Why an outbreak overshoots

Ten thousand people and one infection that each case passes to a few others. Herd immunity is the point where each case infects fewer than one other. Watch where on the curve that point comes, and how many more catch it afterwards.

The rules, all of them. People mix evenly. An infectious person stays infectious for some days and, where nobody is immune, would infect a set number of others in that time (that number is the slider). In each tenth of a day every person who has not had it has the same chance of catching it, in proportion to the share who are infectious. Recovered and vaccinated people cannot catch it again.

The amber line is exact: 1 ÷ (cases per case) of the population still to catch it. The outbreak peaks as it crosses that line, with a large number of people infectious at that moment, and they go on infecting others on the way down. The share who end up catching it solves s = exp(−R × (1 − s)), which this model matches. What it leaves out: households, age, changing behaviour, waning immunity, deaths. It is the simplest outbreak there is, not a forecast of any real one.