Unless polling is way off, and I see no credible reason why it should be, I don't quite understand how Romney is supposed to win.
A Romney NPV win is certainly within the standard error of the polls. However, it would be on the strength of TX vs. CA. For Obama to lose the electoral college, polling averages have to be off by at least 3% at this point, though...
The probability of one poll being an outlier (i.e. outside the margin of error at 95% confidence) is 5%, but the probability that the
average of polls is outside of their "collective" MOE is
considerably lower.The only way Romney could win is if pretty much every poll had large sampling errors (which is what desperate Republicans are trying to argue). Sure it's possible....but Gary Johnson becoming President is also possible.