#148: Computer Models and Chicken Guts

April 10, 2020

During this pandemic, I, like many others, have been watching the daily corona virus update and have been not a little disturbed by the huge role that government has given to computer models in setting public policies. This is frightening.

It’s frightening because it reminds me of steering by chicken guts, among other entrails, in ancient times. Alexander, for example, represented then common practice by consulting his horuscipator before each battle.

Should I do this or should I do that? Is the timing of this battle good for us? Etc. etc etc.

Well, said the horuscipator, Just check the guts, stupid!

Of course, no horuscipator ever called Alexander “stupid,” at least not twice.

We’re rightly unimpressed by this way of making important decisions, studying steaming animal guts just doesn’t seem a very reliable way of choosing direction. Though, to be fair, Alexander did ok for quite while steering by guts. Maybe he was using other decision strategies as well.

Is the use of computer modeling just the modern form of chicken guts?

This question raises two other ones.

Is computer modeling a better decision maker than chicken guts? And

If neither of these methods are any good, why have we ever used either one?

Those of you who have been loyal readers of my blog already know that I am very skeptical of the use of computer models in decision making. In post #43, I introduced the adjective “scientistic” for arguments which have the trappings of science without the actual features of science which make it compelling of belief. That was in 2009. I revisited the theme three more times, the last being post #144 in 2018. I’m nothing if not consistent. The last time I brought it up was in connection with the great climate hoax in which computer models play a central role. [please remember, I am not denying that climate changes are occurring. I am also not denying that human activity has a role in that. I am denying that we know anything about that beyond what our senses tell us.]

Recently, the major player in the daily briefing, Dr. Anthony Fauci, downplayed the reliability of computer models because, as he ostensibly said, there are too many variables. Too many variables?? Duh, Dr. Fauci, we really needed an expert in epidemiology to tell us that?

Who has not watched the weather channel when a hurricane was being mapped for landfall and not seen the computer modeled paths diverge? Which one, the people on shore can be heard to cry, “which one of the models is the right one?”

Well, what all of this watching of computer models has taught us, I think, is this.

The models increasingly converge on the truth the closer they get to the event.

This means that the only computer model that actually gets it right is the one on the day of the fact. Sadly, that’s exactly the time we no longer have any need of an accurate prediction. Going back, the further from the day of the fact, the more the models diverge ever further and we have no way of knowing which one is closer to the eventual outcome than the others.

This means, pardon the French, that computer models are nothing more than today’s steaming chicken guts and absolutely no @#$%ing use in predicting the future!

But why computer models rather than chicken guts? The reason, I think, is not a mystery. Horuscipation by chicken guts is excluded from today’s tolerated bullshit sphere. It’s certainly not that bullshit is no longer tolerated, but that if bullshit is to be tolerated, it must be appropriately dressed. Primitive animistic religious bullshit does not market well anymore; however, modern scientistic bullshit is very popular.

Which leads me to the second question. If neither chicken guts nor computer models have any reliability in prediction making, why do we use them?

I only remember one thing from Hannah Arendt’s The Human Condition; she wrote that the statesman is in the terrible situation of being forced to make far reaching plans and decisions knowing full well at the time he makes them that the conditions which prompt them will change in utterly unpredictable ways from the first moment of their implementation. And that this is true for whatever adjustments he introduces in response to the changes.

This insight on Arendt’s part seemed both true and tremendously important to me at the time and still does. It means that governmental management is at its very best a lucky groping towards the future, not a competent piloted steering towards a clearly discerned goal.

This is a terrible stress on the inner life of any leader who understands his situation and it is understandable that he would reach out for and cling to any straw, no matter how insubstantial, that gives him at least a cover story for what he did if circumstances turn in a bad direction.

Many things have changed from ancient times to the present, but human nature does not. No one wants to be the one left holding the bag, everyone wants a cover story. Computer models are today’s chicken guts.

Fauci made the ultimate political gaffe, he accidentally told the truth: there are too many variables! This means that neither he nor anyone in his cohort had the remotest idea of where the pandemic would lead and, in a panic, they chose the chicken guts with the most extreme predictions to guide their policy suggestions.

Now, countries all around the world struggle with the consequences of policies based on the colors, shapes, and distributions of the steaming guts of innocent chickens.