hampshire! --------

Well, the Iowa caucuses are less than two days away now, so I thought I'd poke my head out of my Canadian hole for a minute to talk about the polling situation there. (Disclaimer: as most of you know, I'm a pretty fervent Edwards supporter. I try not to let it colour my analysis; judge for yourselves.)

So the polling we have on the Iowa caucuses right now is a great big horrible clusterfuck:

Look at that mess. Every day there's three different polls saying three different things; I wouldn't be surprised if we woke up tomorrow morning and Zogby had Mike Gravel in double digits. (My theory is that Iowans, tired of being called up and interrogated all day, are engaged in a mass conspiracy to mess with pollsters' results until they all either quit or go bonkers.)

Let's focus, then, on one in particular: the Selzer/Des Moines Register poll, which was dead-on in 2004 and is widely respected across the field. They call it the gold standard of Iowa polling (which is a complicated business, for all sorts of reasons, and easily screwed up by amateurs). DMR this year has, rather surprisingly, Obama out in front with 32% to Clinton and Edwards at 25% and 24% respectively; this lead is entirely due to what the Register calls a "dramatic influx of first-time caucusgoers", 60% first-timers to be exact.

"Dramatic" is not quite the word I would use. I would choose something more like "fucking insane."

...60% Are they serious? The Iowa caucus is an obscure, ritualized, often intimidating process that requires you to spend upwards of an hour in a gym or (worse) a living room with total strangers at 7 PM on a cold January night. Generally less than a tenth of the electorate shows up. And for this they expect, what, like 80 thousand people who have never done it before to appear and vote en masse for Obama? Get off my lawn.

Yet Obama's strategy is banked entirely on getting these new faces (as well as large numbers of independents, hardly the world's most reliable demographic) out into the caucus rooms. Clinton is more focused on the traditional electorate, though she's still been trying to pump up turnout (especially among women); Edwards has run the most old-school campaign and trained his considerable organizational power mainly on caucus veterans (predominantly older and rural voters, who form the bulk of his base).

So victory in Iowa depends on who shows up. And this DMR poll suggests that Obama's ambitious (dare I say, "audacious") strategy is paying off; now, given the seismic shift it'd take to bring 60% new caucusgoers out on Thursday night, Occam's Razor suggests it's more likely this is just a crap poll. But assume for a moment that the Register is right and Obama really does pull it off. If he can seriously make this fundamental a change in the recalcitrant Iowa electorate, then -- regardless of whatever personal beef I or you might have with his candidacy -- that's a massive political accomplishment that has to be respected in its own right, and it'd be his best argument yet for the nomination.

That's the way I see it. We'll know more on Thursday, of course, and the entrance polls will tell us all we need to know about whose turnout models really paid off. (Less so the final results, which will be affected both by second-choice movements and the Party's disproportionate weighting, though that certainly doesn't mean we should ignore them.)

...And now for some BASELESS AND IRRESPONSIBLE SPECULATION: I don't think Obama's going anywhere. This strategy reeks of bullshit. No, I say Edwards wins Iowa with a strong margin, say 5-6% over Clinton with Obama trailing, and his resultant bounce from undecideds and disillusioned Obama supporters pulls him almost even with Clinton in NH. She gets a narrow win over him there, Obama gets another embarrassing third and drops out (maybe he fights to SC but I doubt it). Edwards then gets a big boost in SC, where his numbers have been lower than they should be (voters there still aren't really paying attention), and the result is a dogfight between Edwards and Clinton all the way through Feb. 5 -- when, in all honesty, Clinton probably wraps it up off New York and California, though a man can dream (and if Edwards' press is good enough, you never know). That's how I see it playing out, anyway -- anyone else want to roll the dice?