Parrondo’s Paradox

Two losing games. One winning combination.

Parrondo’s paradox (Juan Parrondo, 1996) shows that two individually losing strategies can combine to produce a net winning strategy — a result that defies intuition but is rigorously mathematically correct.

Game A is a slightly biased coin toss: you win 1 unit with probability (0.5 − ε) and lose 1 unit otherwise. Played alone, capital drifts down.

Game B uses two coins. If your capital is divisible by 3, you use a “bad” coin (win probability 0.1 − ε). Otherwise you use a “good” coin (win probability 0.75 − ε). The bad coin triggers just often enough to make Game B also a losing proposition on its own.

When you alternate A and B (or mix them randomly), the bad coin of Game B is triggered less often — the games interfere constructively and capital rises. The same principle appears in flashing Brownian ratchets, portfolio diversification, and biological evolution.