Monte Carlo Is No Crystal Ball

When you can’t solve a problem with a formula or equation, you might turn to a Monte Carlo approach in which a computer simulates outcomes for a given situation using a large number of randomly selected data points. This type of forecasting, says a recent article in The Wall Street Journal, tipped off meteorologists that Hurricane Sandy might very well hit landfall rather than stay out to sea as first expected.

The concept was concocted when mathematician Stanislaw Ulam set out to win at Solitaire. He couldn’t figure out a mathematical model to conquer the game, so he decided to use a computer to simulate repeated games until he arrived at a winning hand. Ulam was a key figure in the Manhattan Project, and the approach he developed—named after the famous casino in Monaco— actually led to the development of the atomic bomb.

There are myriad applications for the model besides predicting storms and building bombs, however. The Environmental Protection Agency uses it to estimate adverse effects of health hazards, and pharmaceutical companies turn to the approach for guidance in researching and developing new drugs. The quality of the results generated, however, are only as good as the underlying data used in the model. John Guttag, a professor at MIT who has lectured on Monte Carlo, says “There is no magic to beat garbage in, garbage out.”

The possibilities in the areas of finance and investing are obvious, although it’s important to note that the information generated doesn’t necessarily equate to accurate predictions. “As with any forecasting tool,” the article notes, “the results of a Monte Carlo simulation can’t guarantee the future. But viewing myriad potential outcomes through the lens of probability statistics can help lift the cloud of uncertainty.”