Thanks for posting that one. It gave me something to think about on the my drive back from Southwold to Paris today. On the one hand I thought that the probability argument that it makes sense to change (ie by changing you have a 2 in 3 chance of being right) is right in theory. However, intuitively, once one door has been opened you're into a 50/50 situation as far as the other two doors are concerned, in which case it makes no difference. Even having read your explanation above, I'm still not sure that I go along with the conclusion that it's better to change.
Look at it this way: initially you have a 2 in 3 chance of picking a goat; if you stay with that choice having been shown the wrong door you will still be 2 in 3 to pick a goat. If you swap doors then you transfer the 2 in 3 chance to the car.
