The Ludic Fallacy Life Is Not a Casino
We commit the Ludic Fallacy when we assume that the rules, variables, and boundaries of real life are as known and fixed as those of a board game. In reality, the game might be rigged in ways the model does not capture.
"What is rational under known conditions is irrational otherwise, and the other way around." Nassim Taleb
Consider a coin that has flipped heads ten times in a row. A statistics textbook says the probability of the next flip being heads is still 50%, since each flip is independent. But in real life, there are two possible explanations: either you witnessed an extraordinary random event, or the game is rigged. The correct probability of the next flip being heads is "50% multiplied by (1 + the probability that the game is rigged)" strictly greater than 50%.
The textbook answer is rational only within the bounded world of known rules. In the unbounded world of real life, there are variables outside the model. The Latin root "ludos" game captures the fallacy precisely: in a board game, all rules and payoffs are visible. In reality, they never are. Nassim Taleb introduced this concept to warn against treating real-world decisions as if they were casino games with known odds.
This fallacy explains why behavioral economists have sometimes been wrong in labeling people "irrationally risk-averse." When researchers offer subjects a coin-flip bet with positive expected value, most decline. The economists call this irrational. But the subjects are responding to an expanded payoff space: they know that a $950 loss, even with a positive expected value on paper, can cascade into real-life consequences damaged relationships, depleted savings, lost opportunities that the model ignores.
The practical lesson is to distrust any analysis that assumes all variables are known. When someone presents a clean expected-value calculation, ask: what is outside the model? What would happen if the game is rigged, or if there are payoffs the model does not capture?
Takeaway: Whenever a decision looks obvious on paper, ask what the model is leaving out real life always has variables the spreadsheet cannot see.
See also: Ergodicity Changes Everything | Avoid Ruin Above All | Goodhart's Law Corrupts Every Metric | The Narrative Fallacy Turns Correlation Into Causation