QM323 Final Study Guide

Three Dimensions of Risk

Probability of the Risk Happening Impact of the risk Duration of the risk

=RISKCORRMAT()

RiskCorrmat(Matrix,column,instance) allows a distribution to be associated with a column of a correlation matrix so that rank order correlated values can be generated from several distributions. The optional instance parameter allows the same correlation matrix to be used several times in the one model.

How do these analysis work together?

Sensitivity Analysis generally leads us to do more thorough and important Simulation Analysis.

=RISKMEAN(Cell you want the mean of, Simulation #)

This can give you the value of the mean for a certain output from the specific simulation you desire.

Sensitivity Analysis

quantifies how changes in model parameters or variables affect outcomes. The 1% changes in COGS and how that affects profits.

Value at Risk

the 5th percentile is usually reffered to as the Value at Risk becuse it indicates the worst possible outcome

Simulation Analysis

used to figure out the likely range of variation in the objective function given assumptions about the probability of it happening. How will our NPV Change or likely be as a result of us taking into account the chance a tornado hits our facility.

Deterministic Checks

using fixed values as inputs within your risk model to see if the logic of the model is correct.

Discrete Distribution

A discrete distribution describes the probability of occurrence of each value of a discrete random variable - This works for when no approximation is continious - Probabilities must equal 1

Goal Seek (in terms of what analysis it shows us)

Does Sensitivity Analysis - Shows the percent change as we asl the model to find a new scenario

Scenario Analysis

Find Scenarios that Could happen to the business Determine the impact of them on the business's objective Make a qualitative estimate of the size of the impact Find the probability of that happening Identify Mitigation Strategies (Mostly for Qualitative purposes)

=RAND()

Generates a number that is equally likely to be anywhere between 0-1

=RANDBETWEEN(X,Y)

Generates a random integer that is equally likely between X and Y - This is also a uniform distribution

Benefits of Scenario Analysis

Gets rid of overconfidence because it forces people to actually find the realism in their business. helps people deal with overconfidence. Makes people think about the disribution in which they have put their success or their failure and see the probabilities scaled to more realistic ones.

Priority of Addressing Risk Scenarios

HP - HR HP - LR LP - HR LP - LR

Sensitivity Analysis

How are conclusions affected when the parameters of the model change!? Helps us find the most critical values to consider when planning different outcomes

Flaw of Averages

if a model contains uncertain inputs it can be very misleading to build a deterministic model by using the means of the inputs to predict an output. - The resulting output value will be very different than what's expected from running a simulation with uncertainty incorporated explicitly.

Correlated Input Variables

if they are positively correlated then large numbers will tend to to with larger numbers and smaller correlations with smaller numbers If they are negatively correlated than large numbers tend to go with small numbers and small with large.

Simulation Model

is a computer model that imitates a real-life situation - Shows a distribution of results not a single answer - Takes into account uncertainty

Availibility Heuristic

is a mental shortcut that relies on immediate examples that come to a given person's mind when evaluating a specific topic, concept, method or decision. This limits people to jump to the conclusions most easily thought of in risk simulation.

Triangular Distribution

is broken up by the Minimum possible value, Most Likely Value, and Maximum Possible Value - Shaped like a triangle - Good for continuous input variables

Normal Distribution

is the familiar Bell Curve Shape - Used for continuous input distribution - Allows negative values which may not always make sense in models

Uniform Distribution

is the flat distribution - has a minimum and a maximum and the values in the middle are all equally likely

Scenario Risk Matrix

on the Y Axis it goes from Low impact to higher impact on the X Axis it goes from low impact to higher impact

Triangular Models vs. Normalized Models and How they affect the Results

Triangular models are skewed because they are primarily based on the relationships between the most likely value and the min and the max whereas the Normalized model is a bell curve and thus equally distributed.

Variance (what is it?, Excel, and logical explanation)

Variance is the standard deviation

Probability Distributions for Input Variables

While sometimes called random variables, they are not random rather unknown and given values with probabilities (Probability Distributions) to have as inputs

=RISKSIMTABLE(row of numbers you want used)

enables the use of the same random numbers for all of the simulations you use in the @risk program. - This allows for having multiple simulations with more than just one set of inputs for a variable - Must remember to change the number of simulations to equal the amount of values included.