The story
In 1946, the Polish-American mathematician Stanislaw Ulam, working at Los Alamos, had an idea while recovering from an illness and playing solitaire. He wondered whether the chance of winning could be found better by playing many games and counting than by working out the combinatorics. The same logic could be applied to problems in physics too hard to solve by formula: simulate random outcomes many times and average them. He discussed it with John von Neumann, who saw how it could run on the new electronic computer, ENIAC, and the method was soon used to model how neutrons travel and multiply in nuclear weapons. Nicholas Metropolis suggested the name Monte Carlo.
The method rests on the law of large numbers, a result of probability theory: the average of many random trials approaches the true expected value. Estimating pi by scattering random points over a square and counting those inside a quarter circle is the classic example. Accuracy improves with the square root of the number of trials, which made fast computers essential. A 1949 paper by Metropolis and Ulam introduced it publicly, and in 1953 Metropolis and colleagues published the sampling algorithm that became Markov chain Monte Carlo. The technique now prices derivatives, models particles and climate, and underlies Monte Carlo tree search, used in game-playing AI such as AlphaGo.
Why it mattered
- Simulation became a third way to do science alongside theory and experiment, widely used in physics and engineering.
- Quantitative finance adopted it for pricing options and measuring risk.
- Monte Carlo tree search became part of game-playing AI, including AlphaGo.
- Markov chain Monte Carlo made Bayesian statistics practical for complex models.
Sources
- Hitting the Jackpot: The Birth of the Monte Carlo Method Los Alamos National Laboratory (archived copy)
- A Short History of Markov Chain Monte Carlo: Subjective Recollections from Incomplete Data arXiv
Drafted by AI (Claude) from the sources above and checked against them by AI; no historian has reviewed this page. Spot a mistake? Suggest a correction. How pages are made
Community notes
Loading notes…