Looking back far enough, there are lots of examples of extreme events (high impact, low probability). Numbers we use for risk analysis make better questions than answers, because they never include the full picture. A great book to read to give you a little bit of humility is the story of Long-Term Capital Management, titled “When Genius Fails”. It is an important book to read for those who think they have cracked the code. Thinking in “distributions” (a range of possibility) rather than simple numbers is vital. Howard Marks implores that we, “always remember the six-foot man who drowned in a river that was five-foot deep on average”. It is not just the average that matters. The statistical term is moments. You have to think in moments. How the range of possibility is made up. That includes thinking about tail risk, and what isn’t in the numbers (yet). In scenario testing, you need to plan for what could happen, not just a single path. In the same way as looking at what did happen is not enough. What could have happened? You need to plan for your plan not covering everything. With hindsight, we think we understand the world. We think the lessons we learn through mistakes mean, “if I had done it this way, this is how it would have played out.” In reality, if you had the chance again, everything would be completely different.
No comments:
Post a Comment