Someone please tell me how my explanation isn't relevant here, as there is no trick question.
It is twice as populous. Your chances of having two boys are 1/4. Same for your chances of having two girls. You're twice as likely to have children of different genders. That's where Zit's coin flip example comes in handy.
You had a boy, then you had a girl.
You had a girl, then you had a boy.
You had a boy, then you had another boy.
You had a girl, then you had another girl.
The first two statements unite to form 50% of all probabilities. Remove the last one and you're left with:
You had a boy, then you had a girl.
You had a girl, then you had a boy.
You had a boy, then you had another boy.
It's not about drawing a second card, but about forming a pair. There are 4 types of pairs (BB RB BR RR), 2 of which unite to form 50% of all pairs. Knowing that one card is red only eliminates 1 of 4 possibilities. You're left with 3 possibilities, 2 of which unite to form 66.666...% of all pairs.
I draw two cards for you and hand you one - it's red. You don't know if it was the first or second one I drew. I could've drawn a RB or BR pair... you'll never know.
the problem is you are putting the boy into the equation...the boy has no place in the question...the question does not ask anything about the boy,the question is asking about the other child.....what are the odds on the other child's sex....the boy is irrelevant to the question...
lets equate this to probability at a roulette wheel....the roulette wheel has ONLY red and black on it.....the second sping you got a black....what was the probablity you got black on the first spin?
Geoff, this line of reasoning is wrong b/c you are focusing on the second. If you know the son is older or younger, you would be correct. However, keeping it with roulette. You have (excluding greens) a 25% chance of ending up
BR
RB
BB
RR
If I told you that I spun A (not first or second) black, you remove the possibility of RR, leaving
BR
RB
BB
or 2/3.
Taking it further,
I spun 6 times and grouped them into pairs (spin 1 and 2, 3 and 4 and 5 and 6).
I tell you that I received statistically accurate reflection of two independent events (spins).
you watched the first spin of the first pair and see black
you watched the second spin of the second pair and see black
you watched the first spin of the third pair and see black
how many of the spins that you didn't see have to be red to be a statistically accurate reflection of a pair of independent events?