This is with a +2.5 oversample of Dems.
Dem +4 (Halfway b/w 04 and 08) is probably likely.
Where do you find sampling on Rassmussen?
Where on earth do you get the idea that parity between Republicans and Democrats will be the right turnout model?
Surely you can cite to some objective basis for this right?