George's T-Shirts Addendum

Abstract

This addendum to George's T-shirts (UVA-QA-0346) offers a revised forecast in the form of a cumulative distribution function for the two key uncertainties in the case. It can be used as the basis for introductory risk analysis tools such as Crystal Ball or @Risk.

Excerpt

UVA-QA-0548

George's T-Shirts Addendum

Before placing a 10,000 shirt order for the upcoming rock concert, Lassiter decided to take one last look at his forecasts for grandstand ticket sales and for the percentage of attendees who would buy one of his shirts. A call to the concert promoter revealed that ticket sales were brisk, in fact better than anticipated—10,000 grandstand tickets had already been sold. The very long term weather forecast was still indicating good weather. Informal reactions among his office colleagues to the t-shirt design had been positive.

With this new and rather encouraging information in mind, Lassiter returned to his forecasts. Thinking in terms of only three distinct possibilities for these uncertainties was a nice way to start, but with the decision between 10,000 shirts and 7,500 shirts being rather close, he was troubled by its simplicity. In reality, the number of attendees and the percentage buying his shirts could fall anywhere in a wide range, not just at the values he had assumed. As a result, he decided to change the format of his forecast—from a discrete probability distribution to a continuous probability distribution. His new assessments were the following:

Grandstand Percent of attendees

tickets buying his shirts

. . .

Keywords: risk analysis