The amount of undecided/others in these polls are indefensible, especially from esteemed pollsters.
They aren't all undecided.
Bray (L) 6%
Other 2%
Hayes (C) 2%
Refused 0%
Undecided 11%
Yeah, the sample is just way too heavy on 3rd parties. Considering the sample itself was only Trump 43-42, that's way too many 3rd party
High undecided %s in previous samples among Trump voters (Tillis seems to have a problem with his base) turning into support for conservative third party candidates makes more sense than Cunningham having massive crossover appeal.