d-mac posted:Michael Levitt, a Nobel laureate, has said the models are wrong. There is a March 24th article in the LA Times where he predicted we would get through this much quicker and without as many deaths as predicted. The head of data sciences at Stanford has made similar comments. There are others who have ripped the models for making assumptions off of bad or incomplete data.
Thanks. I'd point out that Leavitt is a chemist, but you'd say I'm quibbling. He is not doing any predictive modelling, he's not looking at cause and effect, or means of disease transmission, he is only looking at the raw data and predicting whether the numbers are going up or down. He was dead wrong about Iran (yes, I said "wrong"). On March 30, he said that New York had turned a corner (Daily Mail article), which it had not.
More to the point, Leavitt says that social distancing works, that it worked for China, and that he agrees with those calling for strong measures, especially social-distancing, immediate social isolation, and more testing and detection. So while he may be "right" in hindsight about total deaths, he doesn't seem to have anything predictive to say. He says that with distancing, it's not as bad as the models predicted - well, that IS exactly what the models predicted.
The point of predictive modelling is to say what you think will happen, to guide future action. Leavitt and everyone else have said that we have to lock down to control this. We did, and it's being controlled. What if the original models had not been so dire? Then no-one might have locked down, and the worst-cases might have come true.