This is the graded finale for the whole course. It pulls the entire arc together: why the perfect optimiser is a fiction (combinatorial explosion, unlistable options, the cost of computation), satisficing as an aspiration level plus a stopping rule and its identity with the reservation-value stopping rule, the maximiser–satisficer psychology and the paradox of choice, fast-and-frugal heuristics and ecological rationality with the less-is-more effect, and — just as important — the honest limits of the model. Several questions carry deliberate traps drawn from the most common misreadings of “satisficing,” so read each stem carefully before you lock in an answer.
How this exam works
This is a final, one-way exam. Questions come one at a time, and submitting an answer locks it for good — there is no going back, no retry, and no restart. Your score stays hidden until the very end, when you will see whether you passed. The pass mark is 70%. Some questions are marked select all that apply and need every correct option checked (and no wrong ones) to earn the point. Take your time on each question, because you only get one shot at it.
What is the central claim of Herbert Simon's "bounded rationality"?
Select an answer to continue.
Course Recap
Big picture
Bounded Rationality & Satisficing — the whole course
- Bounded Rationality & Satisficing
- Why optimisation fails
- Combinatorial explosion (chess ~10^120, travelling salesman), unlistable/open-ended option sets (careers, partners), and the real cost of computation — optimising is impossible in principle, not just hard.
- Simon: Olympian (unlimited-computation) rationality is a fiction; real agents have procedural rationality. The mind as scissors — cognitive limits AND environment structure.
- Satisficing & the aspiration level
- Satisfice = satisfy + suffice: set an aspiration level (good-enough bar), take the first option that clears it, stop. Two failure modes: too low (settle for junk), too high (never stop).
- It IS a reservation-value stopping rule — the aspiration is the reservation value, and it should DESCEND as options deplete and search gets costly (adaptive aspiration).
- Maximisers vs satisficers
- Maximisers chase the best: marginally better outcomes, substantially worse feelings — open counterfactuals, regret scaling with options, inflated expectations.
- The paradox of choice (24 vs 6 jams): past a point, more options paralyse and dissatisfy. Default to satisficing; curate the option set; reserve maximising for the few.
- Fast-and-frugal heuristics
- Take-the-best (one good cue), recognition heuristic (bet on the familiar), 1/N (split evenly) — simple rules that match or beat complex optimisation out of sample because complex models OVERFIT (bias–variance).
- Ecological rationality: a rule has no fixed IQ — it is smart only relative to the environment it exploits, and misfires when the world shifts. Less-is-more, stated carefully.
- Transfer & honest limits
- Triage first: stakes × repeatability × tractability → optimise or satisfice. Set an explicit bar anchored to reality; take the first clearer; let it descend as the runway shrinks.
- Limits: the aspiration must be WELL SET (a lazy bar is settling), some decisions deserve optimising, and a heuristic is only as smart as its fit to the terrain. Satisfice by default, optimise by exception.
- Why optimisation fails
Key takeaways — the whole course
Bounded rationality is the recognition that real minds can’t optimise — combinatorial explosion, unlistable options, and the cost of computation make “find the best” impossible in principle — so they satisfice: set an aspiration level, take the first option that clears it, and stop. That is the same reservation-value stopping rule as optimal stopping, with the bar tuned to how plentiful the options are and how costly the search, descending as the runway shrinks. The maximisers who refuse to stop win a sliver of quality and pay dearly in regret, delay, and the paradox of choice, so the smart default is to satisfice and curate. Simon’s second scissor blade — the environment — is where fast-and-frugal heuristics live: take-the- best, recognition, and 1/N can beat complex optimisation out of sample because complex models overfit, but only where they ecologically fit, and they misfire when the world shifts. Above all, respect the honest limits: the aspiration must be well set (a lazy bar is just settling), a few high-stakes, repeatable, tractable decisions genuinely deserve optimising, and no heuristic is smart outside its terrain. Triage the decision, set the bar, take the first clearer, and stop — good enough, fast, on the many; deep optimisation on the few.