I’ve considered the question of test more, find more as there are both asymptomatic cases and cases not bad enough that people and doctors just recommend stay at home - no need for the test leading to the expectation that real cases may be not just more but much more and therefore the expectation that random testing would return a significant rate (somewhere between a few and a couple tens of percent). For this particular blog the question becomes is there a way to separate the effects of changing rate of testing vs rate of infections. For instance now (May 12th to be specific) the confirmed rate appear roughly flat, but the daily positive rate (ratio of positive to tests) has been clearly on the decline and below the total cumulative positive rate. So it’d appear the total cases is likely on the decline. However, not exactly sure and haven’t come up with a way to separate. I have tried adjusting for the changing test rate (use a sliding 2 week data to determine marginal positive rate and adjust towards assumed if constant test rate) with interesting results, but far from truly conclusive.