In this post with detailed predictions for the Presidential race, we will basically just go down the list of the states that were not automatic locks for one party or the other from the beginning of the general election campaign, starting from those for which there is the least doubt and finishing with the ones which appear to be closest.
The starting point of the discussion are that there are 23 Republican states with 191 electoral votes and 17 Democratic states (plus the District of Columbia) with 217 electoral votes about which there is very little doubt. They include a few states, like Minnesota, New Mexico, and Michigan for the Democratic ticket, and Arizona and Indiana (which Obama won narrowly in 2008) for the Republicans, which may end up having fairly small margins, on the order of 5-7 percent. These states are not realistically in the plans of their opponents, though, and they have not been deluged with advertisements as the other 10 states have. So, we proceed from there, analyzing the other 130 electoral votes which will surely decide the election*:
North Carolina (15 Electoral votes) - It is a sign of the tightness of this election that this, the 10th closest state, is not a sure thing; the polls there have not universally had Romney leading, but most of them do, and both candidates are bypassing the state this weekend, suggesting they both know how it will end up--with Romney (3 pt. margin, about 100,000 votes). Obama, 217-206.
Pennsylvania (20 ) - If Superstorm Sandy is to have a meaningful impact on the Presidential election (apart from Obama's strong handling of the crisis or NYC Mayor Bloomberg's endorsement of Obama afterwards prompted by the storm), it will be in this state (none of the other ones in the affected area on the Eastern Seaboard are close), and that effect can only hurt Obama. It's close enough that a couple hundred thousand votes suppressed by the difficulty of getting to polling places, or by the disruption in people's lives, could turn this state in the Republicans' favor. I don't think it will be quite enough, though, and I think this will be viewed by the Republicans as the big missed opportunity if Romney loses: more effort, sooner, could have allowed him to take this state from Obama (3.5 pt margin, about 100,000 votes). Obama, 237-206.
Nevada (6 ) - This state is usually shown as a toss-up, but Obama is consistently polling narrowly ahead, and turnout--driven by Sen. Harry Reid's political machine--will favor the Democrats. (4 pt margin, about 20,000 votes). Obama 243-206.
Next come the two most consequential states in the election:
Florida (29) - This state is a must-win for Romney--winning is basically impossible for him without it--and would be a great trophy for Obama, the difference between a narrow Electoral college win and a safe one. Obama was leading in the polling at his peak, before the first debate; now he looks to be trailing. Romney (1.5 pt. margin, about 100,000 votes). Obama 243-235.
Ohio (18) - Winning the state is not equal to winning the election, but losing it would put either at the brink of defeat. Romney's biggest problem is that he has never led Obama in the state by any meaningful margin, although he was equal or near-equal at Obama's lowest point (before the second debate). Obama's efforts to help the auto industry, and the somewhat-related fact that its economy is doing relatively well at the present, mean this will not be the closest state Obama wins. (1.5 pt. margin, about 70,000 votes). Obama 261-235.
New Hampshire (4) - This state has gone back and forth, but Obama is leading fairly consistently in the last couple of weeks. (2 pt. margin, about 10,000 votes) Obama 265-235.
Iowa (6) - This state has polled extremely close. There hasn't been that much polling there, so there's considerable doubt, but this is not a particularly strong state for Romney (he narrowly lost the popular vote to Rick Santorum in the caucuses), and it has supported Obama well in the past. (1.5 pt margin, about 20,000 votes) Obama 271-235.
Wisconsin (10) - Obama is polling ahead by an average of 3-5 points in recent weeks, but I think this is one state in which the progressive forces are fatigued, perhaps a bit demoralized, by the continuing partisan battles within the state, which has come out most recently in favor of the Republicans (Gov. Scott Walker's surviving the recall). So, I think this may be the one swing state where turnout will favor the Republicans, closing Obama's margin to a very narrow one (0.7 pt margin, about 20,000 votes). Obama 281-235.
Virginia (13) - The polling has gone back and forth in this state, and no one has it as anything other than a toss-up if he/she has not forced himself/herself to call each state's outcome. Virginia is one of the earliest ones to begin to report its votes, and it does so very quickly. What we will see then, very quickly, is that the race in the state is too close to call. It may go late into the night, or it may not be decided until the next day, but I'm thinking the overall race will already be decided when it is called--for Romney (0.1 pt margin, about 3,000 votes). Obama 281-248.
Colorado (9) - I would say that the polling has been rather consistent for the past month or so in Colorado--it's a tie. Some are suggesting that this state leans Republican in 2012 (analysis of early voting there supports that thesis), but I think it will end up being the clearest case of Democratic turnout affecting the outcome. There is a ballot item to legalize possession of marijuana in the state, and I believe that will bring out additional young voters on election day, and that it will make the difference--for Obama--in an extremely close state. (1 pt. margin, about 20,000 votes). Obama 290-248.
So, that's my prediction--Obama 290 Electoral Votes, Romney 248.
Now, here's a tip for watching on Election Night. There will be some states--Colorado, Iowa, New Hampshire, Virginia--that we should expect will be too close to call even late into the night. Florida and Ohio may or may not be so close, but they are complex political puzzles with areas that are blue, red, and purple, and those may report their votes unevenly through the evening, so we may expect they will take a while to resolve. So, look at these two states instead, which are both generally put as "Leaning Democratic" rather than toss-ups: Wisconsin and Pennsylvania. Wisconsin could be a positive bellwether for Obama. If he wins there fairly comfortably (4 pts or more), that would indicate the turnout is working for him, the white working-class vote in the Upper Midwest is holding, and his firewall states should protect him in the end from any close losses in the nail-biters. Pennsylvania could be a bellwether in the opposite direction: If Obama can not win the state by 2-4 percent, which means that he is not clearly holding the suburban counties of Philadelphia, he could be in big trouble nationally with suburban voters. And, if he should lose the state, it will be very hard for him to win re-election-- Pennsylvania is about as much a must-win state for Obama as Florida is for Romney.
Finally, the question of the popular vote. The national tracking polls in the past two weeks, more often than not, are showing Romney either tied or leading by 1-2 points (sometimes more). It is certainly feasible that Romney could win the popular vote and lose the electoral vote, and I think that outcome would be justice (for 2000)--perhaps justice leading to a greater justice, in which the Electoral College is eliminated (needs a Constitutional amendment) or success of the legislative movement in which 270 or more electoral votes would be guaranteed to the winner of the popular vote.
I do not think that is going to happen this year, though; I think the intrade.com "market" for the event (different candidates win the popular and electoral votes) at 25% is a bit of a sucker bet; I think the chances are more like 10% (Nate Silver's simulations have it at about 5%). Instead, I think this will be one of those elections like 1960, or like 1968, in which a very small popular vote margin is magnified by the all-or-nothing nature of the Electoral College. I will predict a popular vote victory for Obama of about 0.5%, about 600,000 votes. If that is the case, it will go into the books as yet one more case in which the winner of the election receives less than 50% of the people's vote. For the record, I predict Obama 49.4%, Romney 48.9%, Others 1.7%.
*I don't want to forget the fine states of Nebraska and Maine, which award individual electoral votes based on the results of each Congressional district. One of the Nebraska Congressional districts in 2008 went, against the whole state, for Obama, so the state legislature tried to make more certain, through redistricting, that this wouldn't happen again. I think for this election at least they will be successful, and also that both of the Maine districts will go for the state-winner, Obama.