FiveThirtyEight has released its forecast for the 2020 US election. At this point, I’m less interested in the actual forecast (Biden favoured to win, with a 72 in 100 chance), but in the design choices they made to communicate what exactly you can take away from such a forecast and what you shouldn’t. You can clearly see how the reckoning with 2016 has shaped their decisions (the main issue in 2016, at least with FiveThirtyEight, wasn’t that forecasts were wrong, the inherent uncertainty was just poorly communicated). Have a look.

From Weekly Filet #312, in August 2020.