Monte Carlo simulation is a technique that uses the probability distributions specified for the model inputs to produce a probability distribution of a single output of interest. Specialised software is required, although a number of applications are available, such as @Risk (above) and Crystal Ball that operate as add-ins to spreadsheet packages such as EXCEL or Lotus 1-2-3.
Monte Carlo simulation should be used where there are multivariable relationships between key risks and should only be considered in highly complex projects. Project teams that may be required to conduct the analysis should be thoroughly trained in the use of the model, which is susceptible to "garbage in, garbage out" results if inappropriately applied.