The most basic building block of uncertainty is the
concept of an uncertain number. Examples include next month's sales, this
afternoon's temperature, or today's closing price of your favorite stock. In
statistics courses the closely related term random variable
is used in the discussion of uncertain numbers. However, in my
experience, if you use this term in the workplace few people will understand
you, so I will stick with uncertain number.Suppose that you are considering the start up of a new business. Revenues in the first year are quite uncertain. Imagine, for example, that the uncertainty in revenue is the same as if God spins the spinner, and multiplies by $1Million. The downside is that if revenues are less than $200,000, you will go out of business and face personal bankruptcy.
Uncertainty and Risk go hand in hand, but they are not identical. Uncertainty is an objective feature of the universe while risk is in the eye of the beholder. In this case, for example, the uncertainty, determined by the physics of the spinner, is the same for all observers. The risk to you is that revenues will be less than $200,000, forcing you out of business. The risk to your competitors is that revenues will be greater than $200,000, forcing them to continue to contend with you.
Associated with every uncertain number is a shape known as a histogram, which displays the likelihood that the number falls within various ranges of values referred to as bins. Histograms are an important output of simulations.
PUZZLE 1:
In general it is difficult to estimate these shapes, but can you guess the
shape of the histogram for the spinner? Give it a try by opening the file histogram.xls.
Now adjust the bars to show the percentage of spins that you expect to lie in
each of the regions (0.0 - 0.2, 0.2 - 0.4 etc.)

Histogram.xls
When you have made your guess, go on to the following tutorial in which we will simulate the spinner.
To gain a better understanding of this situation, we will perform a Monte Carlo simulation of the spinner, that is, repeat this random situation many times, and analyze the results.
To simulate the revenue of your proposed business, proceed as follows.


The Simulation Settings dialog box will appear as shown below.




Then go to the Graphs tab. Specify 5 bins and 2 decimal places, then click the Histogram button.

You should get a graph as shown below. Note two
important features of this graph:

A. All the bars are about the same height because any number between 0 and 1 is equally likely. That is, the spinner is not more likely to point to some numbers than others. When presented with Puzzle 1, above, many graduates of statistics courses specify bars of different heights (see Reference [3]). In the words of Mark Twain, they have had their schooling interfere with their education. If you got this wrong, make sure you understand why the correct graph is flat across.
B. The bars add up to 100% because there is a 100% chance that the spinner will point to some number between 0 and 1. This assumes that there is no chance the arrow will fall off and land on the floor! The bars of all histograms must add up to 100%.

The mean, mode and median are often misunderstood concepts, which may all be grasped in terms of histograms. For symmetric histograms these three numbers are the same. But for asymmetric ones they will be different as shown in the figure below.
|
The Mean, Average or Expected Value The Median The Mode |
|
A related shape associated with every uncertain number is its Cumulative Distribution, as discussed below.

For example, the percentage of spins less than 0.2 is 20%, so that is the chance of financial ruin in the business example. The vertical line displays the Mean (average) of Revenue at about 0.5.
Note that the percentiles shown on the Statistics tab of Simstats.xls are the numerical equivalent of the cumulative graph.
Uncertain
numbers may be viewed as shapes that depict the likelihood of the possible
outcomes.
Monte
Carlo simulation, the process of bombarding a model with random inputs, can
help us estimate these shapes.