Monte Carlo Analysis
Inputs
Identify variables
Have an effect on output
Have known probability distributions
Standard Deviation
Number of variables
Number should be restricted (to relevant variables only)
Large number of variables can distort the outcome
Relationship between variables
Independence
Independence is not transitive among variables
Uncorrelateness does not always imply independence
Correlation
Must be stable
Assumed as intrinsic characteristic of the distribution
Reasons to choose
We know something about the distribution
We want to determine required operational margins
There are differences of opinion concerning that input
We want to price the risk associated with it
It is "systematic"
It is of concern to some actors, but not others
It is under our control
Excel Techniques
Put all inputs in the same worksheet as the outputs
How to simulate ANY probability distribution
1. Draw from a [0,1] uniform
2. This picks a point in the cumulative distribution function (cdf)
3. Use the inverse of the cdf to find what variable value corresponds with that point
Voilà
Outputs
Understand shape of uncertainty
Graphics
Summaries