Two years ago I remember being very excited because I got a sneak peak at VNS (Voter News Service) numbers at around 2 PM. Of course, those numbers had Iowa and Pennsylvania going for Bush, so I don't place a ton of faith in these instruments. That said, compare and contrast Drudge's info with Joshua Micah Marshall's skinny on early exit polling for crucial Senate elections. Oh, hell, I'll do it for you:

ARKANSAS -- Drudge has Pryor (D) winning "easily"; Marshall has Pryor up by 18 points.

COLORADO -- A shocker. Drudge also has Strickland (D) winning “easily”; Marshall has him up by 20 points.

GEORGIA -- Both Drudge and Marshall have Chambliss (R) up by 4 points.

LOUISIANA -- Drudge has Landrieu facing a December runoff.

MINNESOTA -- Drudge has Coleman (R) up by 3; Marshall has Mondale up by a similar percentage.

MISSOURI -- Drudge has Talent (R) "leading"; Marshall has him up by 10 points.

NEW HAMPSHIRE -- Drudge has nothing on this race; Marshall has Shaheen (D) up by 6 points.

NORTH CAROLINA -- Very interesting. Drudge has Bowles (D) "leading"; Marshall has Dole up 4-6 points.

SOUTH DAKOTA -- Nothing from Drudge; Marshall has Johnson (D) up by 2-4 points.

TEXAS -- Drudge has Cornyn (R) up 8 points, and Marshall has him up by 10 points.

I predicted Strickland winning, but there's no way in hell that a Democrat is going to win in Colorado by twenty points. This should clue you in to how much these numbers will fluctuate over the day. Of all of the races posted above, the one I find the most interesting is North Carolina, where Drudge and Marshall predict a winner contrary to their own ideological preferences.

Developing....

UPDATE: Both Drudge and Marshall now post that the VNS computer has "has somehow broken down or that they themselves aren't trusting their numbers," so that's it as far as exit polls go.