Polls show an Obama victory as history dictates death knell for Romney

http://www.guardian.co.uk/commentisfree/2012/nov/06/obama-polls-victory-romney-death-knell

Version 0 of 1.

The end of the campaign is here. It's been a long journey for many (including your writer). A year ago at this time Mitt Romney was fighting back challengers such as Herman Cain and Newt Gingrich. Romney was eventually able to claw his way to the Republican nomination. Two months ago, he was left for dead by many after a successful Democratic national convention. But Romney came back to near even after the first debate.

It now seems that Romney's luck has finally run out. President of the United States Barack Obama looks destined to claim victory at some point this evening or early tomorrow morning in the 2012 presidential election. Why do I have confidence?

The polling aggregates are crystal clear.

FiveThirtyEight, HuffPollster and Real Clear Politics all have President Obama winning at least 303 electoral votes. The latter two actually have Obama taking 332 electoral votes. Put another way, none of the aggregates really know something the others don't. There's no special formula needed to tell us who will win.

The only state where they disagree is Florida, where FiveThirtyEight and HuffPollster have a small Obama edge, while Real Clear has Romney with a more comfortable lead. As of Monday night, the average of all three gives Obama 303 electoral votes to Romney's 235.

I'd have to have a very good statistical reason to disagree with the average for me go in another direction. I don't have one. I'll go with Obama 303 to Romney 235 in the electoral college as my prediction.

Boosting my confidence is that this result is very close to what Drew Linzer at Votamatic has been forecasting. He's constantly been estimating Obama with 303 to 332 electoral votes and Romney with 206 to 235 electoral votes. If one person gets a lot of credit for a good call for a long time, I hope it's Drew.

I should note that the averages will almost always get at least one state wrong. The best chance for that error will be Florida or Virginia. I'm not sure which it is, and I'm not about to go chasing the "incorrect" state. Still, an Obama electoral vote total of anywhere from 290 to 332 electoral votes is well within the statistical norm.

It also seems unlikely that there will be an electoral/popular vote split. The average of all the aggregates pegs Obama with a 1.6 point popular vote victory, which is what I'll make my final call. Noting the range and past popular vote projection accuracy, anything within a point of 1.6 point margin from 0.6 to 2.6 is perfectly within the margin of error.

I will say that Romney isn't completely dead. I'd put the chance of a Romney victory at somewhere between 5 and 10%. That's not large, yet it's not trivial either.

A Romney win this evening would be quite surprising, however, given past polling accurate. I can see how he gets to ~260 electoral votes through a series of errors. The necessary 270 to emerge victorious is a big stretch. Romney doesn't lead in a single recent non-partisan Nevada, Ohio, Pennsylvania, or Wisconsin survey. Historically, that is a death knell.

The chance of a Romney victory is low enough to ensure that a win by the former governor of Massachusetts would be viewed as a disaster for analysts and pollsters. I'd expect us to have to book an airport Hilton for a month-long special conference on what went wrong in such a scenario. It would also give people reason to question polling for years on end. We can argue over whether that is a good thing or not.

At the end of the day, the only way we'll know if I'll need to book a hotel room is to see the results. So go out there and vote America. It's your fundamental right.