Copyright: 2003
Publisher: Cambridge
ISBN: 0-521-81462-6

Game Theory fascinates me even though I am not nearly a good enough mathematician to claim to be any good at the subject.  Even so when I saw this book, Decision Making Using Game Theory, by Anthony Kelly, I decided to dive in and enjoy!  While it is a bit technical at times, for the most part Dr. Kelly keeps the topic on a very practical level with all of his examples pulled from the business world.  Even though the examples are simplified, they are not contrived.  This is a great book for anyone with even a cursory understanding of game theory who wants to apply their knowledge to the business world.

Definitions

Kelly uses very clear terminology and explains what each term means in context.  Here are a few definitions that helped me understand game theory better.

  • Perfect / Imperfect information.  Describes the knowledge of the other players moves.  If you have incomplete information you don't know what moves the other player has made.
  • Complete / Incomplete information.  Describes the knowledge of probabilities and payoffs in outcomes. 
  • Games of Chance with Risk: You know the probabilities and associated payoffs.
  • Games of Chance with Uncertainty:  You don't know what the probabilities or payoffs are.
  • Utility value: Relationship between your preference for an outcome and a lottery of alternative outcomes.
  • Maximax: Choose the greatest payoff
  • Maximin: Avoid the worst possible payoff
  • Minimax: Avoid the strategy of greatest regret.  (What is the difference between a given choice and the greatest payoff)
    • Find the Minimax by looking for the minimum row & the maximum column
Random Musings

I won't attempt to give a comprehensive overview of everything covered in this book.  The book itself is fairly short and a quick read. These are just some thoughts that jumped out at me as I read.

  • Bernoulli thought that the relationship between money and utility was probably logarithmic rather than linear.
  • Von Neumann - Morgensten point out that players will always try to maximize their utility value rather than their expected value
  • John Kenneth Galbraith - "A wrong decision isn't forever, it can always be reversed.  But the losses from a delayed decision can never be retrieved."  (A Life in Our Times)
  • Minimal social situations are covered by a "win-stay, lose-change" strategy.
    • I want to explore the possibility that UI design is a minimal social situations game.
  • Mixed Nash lead to cooperation with regard to information, which is the opposite of zero sum games.
  • Exploitation games can explain why terrorism is a "rational" choice.  (Rational is distinct from moral here)
  • In Duopoly models, Nash equilibriums aren't usually pareto efficient.
What is Rationality?

In the final section of the book, Kelly covers a number of criticisms of Game Theory.  One such criticism surrounds the differing definitions of rational behavior.  As mentioned in the previous section, terrorism can be considered rational if not moral.  Kant might argue otherwise!

  • Instrumental rationality: All humans act in their own self interest
  • Kantian: If not all players can select a strategy, it is irrational.  (Unselfishness)
  • Bounded: We have limited computational resources so we use simple rules to govern rational behavior
Conclusion

This book is thought provoking at times but mostly it is an intensely practical application of game theory to business management and leadership.  I would recommend this book for any business manager who wants a better understanding of how to quantify decision making.