14

I haven't studied game theory, but it fascinates me. My intuition is that it isn't used by most "enterprise app" developers. However, it is clearly relevant to the large online sites (e.g. recommendation systems), and a huge influence on SO.

Have you applied any principles of game theory in your daily projects? If so, which principles?

Michael Easter
  • 23,733
  • 7
  • 76
  • 107
  • ps. I sincerely think this is an interesting question. SO does not agree, with its "subjective!" warning, so I've marked as community wiki. – Michael Easter Dec 01 '09 at 02:24
  • Um, can you explain how it's relevant to recommendation systems? I studied game theory in the context of economics in particular, and I can't figure out how the things I've studied might be applied to computer science. Granted, I only really remember stuff about Nash equilibrium, maybe other stuff that I don't remember applies more. – Ray Hidayat Dec 01 '09 at 02:45

2 Answers2

2

I am pretty sure that Hal Varian's work on auctions at Google must have a strong flavor of game theory, or at least micro-economics of decisions under uncertainty...

Mathias
  • 15,191
  • 9
  • 60
  • 92
1

I designed & wrote a review system and recommendation system for an online recipe-sharing website a long time ago. The recommendation system didn't require any game theory, just stats and sets, but thinking about the problems I had to solve with the reviews, they were definitely game-theory-y (though I didn't think about it that way at the time).

The problem that the site I worked on faced was that people could be both authors and reviewers, so there was a certain pressure to "pump up" their own recipes' stats perhaps by "pushing down" others. So there was a balance that had to grow organically from our rules to offset that selfish impulse; we did this with a per-user reputation score based on other peoples' reviews of their recipes and how helpful other people thought their reviews were. In retrospect I wish I had more rigorously applied some game theory here.

A research paper which discusses the same. Manipulation-Resistant Recommender Systems through Influence Limits, Resnick and Sami, 2008. Here.

Harsh Nisar
  • 84
  • 1
  • 12
Moishe Lettvin
  • 8,462
  • 1
  • 26
  • 40
  • Jeff Atwood has made blog posts in the past about tweaking the SO reputation system. I don't remember how much detail he got into, but it was along similar lines. – James Dec 01 '09 at 04:00