Nice Analysis

I'm pretty sick of polls at this point, because I'm reasonably certain the polls mean very little. I think most people have made up their minds, a good number (at least 10%) have already voted early, and the only people who would have the time to answer polls at this point are the sorts of people whom I wouldn't trust to pick up my groceries, let alone select the Commander-in-Chief for the next four years.

But this guy actually took some time to do some analysis that's worth reading, about the canard that undecideds break for the challenger. I have no idea if there's a counter to this analysis, but at least he took the time to examine the empirical evidence. It's worth a read.

In addition, Mark Nelson points out Larry Sabato's site, which is usually very down-the-middle in most respects, but also allows you to play around with the electoral vote count as much as possible.