Excellent news! Still within the margin of error, but this may actually be winnable finally.
Gardner would have to be outside the margin of error by election day in order to be decently sure of his election. Buck was ahead by 1 in the last poll, but Bennet won by 2.
I think those assuming that a 3 point error in one election (within the margin of error) is indicative of some systemic polling bias are potentially guilty of a Type I error. Udall underperformed the polls in 2008, which seems just as likely for Udall to replicate as a repeat of Bennet's 2010 upset.
