[Brian,
Wednesday January 26, 2005 at 3:33pm]
Books on Game Theory
In a comment to a prior post, Iain asked for a basic Game Theory book. Larry Levy mentioned:
In fact, that's the book I pulled off the shelf when I wanted to brush up. Along with "How to Lie with Statistics," one of the better math for non-math type reads out there.
But I want something more mathy. Not too mathy, I like English more than formula after formula, but something that assumes a basic technical background. I can do Linear Algebra, program, etc. [At one time, I could do hoary math. But I've gotten my degree so I don't do that anymore]. The problem with Williams book is that it takes 5 pages to tell me what to do a simple formula, and while it gives an alorithm for how to solve these things, it doesn't mention 'why' it works. To me, that's annoying.
So, does anyone have any mid-range books on Game Theory or Developing automated computer opponents ("AI") worth reading? [Most of the AI books I've glanced at focus on pathfinding, shooters, etc. Not strategy games, although the O'Rielly book is interesting, as always.]
Really good websites also appreciated. Please drop them in the comments for everyone.
I still think the best book for what you're looking for is J.D. Williams' The Compleat Strategist. This explains the basics of game theory (mostly two-person zero sum, which covers a lot of interesting cases and is certainly the best introduction to this discipline) and does so in an entertaining and sometimes hilarious fashion. I'm pretty sure the book doesn't use anything more complex than simple arithmetic.
In fact, that's the book I pulled off the shelf when I wanted to brush up. Along with "How to Lie with Statistics," one of the better math for non-math type reads out there.
But I want something more mathy. Not too mathy, I like English more than formula after formula, but something that assumes a basic technical background. I can do Linear Algebra, program, etc. [At one time, I could do hoary math. But I've gotten my degree so I don't do that anymore]. The problem with Williams book is that it takes 5 pages to tell me what to do a simple formula, and while it gives an alorithm for how to solve these things, it doesn't mention 'why' it works. To me, that's annoying.
So, does anyone have any mid-range books on Game Theory or Developing automated computer opponents ("AI") worth reading? [Most of the AI books I've glanced at focus on pathfinding, shooters, etc. Not strategy games, although the O'Rielly book is interesting, as always.]
Really good websites also appreciated. Please drop them in the comments for everyone.
As for a more technical book on Game Theory, you may want to check out Games and Decisions, by Luce &Raiffa, one of the more classic texts in the field. The book covers the classic 2-person zero-sum game problem, but also examines less common topics like utility theory, non-zero-sum games, and multiplayer games. It's intended for math students, but is still pretty clearly written.