For example what if I was to claim there are really two distinct groups of 'unvaccinated' Covid people. Group A and Group B unvaccinated. Group A only contains 1 person and very clearly is not clogging up the hospitals. Someone claims to be 'Group A' unvaccinated and asserts their actions are not a problem because they are not clogging up the system.
A more realistic example is there are 20 smaller groups of people that in total are the same size as the Covid unvaccinated group but these groups have twice the risk as the unvaccinated group. If we make decisions based on group_size * risk then these groups are fine but we see that combined they have a much worse effect on capacity than the unvaccinated group. If we make decisions based solely on risk and not risk * group_size then the 'triage' protocol is going to make more sense.
Practically, you might have to treat groups differently because it's impossible to make efficient rules for a lot of small groups but ignoring the practical costs of rule making it is risk that is important not the size of the groups.