4 ms·
You're going to say it's 1/3. I'm going to say that the problem is not described clearly enough to assert that. For instance, let's say you tell me you have tw
by chowells 7y ago
You're going to say it's 1/3. I'm going to say that the problem is not described clearly enough to assert that.
For instance, let's say you tell me you have two children. One was born on a Tuesday, the other on a Wednesday. I ask about the one born on Tuesday, and you tell me it's a boy. With the information I have available to me at this point, I am correct to judge that the odds you have two boys are 1/2.
This story differs from the one you provided in ways that are insignificant for story telling.
The way you tell the story depends on the reader picking up on a rather tiny detail to reach the conclusion you intend. Specifically, the reader must come to the conclusion that you are someone playing games with them who has intentionally used very careful wording to only eliminate the girl/girl entry from the outcome chart.
That's not the way people communicate naturally. If someone says "I have two kids. One's a boy...", they aren't playing probability games, they're thinking of a child they'd like to tell you a story about. In that case, when they have a distinguished child in mind before making that statement, the odds are in fact 1/2 that the other child is a boy and hence you have two boys. (Modulo all the legitimate criticisms that can be raised that there are more possible outcomes than boy/girl and that even between those two the odds aren't equal. This is a simplification of reality for the sake of a puzzle.)
In the end, problems like this mostly hinge on poor communication that abuses common patterns to convey less information than the reader interprets from their experiences communicating in the real world. As such, they're really rather boring.
- hanoz 7y ago> You're going to say it's 1/3. I most certainly am not. I'm going to say it's 13/27. I agree the imprecise nature of the English language muddies the water somewhat, but the question can be re-framed to remove this aspect and yet still have people give the wrong answer, and as such is not just a "really rather boring" case of "intentional" "poor communication", but offers an interesting insight into the shortcomings of our intuition for probability. So let's set aside the original problem and put this one instead (which may or may not be equivalent to the original depending on your interpretation, but let's take this in isolation anyway): Woman: How may children do you have? Man: Two. Woman: Is at least one of them a boy who was born on a Tuesday? Man: Yes. What now is the probability the man has two boys? The answer is in fact 13/27, because of the unlikely extra information revealed by the Tuesday condition, which I think is surprising and interesting.
- jakeinspace 7y agoThis is a great response. I'll just add that the Wikipedia page [1] gives a good description of this ambiguity with probability notation. [1] https://en.wikipedia.org/wiki/Boy_or_Girl_paradox#Analysis_of_the_ambiguity https://en.wikipedia.org/wiki/Boy_or_Girl_paradox#Analysis_o...