you don't seem to place any importance on logical methodology.
nor do you seem to understand that crooked polls are like broken clocks. They can get a binary result correct even if they have terrible methodology.
polls are not supposed to be forecasting what will happen on nov 8th... they say if you vote today who would you vote for.
As the election approaches vote tallys change. Undecideds and third parties shift. Trump got a higher percentage of republicans.
the key to a good poll is good methodology and an honest template.
Now... there is something else to consider. After the herding some of your crooked polls adopted a reasonable template. closer to d plus 4.
But... they could all be wrong if the turnout does not match 2012.
Because trump received more primary votes than hillary... there is a chance today's turnout is more like 2014 than 2012.
In which case almost all the polls will have underestimated the Trump vote.
But that would not mean the polls with good templates were bad.
They were 3 points off . Other pollsters were closer and had obama ahead months before they even started polling.this year they are taking both sides by having Hillary ahead in one poll and trump in the other.