RISKQuickStart[1]

Step 1: Review the model

Review your model to identify which values you are interested in analyzing and which values

are uncertain.

The values you are interested in analyzing – your “bottom lines” – will become the Output cells

in @RISK. The values that are uncertain will become Input cells in @RISK.

In this model, you are evaluating an investment project. There is an initial investment, followed

by future years of revenues, associated variable costs, and fixed costs. You need to project cash

flows for the next ten years to calculate the key measure of the project's performance: the net

present value, or NPV. The NPV is based on cash flows, the initial investment, and the discount

rate. The model also includes the possibility of a bonus if the NPV is greater than $30,000. The

bonus is the amount by which NPV is greater than the bonus limit. You can look at the Excel

formulas in rows 22-30 of the Model sheet. They are typical Excel formulas and contain nothing

new.

However, there is considerable uncertainty about the costs and future revenues. This means

that you can't be sure about what the NPV and the bonus will really be. This is where @RISK

comes in. As you will see, @RISK can help you assess the probability of negative NPV, positive

NPV, bonus or not, and more. You will also be able to uncover which of the uncertain inputs

contributes most to NPV, information that might help you choose a more profitable strategy.

Now it’s your turn

Take a few moments to look at the formulas in rows 22-30 to make sure you understand the

logic of the model. It is all correct, so you don’t need to make any changes at this point.

Step 2: Add @RISK Outputs

As mentioned in step 1, outputs are the cells you are interested in analyzing. These are your

“bottom line” cells. In this model, the outputs are the NPV and bonus cells.

@RISK won’t tell you exact or certain values for these outputs. This is impossible because the

future cannot be predicted with certainty. However, as you will see in the Results step, @RISK

can report the probability of different values occurring for each output, and that information

can help you make more informed decisions.

@RISK can’t just guess which cells are output cells. You have to designate them as such, but this

is easy.

To define the Output cells in @RISK:

Select the cell with the value you want (such as NPV).

Click the Add Output button.

Type in the name of the output. This name is used in reports.

Note that the formulas in these two output cells have changed. They now include the

RiskOutput function and a plus sign, which is @RISK’s way of indicating that this is an @RISK

output cell. For example, here is the modified NPV formula:

=RiskOutput("NPV")+C26+NPV(C10,D26:I26)

Actually, the RiskOutput function returns 0, so this addition to the formula doesn’t change the

cell’s value.

Now it’s your turn.

Designate the NPV and bonus cells as @RISK output cells.

Step 3: Define input distributions

As mentioned in step 1, there are a number of uncertain values in this model involving the costs

of the project and potential revenues it will produce. In a traditional analysis, you would simply

enter a best guess for uncertain values, or perhaps choose two or three "typical" or "extreme"

values to see what happens.

In @RISK, you don’t have to guess which values to try. @RISK allows you to enter probability

distributions in uncertain cells. These probability distributions indicate the possible values and

how likely they are. The corresponding cells are called @RISK input cells. In this model, the five

values in the green cells are uncertain: the investment cost, the year 1 revenue, the annual

fixed cost, the annual revenue growth rate, and the annual variable cost as a percentage of

revenue.

There are many different probability distributions you can use for @RISK inputs. Here you will

use two common distributions: triangular and normal.

All distributions require you to supply parameters. Parameters are values that describe the

probability distribution, such as its central location, its variability, and its shape.

In a triangular distribution, the parameters are the minimum value, the most likely value, and

the maximum value. The shape of this distribution is literally a triangle, with its peak at the

most likely value. No values below the minimum or above the maximum are possible.

In a normal distribution, the parameters are the mean and the standard deviation. This is the

traditional bell curve. The mean is the average or most likely value. The standard deviation is a

measure of variability around the mean. Normal distributions are symmetric, meaning that

values above the mean are just as likely as values below the mean.

The parameters for this example have been listed in columns E-H for your convenience: the

minimum, most likely, and maximum values for the three triangular distributions, and the mean

and standard deviation for the two normal distributions. Actually, it is not necessary to list

these parameters on the worksheet, but it is often useful to do this for documentation.

A natural question to ask is where these parameters come from. For example, even if you

believe that a normal distribution is appropriate for the annual revenue growth rate, where do

you get the mean and standard deviation for this distribution? The answer might be based on

historical data, it might be based on the opinions of experts, it might be based on your own

subjective feelings about the future, or it might be a combination of all of these. This is always a

difficult decision, but it is an important one, so you should try to choose parameters that are

most in line with your knowledge about the particular situation. That is, you should always try

to choose parameters that are realistic.

To define an @RISK Input:

Select the cell that is uncertain, such as the investment cost cell.

Click the Define Distributions button.

Select the appropriate distribution from the thumbnail gallery, such as Triang for the

investment cost, and click Select Distribution.

Define the parameters. @RISK makes some guesses, but you will usually override these.

You can use cell references to choose the parameters, or you can enter the values directly

in the Define Distribution window, as will be done in this example. Here is the triangular

distribution for the investment cost.

Click OK to accept this distribution with these parameters.

You can now enter distributions for the other inputs in the same way.

Note that there are now formulas in the green input cells. For example, here is the RiskTriang

function is in the investment cost cell:

=RiskTriang(40000,50000,90000,RiskStatic(50000))

Its first three arguments are the parameters of this triangular distribution. There is actually a

fourth RiskStatic argument, which requires some explanation.

