The "cubic model" from @CEA + Kevin Hassett was pretty clearly fit on log(deaths + 1). Which is... pretty dangerous for forecasting.
— David Robinson (@drob) May 5, 2020
Just imagine if they'd fit a quartic model pic.twitter.com/VTlFoD9qvC
The "cubic model" from @CEA + Kevin Hassett was pretty clearly fit on log(deaths + 1). Which is... pretty dangerous for forecasting.
— David Robinson (@drob) May 5, 2020
Just imagine if they'd fit a quartic model pic.twitter.com/VTlFoD9qvC
Code here: https://t.co/vg6S4LmQ67 pic.twitter.com/CRn5mh5T3p
— David Robinson (@drob) May 5, 2020