tags: #publish links: [[Cognitive Biases]], [[Psychology]] created: 2022-02-12 Sat --- # Heuristics That Almost Always Work From Scott Alexander: https://astralcodexten.substack.com/p/heuristics-that-almost-always-work This is closely related to [[Smooth-Sailing fallacy]]. See also [[SLOs and rare events]]. An interesting dynamic: Rare events provide an opportunity for "experts" to convincingly "predict" that they won't happen, using a trivial heuristic (e.g. *always return false*) and no real analysis. This is likely to happen because it's less work for the expert, real prediction is very hard, and there's a reputation cost to false positives. You can build a career on this... not a very good career. **Examples:** - Doctor: *It's probably nothing, take an aspirin.* - Medical advisor: *This probably won't become a pandemic.* - Druidic volcano predictor: *No it's not going to erupt this year.* - Security guard: *That noise is probably just the wind.* When the rare event does happen, people will probably give you a break for getting it wrong, because it was hard to predict! The existence of the heuristic "expert* causes everyone else to under-predict the rare event, too. > ...the point isn’t just “sometimes black swans happen”, but that the existence of experts using heuristics causes predictable over-updates towards those heuristics. > Whenever someone pooh-poohs rationality as unnecessary, or makes fun of rationalists for spending zillions of brain cycles on “obvious” questions, check how they’re making their decisions. 99.9% of the time, it’s Heuristics That Almost Always Work.