A few days ago, I posted a quick and dirty, naive model averaging of the electoral vote probability distributions published by the Princeton Election Consortium and HorsesAss.org. Today, I'm going to post another since there is a week before the election.
Below are the plots of the two probability distributions of Obama's electoral votes that I can access. (Drew Linzer declined to share his EV distribution, and Nate Silver hasn't answered my requests.) HorsesAss is on the top, Princeton Election Consortium is on the bottom. Both predict 303 EV for Obama, but give slightly different odds that he will win.
(To see the plots and read the rest of the analysis, go to Malark-O-blog.)