Timid testers, magic lassos

January 5, 2013

(This article was originally published at Numbers Rule Your World, and syndicated at StatsBlogs.)

That is the title of Chapter 4 of Numbers Rule Your World. When the book came out two years ago, lots of steroids scandals have broken out, and yet many people were in deep denial. Recall the saga of Floyd Landis, the U.S. cyclist, who initially claimed the drug test was wrong, and that he was witch-hunted. There was even a website called Trust But Verify where all his supporters gathered. As we now know, he was guility as sin. He was also the chief whistle-blower that has brought down Lance Armstrong.

If you read the book then, or one of my blog posts here, you'd hopefully have realized the scale of the scandal before all of this transpired. Applying a bit of statistical reasoning to the numbers around us makes us smarter.

The most recent news coming out of that scandal is just maybe Lance will confess (link). What's taken so long? And after all this time, his PR people are still floating a trial balloon? 

At the Guardian, they asked if lie detectors should be used in sports (link). That is the "magic lassos" reference in Chapter 4. Lie detectors and steroids tests both use statistical detection but have different statistical properties. My readers will again have a head start on this next debate. The key point is that a "true negative" in a steroids test is not necessarily a "true negative" in reality because there are a million ways to hide one's doping. The real goal of any anti-doping problem is to catch the dopers, and having zero positive test results is only part way there.

Please comment on the article here: Numbers Rule Your World

Tags: , , , , , ,