Each, not every. Thank goodness.
They won’t call a state until every poll in that state has closed. Some will be open longer than others because anyone who is in line before the official “closing time” will still have the right to vote. That means it could take a while for the polls to actually close even though they would be officially closed.
Slate seems to be down, now – I wonder if their early results just got big coverage somewhere, or if there are lawyers breathing down their necks?
Daniel
I think it’s just getting a lot of traffic. Daily Kos has been up and down all day as well.
Please disregard my previous X, as I realized directly after posting that it was redundent.
Dio:
The NRO has some interesting coverage as well. Folks over there about shit their pants when they saw the first exit poll.
Heh.
That may not be strictly accurate, but it’s pretty close.
My fingers are getting cramped from being crossed all day. Pleasepleasepleaseplease turn out well…
Daniel
Zogby calls it!
Nov. 2, 2004, 5 p.m.:
Bush – 213 electoral votes.
Kerry – 311 electoral votes.
(i.e., a landslide for Kerry)…
with Nevada (5) and Colorado (9) being too close to call.
but (!)
in terms of the popular vote, Zogby finds Bush at 49.4%, Kerry at 49.1% – i.e., 2000 in reverse.
Surreal…only in America.
Go to Brad deLong.
May the force be with us! 
GOP Pollster Frank Luntz is now calling it for Kerry, claiming that he could win between 310 and 320 electoral votes. In the allegedly tight races, Kerry’s ahead by a couple of points.
He could still lose, of course, if something dramatic happens in the late voting, especially in Ohio or Florida. But all in all, looks like Kerry’s on his way to victory.
Luntz, via Wonkette (whose down at the moment), via DeLong.