Just the phrase “Monte Carlo” invites images of glamorous women and debonair men gathered around a baccarat table—ah, the glories of James Bond. While most of us will never see the inside of Casino de Monte-Carlo, it's a safe bet that we will run across a Monte Carlo simulation.

A Monte Carlo simulation is a statistical analysis tool that works well in situations with multiple variables that fall in a known range with equal (or predictable or subject to portion estimation) probability. For example, Ajax Corp. litigates over lost profits on its ever-popular widgets. The widgets require three primary components: Input-A, which costs $2 per unit, but can be purchased at a 10 percent discount in volumes over 1 million units; Input-B, which has a base cost of $5 per unit, but can fluctuate based on the price of oil; and Input-C, which costs $3 per liter with volume discounts, but has special storage requirements increasing warehousing costs for larger orders. Taking advantage of volume discounts would reduce the unit costs of the inputs, but increase warehousing costs and tie-up working capital, a limited resource. Ajax's business plan indicated initial widget sales between 50,000 and 150,000 units per month with growth of 7 to 10 percent per month for the first two years leveling out to 3 to 5 percent, annually, thereafter. As even this relatively simple example illustrates, commercial transactions involve many interrelated variables.

One option is to assume certain values for these variables and present the fact-finder with a simple, point-specific result. For all of its simplicity, the cross-examination questions are self-evident: “You assumed initial monthly sales of X units, but if initial monthly sales were only Y units, your calculations would be wrong, wouldn't they?”