As Election Day approaches it looks like a pretty tight race. The polls tend to favor President Obama but his lead is usually within the margin of error. A lot will probably depend on which side does a better job of turning out the vote. In addition such a close race offers the chance of some rather ugly outcomes, including the following:

• A slow count or recount in which we won’t know the outcome for several weeks. (Think 2000).
• An “Electoral College inversion” in which one side wins the popular vote and the other side wins the Electoral College. (Once again, this happened in 2000.) This is bound to cause resentment on the losing side.
• A 269-269 tie in the Electoral College. In that case the House of Representatives chooses the President and the Senate chooses the Vice President. The House is pretty certain to remain under Republican control while the Senate is pretty certain to remain under Democratic control, so we will end up with President Romney and Vice President Biden. That will be embarrassing to say the least.

On the other hand it is still barely possible that we will have a clear winner soon after to polls close on the East Coast.

In any case, most states are not competitive so the winner in the Electoral College will be determined in a small number of swing states where the race in considered close. If you keep an eye on the swing states you will know the outcome as soon as anybody does.

States where the polls close at 7 PM ET

Florida. Romney seems to be holding on to a comfortable lead, which is good for him because without Florida he’s toast. There is no plausible winning combination for him without the state’s 29 electoral votes.

Virgina. This state looks very close and potentially vital. Whoever gets VA’s 13 electoral votes will have a very big advantage.

States where the polls close at 7:30 PM ET

North Carolina. Probably belongs to Romney but Obama has made a significant effort. Still, if Obama takes NC it probably means that the polls are wrong and he’s going to win easily.

Ohio is where the election is most likely to be decided. To win without Ohio’s 18 electoral votes, Romney would probably have to sweep all the tossup states (CO-IA-WI-VA-NH) while Obama would probably need 4 of them. Obama seems to have a small lead here, but if it comes down to a close race here we may be waiting a long time to know the winner.

Over a quarter of a million people have already voted by absentee ballot and these will not be counted until Nov. 14. (A ballot mailed on Election Day is allowed up to a week to arrive.) Both sides have hordes of lawyers and spin doctors ready to descend on the state if the count is close enough to justify a manual recount.

States where the polls close at 8 PM ET

Michigan. Obama is favored to win the state’s 16 electoral votes, at least partly thanks to the auto bailout.

New Hampshire. First, forget Dixville Notch. It doesn’t mean anything. Most scenarios leading to an Electoral College tie require Romney to take NH. Otherwise the state’s 4 electoral votes probably won’t be decisive.

Pennsylvania. Obama is favored to win here but the Romney campaign has made a significant effort. It is just possible that the aftereffects of Hurricane Sandy might depress turnout enough to give Romney a boost. If Romney were to get PA’s 20 electoral votes that would turn the race around, making it difficult for Obama to come up with a winning combination.

States where the polls close at 7 PM Mountain Time or 8 PM Central

Colorado. a close race with 9 electoral votes.

Wisconsin. Another close race and 10 electoral votes. Assuming that we know who won Ohio at this point, either CO or WI would be in a good position to wrap things up.

States where the polls close at 7 PM Pacific or 9 PM Central

Iowa. If we’re still paying attention at this point the race will be very close indeed. A toss-up state with 6 electoral votes.

Nevada. 6 electoral votes which will probably go to Obama (plenty of Mormons but even more Hispanics.)

Sources

I have relied heavily on 2 sources: Larry Sabato’s Crystal Ball site and Nate Silver’s FiveThirtyEight.

Larry Sabato uses a fairly straight-forward average of polls, combined with his expertise as a long-time observer of U.S. politics.

Nate Silver uses a more complex and controversial mathematical model which combines polling and economic data. Polls are weighted to compensate for “house effects” (some polling firms consistently seem to favor Republican or Democratic candidates) and for differences in methodology (polls that include cell phones get more weight than those that call only land-lines.) He then runs tens of thousands of simulations every day, slightly varying the assumptions for each iteration. This allows him to assign ridiculously precise probabilities to various outcomes.

This approach allowed him to correctly predict the winner in 49 out of 50 states in 2008, but as they say, past performance is no guarantee of future results. For what it is worth, these are some of the probabilities calculated by his model as of Sunday Nov. 4:

Obama wins the Electoral College: 85.5%
Obama wins the popular vote: 50.5% 80.1%
Romney wins the popular vote but loses the Electoral College: 7.3%
Obama wins the popular vote but loses the Electoral College: 0.7%
Electoral College tie: 0.2%
Recount (one or more decisive states within 0.5 percentage points): 8.3%