First of all, my heart goes out to everyone affected by Sandy. I was on the outskirts of it and escaped with nary a scratch. Many have commented on how Sandy will influence the election, with focus on it allowing President Obama to distinguish himself, leaving Romney on the outside looking in. While I believe this analysis is correct, the greatest beneficiary of Sandy is the punditocracy. The MSM has been so focused on making the elections a horserace when all objective data show a clear Obama lead, it was bound to have egg on its face come election night. The question will be asked how they got it so wrong. The answer they will give? "Sandy did it!"
Nate Silver and Sam Wang have done a tremendous job forecasting the election based primarily on statewide polls. They clearly show that the odds against a Romney victory are overwhelming. I believe Nate (or Poblano as I prefer to think of him) is a little too kind to Romney in his analysis by interjecting economic models into the mix. The argument goes that an economy this weak is unfavorable to the sitting President. That is countervailed by memory of where we were when Obama took office - we definitely are in much better shape now, despite an obstructionist Congress.
I definitely do not put myself in the same class as these two analysts, but because I am obsessed by the election and like to tinker with Excel spreadsheets, I constructed my own model based on published statewide polls for both the presidency and senate races, modeling a distribution around the average of the last three published polls in each state, relying on 2008 results for states without any recent polling.
My results are actually on the pessimistic side (see below why they represent the worst I believe Obama can do).
EV: 290
Chance of winning:73%
Senate seates: 52
Chance of retaining Senate: 86%
The reason this is the worst Obama and the democrats can do is the results are based on flawed polling - I (and Nate Silver and Sam Wang) don't cherry pick the polls, and thus include polls with a house bias towards republicans, in particular Rasmussen and Gravis. No other pollsters are so egregiously biased towards democrats. So I believe there is a pretty decent chance of sweeping or coming close to sweeping the swing states, yielding an EV total of at least 303, and probably either 332 (Florida) or 347 (North Carolina). In addition, I would be shocked if we ended up with fewer than 54 democrats in the Senate (including independents caucusing with democrats) based on the same flawed polling.
Even removing the blatant bias among some pollsters, there is also an issue of the likely voter models excluding too many hispanics, which will ensure Nevada and make Colorado also very likely to be blue on November 6. I would also not rule out a sleeper in Arizona.
Many have made the argument that pollsters don't adequately reach cell-phone users, which is true, but I am not smart enough to know if this tilts the results for or against us or is mainly a wash.
But pay attention election night. Chuck Todd and the rest will not be held accountable for their inane projections of a dead heat. They will say it was a dead heat until the Frankenstorm stirred things up and led to an Obama victory.