There are many types of catastrophe. Some of them happen all of a sudden and some of them happen in slow motion. Car crashes and lung cancer are both awful but they both happen under quite different conditions and time scales, though one could smoke while driving under the influence. Prevention and mitigation mechanisms differ, if one is going to breathe or be around cars at all (both might be difficult to avoid). While Nassim Taleb talks about structures and situations that gain from disorder, there is only so much disorder even an anti-fragile system can absorb. Perhaps the body becomes healthier when exposed to colder and warmer temperatures, up to the hottest stone sauna and down to cutting a hole in the ice for a brief plunge. However, it seems unlikely that the body will benefit from exposure to temperatures hotter than such a sauna and colder than the above lake, particularly if exposure happens over a long time.
Back to the car crash and lung cancer. If a car crash were an unacceptable outcome, one would not drive or put oneself in the vicinity of cars. While I hope there are still places in the world where cars have not proliferated, most people seem to be among and within cars from time to time. So car crashes seem to be acceptable, as we are near them and it appears to be possible in being near a moving car that one may experience a crash in some form. However, certain magnitudes and outcomes of crashes are unacceptable. Thus, speeding more than 20 miles per hour over the speed limit may be grounds for arrest and cars are equipped with safety features that make certain types of crash outcomes far less likely (particularly the deaths of those inside cars). Lung cancer is an outcome that seems to become more likely over longer periods of time and under a somewhat more vague series of conditions. There seems to be strong evidence of a causal link between smoking and lung cancer, and so not smoking might be a straightforward way of reducing the risk. Lung cancer victims may not have smoked in their lives, however, and I have not observed complete clarity of the causes and probabilities of experiencing lung cancer from different possible causal factors. If lung cancer were a completely unacceptable outcome, it might be possible to observe groups of people that have an extremely low or perhaps even near a zero incidence of lung cancer and replicate the conditions under which they live. As lung cancer seems to be at least a partially acceptable risk in society (that tobacco companies are still in business appears to me evidence of this), much appears to be invested in curing and mitigating the effects of this form of cancer, in addition to educating people of the causes and preventative measures that can be taken.
The above examples seem to be representative of bounded volatility structures, practices, and expectations. Expectations of volatility levels in a group drive the structures and practices that group will build and propagate in advance of, at the moment of, and after a spike on a given volatility variable’s graph. Finding anti-fragile volatility relationships and investing in them does not appear to be sufficient for a harmonious and efficacious society. Groups (including countries, organizations, and families) might find it useful to share volatility expectations with respect to key variables – heat, cold, air quality, water quality, security all might come to mind, among many other factors that could be important. Insurance companies might even find ways to discover what their policyholders expect on the relevant volatility areas (home safety, car crashes, workplace accidents) and support their customers and members in not only transferring financial risk, but also developing risk management approaches and disaster preparations to help outcomes fall outside the acceptable volatility boundaries less often and make recovery faster when the boundaries are exceeded.
Roman Ryan 