Where did these real numbers come from? And do they represent party registration or a prediction of who will vote this cycle?
That is the national average according to Gallup and that is the normalization that must be applied to the data. Oversampling from any group (data set) introduces error that compounds. Gallup had the distribution in Jan of this year at 29% Dem 26% GOP 45% Indy which has changed with the primary. Trump brought lots of new voters in the GOP which upped their numbers. Applying these 2012 turnout models is horseshit.