Nate Silver backtesting his model
-
-
He was favored 100 days out but I‘m pretty sure I remember the polls tightening in the fall.
Used to be after one convention the polls would swing that way and after the next convention they’d swing the other.
-
I suspect that a Trump will lose big time, but I have witnessed enough new data to come in of late that I’m not so sure that Nate’s models know how to deal with the uncertainty of the next 100 days.
I’m a big fan of data but I’ve also seen it be so much less predictive of late than I thought possible.