Another way to get there is that, given N shots at a 1/N event, you expect 0 hits 36% of the time (1/e). You can divide the population up into into 600 groups

Another way to get there is that, given N shots at a 1/N event, you expect 0 hits 36% of the time (1/e). You can divide the population up into into 600 groups with 500000 people in them, and each of those has a 36% chance of not ever hitting. So seeing no cases is like flipping a 36% coin 600 times and hitting 600 times.