Critical Limitations of Monte Carlo Simulations

Richard RobinsonRisk Management

Monte Carlo simulations have become de rigueur for project risk assessments. There is no doubt the use of monte carlo simulations will provide sound insight into the most likely project outcomes.

However from a due diligence perspective there are major limitations when it comes to the long tail (low probability) distribution (high consequence) outcomes. Consider the situation where three long tail (say 1 in 1000) issues need to converge to cause total project failure.  Collectively, that’s a 1 in a billion chance.  To test for that, at least 1,000,000,000 trials would need to be completed.  To make that statistically significant there would need to be at 10 to 100 billion trials.

No one does that. Further, that assumes that the long tail issues have been accurately described. Usually what happens is that the long tails are just ignored as being hard to know and statistically insignificant anyway.  This means credible (although rare) critical possibilities are just ignored.

The R2A Project Due Diligence process addresses this shortfall.