The binomial theorem in action: if the infected percent in a community is very low, then: if you want 95% confidence of *zero* infected in a meeting, then the meeting size times the number of infected people in the community should be 1/20 of the community size.

Example: we have 10 confirmed ppl currently infected, maybe 100 actual infected, community of 100k. So for 95% confidence of no one in a meeting being infected, (meeting size)(100) = 100k/20 = 5000, so meeting size should be 50.

Many other variables and inputs, but it makes some sense, I think. Shout out this thread to @Noahpinion @politicalmath @NateSilver538