Dear Nate Silver, self-appointed expert on elections:
You said there was a 99% chance of Hillary Clinton would win Michigan.
If you took bets on your forecasts, you would be broke.
The polls this year have been outrageously in error.
The polls missed it by an average of 23 points in the Michigan Democratic primary -- well beyond the wild claims of accuracy to within 5 points.
The Idaho Republican primary on the same day? According to the Real Clear Average, Trump was supposed to romp by 11 -- lost by 17. But that was based on one poll and you did not make a forecast. Good move.
I get that you are a creature of the polls. You feed off them. Maybe you should eat elsewhere. Yours is a 2008 approach to a 2016 race. The polls are as trustworthy as Kim Jong-Un. Consider:
Louisiana? RCP average said Trump would win by 15. He won by 3.
Oklahoma? RCP average said Trump would win by 11. Lost by 6.
Virginia? RCP average said Trump would win by 14. He won by 3.
I won't even go into Iowa. It was a mess. And I am leaving out the romps where Trump's margin (or Clinton's) were understated.
Reporters keep explaining the errors in the polls as someone either falling or shooting up in the final few hours before the actual vote.
But why assume the polls are right? History shows that the vaunted Iowa Poll -- the gold standard according to reporters -- averages a miss of 7.5 points in contested races going back to 1988. Pretty useless.
Given the wide variance in the polls, you might as well be reading goat entrails.
With due respect,
Don Surber, self-appointed critic of polls.