I was pretty pessimistic the night before the 2004 when I saw Rasmussen's numbers showing Bush up only 2. Well, he won by 2.5 and it was only obvious the DAY AFTER that Scott Rasmussen had indeed gottn it right throughout the election year.
Although, really, you can only ever prove whether the final pre-election polls are right or wrong. Just because a pollster nails it in their final poll, it doesn't mean that they were accurately tracking the race before that. Unfortunately, there's just no way to test whether polls taken months in advance of the election are an accurate reflection of where things stand at that time.
