HACKER Q&A
📣 godelmachine

Which are some Covid-19 simulation models that are accurate?


I posted this recently - https://news.ycombinator.com/item?id=22981283

Also, there a model which Trump is following that predicts 80k American lives will fall by August 2020 to COVID-19, which is turning out to be prophetic.

Are there other COVID-19 simulation models which have turned out to be presciently accurate?


  👤 giantg2 Accepted Answer ✓
It depends on your definition of accurate.

Most of the models have been revised as we get new information. Since there is still a lot we don't know about the virus, it is very likely that the models will be updated again when we get that new information. For instance, what if it recurs seasonally?

My opinion is that we won't be able to provide truely accurate estimates for something which is so new, so unstudied, and rapidly changing (policies or virus mutations). We can say stuff like the US death toll for the year won't be measured in millions, and most would agree that would be accurate with what we know now.


👤 buboard
they just released the code for the Imperial model that the UK apparently uses

https://github.com/mrc-ide/covid-sim

Seems like a game of simcity, honestly. With so many parameters that are not very precisely known, it's easy to go into fantasyland. On second look, this model has not been used to reproduce the epidemiology of a previous , real world infection. weird

i think you re more likely to get accurate predictions with typical SEIR models

BTW this is the model the US uses (i think): https://www.medrxiv.org/content/10.1101/2020.03.27.20043752v...

It's statistical in nature, so closer to real world