Wednesday, August 08, 2012

Priorities over Enumeration

Science is about probabilities, not possibilities.  Many smart people don't appreciate this because logic  is often focused on what can be proved true or false, whether something is impossible, or possible.  There are very few 'existence' or impossibility theorems describing social phenomena that are interesting, mainly because it usually takes a pretty small adjustment in assumptions to negate the proof.  For example, I have not met anyone who thinks the first or second welfare theorem was compelling in developing their political beliefs.

In any case, I was struck by this notice of how America's latest psycho killer was on an FBI watch list, along with 420,000 others:
The FBI's apparently had Wade Michael Page on it's radar six years ago, he wound up killing six and injuring several others when he opened fire at a Sikh temple in Wisconsin. 
Figures released last fall put the FBI's watch list total at 420,000 -- including 8,000 Americans. But former deputy assistant director Danny Coulson says you can't watch everyone, even if they are an angry ex-military member with ties to white supremacist groups like Page reportedly was.
I imagine that list of 8000 Americans is pretty simple, they just looked at White Power groups and such. As most of these guys are just disgruntled, there's not much to do. A similar thing happened with the Colorado Batman shooter, where a psychiatrist notified the school the kid was troubled, but nothing was done, as usual.

A major problem with risk management is large organizations are very good an enumerating improbable risks, not so good at priorities.  That's because collectives often lose their common sense when they become a certain size, where those downstream often get little feedback, and so they simply create reports that highlight measurables--enumeration of risk--and their bosses accept this because they too are clueless and have little intuition.  Incompetent managers are a cancer, and they metastasize.  These are inevitable inefficiencies of scale, why the Too-Big-Too-Fail banks don't outperform smaller ones even with their taxpayer funded bond guarantee.

Large organizations should not so much focus on preventing these risks, but rather containing them via a quick response. Knight's initial mistake last week is understandable, stuff happens, but the delayed response--even the NYSE called after only 5 minutes to tell them they had a bug and it took them another 25 minutes to pull the plug--suggests a more profound problem.  

No comments: