Not gonna do any model meta stuff since 1) who cares 2) everyone's forecasts are *pretty* similar. But just one stray comment on this interesting reading from the Economist team on how they think their forecast (Biden 96%) may be overconfident:

https://statmodeling.stat.columbia.edu/2020/11/02/so-whats-with-that-claim-that-biden-has-a-96-chance-of-winning-some-thoughts-with-josh-miller/

The comment is: Having a few elections under your belt helps a *lot*. No matter how much you test things in the lab, there are some things you're going to learn only by seeing how your forecast reacts to real data in real time. (I'm sure this applies to lots of other stuff too.)

I think our forecast has followed an intuitive path this year. It's hedging its bets more in the summer, then it starts to get more confident after Labor Day. But it's been pretty smooth, in contrast to our bouncier 2016 model (though of course the polls were way bouncer in '16).