The problem is that the dopes in 2012 ran the silly unskewed website when in fact the polls were skewed... in Romney’s favor. They were predicting a R +1-2 turnout. Instead it was a D +6.
So now it’s a case of the Republicans who cried wolf.
But the samples this time really are bogus. The polls are averaging D +10 when Obama’s historic 2008 win had D +7 (Obama won by 7) turnout and 2012 was D +6 (Obama won by 4).
There is no way Hillary has near the excitement of Obama. Just look at her and her rallies. Now compare Trump to McCain or Romney.
If the turnout is under D +3 or less, Trump wins. If it’s D +4 or better, Hillary wins.
The thing you’re missing is that nobody weights by party affiliation because that’s a stupid way to poll. You may always think of yourself as a Republican, but a lot of people think of themselves as a Democrat or a Republican based on whether they’re planning to vote Dem or GOP that year. Regardless of what party they’re registered with. So if you try to balance a poll sample with what you think party balance is, you’re just engaging in circular reasoning. If more people are planning to vote for Trump, you will find a higher percentage of self-identified Republicans in your universe. If fewer, less.
You hit the nail on the head. It is all a question of turnout. If the electorate is D+3 or less, Trump will win. If Clinton can get out the vote like Obama in 2008 and 2012, she will win. I just don’t see Clinton getting the Obama coalition reassembled.