Fig. 4.

(A) Select Pareto-efficient testing policies (frequencies for each group) according to median simulated infections. Vertical bars depict the distribution of infections, ranging from the 10% to the 90% quantile over simulation outcomes. The point highlighted in yellow corresponds to the testing frequencies Cornell selected for the fall semester. Policies shown are Pareto-efficient policies from the set of policies where each on-campus group is tested either once or twice a week. The dotted lines are simulation estimates of the expected number of infections if tests are split homogeneously among the on-campus population. (B) Contour plot showing the number of infections in the Cornell community as test compliance and test sensitivity vary under the pessimistic (residential) scenario.