But in the first example we let 49 people with either GB or BG out of the equation early on, because we gave them the choice about the gender of the child they want to mention, and that's why the odds are better in favour of BB. This choice is not there in either your version of the puzzle or the Scientific American's. Sure, if Dan volunteers the information, this choice is maybe implied, but it is not something we'd have to take into account I think. We're only dealing with Dan, after all, not with 196 parents.
This is too confusing
