12/29/2020 09:46

Risk Analytics

My dream, the real dream from this morning not the mataphorical dream of aspriations and vague hopes, was about risk analysis. That is, the dream was a nightmare of being caught in terrorist activity inside an urban church but it was analytical in the sense of systematically highlighting specific risks: structural fire, firearms, cell phone disruption, assault.

Thinking analytically is not the same as performing analysis; the latter is more comprehensive and should exhibit a formal kind of completeness.

Long ago and far away -- specifically, around 1980 in the city of Madison -- a work colleague remarked the the city was vulnerable to total loss of electric power from an attack on transmission lines at a single point in the southwest of town.

That illustrates somebody thinking analytically about risk. (The person talking to me may have been a parrot or the thinker; I've always suspected parrot but can't say.)

My response was to deny the level of risk. There was at least one other point of connection to the transmission system in the northeast of the city. Furthermore at that time Madison Gas and Electric was operating a generating plant on the shore of Lake Monona and the University of Wisconsin operated another on the south side of the campus. So disrupting the single point would not result in total loss of electric power across the city.

That also illustrates analytical thinking. To perform a true analysis would have taken a lot more work.

First off, the premise needs to be examined more fully. How hard a target is the transmission line and the substation? How "tough" are the lines and the substation, how much damage can they absorb while continuing to operate? How resilient is the infrastructure once disruption occurs? And how many different types of damage are possible?

Next, what effect would disruption have on connected systems? If the power could be completely cut off, for example, would part or all of the city go black? What would happen to the power from the other sources? What about effects outside the city? Subsequent regional blackouts proved that the transmission grid of the time was extremely brittle; when faced with sudden change the system engaged in a kind of apoptosis and fell apart. (Apparently nobody had performed the analysis.)

Then we should try to quantify the risks. It is easier to say "it seems unlikely that this would ever happen". In the decades since that conversation nobody has blown up the substation which proves only that such an attack does not occur as frequently as once per decade. It is hard to estimate how likely it is that an attack would be attempted, but a full analysis would also consider how likely it is that any attack would be successful at any specific level of damage.

Just writing this outline of a risk analysis is a lot more effort than either I or my interlocutor were going to invest in the idea. After all, in our actual reality we had our own jobs to do.