Boy or Girl Paradox Simulator
Print PageWhy does Day-of-Week change the probability?
This paradox showcases the extreme sensitivity of conditional probability to additional details.
- When we say "at least one is a boy", the possible gender pairs are: {BB, BG, GB} (removing GG). Out of these 3, only {BB} matches, yielding 1/3.
- When we say "at least one is a boy born on Tuesday", the day-of-week acts as an extra selector. There are 27 possible combinations of child configurations that satisfy this criteria. Out of those, exactly 13 combinations consist of two boys, yielding 13/27 (48.15%).
The Boy or Girl Paradox (also known as the Two-Child Problem, Mr. Smith’s Children Paradox, or Conditional Gender Probability Puzzle) is a classic counter-intuitive probability puzzle popularized by Martin Gardner in 1959. The fundamental premise considers a family with two children (where boys and girls are equally likely, P(Boy) = P(Girl) = 0.5) and asks for the probability that both children are boys. Depending on how the information is specified or acquired, the probability answers range between 1 ÷ 3 (33.33%), 1 ÷ 2 (50.00%), and 13 ÷ 27 (48.15%).
The paradox reveals how linguistic ambiguity in natural language alters the underlying conditional sample space. Unless the exact information gathering process is specified (whether information was revealed by a general statement, an identified child, or a random observation), human intuition conflates distinct probability models.
Our free online Boy or Girl Paradox Calculator provides instant calculations, interactive sample space trees, and simulation tools across all 4 classic formulations:
- Formulation 1: Unspecified Boy (“At least one is a boy”):
P(Both Boys | At least 1 Boy) = 1 ÷ 3 ≈ 33.33%. - Formulation 2: Specified Child (“The older child is a boy”):
P(Both Boys | Older is Boy) = 1 ÷ 2 = 50.00%. - Formulation 3: Random Observation (“You see a boy playing in the yard”):
P(Both Boys | Saw Boy) = 1 ÷ 2 = 50.00%. - Formulation 4: Tuesday Boy Variant (“At least one is a boy born on a Tuesday”):
P(Both Boys | Tuesday Boy) = 13 ÷ 27 ≈ 48.15%.
Master 4-State Two-Child Sample Space Table
For any two-child family where birth order matters (Elder Child, Younger Child), there are four equally likely underlying outcomes, each with a 1 ÷ 4 = 25% prior probability:
| Family State | Elder Child | Younger Child | Prior Probability | Valid under “At Least 1 Boy”? | Valid under “Older is Boy”? |
|---|---|---|---|---|---|
| State 1 (BB) | Boy | Boy | 1 ÷ 4 (25.0%) | YES (Target State) | YES (Target State) |
| State 2 (BG) | Boy | Girl | 1 ÷ 4 (25.0%) | YES | YES |
| State 3 (GB) | Girl | Boy | 1 ÷ 4 (25.0%) | YES | NO (Eliminated) |
| State 4 (GG) | Girl | Girl | 1 ÷ 4 (25.0%) | NO (Eliminated) | NO (Eliminated) |
Step-by-Step Proofs for Formulations 1, 2, 3, and 4
Formulation 1: Unspecified Boy (“Mr. Smith has 2 children; at least one is a boy”)
The information “at least one is a boy” rules out State 4 (GG). The remaining reduced sample space contains three equally likely states: {BB, BG, GB}.
Out of these 3 valid states, only 1 state (BB) contains two boys.
P(BB | At least 1 Boy) = 1 ÷ 3 ≈ 33.33%.
Formulation 2: Specified Older Child (“The older child is a boy”)
Specifying that the older child is a boy eliminates both State 3 (GB) and State 4 (GG). The remaining reduced sample space contains two equally likely states: {BB, BG}.
Out of these 2 valid states, only 1 state (BB) contains two boys.
P(BB | Older is Boy) = 1 ÷ 2 = 50.00%.
Formulation 3: Random Observation (“You see a boy playing in the yard”)
Here, you observed a randomly selected child. Let Saw Boy be event S.
By Bayes’ Theorem: P(BB | S) = [ P(S | BB) · P(BB) ] ÷ P(S).
P(S | BB) = 1.0; P(S | BG) = 0.5; P(S | GB) = 0.5; P(S | GG) = 0.
Total P(S) = (1.0 · 1/4) + (0.5 · 1/4) + (0.5 · 1/4) + 0 = 1/4 + 1/8 + 1/8 = 2/4 = 0.5.
P(BB | S) = (1/4) ÷ (1/2) = 1 ÷ 2 = 50.00%.
Formulation 4: Gardner’s 2010 “Tuesday Boy” Variant
If Mr. Smith states: “At least one is a boy born on a Tuesday.”
Each child has 14 possible gender-day states (7 days × 2 genders). The family has 14 × 14 = 196 total outcomes.
Number of outcomes with at least 1 Tuesday Boy = 27.
Number of outcomes with two boys where at least one is a Tuesday Boy = 13.
P(BB | Tuesday Boy) = 13 ÷ 27 ≈ 48.15% (Adding the day of the week pulls the probability from 33.33% up near 50%!).
Comparative Summary of Boy or Girl Paradox Formulations
Below is a comparative reference chart showing all classic linguistic formulations of the Two-Child Problem, their sample space counts, and exact mathematical probabilities:
| Formulation / Wording Variant | Information Revealed | Reduced Sample Space Size | Calculated Probability P(Both Boys) |
|---|---|---|---|
| Unspecified Boy Statement | “At least one is a boy” | 3 states {BB, BG, GB} | 1 ÷ 3 ≈ 33.33% |
| Specified Older/Younger Child | “The older child is a boy” | 2 states {BB, BG} | 1 ÷ 2 = 50.00% |
| Random Sight Observation | “You see a boy in the yard” | Weighted Bayes’ sample space | 1 ÷ 2 = 50.00% |
| Tuesday Boy Specific Variant | “Boy born on a Tuesday” | 27 states out of 196 | 13 ÷ 27 ≈ 48.15% |
History & Mathematics: 1959 Martin Gardner’s Scientific American
1959 Martin Gardner & Scientific American Column
American mathematics communicator Martin Gardner introduced the Two-Child Problem in his May 1959 “Mathematical Games” column in Scientific American. Gardner presented the puzzle to highlight how subtle ambiguities in how information is phrased change conditional sample spaces.
The 2010 G4G “Tuesday Boy” Revelation
At the 2010 Gathering 4 Gardner conference, statistician Gary Foshee presented the “Tuesday Boy” variant: “I have two children. One is a boy born on a Tuesday. What is the probability I have two boys?” The unexpected answer of 13 ÷ 27 (48.15%) demonstrated how adding seemingly irrelevant specific detail shrinks the denominator toward 1/2.
Popular direct tools:
Frequently Asked Questions (FAQ)
Why is the answer to “at least one is a boy” equal to 1/3 and not 1/2?
Because there are 4 equally likely starting states for two children: {BB, BG, GB, GG}. Eliminating GG leaves 3 valid states: {BB, BG, GB}. Only 1 of those 3 states is BB, making the probability 1/3 (33.33%).
Why does specifying the “older child” change the probability to 1/2?
Specifying which child is a boy (the older one) eliminates both GB and GG, leaving only {BB, BG} (2 states). One out of 2 states is BB, so the probability is 1/2 (50.0%).
What is the “Tuesday Boy” Paradox answer?
The probability that a family with two children has two boys given that at least one is a boy born on a Tuesday is 13 ÷ 27 ≈ 48.15%.