Apps

🧪 This platform is in early beta. Features may change and you might encounter bugs. We appreciate your patience!

Monty Hall Error

Also Known As: Three Prisoners problem Conditional probability neglect
Aspect ID: monty_hall_error

Definition

The Monty Hall error is the failure to correctly update conditional probabilities after structured information is revealed by a knowledgeable agent. In the classic problem, switching doors after the host (who knows where the prize is) reveals a goat doubles the probability of winning from 1/3 to 2/3. Most people — including statisticians — intuitively believe switching makes no difference.

Examples

A game show contestant picks Door 1. The host opens Door 3, revealing a goat. Intuition says the probability is now 50/50, but switching wins 2/3 of the time because the host's knowledge turns his choice into a signal about which door hides the car.

In a fraud investigation, a compliance officer narrows three suspicious accounts down to one flagged transaction. A senior auditor — who already knows which of the remaining two accounts is clean — points to one and says 'that one checks out.' The officer shrugs and says 'so it's 50/50 between the other two.' But because the auditor's reveal was informed, the originally flagged account is still twice as likely to be the source of fraud.

A doctor orders tests for three possible diagnoses. A specialist, having reviewed the full chart, rules out one of the two diagnoses the GP had not initially favored, saying 'it's definitely not condition B.' The GP concludes it must now be 50/50 between the original suspected condition A and condition C — but because the specialist's elimination was knowledge-based, not random, condition A remains far more probable than a naive 50/50 split suggests.

Verification Steps
Verification Steps
Binary yes/no questions that an AI must answer to detect a reasoning pattern in a text.
Each of the 452 aspects has verification steps — simple yes/no questions designed to systematically detect whether a pattern appears in a text. For ad hominem: "Does the argument attack a person rather than their claim?" For false dichotomy: "Are only two options presented when more exist?" This ensures consistent, reproducible analysis.

Binary (yes/no) questions an LLM must answer to identify this aspect:

  1. 1

    Does the argument assume that conditional probabilities remain unchanged after new information is revealed?

    Type: binary
  2. 2

    Is the source of new information non-random (i.e., the revealer knows the correct answer)?

    Type: binary
  3. 3

    Does the analysis fail to update prior probabilities in light of the information structure?

    Type: binary
  4. 4

    Would applying Bayes' theorem change the conclusion?

    Type: binary
Deep Dive
The expandable detail section on each aspect page with examples, psychology, and counter-strategies.
The Deep Dive section provides in-depth information about each aspect: a real-world example showing the pattern in action, an explanation of why it works psychologically, practical advice on how to counter it, alternative names, and links to related aspects.