How to use Monte Carlo simulations for risk assessment in Six Sigma projects?

All answers to this question.

To set this up, you first need to define the probability distribution for each of your input variables (Normal, Triangular, etc.). In Minitab, you use the 'Workspace' or the 'Monte Carlo' tool to run thousands of iterations. The output is usually a histogram showing the spread of potential outcomes. For a non-technical audience, I always focus on the "Percentage of Defective Units" shown in the simulation results. Instead of talking about standard deviations, tell them: "Based on 10,000 scenarios, we have an 85% chance of meeting our delivery window." This makes the statistical risk feel real and actionable for managers who don't understand Z-scores.

   Answered 2024-06-17 by Patricia King

Are you finding it difficult to get accurate distribution data for your input variables? A Monte Carlo simulation is only as good as the data you feed it (garbage in, garbage out). If you don't have enough historical data to define the input distributions, are you using expert "best-case/worst-case" estimates instead, and how are you validating those assumptions?

   Answered 2024-06-19 by Robert Wilson

• Robert, that’s exactly where I'm stuck. For the "Expert Opinion" parts, I’m using the PERT distribution method which weights the "most likely" outcome more heavily. To validate, I’m running a sensitivity analysis to see which variables have the biggest impact on the output. If a variable with an "estimated" distribution is driving the whole model, I know I need to pause and gather more real-world data before presenting the final report.

     Commented 2024-06-20 by David Anderson

Always check the "Capability" part of the simulation report. It will show you the predicted $P_{pk}$, which is a great way to compare the simulated future state to your current baseline state

   Answered 2024-06-21 by Mary Harris

• Excellent advice. Comparing the $P_{pk}$ of the simulation against the current $C_{pk}$ provides a very clear "Before and After" picture for the project's potential success.

     Commented 2024-06-22 by Excellent advice. Comparing the $P_{pk}$ of the simulation against the current $C_{pk}$ provides a v