Breaking news! John McCain is about to make a comeback. Unless, of course, he isn't.

Nearly 20 days have passed since a national poll showed McCain leading Barack Obama. That's good news for the Illinois senator--but it's bad news for the nation's political press corps. Despite the constant complaints of critics on the right--and left--the media isn't systematically inclined to favor liberals or conservatives. It is, however, addicted to drama. That's understandable: the news is what's new. So after three weeks of writing the same basic story--Barack Obama is Doing Really, Really Well!--the journalistic junkies of New York and Washington have begun to scramble for a new narrative fix. And it's no surprise that they've settled on the "comeback" storyline--a natural fit, really, with McCain's long history of Lazarus-like behavior--as object of their obsession.

Consider the evidence. When McCain unveiled a kinder, gentler stump speech yesterday in Virginia Beach--Bill Ayers was out, "scrappy underdog" messaging was in--both Politico's Mike Allen and Time's Mark Halperin decided to frame it as a potential "comeback" moment. Never mind that McCain had said the same stuff at last month's Republican National Convention. Over at the Drudge Report, Matt Drudge has spent the past week treating statistically insignificant one- or two-point gains for McCain in the Rasmussen and Zogby tracking polls as "BREAKING" news; on Sunday, he led his influential site with the headline "READY FOR A COMEBACK." Finally, the New York Times' Adam Nagourney suggested yesterday that "news media's desire for a competitive race and tendency to find the 'underdog is surging' story line irresistible" could still shift momentum back to McCain." That sound you hear is the political press corps chomping at its collective bit.

Now, being a reporter, I'm all for a new narrative. But this is kind of ridiculous. The fact is, there's no data at this point to substantiate a McCain surge--and without tangible, quantifiable proof, the "news media's desire for a competitive race" shouldn't mean squat. Sure, the occasional survey has shown McCain as close as two or three points. But the average national polling gap between Obama and McCain has grown steadily from 2.3 percent on Sept. 23 to 7.3 percent today, and the Illinois senator's estimated lead in the Electoral College has expanded from eight votes to 190 votes over the same period of time. If the election were held today, Obama would win 313 electoral votes from states where he's ahead by an average of five points or more; he'd take an additional 61 from states where he's ahead by less. Any chatter about a McCain comeback should follow--not precede--a sustained reversal of these trends. That's why they call it reporting.

Am I saying that McCain can't battle back? Not at all. But before I report that he has, I'll be watching the numbers--not the news--for signs of momentum.

Here's how I imagine any McCain surge would unfold. First and foremost are the undecideds. Right now, they represent about five percent of the electorate (on average). As the election approaches, they'll begin to break for either Obama or McCain. Historically, undecideds tend to split pretty evenly between the two candidates. But thanks to Obama's unprecedented political profile--and McCain's effort to raise doubts about his character and readiness--an unusually high number of undecideds may revert at the last minute to the more familiar candidate. That would give McCain a boost of about three or four points.

At this point, Obama--who's topped 50 percent in 10 of the last 14 polls--would still lead McCain by at least three percent nationally. The next bloc to watch is older white voters--the demographic group that has "moved most recently into Obama's corner and given the Democrat his big lead." Drawn to Obama for economic reasons, they've always been wary of his inexperience--and could still be lured from his corner. If McCain pries a sizable number of white seniors away from Obama--today's "Pension and Family Security Plan" is designed to help--the frontrunner would presumably dip below 50 percent in the national polls, putting McCain back in play.

From here, I'd turn my attention to the battleground states that voted for Bush in 2004. Given that graybeards represent a key voting bloc in several of these states--which, after all, will decide the election-- I'll be looking to see whether McCain is able to erase his three-to-12-point deficits in Florida (5.0), Ohio (3.4), Colorado (5.2), New Mexico (7.3), Iowa (12.8) and Virginia (6.5). If he can do that--and perhaps pick up some steam in the Kerry states of Wisconsin, New Hampshire and/or Pennsylvania, where he now trails by more than 10 points--we may be heading for a long night on Nov. 4.

Until then, however, I'll leave the comeback chatter to the chattering classes.