Why flawless game theory makes people walk away empty-handed
Imagine a pot with two dollars. You and a partner take turns either grabbing the pot or passing to let it double. If both keep passing, you both walk away wealthy. But standard game theory dictates you should defect on the very first turn. Using backward induction, rational players deduce their opponent will take the money on the final round, triggering a chain of distrust. In experiments, real humans ignore this mathematical logic, cooperate, and leave vastly richer.
The Anatomy of a Missed Fortune
In 1981, economist Robert Rosenthal introduced a thought experiment that exposed a deep rift between theoretical decision-making and human behavior. Known as the Centipede game due to its multi-legged diagrammatic branch structure, the scenario places two participants in a sequential contest over an expanding pile of resources. Two players take turns deciding whether to take the larger share of a growing pot or pass the turn to the other player. Each time a player passes, the total amount of money increases, but the relative advantage shifts, handing the initiative to the rival.
The mechanics are intentionally simple. At round one, a pot containing a modest sum—for instance, two coins—sits on the table. Player A can either take both coins and end the game immediately, or pass. If Player A passes, the pot swells, and Player B faces a similar choice: take the dominant share of the now-larger pot or pass it back. If both players continue to pass until the predetermined final round, both walk away with an exponentially larger sum than was initially available. The structure creates an obvious mutual incentive: continuing cooperation generates immense collective wealth, while terminating the game freezes the prize pool.
Yet, when classical game theory analyzes this setup, it does not recommend patience, cooperation, or mutual trust. Instead, standard formal analysis dictates that Player A should defect and seize the pot on the very first turn, walking away with mere pocket change while entirely foregoing the vast wealth waiting at the end of the line. This stark contradiction between mathematical optimization and common sense forms one of the most famous paradoxes in modern decision theory.