You will note that there is a dice button on the @RISK ribbon. It toggles between static values

when white and random values when orange. It is currently white, so you see $50,000 in the

investment cost cell, the RiskStatic parameter in the formula. If you toggle the dice button to

orange, you will see random values in the input cells. In fact, if you press the F9 recalc key

several times when the dice button is toggled to orange, you will see many different random

values in the input cells.

This latter behavior is the essence of simulation. Instead of getting a single value in an input

cell, you get a range of values determined by the probability distribution you use. For example,

based on the triangular distribution for the investment cost, the most likely value is indeed

$50,000, but there is some probability that the investment cost will be greater than $80,000,

and there is some probability that it will be less than $45,000. In fact, every value from $40,000

to $90,000 has some chance of occurring.

You can decide whether you like the dice button to be toggled to static or random. However, it

has no effect on the eventual simulation.

Now it’s your turn.

Enter distributions for the uncertain inputs, using the parameters suggested. Then check your

formulas in the green input cells.

Step 4: Change the number of iterations

Now that you have designated @RISK output cells and have entered probability distributions

for input cells, you are almost ready to run the simulation. However, before doing so, you

should change at least one simulation setting: the number of iterations. The number of

iterations indicates how many random scenarios you want @RISK to generate. The more

iterations you use, the more accurate your results will be. The only downside is that more

iterations require more computing time, which can be an issue with complex models. For most

models, it suffices to use 1000 to 5000 iterations, but you can experiment.

To change the number of iterations:

Enter a number in the Iterations box in the @RISK ribbon or select a value from the

corresponding dropdown list. For this example, choose 1000 iterations.

Note that there are a number of other simulation settings you can change by clicking the small

Simulation Settings button circled above. No other changes are necessary for this basic model—

the defaults work fine—but be aware that you can make changes.

Now it’s your turn.

Change the number of iterations to 1000.

Step 5: Run the simulation

Running the simulation is the easiest step of all. You just click the Start Simulation button and

watch the progress. However, before doing so, you should be aware of what you will see as the

simulation runs.

First, you will see a progress indicator. Normally, the simulation will run very quickly, but if you

see that the progress is too slow, you can stop the simulation from the progress window and

perhaps reduce the number of iterations.

Second, you will see a chart of the currently selected output cell being built as the simulation

proceeds.

Now it’s your turn.

Run the simulation.

Step 6: Analyze the results

This is the step where you really appreciate the power of @RISK. When the simulation runs,

@RISK keeps track of all 1000 values in the input cells and the output cells, and then it lets you

see these in a variety of ways.

The quickest way to see the distribution of an output is to select its cell and click the Browse

Results button. You see a chart of this output, by default a histogram. This lets you analyze the

output in a number of ways.

You can move the two sliders on the chart to see probabilities or percentiles of this

distribution. For example, to get the probability of a negative NPV, you can enter 0

above the left slider. As you can see, this probability is close to 21%, so the company

ought to think twice before getting into this investment. Alternatively, to get the 90th

percentile, you can enter 10 above the right slider. The result is close to $236,800. There

is only a 10% chance of having an NPV greater than this. A company might look at such

probabilities and percentiles to make a “go-no go” decision.

If you would rather see another type of chart, such as a cumulative ascending chart, you

can click the 4th button from the left to get it. Again you can move the sliders. For

example, the median NPV, that is, the 50th percentile, is slightly greater than $69,900.

If you want to see which of the five inputs has the greatest effect on NPV, you can select

one of the tornado chart options. For example, here is the chart for the change in

output mean option. Each bar indicates how much the mean NPV changes as a

particular input varies over its range. Clearly, the annual revenue growth has by far the

greatest effect. As it varies over its range (and the other inputs remain at their static

values), the mean NPV varies from about negative $35,300 to about positive $260,700.

If you want to see results for another output, you don’t even need to close this chart.

You can simply select another output cell, in this case, the bonus cell. Here you can see

that there is a large probability that the bonus will be 0, but positive bonuses are

possible and some are quite large.

@RISK provides a number of other options for viewing these same basic results:

You can click the Summary button to see quick results for all inputs and outputs. In fact,

you can even drag any of these thumbnail graphs off to its own window, complete with

the interactive features described earlier.

You can get more permanent results, those that can be saved in a workbook, by clicking

the Excel Reports button and choosing one of the report types. Quick Reports works

well, as shown below. (The Output 2 sheet is for the bonus output.)

You can also get detailed statistics on your simulation, all of your simulation data, and much

more analysis in @RISK.

By the way, when you run the simulation, your results will be somewhat different from those

reported here. This is because your random numbers will differ from those generated here.

Now it’s your turn.

Analyze the results, using any of the methods described here.

More @RISK Examples

This completes the application. However, we urge you to learn more. One way is to examine

the wide variety of example models packaged with @RISK. To see them, choose Example

Spreadsheets from the Help dropdown list. This opens a file with many models, grouped by industry and

type of application. To see any of them, just click the corresponding link.

In addition, there is extensive documentation on all @RISK features under the Help dropdown list, there

are many more resources on the Palisade web site www.palisade.com, and Palisade offers a range of

training, consulting, and customization services. We hope these will help you to become one of the

increasing number of expert @RISK simulation modelers in today’s data-driven decision-making world.