4 ms·

>gave me a heuristic for thinking about long-tails and uncertain events If it's possible to summarize and share the heuristic in this forum format, I request y

>gave me a heuristic for thinking about long-tails and uncertain events If it's possible to summarize and share the heuristic in this forum format, I request you to please do. Am always learning from Taleb's books and ideas.

wenc 5y ago
I'm hard pressed to summarize the book because it introduces so many ideas. Perhaps I can point you to summaries other folks have written, like [1]. That said, I can rattle off some heuristics from Taleb that I've applied in real life: 1) Look out for fat tails and rare events in real life. Don't obsess over trying to predict them (you cannot -- rare events are such that they are inherently difficult to predict. They're rare!); instead protect yourself against their downsides by designing the appropriate defenses for when they do happen (e.g. having some form of insurance, setting up legal clauses that limit liability, etc.) For instance, contracts can be set up to limit liability from very rare events that have damages that scale infinitely. This can be done by adding a clause that says if said rare event happens, one will only pay damages up to $x, which contains the downside. (Insurance companies do this a lot) 2) When to take risks? In situations where there are low downsides if you're wrong and high upside if you're right. When to not take risks? When the opposite is true. 3) Don't mistake individual risk for systems risk. The startup community valorizes risk-taking (after all, nothing ventured nothing gained). However never fall into the trap into thinking that risk-taking makes everyone stronger. When individuals fail, they don't always get stronger (many are destroyed) -- instead their failures make the system/ecosystem stronger. It's like airplane crashes, which are themselves unrecoverable failures. But each airplane crash actually improves airline safety for everyone through learnings obtained from post-mortems and blackboxes, which lead to new FAA guidelines. 4) Don't obsess over probabilities alone -- pay attention to payoffs (or better yet, focus on "expectation", which is the sum/integral of probabilities times payoffs). (everybody learns this in statistics class but no one has thought about how to apply this). You want to maximize expectation, not probability. [1] https://tylerdevries.com/book-summaries/fooled-by-randomness/ https://tylerdevries.com/book-summaries/fooled-by-randomness...
- reactspa 5y ago
  Thanks -- good distillation of Taleb.

wenc 5y ago

I'm hard pressed to summarize the book because it introduces so many ideas. Perhaps I can point you to summaries other folks have written, like [1]. That said, I can rattle off some heuristics from Taleb that I've applied in real life: 1) Look out for fat tails and rare events in real life. Don't obsess over trying to predict them (you cannot -- rare events are such that they are inherently difficult to predict. They're rare!); instead protect yourself against their downsides by designing the appropriate defenses for when they do happen (e.g. having some form of insurance, setting up legal clauses that limit liability, etc.) For instance, contracts can be set up to limit liability from very rare events that have damages that scale infinitely. This can be done by adding a clause that says if said rare event happens, one will only pay damages up to $x, which contains the downside. (Insurance companies do this a lot) 2) When to take risks? In situations where there are low downsides if you're wrong and high upside if you're right. When to not take risks? When the opposite is true. 3) Don't mistake individual risk for systems risk. The startup community valorizes risk-taking (after all, nothing ventured nothing gained). However never fall into the trap into thinking that risk-taking makes everyone stronger. When individuals fail, they don't always get stronger (many are destroyed) -- instead their failures make the system/ecosystem stronger. It's like airplane crashes, which are themselves unrecoverable failures. But each airplane crash actually improves airline safety for everyone through learnings obtained from post-mortems and blackboxes, which lead to new FAA guidelines. 4) Don't obsess over probabilities alone -- pay attention to payoffs (or better yet, focus on "expectation", which is the sum/integral of probabilities times payoffs). (everybody learns this in statistics class but no one has thought about how to apply this). You want to maximize expectation, not probability. [1] https://tylerdevries.com/book-summaries/fooled-by-randomness/ https://tylerdevries.com/book-summaries/fooled-by-randomness...

reactspa 5y ago
Thanks -- good distillation of Taleb.