Days ago Dick Morris predicted a 400+ EV Romney landslide. Now he thinks Sandy may undo Romney.
Let's translate this from CYA-ese: I was right. Romney was winning. But that was before Sandy, the Mitt-slaying anti-Unicorn which changed everything. So, just to be clear, I was right, Romney wins. Except then there was Sandy, which no one could have predicted, so he doesn't win. Bummer.
This is the trend chart for the national horse race going back through October. As we know and can see, a series of events in early October robbed Obama of the lead. I say a series of events because if you hover over the graph, the tightening actually precedes the debate, though there's little doubt the debate was the major factor in pushing Romney ahead. Since about the middle of October the race has been exceedingly tight and pretty stable. Obama seems to slowly regain a lead, with the crossover coming just before Sandy. But the margins are so tight that it's probably best to say the national vote has just been a tie for the last three weeks.
(Note: PollTracker's 2012 methodology reports Internet polls but does not include them in its averages. So this dataset does not incude Ipsos-Reuters or Rand -- among Obama's most favorable polls. Also note that Gallup stopped reporting during the storm. And they were by far the worst for Obama. So that's helped his numbers a bit.)
This all gets pretty deep in the weeds. But the big picture is that there's simply no polling data to support the Sandy Unicorn theory. There has been some trend back to Obama in critical swing states like Virginia, Florida and North Carolina. Small. But all the margins are so small that small changes can matter. Yet those too predate Sandy.
The reason Obama's chances look better today than they did a week ago is simply that we're 7 days closer to the election. And 3 days out polls are much more predictive than 10 day out polls, even if they say the same thing.