Not a bad result for a +6 Republican sample. No one should be surprised if the Dem internals end up being accurate again. Democrats just don't mess around when it comes to internals, they can't afford to.
Yeah, because internals aren't released to look favorable to the candidate to rile up their base or anything to make it look like they have a chance.
All the non-partisan Senate polls are wrong. (Except those in Georgia, North Carolina and New Hampshire that show the Democrat ahead, right?)
Again, in 2012, the internals were better at measuring the actual turnout, because a lot of the public pollsters were assuming a bigger drop off in turnout which ended up not happening.
Again, 2012 was a Presidential election year. A good number of pollsters still got it right. This is a six-year election for a two-term President. Off-year elections favor Republicans. Six-year elections are usually bad to the party of the two-term President.
Could Udall be ahead in Colorado right now? Probably not. He's behind in most polls from non-partisan sources with the exception of the YouGov poll. These polls have Gardner ahead at the fringes of the margin of error.
If you cite a D+3 poll in Iowa, that's one thing, it's around the MOE and plausible. But Udall +3 in Colorado? It doesn't match the other margins. Most reasonable minds would have to conclude that this is your typical internal poll where you are trying to rally your base.
And note: I am not saying that Udall can't pull this off. The status of the race on October 18 suggests otherwise, however.