WS MoreOrLess: What is "rare"?

Episode Summary

Executive Summary: The episode examines how vague words like "rare" and "unlikely" map imperfectly onto probabilities, showing that everyday language often conveys meaning beyond numbers. It then turns to Nassim Taleb’s argument that we should design systems to be anti-fragile—able to benefit from shocks—rather than merely trying to predict rare, high-impact events.

Main Topics: The meaning of "rare" in everyday language (Priority: 5/5): The show explores whether "rare" has a precise numerical meaning, prompted by a listener with motor neuron disease. It contrasts ordinary language with statistical definitions and asks what speakers intend when they use such terms. Psychology of probability words (Priority: 5/5): Neil Stewart explains that people interpret verbal probability terms differently, with research estimating typical percentages attached to words like "rare," "never," and "always." Formal risk lexicons in science and policy (Priority: 4/5): The IPCC’s confidence language is presented as an attempt to standardize phrases such as "very high confidence" and "medium confidence" into fixed probability ranges. When numbers are more useful than words—and vice versa (Priority: 4/5): The program argues that the best choice between numbers and verbal descriptions depends on context and on how much supporting information the listener already has. Nassim Taleb on black swans and anti-fragility (Priority: 5/5): Taleb explains that rare, catastrophic events are hard to predict but should motivate building systems that can absorb or benefit from shocks instead of trying to forecast every extreme event. Contagion and systemic resilience (Priority: 4/5): The interview extends Taleb’s ideas to airlines and banks, emphasizing that resilient systems prevent failures from spreading and avoid making the next failure more likely.

Key Arguments: There is no universally agreed precise meaning of "rare" in everyday language, even though "rare disease" has a formal definition. People interpret verbal probability terms as approximate percentages, but with substantial variation across individuals. Scientific bodies like the IPCC try to standardize risk language, but verbal categories can still imply different numerical ranges and ambiguity. Words can communicate significance and inference, not just numerical magnitude, so they may be more useful than exact numbers in some contexts. Numbers are only helpful when interpreted alongside context; isolated figures can be misleading or useless. Taleb argues that the goal should not be perfect prediction of black swans, but designing systems that withstand or benefit from them. Anti-fragile systems use stress, error, or disruption as fuel, while fragile systems suffer from it. Systemic failures can spread through contagion, so resilience requires preventing one collapse from making others more likely.

Data Points: Rare disease threshold: fewer than 0.05% of the population - Definition used by UK authorities and the European Union Rare disease threshold in ratio form: 1 in 2,000 people - Equivalent formal definition of a rare disease Typical meaning of "rarely": 8% chance - Neil Stewart’s sample mapping verbal probability terms to percentages Range for people interpreting "rarely": 0% to 20% - Shows variability in how individuals interpret the term "Toss-up" probability: 47% - Position in Stewart’s verbal probability scale "Always" probability: 94% - Top end of Stewart’s sample list of probability words IPCC "very high confidence": at least a 9 out of 10 chance - Example of a formal mapping from language to probability IPCC "medium confidence": 5 out of 10 chance - Another standardized confidence term Air travel crash example: 70 persons die - Taleb uses this as an example of a catastrophic event affecting system learning Listener’s disease context: 5,000 sufferers and 1,500 deaths per year - Used to illustrate the tension between rarity as language and rarity as statistics Plane seats example: 20 seats left - Illustrates that a number can be unhelpful without additional context Black swan events: very hard to predict; massively large consequences - Taleb’s definition of black swans

Pivotal Quotes: "What does rarity really mean?" — Les Halpin (email read by actor): The listener’s core question about the mismatch between language and numeric precision "They're telling you the significance of that number as well as giving you a number." — Linda Moxie: Explaining why words like rare and common do more than encode imprecise quantities "The mistake people made is to say, well, you know, let's try to predict them, right? You're not going to get there." — Nassim Taleb: On black swan events and the limits of prediction

Implications: Listeners should treat probability words as context-dependent, not exact measurements. For policy, science, and media, the best practice is often a mix of numbers and plain language. For institutions, resilience matters more than prediction when facing extreme uncertainty.