In reply to Andy Hardy:
> I have just written out a truth table to look at this and whilst I get that there are different numbers of permutations if you know that 1 boy was born on Tuesday, it feels like that spoof proof that 1 = 2. So to put my mind at rest please could you let me know the answer to the following questions.
To clarify, the interpretation of the problem is the standard: What proportion of all the two child families with at least one boy born on a Tuesday consist of two boys?
The answer is 13/27=0.481. It at first seems unintuitive that this should be different from the answer without the Tuesday bit, which is 1/3=0.333 (Only BB out of the equally likely BB, BG, GB families consists of 2 boys). The obvious question is whether it is significant that the answer is only a little bit les than 1/2.
> 1. I have 2 children, 1 of the them is a left handed boy, born on a Tuesday. What is the probability my other child is also a boy?
A table quickly gives the answer as 27/55=0.491, which is even closer to 1/2.
> 2. I have 2 children, 1 of them is a brown-eyed left handed boy, born on a Tuesday. What is the probability my other child is also a boy?
We need to know the proportion of children with brown eyes. If it is, say, 1 in 5, then the answer is 139/279=0.498
The less likely it is for a given boy to meet all the conditions, the closer the answer gets to 1/2. In fact, if the probability of a given boy meeting all the conditions is 1/n, then a bit of thought about the table shows that the answer is (2n-1)/(4n-1) = 1/2 - 1/(8n-2). For the original Tuesday n=7, include left handedness and n=14. For a left handed boy born on a Tuesday on the first of January, n=2 x 7 x 365 = 5110 and the probability is 0.499976.
> If the additional info about the boy makes no difference to the probabilities, why should the day of birth?
Here is how to make it more intuitive that it does make a difference:
A BB family has had two shots at having a boy born on a Tuesday, so nearly* twice as many BB families will have a boy born on a Tuesday than will BG or GB families. So say we have 70 BG, 70 GB and 70 BB families. About 10 of each of the BG and GB families will have a boy born on a Tuesday but nearly 20 of the BB families will. So of the nearly 40 families with a boy born on a Tuesday, nearly 20 (nearly 1/2) will consist of two boys.
* "nearly" because a rare family will have two boys born on a Tuesday, in a sense getting more than their fair share and depriving another family of a boy born on a Tuesday. Adding extra information makes such families rarer and rarer and so makes the probability closer and closer to 1/2.
> If the additional info about the boy makes a difference, how many additional pieces of info would you need to be 99% certain of my other childs gender?
The probability will never exceed 1/2 however much additional information is given!