two issues:
1) conservatives should not put up with second rate products or services just because the supplier is “conservative”
2) the polls in aggregate were wrong, the final average of the real clear polls was more than 3 standard deviations from the actual result
the thing about polls is that the non-sampling error is larger than the sampling error. the big errors are from non-response bias, likely voter identification, party weighting etc. these things are guesses and this election cycle’s hero will be next cycles goat
someone like Nate Silver can come up with an overfitted model that fits the past data well. its like economic forecasting, as long as future elections are similar to the past, he will be accurate. when there is a dramatic shift (like 1980) Nate Silver will never catch it, just like all the “blue chip” forecasters don’t ever forecast things like financial crises.
Because he factors in voter fraud ...pure and simple