So, does the No Free Lunch-like theorem only predict these results, or something stronger than this?
You have it. (Roughly speaking.)
Of No Free Lunch relevance is the additional consideration that most decision theories perform poorly at tasks other than maximising expected utility according to the supplied values. That is, No Free Lunch tells us that decision theories are very bad at making bad decisions.
You have it. (Roughly speaking.)
Of No Free Lunch relevance is the additional consideration that most decision theories perform poorly at tasks other than maximising expected utility according to the supplied values. That is, No Free Lunch tells us that decision theories are very bad at making bad decisions.