Things on my mind. George Drapeau's Weblog

This is an entry I've been meaning to write for a while, on the topic of how bad most of us are at estimating within a given range, even when that range is generous.  Let me explain:

In the book "The Psychology of Judgment and Decision Making" by Scott Plous, he provides a quick self-test that shows that we are often over-confident about our ability to estimate, even when we're given every opportunity to define the margin of error ourselves.  In software engineering, this comes up all the time: a developer is asked to estimate how long it will take to implement a particular feature.  You know what happens: the developer says "Oh, 2 weeks," and the feature ends up taking 6 weeks to finish.  Why did the developer say "2 weeks"?  Maybe it's because s/he was afraid to say a date that sounded too long, but often it's just a matter of being overconfident about our own estimates.

Don't believe me?  Try this test yourself.  Here are the rules:


For each of the following ten items, provide a low and high guess such that you are 90 percent sure the correct answer falls between the two.  Your challenge is to be neither too narrow (i.e., overconfident) nor too wide (i.e., underconfident).  If you successfully meet this challenge you should have 10 percent misses -- that is, exactly one miss.

(for each of the following items, list both a LOW and a HIGH number; you're 90% confident that the real answer lies between your LOW and HIGH number for that item)

1. Martin Luther King's age at death
2. Length of the Nile River
3. Number of countries that are members of OPEC
4. Number of books in the Old Testament
5. Diameter of the moon in miles
6. Weight of an empty Boeing 747 in pounds
7. Year in which Wolfgang Amadeus Mozart was born
8. Gestation period (in days) of an Asian elephant
9. Air distance from London to Tokyo
10. Deepest (known) point in the ocean (in feet)

Okay, have you done it?  You can find the answers here, along with a little wrap-up of what this all means.