Reason.tv: Sam L. Savage on "The Flaw of Averages"—Why we underestimate risk in the face of certainty

September 11, 2009

Have you heard the one about the river-crossing statistician who drowns after determining that the water is, on average, only three feet deep? This, says author Sam L. Savage, is just one example of the flaw of averages.

But don't despair, Savage writes in The Flaw of Averages: Why We Underestimate Risk in the Face of Uncertainty. There are sensible ways to make judgments involving uncertainty and risk. Reason senior editor Michael C. Moynihan met up with Savage to discuss the "seven deadly sins" of averages and how a greater understanding of these flaws could prevent future financial meltdowns.

Approximately 4:30. Shot by Dan Hayes and edited by Meredith Bragg.

Scroll down for embed code, and iPod and audio versions.

Subscribe to Reason.tv's channel there today.

Comments

not the real jb|9.11.09 @ 11:13AM|#

Holy metaphor, Batman!

I specialize in quantitative analysis of social data and while these cautions are important, we mustn't let them paralyze us into abandoning all techniques that are based on averages or other measures of central tendency. In fact, every example he gave in the video exemplified a single fallacy: mistaking the model for reality.

You wouldn't go into a restaurant and eat the menu; likewise you should remember that all mathematical - and verbal - models are only abstractions of reality that provide a limited set of detail.

And his analogy of looking down the barrel of a loaded gun is ridiculous. Running a goodness-of-fit test is not risking your life and will estimate the degree to which your model fits the data. It's like cracking the breach on a shotgun to see if it's loaded.

|9.11.09 @ 12:00PM|#

Are you certain about that headline?

|9.11.09 @ 1:33PM|#

As far as financials go, you just have to remember that people are irrorational, and past perforance is not necessairly a predicator of future returns.

Also, I wonder how this risk analysis applies to climate change :P

|9.11.09 @ 7:00PM|#

RC'z law strikes again. Krone, are you a touch typist or are you a huntin' pecker?

GrilledCS|9.12.09 @ 1:43AM|#

The first thing I thought today when I saw this headline on Reason.com was that there would be a reference to Nassim Taleb's book "The Black Swan" (a book that is very interesting, when it isn't unnecessarily interrupted to serve Taleb's unapologetic egotism.)

Just a surface impression, but this video looks like it has been edited to hide frustration with its subject, Mr. Savage. Mr. Moynihan looks like he's on the brink of bursting out into frustrated laughter, and the cuts are oddly placed.

Maybe this is some kind of karma for Moynihan's article suddenly criticizing his supposed libertarian brethren, while somehow lacking the perspective that the government-control Left, and the religious Far Right have never held any such criteria to themselves, but entirely molded American policy.

http://www.reason.com/news/show/135901.html

supra shoes|9.12.09 @ 2:51AM|#

Your article is very interesting, I have introduced a lot of friends look at this article, the content of the articles there will be a lot of attractive people to appreciate, I have to thank you such an article.

diesel jeans|9.12.09 @ 2:56AM|#

Nice work guys!
this is just Amazing!
Thanks

jordan 6 rings|9.12.09 @ 2:57AM|#

Thank you very much. I am wonderring if I can share your article in the bookmarks of society,Then more friends can talk about this problem.

|9.13.09 @ 12:02PM|#

What he's talking about is a lay person's mistake--one that good statisticians have been trained out of. It isn't hard to make mathematical models that don't make these mistakes.

air jordan shoes|9.14.09 @ 2:17AM|#

the post is very good. thank you for sharing.

|9.14.09 @ 9:22AM|#

Don't be fooled by how simple this sounds. It is a real problem especially in financial modeling where all you ever look at is an average input.

Also, a statistician might be able to understand this, but I have yet to work in a place where this issue was actually taken seriously into account.

Robert Brown|9.16.09 @ 2:03PM|#

The Flaw of Averages is a very good book. Application of the flaw is an example I run into with business analysts all the time. And I can't count the number of financial and engineering mistakes I've witnessed as a result of the flawed way of thinking.