Homeland Security Is Not Statistics Driven

The NY Times has an interesting op-ed from Dr. Lawrence Wein, a professor of management at Stanford University. It's a good example of why we shouldn't use statistics to drive policy but rather inform us as to the possible and improbable. He wants to show how easy it is to populate the country with radiological monitors, which I think is pretty damned stupid. But I didn't run the numbers.

As the accompanying chart illustrates, there are a dizzying number of paths that terrorists could use to transport a foreign-built weapon to an American target city — 132 variations, in fact, taking into consideration all four likely modes of transport: commercial airplane, cargo airplane, container ship and cruise ship.

So, how do we decide which route the terrorists are most likely to choose and which path we the are most vulnerable to? Game theory implies that we should maintain an equal chance of detecting fissile material along each of the 132 paths because if we harden one path too much, the terrorists will simply choose an easier one. On top of it all, the agency needs to consider cost-effectiveness: if certain sets of nodes along the transportation network are much more cost-effective to reinforce than others, then the best defense may not come from allocating resources equitably across the system.
------------
The seaborne route is even more worrisome. The Coast Guard is undertaking a three-year pilot project aimed at securing maritime routes, but faces daunting challenges in both identifying suspect vessels and detecting fissile material amid the background radiation present at sea. This pathway will perhaps be the weakest link in our border defense for the next several years, and should be one of the highest priorities of the Domestic Nuclear Detection Office.

Last, assuming the terrorists aren’t planning to detonate the device at the point of entry, they must move it to the target city. They could do this in several ways: with a truck, a small airplane or, for coastal cities, a small boat. As we have no idea which is most likely, our goal should be to ensure an equal chance of detection no matter which form of transportation is used.

Essentially, he's saying that because terrorists could run down any number of paths, we need to put radiological monitors and forces at each one - every border crossing, every airport, every seaport, and every city. Maybe in an ideal world where money is no object, this might be possible. But even if it were, would it not be smarter to interdict the terrorists where the radiological material is? Stop it at the source, or stop the terrorists as they are building/moving the bomb. This idea of trying to protect everything is just wrong. 

Wein thinks he's making the case for a better monitoring process. I suggest that instead, he is making the case for why we ought to think twice - and maybe thrice - and decide against such a plan.

Comments

[Ray]

Well, the Doctor's Idea might increase numbers in some areas of the Security Industry, but consequently would raise ever higher tensions among a public that is becoming weary of the Whole Thing, in the ever growing security interruptions in the bustle of their daily lives. However, it would be safer for the terrorist therefore to obtain the material inside the country.
Better to listen for whispers in the relevant areas of unlawful activities - and keep track and monitor those places that have fissile material.
An incredulous public is the last thing that should be encouraged- it is from them that useful information may be obtained.
R.

[J.]

Yes, thanks Ray, I forgot to point out the self-obvious issue of how radiological material could be easily obtained here in the United States, thus bypassing all these monitors. Much smarter to take steps to minimize civilian facilities holding radiological material (open Yucca Mountain) and to watch over the few countries that actually can make nuclear bombs.

[George Smith]

Wein shows up in the NY Times op-ed page every few years. His purpose is to deliver terror porn, usually condensed from a scenario in a recent paper of his dealing with a presumed method of terrorism, and his recommendations.

A few years ago he tried to run with the idea that terrorists could kill by poisoning the milk supply with botox. He had a flurry of suggestions: require security checks for everyone working on milk, more locks and barriers, increasing the temperature of pasteurization.

The dairy industry used a critique Milton Leitenberg and I made to chase him off.

This outlines it at fas.org

[BGG]

Oh, god, the milk guy. I knew I recognized that name. Why does this guy keep writing about things that are outside of his expertise?

[Kilo]

"Game theory implies that we should maintain an equal chance of detecting fissile material along each of the 132 paths because if we harden one path too much, the terrorists will simply choose an easier one."

They'll simply do that eh. If that's the case, game theory would appear to be laughably useless when it comes to predicting the actions of terrorists.

Take a look at the foreign terrorist attacks on US soil that you're actually talking about. Then try to justify bringing reasoning anywhere near your predictions, let alone assigning strategic planning.

In the first attack the fkwits in question arrived at the airport with the bomb manuals in their luggage, got detained, then later claimed the same bomb making manuals back from customs so they could build their bomb.
Fkers got caught trying to claim the deposit back on the truck they blew up.

In the second, well, was their any law enforcement agency they didn't waive a red flag at and were promptly ignored ?

I mean, don't know that the 9/11 hijackers sold crack to kids while they were here, but if the DEA didn't drop the ball on them too, they'd be the odd ones out.
Those guys were reporting muggings to the cops. Getting pulled over for speeding prior to the attacks. They arrived on their own terrorist-named passports.

So when it comes to the task of planning to prevent terrorist plans, there really is no call to outwit these people. You just need to stop hiring retards in your security services.

This also raises the question what terrorists? There wouldn't appear to be any such terrorist threat as the one being described. Now, I'm happy to hear arguments that the Shining Path are working these kind of intelligent risk assessments and vulnerability targetting into their plans, but you sure as shit ain't describing al Qaeda.

You are simply describing a fictional threat.