Their analysis depends upon the assumption that the terrorist needs to smuggle his own weapons through the check points in his own luggage.

And easier way to defeat this system is to have the terrorists go through the checkpoints WITHOUT their weapons. That way, no amount of searching them will show any reason not to allow them to fly.

The weapons will be smuggled aboard by the girlfriend (cute, white, with a mid-western accent) in her luggage (which the terrorist boyfriend had helped pack and hidden the weapons in).

The focus will be on the terrorists which will minimize the likelyhood that the girlfriend will be checked for weapons.

Since she doesn't know she's carrying weapons or anything, she will not show any fear or anything during routine questions.

Now, combine the above scenario with their Carnival Booth approach and you won't even need the girlfriend. The person with the low score will smuggle the weapons through security while his high-scoring associates take up the limited amount of security.

As I've said before, profiling is stupid.

Once someone knows that you're profiling, your system becomes less secure than random searches. As shown in their paper.

And by "less secure", I mean "totally insecure". Using their techniques for determining high/low scores and having a low score smuggle the weapons, there is only a minimal chance of the attack being found.

And if the attack is discovered PRIOR to take off, the high scores go free.