Yesterday, fellow Forbes contributor Kashmir Hill pointed out that there are giant problems with the NSAs Utah data center. This is a bad thing, right? Color me unconvinced. Ill refer you to Glenn Greenwald for in-depth discussion of the legal issues surrounding the NSAs encroachments on freedom and turn your attention to another, perhaps more practical issue: under what conditions will NSA screening of communications actually work?
For this, we turn to statistics and employ Bayes Theorem. In August, we applied Bayes Theorem to baseball. Now, lets apply it to something more serious: catching terrorists.
Suppose the NSA runs your email through a program that is 99% accurate, meaning that it only makes a mistake one out of every 100 times it runs, and flags you as a potential terrorist. How likely is it that youre a terrorist?