Read the notes, then try the practice. It adapts as you go.When you're ready.
Session Length
~17 min
Adaptive Checks
15 questions
Transfer Probes
8
Decision theory is the interdisciplinary study of how rational agents select actions among available alternatives when outcomes are uncertain. Drawing on mathematics, philosophy, economics, psychology, and statistics, it provides formal frameworks for analyzing choices by modeling preferences, beliefs, and the structure of decision problems. At its core, decision theory asks what it means to make a good decision and whether there are systematic principles that rational decision-makers should follow.
The field divides into two major branches: normative decision theory and descriptive decision theory. Normative (or prescriptive) decision theory investigates how decisions should be made under ideal rationality, producing frameworks such as expected utility theory, Bayesian decision theory, and minimax strategies. Descriptive decision theory, by contrast, examines how people actually make decisions, incorporating findings from psychology and behavioral economics about systematic departures from rationality, such as those documented by Prospect Theory and satisficing behavior.
Decision theory has far-reaching applications across disciplines. In economics, it underpins models of consumer choice and market behavior. In artificial intelligence, it guides the design of autonomous agents that must act under uncertainty. In medicine, it informs clinical decision analysis and evidence-based treatment selection. In philosophy, it raises deep questions about rationality, free will, and the nature of preference. The field continues to evolve through connections with game theory, information theory, and computational approaches to bounded rationality.
One step at a time.
A normative framework holding that a rational agent should choose the action that maximizes the weighted average of utilities across all possible outcomes, where weights are the probabilities of each outcome occurring.
Example: When deciding whether to carry an umbrella, you implicitly weigh the utility of staying dry (times the probability of rain) against the disutility of carrying it unnecessarily (times the probability of no rain).
An approach that combines Bayesian probability (updating beliefs based on evidence) with utility theory to determine optimal actions. The decision-maker assigns prior probabilities to uncertain states, updates them with new evidence, and chooses the action maximizing expected utility.
Example: A doctor uses Bayesian reasoning when updating the probability of a disease after receiving test results, then deciding whether to treat based on the expected utility of treatment versus no treatment.
Decision under risk involves known probability distributions over outcomes, while decision under uncertainty (or ambiguity) involves unknown or imprecise probabilities. Different decision rules apply to each: expected utility works under risk, while maximin or minimax regret may apply under uncertainty.
Example: Choosing between lottery tickets with known odds is decision under risk; choosing whether to invest in an entirely new market with no historical data is decision under uncertainty.
Conservative decision rules for situations of deep uncertainty. Maximin selects the action whose worst-case outcome is best (maximizing the minimum payoff). Minimax regret selects the action that minimizes the maximum possible regret across all states of the world.
Example: A general planning a military operation under complete uncertainty about enemy positions might use maximin, choosing the strategy that guarantees the least-bad outcome regardless of what the enemy does.
A foundational result proving that if a decision-maker's preferences satisfy certain axioms (completeness, transitivity, continuity, and independence), then their preferences can be represented by a utility function, and they will act as if maximizing expected utility.
Example: If you consistently prefer coffee to tea, and tea to juice, and your preferences are stable when lotteries over these options are introduced, then a utility function exists that captures your preference ordering.
The framework assuming that individuals make decisions by selecting the option that best satisfies their preferences given their beliefs and constraints. It provides the theoretical backbone for much of economics, political science, and sociology.
Example: A consumer allocating a fixed budget among goods chooses the combination that maximizes satisfaction, considering prices and personal preferences.
A decision strategy introduced by Herbert Simon in which agents seek a solution that meets a minimum acceptability threshold rather than optimizing. This accounts for the cognitive costs of exhaustive search and computation.
Example: Rather than researching every possible apartment in a city, a renter sets criteria (under $1500, near transit, allows pets) and takes the first option that meets all thresholds.
A thought experiment demonstrating that people prefer known risks over unknown risks (ambiguity aversion), violating the independence axiom of expected utility theory. It reveals a systematic preference for betting on urns with known compositions over urns with unknown compositions.
Example: Given an urn with 30 red balls and 60 balls that are either black or yellow in unknown proportion, most people prefer betting on red (known probability) over black (unknown probability), even though expected utility theory says they shouldn't care.
More terms are available in the glossary.
Choose a different way to engage with this topic — no grading, just richer thinking.
Explore your way — choose one:
See how the key ideas connect. Nodes color in as you practice.
Walk through a solved problem step-by-step. Try predicting each step before revealing it.
This is guided practice, not just a quiz. Hints and pacing adjust in real time.
Small steps add up.
What you get while practicing:
The best way to know if you understand something: explain it in your own words.
More ways to strengthen what you just learned.