John Conway had a teaching technique where he would appear to make an error in a talk, act slightly confused, then restart the problem now that a faulty assumpt

John Conway had a teaching technique where he would appear to make an error in a talk, act slightly confused, then restart the problem now that a faulty assumption (for example) was made explicit. Second time around greater insight would be reached. You have to be very good to pull off that kind of move.