Tip of the Week #75                    Tip Index

Go to the Prior Tip The Greatest Inventions of the Past 2,000 Years
Go to the Next Tip "Making Decisions with Precision" (optimization software)
Return to MaxValue Home Page

Not "Estimating" Chance of Success

Recently, while writing an article and later when revising Paul Newendorp's book, it struck me as illogical to talk about "estimating chance of success." I've used the phrase for many years, and now I don't like it.

Consider, for example, an exploration testwell. The well's possible outcomes are "Success" or "Failure." We commonly use the computer science notation of 1=Success (or True) and 0=Failure (or False). Thus, the outcome of nature is either 0 or 1. (Nature knows already; we don't yet know.)

"Estimating" is something one does in appraising or forecasting a value. Assigning a probability to a binary chance event, which will happen or not, doesn't seem like an estimation. Rather, the probability is representing the expert's degree of belief about whether the event will happen. As a probability, this forecast—if one can call it that—will always be wrong!

I suggest, instead, that you "assess" or "judge" a chance of success (or probability of success).

If your job involves assessing chances of success, I recommend that you keep a journal. Compare your judgments with what actually happens, whenever possible. Analyze the data from time to time. This feedback is perhaps the best single thing you can do to improve your risk assessments.

—John Schuyler, October 2000.

Copyright 2000 by John R. Schuyler. All rights reserved. Permission to copy with reproduction of this notice.