Thursday, Aug. 6A general problem with this pandemic is estimating the number of infections for a given set of test results. We know we are missing a lot of cases, so how can we estimate that which we missed?Today, I present a comparison of deaths per capita and cases per capita, and states with high test positivity are colored blue.They are not biased to the left of the diagonal on this plot, which indicates that higher test positivity likely no longer indicates more missed infections, which is a good thing. The variability in test rates currently seen in our states is not helping us find more cases, and we should not get alarmed about high test positivity anymore. We can simply get alarmed about cases per capita instead.This plot suggests the deaths per case in the United States runs at about 2 percent right now, and if the infection fatality ratio is 0.5 percent, that means we miss identifying three cases for each one we find. These missed cases are predominantly people under the age of 40, and there is a good indication that some of these people are quite capable of transmitting the virus.These are ballpark numbers, and epidemiologists will spend a lot of time and effort getting the details correct.For this analysis, I used a 28-day window in which to sample deaths and cases, and time shifted the deaths by 20 days relative to the cases. All the data comes from: covidtracking.com/.Here is the graph:As always, if you would like the code to anything I post here, feel free to email me at dblake@augusta.edu. Contact Dave Blake.