True or False: If McCain fails to win Pennsylvania, he has lost the election.

[J. J.]

False, but PA is one of two paths to victory (not the one I predicted).

[J. J.]

No, Bush 2000 minus NH.  It actually has McCain running not quite as well as Bush 2000 did.

There is a third way, but the odds are even higher on that.  That one is the IA hold with a loss in NH and CO.

I think the two main paths are a Bush 2000 minus NH pattern or a PA pattern.  

Basically a victory in PA would give McCain a sizable cushion where he could lose NH, VA, CO and every other state Gore carried.  

I've never expected anything but a close election.

[J. J.]

PA is the ONLY path for McCain.  Colorado is already gone.  McCain has 2 days not 2 months.

Not the "ONLY" path.

[J. J.]

McCain has no realistic path that adds to 270. McCain can talk all he want about how competitive this race is. It doesn't make it so. We're looking at a 1996 style map with a few states flipped in each direction.

A McCain victory basically requires public polling to utterly fail. Obama has lead in the past 20 polls in Colorado as well as the past 46 polls in PA.

It's called a "Bradley Effect."  How much, if any, that will be is questionable.  We'll find out in about 48 hours.

[J. J.]

It's called a "Bradley Effect."  How much, if any, that will be is questionable.  We'll find out in about 48 hours.

Okay, so a McCain victory requires public polling to utterly fail due to an unproven phenomena that must occur with a multiple point magnitude in at least three different states.

Well, it did occur arguably in 1982 and in 1989.  It also seems to have occurred in three states in 2006.  Sorry, but there is evidence that it does occur.  The question is, does it still exist and how great is it?