Based on a idea from Nick Bostrom.

Suit A: "Welcome to our futurology meeting extravaganza, where we are going to do a complete analysis of the future using... drumroll... Scenario analysis!"

All: "All hail mighty scenario analysis!"

Suit A: "So, what are the big risks in the future?"

Suit B: "Global warming? I heard that's bad."

Suit A: "Indeed it is. What else do we have that's bad?"

Suit C: "How about obesity?"

Suit B: "I still think global warming is rather more important, it's getting hot and..."

Suit C: "Well, my grandfather was fat, and he suffered and died because..."

Suit A: "No need to argue, gentlewomen! We'll simply do a scenario analysis with both variables. So here we have the Sweaty Fat quadrant... Let me put it up on the board:"

Suit A: "Now let's give each scenario a thorough analysis!"

Suit D: "Isn't fat an insulant?"

Suit A: "That's the kind of incisive commentary we need!"

...

...

Much later:

Suit C: "So we have an ideal strategy: keep an eye on sweat pants purchase, and adjust our investment accordingly."

Suit D: "What about our social responsibilities?"

Suit A: "Good point."

Suit B: "Well, then we can track the size of suits and ice cream consumption and adjust health spending and gas subsidies in function of these."

Suit A: "Well, I think we've done a fabulous job today; really. No-one could have done a better job predicting than us. And it's all thanks to... Scenario analysis!"

All: "All hail!"

 

(very tangentially connected to the problem of models that are over-precise in narrow areas)

New to LessWrong?

New Comment
7 comments, sorted by Click to highlight new comments since: Today at 9:34 AM

I know this is meant to be parody, but how closely does it resemble scenario analysis in the corporate world? From what I've read about the actual use of scenario analysis (e.g., at Shell), the process takes much longer (many sessions over a period of weeks).

Second, and more importantly: suits are typically not quants, and have a tendency to misinterpret (or ignore) explicit probabilities. And they can easily place far too much confidence in the output of a specific model (model risk). In this context, switching from full-on quant models to narrative models (as scenario analysis entails) can increase accuracy, or at least improve calibration. This is a classic "roughly right vs. precisely wrong" situation.

I feel like you're straw-manning scenario analysis. Here's an actual example of a document produced via scenario analysis: Global Trends 2035.

I feel like you're straw-manning scenario analysis.

I am.

But if we're going to analyse scenario planning seriously - which I certainly didn't here - we need to look at older scenario planning attempts, and see how useful they were.

If you admit that this is an unfair strawman, then why are you bothering to post it?

Caricatures can be highly useful as pointers at particular failure spots. Caricatures and parodies are, by definition, "unfair strawmen".

It's a partially unfair strawman. And it is labelled a parody.

The moral, for me, is that you would never entrust your future to a thought process narrow enough to pass through the heads of nine suits in the space of an hour. At some point, some person or model will have to carry out an effulgence of inarticulably intricate, nuanced computations that you'll never understand, and you're going to have to learn to just trust its conclusions.