There are 3 full days until the election, or 3 days, 11 hours, 18 minutes, and about 10 seconds as I type this sentence around 4:41pm EST.
And Mitt Romney is in deep shit.
It's come to that. That is, believe it or not, still non-partisan analysis. Something people tend not to take into account is that an average lead of X points in a given state two weeks out is probably twice as robust when it's only 3-4 days out from the election. I don't know exactly how to calculate this, but Nate Silver has said that a lead of 2.6 points with about six days left "should convert to a victory about 80 percent of the time". With a little over half that time left, that 80% figure surely has increased. By how much, I'm not sure. But Silver's model puts Obama at 80.5% to win Ohio this afternoon, and his average lead is around 2.4 points. The loss of 0.2 points probably kept his chances of winning the state from moving much higher.
But the time factor can't be expressed in strong enough terms. It takes time to moves polls and there's precious little time left for Mitt Romney to make up ground in Virginia, Colorado, Nevada, Iowa, Ohio, New Hampshire, and protect his crumbling lead in Florida.
To illustrate this, consider that Mitt Romney had a 1.8 point average lead in Florida on October 24th, 1.2 point lead on the 26th and 27th, a 0.4 lead on November 1st, and now has a 0.2 point deficit to Barack Obama this afternoon. It took Obama ten full days to turn a 1.8 point deficit into a 0.2 point lead, a two point swing. That's about two tenths of a point per day, although the polls didn't move every day or in increments nearly that small.
Now consider that Romney is trailing Obama by 2.4 points in Ohio, a must-win state for both men, but certainly more must-win for Romney. Romney has virtually no paths to electoral victory without Ohio, while Obama has a few, but they are very low percentage. I think Silver put Romney at like 2% to win without Ohio, and Obama at 8%. Based on the movement of Florida, Mitt Romney would need about 12 days to bring Ohio into a tie, yet there are only about 3.5 days left in this contest.
There are exceptions, because polls can move quicker and in larger increments than that based on noise and malfeasance. Some are real and some aren't. Obama jumped from a 0.8 point average lead in Colorado on the 24th of October to a 2.2 point lead on the 25th, a lead he sustained for three days, until falling back to 0.8 on the 1st of November, and then moving back up to +1 today. But Ohio is being polled far more frequently than Colorado and it has more electoral votes, so that will smooth movements out and make it harder to make leaps like that.
It gets worse for Romney when you consider how many states he's trailing in that he needs to win, and worse still when you look at the past 10 day trend which shows Romney falling behind in nearly all of them.
Mitt Romney needs about 12 days to reach a tie in Ohio with 3.5 days left, and he did gain 0.2 points there today, to trailing by 2.4 points. But he also needs more than 12 days to catch up in New Hampshire, where Obama leads by an average of 3.6 points. But he's not catching up there, he's falling behind:
10/24: Obama +1.4
10/25: Obama +1.4
10/26: Obama +2.2
10/27: Obama +2.2
11/01: Obama +3.6
11/02: Obama +3.6
At the pace that Obama caught Romney in Florida, Romney would need 18 days to catch Obama in New Hampshire at a 3.6 point deficit. Here's Virginia:
10/26: Obama +0.2
10/27: Obama +0.2
11/01: Obama +1.4
11/02: Obama +2.0
Romney needs about 10 days to bring Virginia back into a tie. And Iowa:
10/24: Obama +2.2
10/25: Obama +1.2
10/26: Obama +1.2
10/27: Obama +1.2
11/01: Obama +2.2
11/02: Obama +3.2
Obama expanded his lead by two points in two days, so Romney can too. The pace that Florida changed is not some hard rule. If Florida is an example of how long it can take to flip a close state, then Iowa is an example of how quickly it can get away from you, or come back to you.
If the few days left argument isn't persuasive, and it's not a lock by any stretch of the imagination, then the trends are probably what have finally ended Mitt Romney's 2012 campaign.
Between October 24th and November 1st, Barack Obama has expanded his leads in Colorado, Iowa, Nevada, New Hampshire, and Ohio, taken over the lead narrowly in Florida and (this late) substantially in Virginia, and narrowly cut into Romney's lead in North Carolina (+2.8 to +2.4).
