Reading about game theory in my Software Engineering book Algorithms to Live By by Brian Christian & Tom Griffiths made me remember a section about the prisoner’s dilemma and the implications of it in game theory that I needed to mention due to its importance. Algorithms to Live By is a book that I read in Software Engineering and Max had also read the same book and talked about the first chapter. You can view it here. While Max spoke about the first chapter, I will discuss the last one, game theory.

Last time I focused on the math side of the prisoner’s dilemma. This time, I will focus on the psychological and its implications on human nature. Going back to the actual dilemma, the prisoner is given two choices, either defect or stay silent. When we think about this problem, surveys initially conclude that a majority of the population would stay silent. (I’ll get back to the reason why in a bit.) But after listening to that most people would stay silent, when given a choice again, people then chose to defer to benefit their own selves. This then gets implemented into an algorithm. In Algorithms to Live By, we learn that an algorithm would also say that deferring is the optimal, or dominant solution. This is because usually, the benefits of deferring are enough to ignore the negatives. But that is if people know that the other person will likely stay silent. In the actual dilemma, both parties would know that most people are likely to stay silent and they would both defer leading them to failure and demise.

Why is this the case? And why is it that initially, people tend to stay silent? These questions are answered by our current state of society. Most people want to be ahead of a group of people (for example an employee in a workplace) because that will allow them to gain the most rewards. This is shown in the prisoner’s dilemma when people defer when they know the other person will stay silent. They will gain the most rewards. This type of mental process is a self-serving process. But as we learned from the survey, people initially chose to stay silent. Why would people choose to stay silent at the risk of not getting the rewards (or at the risk of getting punished)? This is because the world we live in is a communal place, we don’t work by ourselves to further development, we gain the most by working as a society. Since most people subconsciously believe in the idea of working together, they would choose to stay silent in the prisoner’s dilemma. In psychology, we learn the term operant conditioning. Operant conditioning is a method of learning that gives rewards and punishments based on the behavior you are exhibiting. To find an example of this conditioning, we don’t have to look further than school. In high school, we are given a lot of group projects and assignments. When you contribute more to the group assignment and the quality of the end product is higher, your grade will also be higher. The behavior in this example is the group project, and the reward is your grade. The positive reinforcement you gain from getting a good grade makes you realize that you should work to better your group. This is backed by the fact that when working after school, you most likely are working with others. In order to maximize the growth of your workplace, working together cohesively is important. 

So, going back to the prisoner’s dilemma. Based on the demographic of the survey, we can see why these people chose to stay silent. The people who were surveyed were above the age of 18, and most likely have graduated high school. These people have most likely been conditioned to believe that working in a group is better. This type of mental process is a society-serving process. What we have learned is that people want to work together in a group, but they also want to be one step ahead of the group to gain the most rewards. In Algorithms to Live By, this mental process of self-serving versus society-serving is elaborated. Consider a street in downtown Naperville. Suppose there are 3 stores that sell bagels of equal quality on street. These three business owners agree to only work 6 AM to 3 PM on Mondays through Saturdays. This is ideally

a society-serving process because everyone gets paid an equal amount and bagels are getting sold to everyone. But then a fourth store opens up, and this business owner has a self-serving mental process. This business owner finds out that there is a lot of demand for bagels between 5 AM and 6 AM. So the owner opens up their store an hour earlier than the other stores. This breaks the society-serving pact the other owners made, and all of the sudden the balance of profit is skewed. This is something an algorithm would suggest to do because it benefits you, but the algorithm doesn’t account for the society it tears down along with it.

Algorithms to Live By indicates that the best way to find answers to these dilemmas is by both looking at the algorithm and what you believe based on your experiences. Algorithms are tools to help us learn about ourselves and our behaviors, and game theory is just one example of a mix between algorithms and human thinking (math and psychology). The quote that ends the chapter on Game Theory sums it up perfectly: “Seek out [problems] where honesty is the dominant strategy. Then just be yourself.” Algorithms that allow you to be yourself are always the best strategy. 

Ideas inspired by Algorithms to Live By by Brian Christian & Tom Griffiths