I feel incredibly confident that Obama will win tonight, just as I felt incredibly confident he would win in 2008. In both cases, that's because of the incredible analysis of Nate Silver (as well as the other statistics junkies out there who run aggregation sites).
Thinking back to 2004, I remember feeling cautiously optimistic about John Kerry's chances, but in large part that was because I bought the spin that undecideds would break for the challenger.
I don't remember anything close to the sort of aggregation we see on the Web these days. I was a steady Daily Kos reader back then but the positive diaries I read were never as data-based as the ones we see today.
So I wonder... if Nate had been dissecting the polls in 2004 the way he does now, would he have shown Bush with a strong chance at victory? 60, 70, even 80% or above?