Narrative Fallacy
Nassim Taleb describes the narrative fallacy: our tendency to construct stories that explain sequences of facts, even when the facts are largely random. We see a company succeed and invent reasons — visionary leadership, superior culture, strategic insight. We ignore that equally visionary, cultured, insightful companies failed. The story isn’t false exactly — it’s incomplete in ways that create the illusion of understanding.
Stories compress and order. They take the blooming confusion of events and impose beginning, middle, causation. This is useful for memory and communication. But the compression loses information. The story privileges facts that fit and discards facts that don’t. What remains is coherent but not representative.
The fallacy interacts with hindsight bias. Once we know the outcome, we construct narratives that make the outcome feel inevitable. The successful entrepreneur was always destined for success — we find the childhood signs, the pivotal decisions, the defining traits. We don’t tell the same story about the failed entrepreneur with the same signs, decisions, and traits.
The fallacy also amplifies survivorship bias. We only hear stories of survivors. Their narratives become templates: “do what I did.” But we never hear from the failures who did the same things. The narrative is drawn from an unrepresentative sample, but the sample bias is invisible.
Stories remain essential — they’re how we think. The trap is mistaking narrative coherence for causal understanding. A story that makes sense can still be false. The world runs messier than any narrative; luck plays a larger role than stories suggest; patterns we see may not generalize.
The antidote isn’t to stop telling stories but to hold them lightly. To ask: “What would have to be true for this story to be wrong?” To look for the stories not being told. To remember that the most compelling narrative is often the one that flatters our need for coherence, not the one that reflects reality.
Related: hindsight bias, survivorship bias, signal and noise, epistemology, map and territory