> Aside: Three years? I wonder how long before there's a reasonable chance that some of the purchasers died before the release?
Assuming all his customers are roughly 20-40 years old, the annual mortality rate is ~0.1% IIRC on average (mortality rates, because of the exponentiality of the Gompertz curve, only begin substantially increasing post-40 years, up to the ~90s with annual mortality rates like 10-50%). So then the 3-year mortality rate would be 1-(1-0.001)^3=0.002997001 or ~0.3%. So roughly you would need ~334 customers to expect 1 death over 4 years (100/0.3, since 333.3333333 * 0.002997001 ~> 0.9990003332), with the full binomial distribution covering 2/3/4/5/6 deaths as well (but not much higher with any meaningful probability). I forget where I read it, but I feel like I remember that at least 1000 people pre-ordered the course, in which case the mean number of deaths becomes more like 3 with up to 9 deaths, etc.
Assuming all his customers are roughly 20-40 years old, the annual mortality rate is ~0.1% IIRC on average (mortality rates, because of the exponentiality of the Gompertz curve, only begin substantially increasing post-40 years, up to the ~90s with annual mortality rates like 10-50%). So then the 3-year mortality rate would be 1-(1-0.001)^3=0.002997001 or ~0.3%. So roughly you would need ~334 customers to expect 1 death over 4 years (100/0.3, since 333.3333333 * 0.002997001 ~> 0.9990003332), with the full binomial distribution covering 2/3/4/5/6 deaths as well (but not much higher with any meaningful probability). I forget where I read it, but I feel like I remember that at least 1000 people pre-ordered the course, in which case the mean number of deaths becomes more like 3 with up to 9 deaths, etc.