Newcomb's problem
On this page
A decision puzzle in which a reliable predictor has filled an opaque box according to what it predicted you would choose; dominance reasoning says take both boxes, expected-utility reasoning seems to say take only one.
The problem at a glance
A guide to the entry below, quoting its own words. Skip to the full entry and its sources.
- The reasoning that leads to it , Standard form of the problem, as given in the entry; the entry does not say who numbered the steps
-
Structure of the conflict:
- whatever the opaque box contains, taking both boxes yields 1,000 dollars more than taking one, so the principle of dominance favours two-boxing;
- given the predictor's reliability, choosing one box is strong evidence that the million is there, and choosing both is strong evidence that it is not, so computing expected utility with conditional probabilities favours one-boxing;
- Result: dominance and expected-utility maximisation, the two principles Nozick set against each other, appear to conflict.
Objections and responses
- Objections and replies , Raised by other thinkers and critics, as reported
-
One-boxers' main argument is that one-boxers end up richer; evidential theorists such as Horgan, Horwich, and Hare and Hedden favour one-boxing on this basis. Causal theorists reply that the problem simply rewards irrationality, so that prospering by one-boxing does not make it rational. Within evidential decision theory, Richard Jeffrey introduced the requirement of ratifiability, and Eells's tickle defence argues that once the agent's own introspectible beliefs and desires are taken into account they screen off the evidential link, so that evidential theory too recommends two-boxing; Horwich, Sobel and Lewis objected to the tickle defence. Weirich notes that causal decision theory may concede that cultivating a disposition to one-box is rational, although one-boxing itself is not.
Coordinate 2.6Coordinate 6.5Robert NozickRobert StalnakerDavid LewisRichard JeffreySources and primary textsRead the full entry
What it means
Newcomb's problem is a decision puzzle in which two plausible principles of rational choice give opposite advice. Robert Nozick presented it in 1969, naming it after the physicist William Newcomb, who first formulated it; Paul Weirich's Stanford Encyclopedia entry on causal decision theory treats it as a central case for that theory.
The case: you may take an opaque box alone, or both the opaque box and a transparent box that visibly contains 1,000 dollars. The opaque box contains 1,000,000 dollars if a reliable predictor predicted that you would take only the opaque box, and nothing if it predicted that you would take both. The prediction has already been made.
Structure of the conflict: (1) whatever the opaque box contains, taking both boxes yields 1,000 dollars more than taking one, so the principle of dominance favours two-boxing; (2) given the predictor's reliability, choosing one box is strong evidence that the million is there, and choosing both is strong evidence that it is not, so computing expected utility with conditional probabilities favours one-boxing; therefore (3) dominance and expected-utility maximisation, the two principles Nozick set against each other, appear to conflict.
Purpose. The case forces a choice between two ways of evaluating acts. Robert Stalnaker (1972) proposed computing expected utility with probabilities of subjunctive conditionals rather than conditional probabilities; since the prediction is already fixed, the act does not change its probability, and two-boxing comes out best. Gibbard and Harper (1978) developed this proposal and distinguished causal decision theory from evidential decision theory. David Lewis argued that the prisoner's dilemma played against a psychological twin is a Newcomb problem and developed a version of causal decision theory based on dependency hypotheses; Skyrms and Joyce offered further formulations.
Objections and replies. One-boxers' main argument is that one-boxers end up richer; evidential theorists such as Horgan, Horwich, and Hare and Hedden favour one-boxing on this basis. Causal theorists reply that the problem simply rewards irrationality, so that prospering by one-boxing does not make it rational. Within evidential decision theory, Richard Jeffrey introduced the requirement of ratifiability, and Eells's tickle defence argues that once the agent's own introspectible beliefs and desires are taken into account they screen off the evidential link, so that evidential theory too recommends two-boxing; Horwich, Sobel and Lewis objected to the tickle defence. Weirich notes that causal decision theory may concede that cultivating a disposition to one-box is rational, although one-boxing itself is not.
Status. Weirich describes decision theory as an active area of research. Causal decision theory is the prevailing view among those who distinguish the two theories, but Ahmed defends evidential decision theory, Spohn argues for one-boxing on a causal model, and functional decision theory permits one-boxing.
On this site it belongs to Coordinate 2.6 (Rationality); it also bears on Coordinate 6.5 (Philosophy of Economics).
See also the glossary entries Prisoner's dilemma, Interpretations of probability, Thought experiment and Allais paradox.
The questions it belongs to
- 2.6 RationalityIs it rational to believe what the evidence supports, or what serves you?
- 6.5 Philosophy of EconomicsAre economic models describing people, or a convenient fiction about them?
Thinkers
To read
- Anarchy, State, and Utopia — Robert Nozick
- On the Plurality of Worlds — David Lewis
- Subjective Probability: The Real Thing — Richard Jeffrey
Sources and further reading
Cite this page
Responsibility: Coordinates of Thought Editors. Each statement was drafted from the sources listed above and checked against them in a separate pass before publishing; this page is not peer-reviewed by outside specialists. Source-checked 11 November 2026. How to cite this site · How we write and check.
APA (7th edition)
Coordinates of Thought Editors. (2026, November 11). Newcomb's problem (Arguments). Coordinates of Thought. Retrieved [date accessed], from https://coordinatesofthought.com/arguments/newcombs-problem/
MLA (9th edition)
“Newcomb's problem (Arguments).” Coordinates of Thought, 11 Nov. 2026, coordinatesofthought.com/arguments/newcombs-problem/. Accessed [date accessed].
Chicago (17th edition)
Coordinates of Thought Editors. “Newcomb's problem (Arguments).” Coordinates of Thought. Source-checked November 11, 2026. https://coordinatesofthought.com/arguments/newcombs-problem/.
BibTeX
@misc{cot-glossary-newcombs-problem-2026,
author = {{Coordinates of Thought Editors}},
title = {Newcomb's problem (Arguments)},
howpublished = {Coordinates of Thought},
year = {2026},
month = nov,
url = {https://coordinatesofthought.com/arguments/newcombs-problem/},
note = {Source-checked 11 November 2026}
}
RIS
TY - ELEC AU - Coordinates of Thought Editors TI - Newcomb's problem (Arguments) T2 - Coordinates of Thought PB - Coordinates of Thought PY - 2026 DA - 2026/11/11 UR - https://coordinatesofthought.com/arguments/newcombs-problem/ LA - en ER -
For reference managers (Zotero, Mendeley, EndNote): BibTeX file · RIS file