the math (or florida, again)

 

(Because I am a sucker for data...)

I know some people make their living at this, and have more sophisticated models. Me, I prefer to keep things simple, because you're nearly as likely to arrive at a good result in much less time than it will likely take to come up with a slightly better result.

I looked at aggregated poll data (primarily here...though it is not likely the most up-to-date, it is among the easiest to sift through quickly). I broke the states (plus DC) into seven categories: those that are "safe" for each candidate (leading by a margin of 10 points or greater), those that are "likely" for each candidate (margin between 5 and 10 points), and toss-ups favored one way or another by a margin of less than 5 points. One state--North Carolina--is currently considered a dead tie. 

In the simplest scenario, we take both safe and likely votes for granted, and assume that only the middle three categories are really up for grabs. In this scenario, Obama starts with a huge advantage: of the 133 votes up for grabs, he only needs 29. Romney, on the other hand, needs 106.  To pass that mark, he needs all of the states currently leaning his way, plus North Carolina, plus at least two states (totaling at least 17 votes) of those currently leaning towards Obama.

Here's why I think the race is actually much closer than those numbers would indicate. 

1. Romney's "safe" states really are safe. Only one state--South Dakota--falls in Romney's "likely" category. He has a comfortable lead in the rest. He doesn't really need to worry about Obama making any inroads in states he can take more or less for granted. He doesn't need to spend a dime or any time there, really.

Obama, on the other hand, really can't afford to take CT, MI, MN, OR, and PA (his "likely" states) for granted.

2. Romney's "toss-up" votes are geographically concentrated. Florida, Georgia, and Tennessee are contiguous, and in the Chattanooga and Jacksonville areas, even share media markets. And of course, North Carolina is right next door. I'm glad I don't live in any of those places right now. (Naturally, Ohio is always in contention in presidential races.)

3. My gut. I just have an easier time imagining CO, IA, NV, NC, VA, and/or WI breaking Republican  than I do imagining GA or TN breaking Democrat.

The only prediction I'm really comfortable making at this point: if Obama wins Florida decisively, it's going to be an early night. Otherwise, it will be very, very close, and probably will come down to Colorado and/or Ohio.

Update/footnotes: I have the above data in a spreadsheet, so that when newer/better polls become available, I can update as needed. If you are aware of any I should look at--especially in the 10 current "toss-up" states--please feel free to put a link in the comments section.

Also, it is worth mentioning that ME and NE apportion their electoral votes by congressional district. My sense is that any deviation from the rest of the state in one is likely to be cancelled out by the same thing happening in the other. So I'm just treating them like the other states for now. But...if you've got data, I am happy to plug it into the model.