stats nerds: here is a simple worksheet showcasing the probability of detecting X or more IgG/IgM+ individuals in SF, based on assumed infection rate of 1,2,3,5,10x the current PCR+ reported rate, and test sample size. does this look right? (pls DM) docs.google.com/spreadsheets/d…
if so, the probability of detecting 1 or more people with 1,000 IgG/IgM tests in SF is 32%, 54%, 68%, 85%, 98% assuming 1,2,3,5,10x the current reported infection rate of the population (its .04% as of this morning).. we are going to need a LOT more tests! but will keep going..
even with 1000 tests, going to be hard to draw any hard conclusions about % of asymptomatic/untested patients, and multiple detections unlikely until 7000+ tests.. we're currently in 100s. lots of tests go "missing in the wild" but the data we're getting seems clean & reliable
