[MUD-Dev] Decision making...

[David Kennerly]

J C Lawrence wrote:
>  http://msnbc.msn.com/id/5570554/site/newsweek/ wrote:

>>   As predicted by the theories of mathematician John Nash
>>   (subject of the movie "A Beautiful Mind"), A makes the most
>>   money by offering one dollar to B, keeping nine for himself,
>>   and B should accept it, because one dollar is better than none.

...

>>   People playing B who receive only one or two dollars
>>   overwhelmingly reject the offer.  Economists have no better
>>   explanation than simple spite over feeling shortchanged.

Nash equilibrium for the repeated game version better explains human
behavior.  If B rejects always, then A has played suboptimally, too.
For example, just run the experiment 10 times.  Then B has the
opportunity to "teach" A what B will accept as fair.  Even if B
receives suboptimal results, so does A, and A is the one cutting the
cake.  So A's optimal cut, against an optimal B is closer to, but
not necessarily equal to, a 50/50 split.

The single game version is a silly version of the experiment, if
discusing human behavior.  Humans play repeated games much more than
they play nonrepeated games.  Even among strangers that they'll
never see again, a thread of repetition is included in the
interaction, because if you fuck someone over (e.g., hit a
pedestrian with a car), society provides methods of recourse to fuck
back (e.g., charges of attempted manslaughter and civil suits).

Axelrod made a good game theoretical argument for this (Evolution of
Cooperation), and, in (my biased opinion!), Matt Ridley went even
one further (through findings in anthropology and evolutionary
sciences) to state that government is largely unnecessary and
inefficient at creating civility--only repetition of interactions is
necessary for civil interaction (The Orgins of Virtue).  All other
things being equal, by civility, I mean closer to a 50/50 split.

David
_______________________________________________
MUD-Dev mailing list
MUD-Dev at kanga.nu
https://www.kanga.nu/lists/listinfo/mud-dev