Fooled by Randomness, Risk Perception, and Risk Management
Fooled by Randomness, authored by Taleb, is one of my favourite books.
At the center of his thesis – Pareidolia – the tendency of humans to see patterns in random sequences.
We’re programmed, all humans (with possible debilitating exceptions), with the ability to make sense out of randomness. How else would you be able to recognize faces or learn language from birth? One of my favourite passtimes is to sit at the roulette and craps tables and listen to the fantastic theories of how the wheel and dice behave.
Paired with that, or almost implicit in that, is the tendency to under-estimate risk. In fact, so many of us are programmed with a case-study mentality that it becomes even harder to estimate risk.
I’m going to augment Taleb’s thesis by arguing that a key aspect in risk perception is your degree of controllability.
You have some people who had a perception of absolute controllability over what they were doing, and they drastically underestimated risk. When you take a look at the CDO market, and at some of the people who were authoring those CDO’s, well, I think they didn’t properly perceive risk, over and above the problem of Pareidolia. Specifically, you had people writing insurance contracts against events which the authors believed would never happen. These were Six Sigma events in their minds.
It’s also part of the reasons why people are more afraid, generally, of poor maintenance on airplanes than their own driving habits on the roads. Even though the odds of death on the roads is one or two orders of magnitude higher. Controllability, or perception of controllability, reduces risk perception.
I can think of another company that tried to neuter risk. LTCM.
There’s an important lesson here for me, at least.
I’m a manager of risk. Most web analyst managers, directors, and owners are also active managers of risk.
The same way that sometimes analytics data can fool me, patterns out there in the world, and with staff and revenue can also fool me. I shouldn’t be lulled into a sense that I can properly perceive risk just because I can control some, or, all of it.