Transcript Overview of Monte-Carlo Simulation

Desktop Business Analytics -Decision Intelligence
Time Series Forecasting
 Risk Analysis
 Optimization

Current Products

Crystal Ball®


Crystal Ball Pro


Integrated Optimization and Monte Carlo
simulation
CB Predictor


Excel-based Monte Carlo simulation
Integrated Time-Series Forecasting with Monte
Carlo
CB Turbo

Distributed Processing capability to speed up
simulations
Monte Carlo Applications
Capital Budgeting
 New Venture Planning
 Manufacturing Planning
 Marketing Planning
 Quality Design
 Environmental Risk
 Petroleum Exploration

Spreadsheets - Pros
Easy to use
 Popular
 Flexible model-building tool

What-if Analysis

Methodically entering even increments
of values to view the projected
outcomes
Pros: Reveals incremental range of possible
outcomes
Cons: Time-consuming, Results in a mountain
of data, Reveals what is possible, not what is
probable
What is missing?

The ability to know the range of possible
outcomes and their likelihood of occurrence

As a result, we use Monte Carlo Simulation
as a system that uses random numbers to
measure the effects of uncertainty on our
decision-making process
What is Simulation?
Modeling a real system to learn
about its behavior
 The model is a set of mathematical
and logical relationships
 You can vary conditions to test
different scenarios

Advantages of Simulation
Inexpensive to evaluate decisions
before implementation
 Reveals critical components of the
system
 Excellent tool for selling the need for
change

Disadvantages of Simulation

Results are sensitive to the accuracy
of input data



Garbage in, Garbage out
Intelligent agents using secret rules
Investment in time and resources
The Five Steps of
Model Development
1. Develop a system flow diagram
2. Write an Excel spreadsheet to model the system
3. Use Crystal Ball to model uncertainty
4. Run the simulation and analyze the output
5. Improve the model and/or make decisions
Crystal Ball Demonstration
2+2 = 4 ?
Crystal Ball Pro


Decision Intelligence
Includes




Crystal Ball
Optimization
Extenders
Developer Kit
Optimization Model

Decision Variables
 Quantities
over which you have control
(Accept or reject each project)
– Upper and lower bounds
– Continuous or discrete
Optimization
X
Function
Find the possible input values that make
the output as large or as small as possible
F(X) = Y
Project Selection
Project Mix
Model
Find the project mix that generates
the highest combined NPV
Combined
NPV
A Realistic Model


Uncertainty analysis
Constraints and Requirements

We will us the simplifying assumption of applying a
budgetary constraint to limit investment
The ‘Flaw’ of Averages
“Never try to walk across a river just
because it has an average depth of four
feet.”

Milton Friedman
Academic v. Real World

Professors and students have used
many techniques
 Inaccessible
 Difficult
to implement
 Clients do not understand the results

Decisioneering makes Monte Carlo easy
to use in everyday spreadsheet
modeling.
How are you handling
uncertainty?
Do you use low, middle and high
values?
 Do you do What-if analysis?

Multiple What-if
scenarios confuse as
much as enlighten...
A Picture is Worth...

A thousand What-ifs
Decisioneering, Inc.
Provider of Analytic Tools since 1986
 Headquartered in Denver, Colorado,
USA
 More than 70,000 Users
 85% of Fortune 500 Companies
 45 of Top 50 Business Schools
 65% CAGR over 3 Years

Monte Carlo

Random number generation simulates
the uncertainty in the assumptions. The
program selects a value for the
assumption, recalculates the
spreadsheet, plots the forecast and
repeats.
Deterministic v. Stochastic
Deterministic
Fixed
Data
Fixed Outcomes
$1,200,00
7%
Stochastic
Variable
data
Variable
Outcomes
M onthly S a v ings
Forecast: Scenario A Retirement Portfolio
500 Trials
Frequency Chart
6 Outliers
.09 4
47
.07 1
3 5 .2 5
.04 7
2 3.5
.02 4
1 1 .7 5
M e an = $6 46,19 8
.00 0
3 5 0 .0 0
4 2 5 .0 0
5 0 0 .0 0
5 7 5 .0 0
6 5 0 .0 0
0
$3 00,00 0
$5 25,00 0
$7 50,00 0
D o l l a rs
$9 75,00 0
$ 1 ,2 0 0 , 0 0 0
Statistics

Normal Distribution, Mean and Standard
Deviation
M o n th ly S a v in g s
Mean
Standard Deviation
3 5 0 .0 0
4 2 5 .0 0
5 0 0 .0 0
5 7 5 .0 0
6 5 0 .0 0
Retirement Example
Monthly Dollar Saving
Number of Years
Annual Growth Rate
$
Value at Retirement
$
Uncertainty
500
20
12%
432,315
Define Assumptions
Retirement Example Assumptions
Retirement Example
Assumptions
Retirement ExampleForecasts
Retirement Example
Forecasts
Communicating Results
Get the client to understand alternatives
 Take action

Uncertainty over time
Compare Alternatives