The breakdown on election day will be around 40D-37 or 38R. (History suggest we will not see a voter turnout showing more Republicans than Democrats).
I can agree with you here. We really don't have real way of knowing what turnout will be until after Nov 2nd.
However, to suggest that we will see such a reverse in voter weighted turnouts than wew ever have before is also a stretch.
And perhaps 2002 was the 9-11 vote. In that, that was the first time Americans were able to go to the polls and vote regarding what happened.
Lastly, I still think GW is in decent shape. This election was always going to be close. Those suggesting otherwise were merely suggesting this for their own hope and psyche.
And those suggesting it was fine for Newsweek and others to oversample (or over weight voter turnout models) in favor of Republicans - Because of the notion that more Americans were identifying themselves as Republicans at that given time - Well, then it is also fine for Newsweek (CBS, Gallup)) and the rest to do the same thing now (if more are identifying themselves as Democrats after the debate).
My point is a consistent voter turnout model is the better way to gage where the race is at.
And with that, we have a very tight race.