Read on Twitter
Not necessarily. Election forecasting isn't a binary thing, or else Helmut Norputh was the "most correct" in 2016 because he predicted a Trump win, but Norputh ALSO said that Trump would win 52.5% of the popular vote, which 6.4 points off https://twitter.com/Ignacioo731/status/1297176516657324033
Meanwhile, Silver said Trump would only win 44.9%, which was only 1.2 points off. The 5.2-point difference between Norputh and Silver is obviously huge, and Nate was 5.3 times "more correct" if you want to view it that way
However, the gap between the projected margin between The Economist's and 538's models isn't *as* huge
Silver projects that the PV margin will be D+7.0 while Morris projects that the PV margin will be D+8.2. 1.2 points is still a big difference in projections
Silver projects that the PV margin will be D+7.0 while Morris projects that the PV margin will be D+8.2. 1.2 points is still a big difference in projections
Morris vs. Silver is a battle of certainty too. The aggregate conclusion may not be too different, but when you go state-by-state. Morris is way more certain on "SOLID" states for either party, especially ones that favor Biden
A good test will be if Trump wins LEAN D states like MN, NH, ME, or NV. Trump can still win some LEAN D states while losing overall via Biden winning Sun Belt. These sort of "one state away" maps can manifest, and we need to be ready for these potential outcomes
If orderings like these happen -- whereby, simultaneously, LEAN R states go Biden and some LEAN D states go Trump -- Silver's theory of "uncertainty" will be proven right, and Morris's model will be "less" correct even if Biden ends up winning
Lol sorry for this thread @Ignacioo731. You made me go on a tangential rant
