6 ms·

Great read. My two favorite quotes: The Prisoner’s Dilemma is the obverse of Adam Smith’s theory of the invisible hand, in which the free market coördinates th

by hy3lxs 17y ago

Great read. My two favorite quotes: The Prisoner’s Dilemma is the obverse of Adam Smith’s theory of the invisible hand, in which the free market coördinates the behavior of self-seeking individuals to the benefit of all. Each businessman “intends only his own gain,” Smith wrote in “The Wealth of Nations,” “and he is in this, as in many other cases, led by an invisible hand to promote an end which was no part of his intention.” But in a market environment the individual pursuit of self-interest, however rational, can give way to collective disaster. The invisible hand becomes a fist. and Limiting the development of [complex] securities would stifle innovation, the financial industry contends. But that’s precisely the point. “The goal is not to have the most advanced financial system, but a financial system that is reasonably advanced but robust,” Viral V. Acharya and Matthew Richardson, two economists at N.Y.U.’s Stern School of Business, wrote in a recent paper. “That’s no different from what we seek in other areas of human activity. We don’t use the most advanced aircraft to move millions of people around the world. We use reasonably advanced aircrafts whose designs have proved to be reliable.” The second seems quite applicable to software engineering as well.

Perceval 17y ago
It's unfortunate that the article uses the Prisoner's Dilemma as its example, because the Prisoner's Dilemma is actually not the best game model to set up as the opposite of the Invisible Hand. A more appropriate model would be the Tragedy of the Commons, which was proposed by a fellow named Garrett Hardin in an article in a 1968 edition of Science magazine. If you are interested in formal modeling of coordination problems, an interesting overview can be found in Elinor Ostrom's book Governing the Commons. In chapter one she gives an overview of the Prisoner's Dilemma, the Tragedy of the Commons, and the Logic of Collective Action (Mancur Olson's theory). After reviewing these abstract formal models, she looks at other possible institutional solutions for coordination problems. Here are some links: http://en.wikipedia.org/wiki/Prisoners_dilemma http://en.wikipedia.org/wiki/Prisoners_dilemma http://en.wikipedia.org/wiki/Tragedy_of_the_Commons http://en.wikipedia.org/wiki/Tragedy_of_the_Commons http://en.wikipedia.org/wiki/Logic_of_Collective_Action http://en.wikipedia.org/wiki/Logic_of_Collective_Action http://en.wikipedia.org/wiki/Coordination_game http://en.wikipedia.org/wiki/Coordination_game http://en.wikipedia.org/wiki/List_of_games_in_game_theory http://en.wikipedia.org/wiki/List_of_games_in_game_theory
- DanielBMarkham 17y ago
  Any quick one-liners or easy-to-understand wisdom that you've gained from the material? The entire subject area looks immensely fascinating, yet very dense at times.
  - Perceval 17y ago
    Yes, game theory can get inscrutably complex, even through the initial models are pretty simply to set up and understand. The basic components of any coordination/cooperation problem are the same: you have actors with preferences, you have a situation with rules that they can't change, and you have outcomes with payoffs/penalties depending on whether the actors choose to cooperate or not. Most of the problems revolve around how to get people to cooperate, when defection is attractive for some reason. Most of the solutions involve making sure situations are not one-off interactions, that all actors have information about what each other is doing, and that in the case of multiple actors using common resources, you need to have some form of monitoring and enforcement. If actions are one-off, the temptation to act selfishly is higher than if actors know they are in a long-term situation in which they must choose to cooperate or defect time and time again. In this way information such as reputation can then function as a signaling device to other actors. Games such as the Prisoner's Dilemma work only when presented as a one-off situation in which both actors have no information about what they other is going to do. So the solution is 1) make sure you're not setting up one-off interactions, and 2) make such actors are in an information rich environment wherein they can make appropriate signals about their intentions and past behavior. Most of the other games, such as Tragedy of the Commons or Battle of the Sexes, have stable solutions if you repeat them and increase the information available to the actors. The equilibrium solutions are called Nash Equilibriums, after John Nash, the paranoid schizophrenic mathematician featured in the movie A Beautiful Mind. http://en.wikipedia.org/wiki/John_Forbes_Nash,_Jr http://en.wikipedia.org/wiki/John_Forbes_Nash,_Jr. A bit more than a one-liner, but hopefully this helps a bit.
    - alecco 17y ago
      Beware of simplistic conclusions of very basic game theory, especially if it's coming from RAND. http://en.wikipedia.org/wiki/Tragedy_of_the_anticommons http://en.wikipedia.org/wiki/Tragedy_of_the_anticommons This documentary enlightens a bit on the subject and is quite accessible. http://en.wikipedia.org/wiki/The_Trap_%28television_documentary_series%29#1._.22Fuck_You_Buddy.22_.2811_March_2007.29 http://en.wikipedia.org/wiki/The_Trap_%28television_document... http://video.google.com/videoplay?docid=404227395387111085# http://video.google.com/videoplay?docid=404227395387111085# (AFAIK not released outside of Britain due to footage rights issues, a real shame.)
anamax 17y ago
One difference between designing airplanes and govt regulation is that designing airplanes isn't prone to folks making design decisions to benefit to friends and family. Political corruption is pretty much a fact of life whenever govt gets involved in money. I don't see why folks who complain about corruption involving govt contracting believe that financial regulation will be any better.

