3 ms·

> many models predicted over 2 million deaths by now citation needed. Here are many models that we can backcheck: https://projects.fivethirtyeight.com/covid-f

• vernie 6y ago
  I wish the historical view on this site projected forward to the present. I'd like to know which model was the least wrong from the beginning.
• jjcon 6y ago
  > without action by the government and individuals to slow the spread of coronavirus and suppress new cases, 2.2 million people in the United States could die https://www.nytimes.com/2020/03/16/us/coronavirus-fatality-rate-white-house.html https://www.nytimes.com/2020/03/16/us/coronavirus-fatality-r...
• analog31 6y ago
  In addition to the many articles that have been published, it has been possible since early March for people to extrapolate the graphs that were widely available. Those graphs, and their potential shortcomings, were widely discussed, and a "citizen scientist" could interpret them with reasonable confidence. My own analysis at the time was something like this: If the graph keeps going like the way it looks, it will consume effectively the entire population by month X, with Y percent dying. While this was not a formal epidemiological model, it allowed two tentative conclusions: 1. If the curve (a straight line on a semi-log graph) shows signs of turning over, I will breathe a huge sigh of relief, the sooner the better. 2. The curve was a "sniff test" for whether more advanced predictions were believable, notably, any prediction that did not roughly follow this baseline needed a darn good explanation.