Beating the Odds in Monte Carlo

Beating the Odds in Monte Carlo

What does science have in common with games of chance? Is RTB now hooking up with gamblers to establish its creation model? You may be surprised to learn that many of the most difficult problems facing the scientist can be solved using the same approach a gambler takes in trying to improve his odds at the craps table or the roulette wheel. This technique is known as the Monte Carlo method and was invented in the late 1940s by mathematicians Stan Ulam and Nicholas Metropolis.1

By way of illustration: We’ve all experienced the frustration of waiting in lines to obtain some service, like cashing a check at the bank. On average the bank knows that it must service, say, one customer per minute, based on its long-term average of customers. However, at any particular time there may be more or fewer customers waiting in line because of the random nature of their arrival at the bank. If you wanted to know what the peak wait time would be, you could sit to the side for a long time and keep track of the wait times using a stopwatch.

Another approach is to set up a Monte Carlo simulation by programming a spreadsheet running on a personal computer. This simulation process makes use of a random-number generator (the computer equivalent of throwing a die to get a random integer between 1 and 6) by designating the random number to correspond to the time during the day when a given customer arrives (e.g., in the range 10 a.m. – 6 p.m.). An accurate peak wait time can be derived because of long-term statistical averages. For a more sophisticated simulation, one could use a second, independent random-number generator to estimate the time the teller takes to process a customer. Before the advent of digital computers, performing such a simulation would have been extremely difficult, but now it is fairly easy.

Scientists in all fields tackle a variety of problems with this method, many of them difficult or impossible to solve in any other way.2 One of the biggest breakthroughs using Monte Carlo simulations has come in astrophysics.3 Scientists have been able to simulate the conditions of the early universe (that is, just a few minutes after the big bang event). During this time, it is thought, the various light elements beyond hydrogen (deuterium, helium, and lithium) condensed out of the energy field as the universe expanded and cooled. The Monte Carlo simulations estimated these primordial element abundances, and the results were in remarkable agreement with observations, providing an early confirmation of big bang cosmology.

Even though the name derives from a resort casino, Monte Carlo simulations offer far better odds and results. As scientists employ this technique to understand the universe, the odds that RTB’s cosmic creation model is correct stand to improve.

  1. Nicholas Metropolis and S. Ulam, “The Monte Carlo Method,” Journal of the American Statistical Association, 44 (1949): 335-41.


  3. Gary Steigman, “Primordial Nucleosynthesis: Successes and Challenges,” International Journal of Modern Physics, E15 (2006): 1-36.