> Nothing you say changes this point - if you make it easy for 80 year olds to get through security, 80 year olds will be the ones with bombs.

The whole point of my comment is to refute this idea, so I guess I'm in trouble. Let me try to phrase things differently.

For every person in the world, there is a cost to recruit them into your organization as an attacker, which I postulate is low for young Arab men and high for Laotian grandmothers (and also high for 80-year-olds generally). Call this cost function COST(x).

There is also a likelihood that if they attempt an attack, they will succeed. Call this EFFECTIVENESS(x).

I can only interpret "if you make it easy for 80 year olds to get through security, 80 year olds will be the ones with bombs. Whether by recruiting, coercion, or anything" as saying that a terrorist group sending an attacker t will strictly try to maximize EFFECTIVENESS(t). But that's wrong. A terrorist group sending an attacker t will try to maximize EFFECTIVENESS(t)/COST(t).

Imagine some categories like so:

    Age Sex Ethnicity  EFFECTIVENESS      COST
    19  M   Syrian     10%              $20,000
    80  M   Egyptian   60%             $150,000
    80  F   Laotian    80%             $250,000

Obviously, it's easier for the 80-year-old Laotian female to get through security (or they're just generally more competent, or whatever). But it's cheaper (80% of the cost), and more effective (25% more expected successes), to send 10 19-year-old Syrian men. So, in a world where these numbers were accurate, we'd expect to see terrorist organizations using young Arab men to make their attacks even though they're eight times less likely to clear security than 80-year-old Laotian women. That means the correct thing to do with extra security is to completely ignore the high success rates of 80-year-olds and try to drive the effectiveness of Arab teens even lower.

> your argument about recruiting rates, while possibly correct, is, well, irrelevant :)

I'm not making an argument about recruiting rates; point me to where I mentioned the concept. I'm making an argument about recruiting costs. Doesn't matter why the cost is what it is, or whether your attackers are there voluntarily.

"For every person in the world, there is a cost to recruit them into your organization as an attacker, which I postulate is low for young Arab men and high for Laotian grandmothers (and also high for 80-year-olds generally). Call this cost function COST(x)."

If you are including the cost of coercion in COST(X), then i strongly disagree with your theory that there is a huge cost differential between young arab men and 80 year old laotian grandmas.

Maybe you can explain why you believe there is?

It seems, for example, that holding the family member of an 80 year old grandma hostage is cheap and effective as a mechanism of recruitment.

Well, the first thing we can observe is that we're not getting attacks by these highly nonsuspicious types, so we can safely assume that for whatever reason they're not cost-effective (we do know that in fact nonsuspicious types get much less screening than suspicious types, so, theoretically, they should be 100% of attackers).

The real beauty of this logic is that it doesn't actually matter if it's correct. If it is, great -- we need to harass suspicious types even more than we already are. But maybe it isn't. If terrorist groups are sending suspicious attackers for irrational or idiosyncratic reasons... that doesn't matter to us! We're trying to defend against the people they do send, not the people we think they should send. So we need to harass suspicious types again.

That said, there are plenty of heavy costs associated with kidnapping foreigners from around the world and using them as hostages:

1. The language barrier. If you want to coerce someone, you need to be able to make them understand what you want them to do.

2. Security. Your target country will hate you and make great efforts to root you out. These groups survive in countries where they have popular support. Pakistan might be willing to look the other way while you hang out and make trouble for the US; they're much less likely to look the other way while you kidnap and threaten to kill Pakistanis.

3. Public relations. Again, these groups survive where they have a certain level of popular support. But the same populace that doesn't really care when bad things happen to the Great Satan might not feel the same way about randomly kidnapping and killing bystanders from around the world.

4. Morale. The internal reflection of public relations. Most of the people in the group are there because they think it's the right thing (or can be talked into it). Kidnapping and killing peaceful foreigners from around the world could be a blow to that. Telling 80-year-old women to kill themselves just doesn't feel like the right thing.

"Well, the first thing we can observe is that we're not getting attacks by these highly nonsuspicious types, so we can safely assume that for whatever reason they're not cost-effective"

[citation needed]

For example, there were plenty of IRA bombings by older folks, etc. Even if you only consider airplanes, there have been hijackings, suspected bombings, etc by older folks.

Just because today's media focuses on certain ethnotypes and attacks does not make them actually predominant in actual attacks.