Review of “The Joy of Game Theory”
The first official book I finished in my 100 Books in 2018 goal is The Joy of Game Theory by Presh Talwalkar.
One might think that a book squarely focused on the topic of strategic thinking and game theory must have a lot of math and arcane algorithms to parse. This book doesn’t introduce anything that would require an advanced math degree to understand. It did have me wanting to rent “A Beautiful Mind” again after reading about the Nash Equilibrium.
For me, the most interesting theories and explanations that the book offered were around the Nash Equilibrium, The Prisoner’s Dilemma, and Mutually Assured Destruction. Reading through this book it was fascinating in that I had never considered the concept of dominant strategies or, rather, never knew what they were as officially referenced in game theory.
Let’s take the prisoner’s dilemma as that one is often seen in TV dramas. Two suspects are placed into separate rooms. They are told they will be rewarded or penalized based on how each person acts. The scenarios from the book:
—If both suspects conceal information, each will serve a 1 year sentence
— If both disclose information and confess, both will be convicted and serve 3 years
— If only one discloses, that suspect will be set free and the other will serve a 4 year sentence
In a Nash equilibrium, it’s about each person playing their best response to the other person. Looking at the above, you have to identify what the best strategy would be for the other suspect in this situation, and they would play the dominant strategy which game theory says you should always play. The right choice is to disclose, which gives you either the possibility of a 3 year sentence, or to be set free if your partner conceals.
If I were writing the examples, I may have placed a harsher penalty on the last choice to emphasize the stark difference between them. Therein lies a fallacy that I didn’t understand at the time. Given the choice between 3 years of prison, and 4, it’s a tougher choice. If the penalty for not confessing was longer, perhaps 8 years, the dominant choice would always be to confess.
The text goes on to talk about several situations where the prisoner’s dilemma is applied in life outside of possible incarceration.
All in all, I enjoyed this book, and it was a good introduction to the some of the concepts in game theory that can be useful in real life. It suffers a bit at times from a lack of payoff at the end of describing a theory. I learned later that the book was compiled from a lot of blog articles that the author had written over some span of time. It helps describe the “so what” feeling I had after some of the sections.
If you’re interested in game theory at all, this is a quick read, and I recommend it.
