Sorry, in advance, if this gets preachy. I've wrestled with these ideas for 15 years and still haven't really found my voice. Sorry too for my feeble ascii art skilz.
On Sep 2, 7:32 pm, Natalie Pang <[email protected]> wrote: > People are playing the prisoner's dilemma - they have no idea what others > are bidding so they put in bids as high as they could afford in the attempt > to secure them. A quibble: The term "prisoners dilemma" has come to be used for all manner of interactions in which there is either a sucker's payoff or a failure in communication or just generally where the observer thinks a suboptimal solution has been reached. But in its inception the Prisoners' Dilemma was never so wide ranging. If there is iteration, it's a different game. If there is any possibility of communication, it's a different game. The Prisoners' Dilemma not only includes a "sucker's payoff" in the payoff grid, but absolutely precludes iteration and communication. Axelrod did a great disservice by not being more precise in his naming of his tournament. By so doing he popularized a term severed from its proper meaning and thus muddied an already confusing scenario. Nor is he alone. In "Two-Person Game Theory" no less a luminary than Rapoport himself launches into a discussion of a game with a sucker's payoff but without mention of the initial hypothetical scenario from which the name of the game is derived (Game 34, Chapter "Nonnegotiable Games"). Rapoport offers the following grid in Game 34, which he labels, "The Prisoners Dilemma": C2 D2 __________________ | | | C1 | 5,5 | -10,10 | | | | -------------------------------- | | | D1 | 10, 10 | -5,-5 | | | | -------------------------------- But the original scenario would look more like this: Player A Confess NotConfess __________________ P C | | | l n | -24,-24 | -240,0 | a f | | | y s -------------------------------- e | | | r N | 0,-240 | -6,-6 | o | | | B t -------------------------------- It is a conceit of game theory that it treats of mathematics and has no truck with messy human behavior. To the extent this is true, fine. But since the primary interest in game theory seems to be the desire to predict behavior (or at least to evaluate it so we can argue for or against some given policy) it is only proper to take into account those qualitative matters which affect human decision making. Playing for abstract points will trigger different evaluative processes than playing for jail time. But even when playing simply for abstract points, point maximization criteria can yield different choices than, say, point difference maximization (that is, it is possible, depending on the payoff grid, to maximize one's points but lose overall, likewise to win by accepting fewer points because in so doing one has limited the opponent even more.) Even more damning in Axelrod's work is that, as others before him, he attaches labels to the moves which strongly prejudice in favor of a given result, and this act of labeling is a prime source of confusion about the game in its various forms. Where Axelrod and others use the labels, "Cooperate" and "Defect", the original uses "Confess" and "NotConfess", and in so doing raises an emotional bias against the option of NotConfessing. Some discussions of the game bring a paradoxical feeling that one does best by going against one's interest, or, put in the converse, pursuing one's interests is against one's interest, but this is primarily an epiphenomenon of using wildly emotive terms like cooperate and defect. Worse still, using Axelrod's work, simply by holding the payoff grids stable but changing the labels one could as easily write a book called, "The Evolution of Defection", or "...of Corporate Malfeasance". Notwithstanding the above failings, which are found in most discussions, there is another major flaw in every single analysis I have seen: Inapt Visualization. Use of a grid in the accepted manner leads to an illusion of interdependence which is wholly lacking in the actual game as envisioned by Tucker and from which the name of the game derives. A better visual model might be: Prisoner A | | Confess NotConfess | | | | 0 -24 -6 -240 The prisoner is alone in a cell. If she confesses she will either serve no time or serve two years. THERE IS NO WAY TO PREDICT OR INFLUENCE WHICH RESULT OBTAINS. If she does not confess, she will either serve six months or twenty years, again, with no means by which she can predict or influence which result obtains. Efforts to predict yield an infinite progression of "He'll think that I think that he thinks that I think..." The prisoner might as well be tossing coins. The only real choice in the matter is which of two coins the prisoner will choose to toss. But the grid visualization makes this very hard to see, instead promoting that infinite progression of "he'll think I think he thinks..." Price wars may (or may not) be the result of failure to rationally achieve and optimal result, but they bear little resemblance to the Prisoners' Dilemma. > While this may not be purely a case of a wrong game (since the perception of > COEs is that they're scarce, and in reality it is based on how many the > government wishes to release into the market), I wonder if there is a certain > critical threshold by which it becomes a wrong game. Something this does remind me of is governmentally created false scarcities in intellectual property usage rights. Given modern technology it is child's play to create an effectively limitless supply of electronic copies of various cultural artifacts, distribution of these artifacts is similarly simple and robust. Where supply approaches infinity, one would reasonably expect costs to go towards zero. However, in the U.S. copyright and patent law prohibit unauthorized duplication, creating a false scarcity, and propping up prices for an industry desperate to stave off the clear course of history. To my eye that is very much a "wrong game". -- You received this message because you are subscribed to the Google Groups "CooperationCommons" group. To post to this group, send email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/cooperationcommons?hl=en.
