Monte Carlo Analysis
As a rule, the most influential model parameters found in the sensitivity analysis are considered for Monte Carlo modelling.
There is a big number of model simulations performed within a model framework during both simultaneous and random change of the set parameters.
Defining the law distribution is, when possible, based on historical data of the set value. (for this another method is available Distribution adjustment).
User specifies the number of method simulations for each chosen parameter, and variety range and distribution law of the related random value for each parameter.
If chosen parameters are mutually dependent then it is necessary to set the correlation coefficient between them.
Algorithm for Monte Carlo Analysis:
 Choose a set of changing parameters(risk factors)
 Set the variety range and distribution laws for each factor
 Specify correlation matrix for mutually dependent factors
 Launch calculation for a necessary number of experiments (usually not less than 1000)
Parameters setting forms
Monte Carlo Analysis helps to understand the uncertainty degree of the final value (for example by 90% confidence interval).
Also a user can focus only on set chosen values, instead of probability % when setting the interval.
Results table shows information in percentiles.
Percentile is a percentage measure or probability that the value of the prognosis is below or equates the value for the given percentile.
Calculation results in the Excel sheet
A lot of experts believe Monte Carlo Analysis to become a necessary tool for all analysts, and their reports not only shall include general reasoning about “potential risks” but also a separate section with case studies of Monte Carlo Analysis.
