C[omp]ute

Welcome to my blog, which was once a mailing list of the same name and is still generated by mail. Please reply via the "comment" links.

Always interested in offers/projects/new ideas. Eclectic experience in fields like: numerical computing; Python web; Java enterprise; functional languages; GPGPU; SQL databases; etc. Based in Santiago, Chile; telecommute worldwide. CV; email.

© 2006-2013 Andrew Cooke (site) / post authors (content).

Bayesian Analysis of NSA Surveillance

From: "andrew cooke" <andrew@...>

Date: Sat, 15 Jul 2006 10:54:55 -0400 (CLT)

This is from the latest Cryptogram -
http://www.lewrockwell.com/orig7/rudmin1.html

Is it right?  The basic maths seems OK.  But what worries me is the
statement "What is the probability that people are terrorists given that
NSA's mass surveillance identifies them as terrorists? [...] If the
probability is fifty-fifty (p=0.50), that is the same as guessing the flip
of a coin."

When I read that, I understood: "If p=0.5 then we may as well flip a coin
to decide whether people are terrorists".

That reading is incorrect.  The article assumes that the chance of picking
out a terrorist by choosing someone at random is 0.000003.  Effectively
impossible.  If the NSA can shift that to 0.5 then the odds have been
decreased hugely.  Knowing that someone is 50% likely to be a terrorist is
fantastically useful.  A candidate with that level of confidence is worth
spending many thousands of dollars on to improve things further and make a
final decision.

Now the numbers don't come out to 0.5 anyway.  But they are closer to
being useful than this rather misleading argument suggests.

Andrew