Haugh's support will of course collapse mightily by Election Day. Normally I'd say that's good news for Tillis, but I'm surprised to see Hagen actually running slightly ahead as Haugh voters' second choice.
I think the CW that people who vote libertarian would ALWAYS overwhelmingly break Republican is starting to falter.
http://www.nytimes.com/projects/elections/2013/general/virginia/exit-polls.html
Sarvis got more liberals than conservatives, and did best among moderates, which solidly favored McAuliffe. This doesn't conclusively prove they would've broken for McAuliffe (other demographic crosstabs suggest otherwise), but it shows that it isn't nearly as clear cut as many pretend it is.
Good points all around, Icespear. Maybe libertarian-leaning southern voters are more receptive to considering Democrats as social conservatism is more strongly emophasized by southern Republicans than elsewhere?
Still, I say the most accurate way to poll a race is to not mention 3rd party candidates by name as an option but include those respondents who volunteer a 3rd party candidate. Exceptions should only be made for MAJOR 3rd party candidates like Cutler in ME.