Simon on satisficing
📚SOURCE
💎PERMANENT NOTES
🧵SYNTHESIS
Simon starts from a complaint about the economic man of the textbooks, who knows every option, can work out what each one leads to, and ranks all the outcomes on one scale. No actual person or firm has that information or that arithmetic. So he asks what a rational procedure looks like for a chooser with limited knowledge and limited ability to compute.
His answer has three moves. First, simplify the payoff: sort outcomes into acceptable and not acceptable, without scoring each one. Second, search in sequence: options turn up one at a time and looking has a cost. Third, set an aspiration level and take the first option that clears it. If acceptable options are easy to find, the level rises. If the search drags, it falls.
This is 🧩 Satisficing, although the word 🔤 Satisfice is not in this paper. He coined it the following year. What the 1955 paper gives is the machinery, and the insistence that this is not a lapse from rationality. Given the cost of search, stopping at good enough is the rational thing.
✨KEY IDEAS
The chooser of classical theory needs knowledge and computing power that no real organism has.
Replace a full ranking with a simple test: does this outcome meet the standard or not.
Someone selling a house sets a price they will accept and takes the first offer above it.
The aspiration level moves: up when good options come easily, down when they do not.
A chess player does not examine every line, only until a satisfactory move is found.
📝READING NOTES
Paraphrase, not quotation. Quarterly Journal of Economics, 1955. Hard going: set notation, and a writer who assumes you have the textbook he is arguing with.
What I understood
- The paper is about procedure. Classical theory describes the best outcome. 🏛️ Herbert A. Simon describes a way of choosing that a person could carry out.
- The limits are on both sides: what the environment lets you know, and what you can compute.
- The stopping rule does the work. Without one, search never ends.
What I did not
Whether satisficing can always be rewritten as optimising with a search cost added. Economists later said yes. I think he would say that misses it: you do not know the cost of search in advance either.
My argument with it
The model is silent on where the first aspiration level comes from. In his examples it is given. In my life that is the entire difficulty. The sabbatical fund, the newsletter draft, the roadmap: in each case the search is easy and choosing the bar is hard. Left alone, my bar only moves one way, which is the opposite of what his rule predicts and exactly what 📖 Housel on enough describes.
So I take the mechanism and add a step he does not need: write the bar down before looking.
Uses this year
- Three laptop options in an afternoon would have been an evening of thirty tabs. Noted for next time something dies.
- Noor's three comments on a draft are a satisficing rule for editing: fix those, send.
- Became 💎 Good enough is a rational stopping rule, which sits beside 💎 Enough is a number I have to choose.
Read the same weekend the flat weeks ended, and four days before I asked Katarina for the role. Not unrelated. I had been waiting for a best moment, and there is only a good enough one.
🗣️IN SHORT
- In : A Behavioral Model of Rational Choice, Quarterly Journal of Economics, 1955
- Simon argues that : a real chooser cannot optimise, and can still be rational by searching until an option meets an aspiration level
- by stating that : knowledge and computation are limited and search is costly, so a simple acceptable or not test, applied to options as they arrive, with a bar that adjusts to how the search is going, is a sound procedure