Fig. 10.

Length of pandemic vs. logarithm of Average COVID-19 deaths per day (per million). Both graphs exhibit a very peculiar and complex dependence of the two variables; (a) presents data from basic behavioural scenarios; two main clusters can be observed one from the scenarios with functioning healthcare system (without collapse) and the other from the scenarios in collapse. Both the clusters happen for very short pandemics but with very different death rates. (b) presents data from special groups behavioural scenarios; only one cluster can be observed related to the scenarios in collapse while the data related to the functioning healthcare system (without collapse) is somehow shifted towards the same functional dependence of the scenarios in collapse. Another important difference is that in (b) there are not scenarios in collapse where the pandemic is longer than 450 days while in (a) there are several even above 600 days in length. Meaning that for the special groups behavioural scenarios all the simulations where the pandemic lasts the longest (the infection curve is successfully flattened) have never reached healthcare collapse.