"Matt Whiting" wrote in message
...
wrote:

His point was that if you profiled pilots instead of Arabs, you would
have had a greater proablility of a hit on a 9/11 hijacker.


And if you had profiled Muslims, you'd have had virtually a 1.0
probability.

No, you're addressing a different probability. CFII and I were talking about
the probability that a given passenger screened for trait X would turn out
to be one of the 9/11 hijackers. That probability is higher for X=pilot than
for X=Muslim (though miniscule in both cases).

You're addressing instead the probability that one of the 9/11 hijackers
exhibits trait X. That probability indeed equals 1.0 for X=Muslim, or
X=male, or X=non-Asian, and many more. Obviously, then, that probability is
not the right criterion for identifying an effective profiling trait.

--Gary