Adding a D+3 Model
I think you know what I think the likelihood of a D+3 turnout is, but since some many pundits have decided the D+3 "feels right" without any evidence to back it up, I'm going to add it to the model. This also captures somewhat the "conservative thought" of the Romney campaign, since his pollster is on record saying he expects a D+3 turnout. I think that is not their real expectation, but it is a data point.
All polling reweights and averages going forward will include this model. My D/R/I is 38/35/27 since that captures a small decrease for Democrats off their 2008 results and a modest increase for Republicans from 2008.