>Citations needed. Well, like you point out it depends on how many of the rare effects are determined by variation from person to person, and how many are dete

>Citations needed. Well, like you point out it depends on how many of the rare effects are determined by variation from person to person, and how many are determined by variation from time to time. If it's all person-to-person then your chance of a rare effect drops to zero once you have a single dose, but if it's time-to-time, your chances of making it through without something weird happening drop as (P_safe)^n, i.e., exponentially. Without any studies I know of that would tell us which was happening, we could assign a neutral prior of 50% to both cases. That leads to: P(safe) = 1/2(1-0) + 1/2 (P_safe)^n. Obviously this is an oversimplification, but it still leads to an exponential decay, just not to 0.