Nate Silver @NateSilver538
The Ray Fair model that @nytopinion highlights as an example of forecasting prescience was actually off by *10* points. It had Trump winning the popular vote by 8 points and he lost by 2. About twice as bad as the ~5-point polling error in Wisconsin. — PolitiTweet.org
Scott Lemieux @LemieuxLGM
But they have a model that got within *8 points* of correctly predicting the 2016 presidential election! I hardly t… https://t.co/YdaeJb5GAh