These judgements are not from personal experience. If I'm correct, Hods has taken several hundred or thousand pieces of data and used them to determine the likelihood of getting certain items. His explanation seems good enough for me.