Talk:Thought experiment

From RationalWiki
Jump to: navigation, search

Newcomb's Problem[edit]

Is it just me, or is Newcomb's Problem not really a paradox at all? It seems to me in order to make it a paradox, you have to make the assumption that the "intelligence" is actually not able to take into account your future state, which isn't stated in the problem. Presumably, if the "intelligence" is able to account your knowledge of its decision making process, it should still have an accurate prediction. Then, if you know the accuracy P of the intelligence, and the money in box A and box B, you can mathematically make a best decision. Somebody please prove me wrong, because I feel safer knowing that I'm the only one on stupid pills rather than everybody else taking them. --ShadowofLords (talk) 02:27, 12 March 2013 (UTC)

Replace "paradox" with "dilemma"? Scream!! (talk) 02:54, 12 March 2013 (UTC)
I guess. But even then, it's a solvable "dilemma". If the "intelligence" is able to predict with an accuracy of P (including taking into account how you will think in this future state), then by taking B you should expect a return of P*B. If you take both, you should expect a result of (1-P)*B+A. It seems less like a "dilemma" or "paradox" and more like a "this only is a problem if you are vague about the knowledge and accuracy of the intelligence"... --ShadowofLords (talk) 14:56, 12 March 2013 (UTC)
Whether it's a paradox or not is disputed, but it is from a Game Theory perspective as two equally valid strategies to solve it produce conflicting results, which apparently shouldn't happen. Scarlet A.pnggnostic 11:53, 13 March 2013 (UTC)
Note that the problem isn't asking you to assess the intelligence of the predictor. You assume that it is perfect. Thus is the nature of thought-experiments. Scarlet A.pngbomination 11:57, 13 March 2013 (UTC)