Okay, looking at it logically... the second, unknown child has a 50% chance of being a boy, assuming no outside factors. However, we haven't accounted for the first (boy) child yet. It should come out to 75%, but I'm sure I'm missing something.