Two identical envelopes lie on a table. One holds twice as much money as the other. You choose one, without looking inside, and are offered the chance to exchange it for the one you left behind. Before either envelope is opened, the choice seems almost empty: there is no visible difference to guide your hand.
Act I: The choice
The setup and the puzzle
The problem is known as the two-envelopes problem, or the exchange paradox. Its first instruction is simple: pick an envelope. Its second is stranger. You are asked to consider whether, having picked, you should switch. If the envelopes are identical and your choice was random, symmetry seems to settle the matter. The other envelope has no visible claim to be better than yours, and yours has no visible claim to be better than the other.
That answer has the calmness of a catalog entry: two objects, one choice, no distinguishing mark. The puzzle changes when you name the amount in your envelope A. The other envelope must hold either twice that amount, 2A, or half that amount, A/2. The familiar switching argument assigns a one-half chance to each possibility. It then makes the choice look less balanced than it first appeared.
This is a variant of an older puzzle called the necktie paradox. The name offers a trace of its history, though the name alone does not tell us how the earlier version was framed. What carries over is the shape of the difficulty: a choice that appears symmetrical, followed by reasoning that seems to reverse the verdict.
The switching argument
Suppose you keep the envelope marked A. If the other envelope contains 2A, switching gains you A. If it contains A/2, switching costs you A/2. Give those outcomes equal weight and calculate the expected amount in the other envelope: one-half times 2A, plus one-half times A/2. That is A + A/4, or 5A/4.
Since 5A/4 exceeds A, the calculation seems to offer a quarter more than staying. It also seems to offer the same advice no matter what number is inside the envelope. The argument is short enough to carry in the mind, and each step appears familiar: name the amount, list the two possibilities, average them.
Then the advice threatens to repeat. If you switch and call the new amount B, the same reasoning appears to say that the envelope you left holds either 2B or B/2, again with equal chances. Switching back seems beneficial too. The table has not changed, and neither have the envelopes, but the arithmetic keeps pointing away from the object in your hand.
Act II: What the calculation asks us to believe
Where the intuitions collide
The symmetry intuition and the expected-value calculation are not merely two moods about the same choice. They make claims that seem to conflict. Symmetry says the two envelopes have equal standing before either is opened. The calculation says that, given an amount A in your hand, the other envelope has a higher expected value. If both claims apply to the same situation, they cannot both guide the decision as stated.
The arithmetic itself is not the obvious place to find a mistake. One-half of 2A plus one-half of A/2 does equal 5A/4. The pressure falls on the assignment of those halves. Why should the chance of finding 2A in the other envelope equal the chance of finding A/2, once the amount in your own envelope is called A? The envelopes were selected at random, but that fact by itself may not answer the question about the amount now observed.
To answer it, one must say something about how the amounts in the pair were chosen. If you know only that one amount is twice the other, the values of the pair could still vary. How those possible pairs are represented, and what information counts as evidence, matters to the probabilities assigned after the choice. The calculation can look complete while carrying an assumption about those probabilities in its pocket.
Amounts have a history
Here the puzzle turns from counting to modeling. Before an envelope is opened, uncertainty concerns which envelope holds which amount. After an amount is observed, uncertainty concerns what that observation tells you about the pair. Those are related questions, but they are not automatically identical. Naming the amount A does not, by itself, establish that the two alternatives for the other envelope are equally likely.
This is why the problem has drawn attention in decision theory and in Bayesian accounts of probability. A Bayesian approach asks how belief should change in light of the information available, including assumptions about the amounts that could have been placed in the envelopes. The model of those possibilities matters. The expected-value rule does not supply that model on its own; it works with probabilities that must already be justified.
That observation points toward a place to investigate, not a complete verdict on every version of the paradox. Proposed responses include Bayesian resolutions, approaches from mathematical economics, and a non-probabilistic variant associated with Smullyan. Philosophers disagree about which account is best. The different approaches are a reminder that the puzzle is not exhausted by spotting a suspicious line in a calculation.
What remains on the table
The two envelopes are still a useful test of how reasoning handles hidden assumptions. An expected value can be calculated correctly from its inputs and still mislead if those inputs do not fit the situation. Symmetry can also feel decisive while leaving open what changes when new information arrives. The puzzle asks us to keep both questions visible: what do we know, and what probability model makes that knowledge usable?
No practical rule about switching follows from the bare setup alone. The amounts and the way they were selected would need to be specified before a decision could be assessed. A fuller account of the proposed solutions and their disagreements appears in the two-envelopes problem entry. The compact paradox leaves a larger question behind: when a neat calculation promises an advantage, have we learned something about the choice, or only about the assumptions we smuggled into it?
Sources
- Two envelopes problem - Wikipedia
- https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQH3i8kYkzyGpfYBUkG316ZQrFvthh8fHvk9UqsyZfgptApILJqkco8miXqpe8hAKx2daQN6MDnd5V2x5CajJ1qXNSWYTaq4gvTnNZSy1N5iIlx_UtA-7fYvU1NNqfaJ5HkEW_2JR-p1_f5vCCHi7mtjWAaYMkPCqymvGIXuD2itpEEH
- The Two-Envelope Paradox | [Optional: The Two-Envelope Paradox] | 24.118x Courseware | MIT Open Learning Library
- Two Envelope Paradox
- Variables in the Expectation Formula and the Two Envelope Problem
- https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQEtC4zFFTLbkEW_CygJdzQK1OPUMGgAlCAix_iOe4f3afCVO4macCCDauS78lZnTd1gSLPsZ0vaiMb0M7dvNSBVBpTlAn65r6JvjDuo03vGtIBSTSFHXtiRjIWh_lCuaQ9RZmCA-bWsRqKKOBe-AZ_CiePMU5QAxSwk-A_1p4d5QshngA==
- [2101.03484] Resolving the two envelope paradox
- The Two-Envelope Problem: A Numerical Simulation | LUP Student Papers
An exceptional writer at Deeperpectives, she covers public memory, family archives, and the histories held in ordinary objects. Before joining the desk, she cataloged oral histories for a neighborhood museum, where she learned to treat a small detail as a door, not a conclusion.
