Bad tests worse than no tests? No, actually, probably not.

• In response to a comment that inaccurate testing could be worse than no testing, I came up with the following as a response.

Well, yes, to some extent, but as Paul Romer has argued (https://paulromer.net/), this is not a black and white issue. Even tests with some degree of systematic inaccuracy, while potentially pretty bad for an individual affected, can be almost as useful as non-existent "perfect" testing in getting the critical infection rate factor R0 ("R-naught") down to ~1 or below. Think of it this way. Unchecked, an infected person will infect about 2.5 others (R0 ~ 2.5). If even somewhat flawed (say 90% accurate) testing, with follow up contact tracing and isolation of positives and close contacts of positives, were aggressively pursued, it should be possible to get the real R0 down to something like 0.7, meaning each person on average infects less than one person; 10 infections lead to 7 new infections. That process leads to complete eradication of the epidemic, fairly quickly, even though that may seem counterintuitive. So, in other words, yes, of course, they need to strive for the most accurate testing possible, but even a moderately flawed testing regime could be effective in eliminating the virus from the population. Whereas our current inadequate, in fact nearly nonexistent, regime CERTAINLY will not. As things are now the ONLY things that will end the epidemic are 1) it burns through the population, killing 2 million or possibly many more people, and disappears on its own in a year or two; 2) we develop effective antivirals or vaccines, and maintain social isolation... again for a year or two. We could have some unsatisfactory in between, which would be less bad but still really bad and would still last well over a year, or we can mobilize and do the necessary millions of tests, contact tracing, and isolation, and actually suppress this thing in a few months. Seems to me there is no real basis to argue for any other course of action