Well, we know that: Expected payoff = Probability * Impact. Extreme activities have a low probability of resulting in anything bad (especially with the gear a

Well, we know that: Expected payoff = Probability * Impact. Extreme activities have a low probability of resulting in anything bad (especially with the gear and protocol we have today), but the impact can be high when it happens -- though if we take the summation across the population the overall impact is low because only a small sliver of society will participate. Sugar consumption has medium probability of leading to poorer health outcomes, and impact is also medium, but I suspect the overall payoff (or harm in this case) is much higher when summed across the population. So yes, from a purely rational perspective, we should try to stop people from doing things that harm the system as a whole. If they want to do things that harm themselves at a localized level (with no 2nd or 3rd order effects -- though this is rare), then on balance policies should allow them to.