Presidential Election Predictions: Nate Silver of FiveThirtyEight Predicts 91 Percent Chance of Obama Win; Does He Have the Electoral College Locked?

I've covered a lot of different polls and predictions here for you. Sometimes they agree, and sometimes they diverge wildly.

But they usually fall within a relatively narrow range. Neither President Obama nor Republican challenger Mitt Romney have been 10 points up in the polls, for example.

And in a race like this one, where the expected results fall on both sides of 50 percent, the election becomes very difficult to call.

When one candidate wins by 20 points, it doesn't matter if the polls were off by 5; they'll still correctly predict the winner.

But this year, even error as low as a single point could completely throw off predictions. Hence the cry of "too close to call!" echoing from every newsroom.

But Nate Silver at FiveThirtyEight thinks the election is in the bag for Obama, and his reasoning is quite compelling.

FiveThirtyEight analyzes the same polls that RealClearPolitics does-the same polls I use.

But Silver runs an extensive statistical analysis rather than a simple average.

And while he gives Obama a slight edge in the popular vote (just like RCP and me), he says Obama has a 91 percent chance of winning the Electoral College, and therefore, the presidency.

Now, you may remember how pessimistic I've been about Romney's chances to win the election if he can't win Ohio. While there are technically ways he can cobble together the 270 electoral votes he needs, personally I don't see any of them as probable.

Silver takes this pessimism one step further. His model assumes that if Romney can't win Ohio, that he won't win any of the states where he's polling worse than he is in Ohio.

Basically, if all the energy and money the candidates are pouring into Ohio doesn't convince voters, then nothing will.

I've talked about the similarities between Ohio and the other Rust Belt swing states, and I essentially agree with Silver.

I can't see Romney doing poorly in Ohio but also doing well in Pennsylvania, Michigan or Wisconsin.

Conversely, if Romney performs well in Ohio, he's likely to do better in those states. But since his support is lower in the other three states than it is in Ohio, a win in Ohio for him doesn't necessarily translate to wins elsewhere in the way that a loss does.

And since Obama has led (not tied, led) in every poll in Ohio except those coming from Rasmussen (whose robo-calling practices tend to reach more conservative-minded voters), Silver gives him about a 91 percent chance of winning the state, and therefore, the Electoral College.

Interesting analysis. And it's important to note that Silver isn't predicting an Obama landslide, just a victory-albeit a very likely one.

I won't be surprised if Silver is gloating on every talk show and pundit outlet on television tomorrow morning.