News

Loading...

Tuesday, June 19, 2007

More Game Theory

Here's what I mean: A game theorist can set up a game and say "If a real situation exactly duplicates this game, and a fully rational actor in the situation is concerned solely with monetary [for instance] payoffs, then he will employ this strategy." The theory itself cannot say how closely any real world situation does match its structure, nor that an actor in the situation should be concerned only with what the game describes as the payoffs. For example, we sometimes hear something like, "While fully rational actors would bet this way at Game X, real humans are less than fully rational and, through their [love of risk / fear of loss /etc.] bet differently. But why is it "irrational" not to seek solely the reward that the game theorist has stipulated, and instead seek, say, the thrill of risk taking? Other than the theorist having mistaken his model for reality, there is no reason at all.

6 comments:

  1. Keep this sort of stuff coming. I love it.

    ReplyDelete
  2. Well, we could say "If a person is rational AND has the assumed preference ordering AND doesn't find disutility in the mental process of thinking through the likelihood and value of different outcomes AND never makes an error in that thought process..." and avoid this mistake. But economists want to believe that they are saying something about the real world, something that specifies a range of things that could actually happen and a range of things that could not actually happen. Relaxing the unrealistic assumptions makes it much harder to make those kinds of concrete claims.

    Most economists would rather say something false than say nothing, so we make unrealistic assumptions with the understanding that our conclusion will be an educated guess at what is likely, not a certain claim as to what is true.

    The real problem isn't the auxiliary assumptions (besides rationality) in game theoretical models. The problem is that too many economists want to take a tool that is moderately helpful for anticipating how the circumstances of the present will affect the future and use that tool for a purpose to which it was never suited.

    ReplyDelete
  3. "The real problem isn't the auxiliary assumptions (besides rationality) in game theoretical models."

    I quite agree! It's using the models to start to judge real-world people as "irrational" that I object to. (I think I'm agreeing with you here.)

    ReplyDelete
  4. The theory itself cannot say how closely any real world situation does match its structure

    That's exactly my complaint about the Austrian School's a priori view of human action.

    ReplyDelete
  5. That may be your complaint, but in the case of the basic theory of human action it does not apply, in that, for instance, to say that someone acts to achieve a situation that they prefer to one the one that would hold if they did not act is simply part of what it means to act. This complaint carries more weight in reference to, say. the Austrian Theory of the Business Cycle, which may more or less apply.

    ReplyDelete
  6. when you haveLOTRO Gold,you will have more
    Lord Of The Rings Gold,because you
    buy LOTRO Gold,that is so funny.At last,we sell
    cheap Lord Of The Rings Gold for future.

    ReplyDelete