Monte Carlo step by step
Each iteration simulates the whole project, from start to finish: one possible version of reality, in which chance decides which risks occur and how much each one costs. In real life the project runs only once; here we can repeat it thousands of times and see what happens overall. The first iterations run slowly so you can see where every number comes from.
The risk register
Project: ERP implementation. We simulate how much its risks could cost, in thousands of euros (k€).
| Risk | Probability | Minimum cost | Most likely | Maximum cost |
|---|
The probability tells us whether the risk occurs. The three cost values define a PERT distribution, which gives the cost of the risk when it occurs.
What happens in one iteration
“Next step” moves at your own pace. “Automatic” starts very slowly and speeds up on its own.
The histogram
Each square is one iteration: one simulated project. It goes into the column for its total cost.
With very few iterations, no conclusions can be drawn yet.
Iterations table
Which risks occurred in each iteration, how much they cost and the total. Most recent at the top.
How we decide whether a risk occurs: we draw a random number between 0 and 1. If it is lower than the risk’s probability (for example, 0.40 for 40%), the risk occurs. This is a Bernoulli variable.
How much it costs: we draw a value from a Beta-PERT distribution between the minimum and maximum cost, peaking at the most likely value.
P50 and P80: the cost that is not exceeded in 50% and in 80% of the iterations.
Licensed under CC BY 4.0: you may use and adapt it with attribution to ProjectWorkLab.