by hy3lxs 17y ago

Perceval 17y ago

It's unfortunate that the article uses the Prisoner's Dilemma as its example, because the Prisoner's Dilemma is actually not the best game model to set up as the opposite of the Invisible Hand. A more appropriate model would be the Tragedy of the Commons, which was proposed by a fellow named Garrett Hardin in an article in a 1968 edition of Science magazine. If you are interested in formal modeling of coordination problems, an interesting overview can be found in Elinor Ostrom's book Governing the Commons. In chapter one she gives an overview of the Prisoner's Dilemma, the Tragedy of the Commons, and the Logic of Collective Action (Mancur Olson's theory). After reviewing these abstract formal models, she looks at other possible institutional solutions for coordination problems. Here are some links: http://en.wikipedia.org/wiki/Prisoners_dilemma http://en.wikipedia.org/wiki/Prisoners_dilemma http://en.wikipedia.org/wiki/Tragedy_of_the_Commons http://en.wikipedia.org/wiki/Tragedy_of_the_Commons http://en.wikipedia.org/wiki/Logic_of_Collective_Action http://en.wikipedia.org/wiki/Logic_of_Collective_Action http://en.wikipedia.org/wiki/Coordination_game http://en.wikipedia.org/wiki/Coordination_game http://en.wikipedia.org/wiki/List_of_games_in_game_theory http://en.wikipedia.org/wiki/List_of_games_in_game_theory

DanielBMarkham 17y ago
Any quick one-liners or easy-to-understand wisdom that you've gained from the material? The entire subject area looks immensely fascinating, yet very dense at times.
- Perceval 17y ago
  Yes, game theory can get inscrutably complex, even through the initial models are pretty simply to set up and understand. The basic components of any coordination/cooperation problem are the same: you have actors with preferences, you have a situation with rules that they can't change, and you have outcomes with payoffs/penalties depending on whether the actors choose to cooperate or not. Most of the problems revolve around how to get people to cooperate, when defection is attractive for some reason. Most of the solutions involve making sure situations are not one-off interactions, that all actors have information about what each other is doing, and that in the case of multiple actors using common resources, you need to have some form of monitoring and enforcement. If actions are one-off, the temptation to act selfishly is higher than if actors know they are in a long-term situation in which they must choose to cooperate or defect time and time again. In this way information such as reputation can then function as a signaling device to other actors. Games such as the Prisoner's Dilemma work only when presented as a one-off situation in which both actors have no information about what they other is going to do. So the solution is 1) make sure you're not setting up one-off interactions, and 2) make such actors are in an information rich environment wherein they can make appropriate signals about their intentions and past behavior. Most of the other games, such as Tragedy of the Commons or Battle of the Sexes, have stable solutions if you repeat them and increase the information available to the actors. The equilibrium solutions are called Nash Equilibriums, after John Nash, the paranoid schizophrenic mathematician featured in the movie A Beautiful Mind. http://en.wikipedia.org/wiki/John_Forbes_Nash,_Jr http://en.wikipedia.org/wiki/John_Forbes_Nash,_Jr. A bit more than a one-liner, but hopefully this helps a bit.
  - alecco 17y ago
    Beware of simplistic conclusions of very basic game theory, especially if it's coming from RAND. http://en.wikipedia.org/wiki/Tragedy_of_the_anticommons http://en.wikipedia.org/wiki/Tragedy_of_the_anticommons This documentary enlightens a bit on the subject and is quite accessible. http://en.wikipedia.org/wiki/The_Trap_%28television_documentary_series%29#1._.22Fuck_You_Buddy.22_.2811_March_2007.29 http://en.wikipedia.org/wiki/The_Trap_%28television_document... http://video.google.com/videoplay?docid=404227395387111085# http://video.google.com/videoplay?docid=404227395387111085# (AFAIK not released outside of Britain due to footage rights issues, a real shame.)

anamax 17y ago

One difference between designing airplanes and govt regulation is that designing airplanes isn't prone to folks making design decisions to benefit to friends and family. Political corruption is pretty much a fact of life whenever govt gets involved in money. I don't see why folks who complain about corruption involving govt contracting believe that financial regulation will be any better.