The two strongest arguments for Dems tonight are the closed enthusiasm gap and Dems performing well (and polls for that matter) in the special elections this cycle.
I don't think it is likely, but it is the best case for a good Dem Midterms. Conversely the fundamentals are terrible this cycle.
Frankly key races are close enough anything is possible this Midterms- good and very bad. And as I've said a few times now, historically average polling bias is effectively zero, and not consistent election to election.
We have fewer polls this cycle and more of them are from partisan pollsters. The one exception is that there are more generic ballot polls this cycle and most of the generic ballot polls are from non-partisan pollsters.
Not really possible to predict polling error or bias ahead of time, but could very well have a divergence between the generic ballot and horse race polls as a result of this. The generic ballot average is R+1 at this point.
The generic ballot is actually not terribly inaccurate to the eventual House popular vote. Average error is ~2-3% historically. Though it does tend to overestimate Dems historically on average (not all the time).