Why Dean is absolutely right about security - it's no better than before

[Snow]

9/11. Not real sure where to put this, so when in doubt try
the Lounge. This is not an exciting topic in that it includes
math, but it shows a basic flaw in our security screening that
I've seen the press address only once. Here goes. Let's say
you want to screen out bad guys. The bad guys occur at a rate
of 100 per million. You have a great screen that catches 99%
of the true positives (ie the bad guys) while correctly
labeling 99% of the true negatives (the ordinary travelers).
Here's how that looks tabulated (bear with me; we
epidemiologists love to tabulate stuff).

                            Truth

               Bad Guy       Ordinary Traveler   Total
  ----------+-----------+----------------------+------------
  Screen  |             |                      |
 ---------+             |                      |  
  Bad Guy |  99         |       9,999          | 10,098
          |             |                      |   
 ---------+-------------+----------------- ----+------------
          |             |                      |   
Ordinary  |    1        |     989,901          | 989,902
Traveler  |             |                      | 
 ---------+-------------+----------------------+------------
Total     |   100       |     999,900          | 1,000,000
          |             |                      | 
          |             |                      | 

Okay, see the problem? Your screen, which is really
fantastically accurate, has caught all but one of the bad
guys. But it's also labelled 9,999 ordinary travelers as bad
guys - those are your false positives - but to you, based on
your screen, they look just like the bad guys. You have no way
of telling them apart. You can play with the numbers, base it
on passengers travelling by plane per day, change my
assumptions about the proportion who are terrorists, maybe
even see what happens with a not-so-good screen (the false
positive rate gets worse). But this is why we truly are not
any safer than before and why the idea of screening a large
population for a rare condition is pointless in this case. The
foundation of what we're trying to do with airport security is
flawed.

edits to make the table work....