Decision Theory: Great Expectations

Core idea

Decision theory is the branch of logic that studies practical reasoning — reasoning about how one should act when the outcomes are not fully under one's control. Its central tool is the expectation (also called expected value or expected utility) of an action.

How expectation is computed

When you act, several outcomes are possible. You can usually estimate two things about each: how likely it is, and how much you value it. Assign each outcome a value on an open-ended scale — positive numbers are good, negative numbers bad, zero is indifference. The expectation of an action is the sum of each outcome's value, weighted by its probability.

Priest's argument: "Expectation" here is a technical term with virtually nothing to do with its ordinary English meaning — it is a probability-weighted average of how much you stand to gain or lose.

The rational-choice rule

Once you can compute the expectation of every available option, the rule is simple: choose the action with the greatest expectation. If two tie, pick at random.

Why it matters

Decision theory turns vague deliberation into arithmetic. It underlies economics, game theory, insurance, and risk management — anywhere choices are made under uncertainty. It also sharpens famous philosophical arguments. Pascal's Wager claims you should believe in God because non-belief risks an infinitely bad outcome that swamps every other figure. Decision theory exposes the flaw: the Wager omits relevant possibilities — there are many possible gods, several jealous, so once the full table is drawn up, no theistic belief comes out best.

Key takeaways

Mental model

Copy sourceZoom

Practical application

Example

Should you buy travel insurance? It costs $40. Without it, a lost-bag mishap costs you $600; with it, the same mishap costs $40 excess. You estimate the probability of a mishap at 0.1.

• Buy insurance: mishap outcome 0.1 x (-80) plus no-mishap outcome 0.9 x (-40) gives an expectation of -44.
• Skip insurance: mishap outcome 0.1 x (-600) plus no-mishap outcome 0.9 x 0 gives an expectation of -60.

Since -44 is greater than -60, decision theory says buy the insurance — the small certain loss beats the large probable one.

Related lessons