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.
1.) Unless you're on a Marist-only polling diet, the race has been essentially tied for the entire campaign.
2.) You're right about the lessons learned from Buck/Bennet. With the exception of the low-turnout off-year 2013 conservative fiesta, the GOP in Colorado has frequently underperformed on Election Day. Maybe 2013 was a signal of a realigned, GOP-leaning electorate, but otherwise I have to assume that the GOP will probably not do quite as well as the polling suggests. We'll see.
To address 1), Gardner was doing fine enough, but with Udall barely ahead, it was bound to be Udall favorable, and to be fair, I wasn't really sure Gardner had what it takes. I'm still not, but he's proving me wrong.
That would line up with Gardner losing by a couple of points, same as Buck did.
The last PPP poll in Colorado that year had Buck up 1. Gardner up 2 suggests an incredibly narrow victory for Udall. If Gardner is up 3 or more by election day, he probably wins it.