IMO this is pretty fundamentally irreconcilable

538, Aug 26 2016: Clinton +4, 18% other/undecided, 81% to win
538, Aug 26 2020: Biden +9, 7% other/undecided, 70% to win

Here are the numbers for the "polls-plus" model from 2016, which look better but still don't make the most sense to me. If uncertainty is the same like Nate claims, I don't see how 2020 D prob is lower than 2016

A bit of clarification: If you calculate the implied standard deviation in electoral votes for 2016 and sample from a normal (or even student T) distribution for 2020 you get a Biden win % closer to 80 than 70. This is what I mean when I say something doesn’t make sense to me.