>You might be right that there could be large numbers of people with the virus who are doing just fine right now, but we know from nearly 200,000 cases (that's plenty) that a non-trivial portion of cases require hospitalization. That's the whole problem, right there.

You're not applying the logic all the way through.

If a great number of people are already infected and do not require hospitalization it grossly affects the math on the potential for overwhelming our medical system.

Those infected people unnoticeable to minor symptoms have the potential to be the herd immunity we'd want from a vaccine.

> All that is to say that the situation can only be worse than we already know. I'm not sure more data would lead us to isolate less at all.

More infected people who are unnoticed because they don't show severe symptoms and require hospitalization actually makes the situation better. It would lower the hospitalization rate and alter the projections drastically in the direction of less epidemic.

We are being cautious right now. That, by definition, means we're likely overreacting.

I was wondering if there’s an efficient way to screen a large population with low prevalence with limited tests. I’m thinking of the 1000 prisoner wine riddle with one drop of poison if that rings a bell

I thought that, but then how do you explain situations like Italy? Such a tremendous number of people are suffering.

> The trouble is we can only get "real data" by testing more people.

We get "real data" by properly testing, not necessarily by testing more people.

Currently we are mainly testing people arriving with symptoms.

This results in underestimate of the number of infected and an overestimate of mortality and morbidity rates.