Each week I've been taking polling data from Andrew Tanenbaum's www.electoral-vote.com and running Monte Carlo simulations of the election. I use two different margins of error, a 4% margin that is common to most state polls, and another I computed by running a regression of polling errors in 2004 data compared to the final election results. The 4% margin of error is a snapshot of who would win if the election were held today. The other simulation now uses a 10.6% margin of error, which was an attempt to allow movement before the election.

Obama wins almost all the time in both simulations this week/
4% Margin of Error:
Obama wins 99.99%, averages 340.9 EV (low 269, median 344, high 375)
McCain wins 0.00%, averages 197.1 EV (low 139, median 194, high 269)
Electoral tie 0.01%

10.6% Margin of Error:
Obama wins 97.86%, averages 322.6 EV (low 230, median 323, high 419)
McCain wins 1.99%, averages 205.4 EV (low 119, median 215, high 308)
Electoral tie 0.15%

This is easily Obama's most commanding lead of the year, in both simulations. The 10.6% margin of error clearly overstates Obama's chances of winning, because it is assuming each state varies independently. If McCain makes a comeback, he's likely to do so in most states at the same time.

Needless to say, this week's state polling data were quite favorable to Obama. He's probably moved ahead in Florida (leading in 4 of 6 polls, with one tied, and the other giving McCain just a 1 point lead), and is polling a little stronger than McCain in Virginia (leading in 3 of 5 polls), Ohio (4 new polls split, with each candidate leading 2, but Obaa's leads are larger). North Carolina had two new polls, with each candidate showing a 3 point lead, so that state may well be a toss-up right now (McCain had led by double-digits immediately after the GOP convention). McCain gave up on Michigan, but Pennsylvania has shifted more solidly blue also, where the closest of 4 new polls was a 5 point lead, while another poll showed Obama ahead by double-digits.

And while it didn't affect the 4% margin of error simulations, McCain's leads in some traditionally red states shrunk also, with Georgia, Mississippi, and West Virginia all showing McCain leads under 10%. In the simulations, McCain was more likely to lose West Virginia than he was to win Pennsylvania. Yes, Obama now has more very small leads, but adding up the states where he's ahead now puts Obama at almost 340 electoral votes, so he can afford to lose several big, close states and still win.

It's just a month after McCain was rated as more likely to win than Obama, and there's still one month to go, so it's possible that things could swing back the other way.  Only four states(West Virginia, Wisconsin, Michigan, and Maine) which last showed a lead of under 10 points didn't have a new public poll this week. McCain's decision to pull out of Michigan implies his internal polling was not favorable.

The state polls don't yet include reactions to Thursday's vice presidential debate, but it doesn't seem like it was the game-changer McCain needs to make this competitive. Obama is now in a very strong position, and while these numbers overstate the chances, he is now highly likely to be our next president.