Bayesian Decision under Uncertainty
How this teaches game theory
This lab shows how rational decisions depend on beliefs as well as payoffs. A signal updates prior probabilities through Bayes' rule, and the decision maker should choose the action with the highest expected utility rather than the action that would have been best after seeing the hidden state.
Complete scenario, mode, level, and concept guide
Scenarios
Weather-sensitive event
Choose an outdoor or indoor plan after receiving a weather forecast.
Cybersecurity investment
Choose whether to invest heavily or lightly after observing a risk signal.
Product launch
Choose a large or small launch after receiving imperfect market research.
Modes
Human decision maker
You interpret the signal and choose an action without a recommendation.
Human + recommendation
The computer displays the Bayesian expected-utility recommendation, but you still make the final choice.
Computer decision maker
The computer chooses the expected-utility-maximizing action automatically.
Levels
Level 1 — Clear signal
Signals are highly informative, so posterior beliefs move strongly.
Level 2 — Noisy signal
Signals are less reliable, making prior beliefs more important.
Level 3 — Costly information
Observing the signal incurs a cost, introducing the value-of-information question.