Of eight recognized battleground states, Obama has improved his numbers in all eight over that period while Romney has necessarily fallen in every single one. This is what America looks like this afternoon:
Now, that map is extremely tenuous. Obama's average 0.2 point lead in Florida is a tie no matter how you look at it. But remember, Romney had an average 1.8 point lead there just a little over a week ago and now has lost it. As far as tie breakers go for projections, I think the guy that has gained two points in ten days ought to have it. Quibble all you want with that, it really doesn't matter in the final analysis, because Obama doesn't need Florida to win the election.
Mitt Romney absolutely cannot win without it.
I find this map just as likely as the one above here:
Colorado (O+1) and Florida (O+0.2) could easily go to Romney, but as you can see, Romney is simply too far behind for either state to matter if those are the only battleground states he wins. Of course they'd matter if he also won Iowa and Virginia.
Let's work our way from Obama's smallest lead in the battleground states, and the electoral votes Romney has, to the biggest lead, to show Romney's easiest path to victory. Romney begins with 191 electoral votes to Obama's 247, and our tossups are Nevada, Colorado, Iowa, Ohio, Virginia, New Hampshire, North Carolina, and Florida which have 110 electoral votes between them. You need 270 to win.
North Carolina (15 electoral votes)
Winning gives him 206 EVs.
Florida (29 electoral votes)
Winning gives him 235.
Colorado (9 EV)
Winning gives him 244.
Virginia (13 EV)
Winning gives him: 257.
Ohio (18 EV)
Winning gives him 275.
Some of these states can be swapped, Nevada and Colorado for example. But they are sorted this way because that's the easiest path for Romney. Obama leads by an average of 3.2 points in Iowa, and 3.6 in Nevada and New Hampshire. Romney has great odds to win North Carolina, split odds to win Florida, low but still within reach odds with Colorado, but each state up that list is a steeper climb. And he has to have all of those without swaps.
Here's Obama's list, and he starts at 247 already because I'm calling bullshit on Wisconsin, Michigan, and Pennsylvania being in play. Not this year, and not this election. Maybe when Arizona is, down the road.
New Hampshire (4 EV)
Winning gives him 251.
Nevada (6 EV)
Winning gives him 257.
Iowa (6 EV)
Winning gives him 263
Ohio (18 EV)
Winning gives him 281.
The closer Florida gets to Obama, the less Ohio (and all these states) matters, because Ohio matters primarily to Obama and Romney as a tipping point with a coalition of other swing states lined up behind them. But Florida will push Obama over 270 all by itself. If they call Florida for Obama on election night before any of the other seven or so battleground states add up to 270 for somebody, then that's it.
The election is over.
But that won't happen because Florida is large and extremely slow at returning results. So we'll all be waiting on Ohio and Virginia, where Obama leads by at least 2 points each.
In summary, every state that matters to Mitt Romney is moving away from him and he was already polling behind in too many of them to win as it was. The story of the past ten days is that the chances of Obama winning narrowly in the electoral college are shrinking, while his odds of winning very comfortably, between 300 and 330 electoral votes, are increasing.
You want to be the one getting stronger at the very end, and that's Obama. You also want to be ahead near the end so that you can afford to fall, and not lose. And that's Obama as well.
Put another way, the election is much closer to the way it was after the conventions, when Obama was an 80-87% favorite to win, than it is close to the way it was after the Denver debate. Obama didn't get a debate bump like Romney did, and this resurgence took a lot longer. But the end game is that Obama got almost everything back that he lost, and did it with like 5 seconds left on the clock.
Update - The two electoral collage maps have been replaced as of Nov 03. They erroneously showed Barack Obama winning Arizona due to an error on my part converting the state from "leans Romney", a light shade of red, to "solid", a dark red, so that all colors matched to present the maps as a final projection. I don't believe any part of the story contained the error, other than the images. (Thanks Chasing for being the only person to notice, and consequently for speaking up.)