My Mini Course in Epistemic Game Theory

Home

Introduction

I have developed a mini-course in epistemic game theory.

The purpose of this course is to introduce Master students, PhD students  and researchers to the world of epistemic game theory.

Epistemic game theory is not a specialization within game theory, but rather a particular way to look at game theory. The purpose of epistemic game theory is namely to model, and analyze, possible ways in which a player in a game may reason about his opponents. This reasoning process is really fundamental in game theory: before a player makes a choice in a game, he must form a belief about what his opponents will do, which in turn is based on what he believes that his opponents believe that others will do. That is, before making a choice, a player must reason about his opponents, and subsequently base his choice on this process of reasoning.  

My approach in this course

In this course, I use the following four-step approach:

1. We first identify an intuitive way of reasoning about your opponents, by means of one or several examples.

2. In order to fully understand the behavioral consequences of this way of reasoning about your opponents, it is necessary to build a formal mathematic model that precisely describes the way of reasoning.

3.  Once such a model has been delivered, we can logically derive the choices that a player can rationally make if he follows this pattern of reasoning.

4. If possible, we present an algorithm that computes, for a given player, all the choices he can rationally make if he follows this pattern of reasoning.

About the course

The course consists of six lectures. Each lecture will take approximately two hours. To every lecture there is a set of exercises that students and researchers can work on.

Until now given, I have given the course at the following places:

bulletMax-Planck Institute at Jena (Germany), Strategic Interaction Group, May 8th and May 9th 2007. (In compressed form).
bulletMaastricht University (The Netherlands), May 15th and May 16th 2007. (In compressed form).
bulletUniversidad Carlos III de Madrid (Spain), Department of Economics, November 8th and November 9th 2007. (In compressed form).
bulletUniversity of Amsterdam (The Netherlands), Institute for Logic, Linguistic, and Computation. February 2008 until April 2008. (Full course).
bulletLausanne (Switzerland), at the Second Annual Meeting of the Swiss Graduate Society of Logic and Philosophy of Science, June 17th and June 18th 2008. (Part I and III).
bulletUniversity of Aarhus (Denmark), Center for Algorithmic Game Theory, April 15th, 16th and 17th 2009.  (Complete mini course).

 

Contents of the course

Chapter 1: Introduction

1.1. What is game theory about?
1.2. Epistemic models
1.3. My approach in this course
1.4. Outline of this course
1.5. Surveys on epistemic game theory
1.6. References

Part I : Standard beliefs in static games

Lecture 1: Common belief in rationality

Chapter 2: Common belief in rationality

2.1. Example: Where to locate my pub
2.2. Example: Going to a party
2.3. Beliefs diagram
2.4. Epistemic model: The main idea
2.5. Epistemic model: Formal definition
2.6. Common belief in rationality
2.7. Algorithm: Iterated strict dominance
2.8. Related models
2.9. References

Exercises to Lecture 1

Lecture 2: Epistemic foundation for Nash equilibrium

Chapter 3: Epistemic foundation for Nash equilibrium

3.1. Definition of Nash equilibrium
3.2. Example: Teaching a lesson
3.3. Self-referential beliefs
3.4. Epistemic foundation for two players
3.5. Example: The lazy professor
3.6. Extension to more than two players
3.7. Related models
3.8. References

Exercises to Lecture 2

Part II : Lexicographic beliefs in static games

Lecture 3: Common weak belief in rationality

Chapter 4: Common weak belief in rationality

4.1. Example: Should I call her or not?
4.2. Lexicographic beliefs: The main idea
4.3. Lexicographic beliefs: Definition
4.4. Optimal choices
4.5. Epistemic model
4.6. How to define belief in opponent's rationality?
4.7. Common weak belief in rationality
4.8. Example: Teaching a lesson
4.9. Algorithm: Dekel-Fudenberg procedure
4.10. Related models
4.11. References

Exercises to Lecture 3

Lecture 4: Proper rationalizability & An epistemic foundation for iterated weak dominance

Chapter 5: Proper rationalizability

5.1. Example: To which pub shall I go?
5.2. Respecting the opponent's preferences
5.3. Proper rationalizability
5.4. Example: To which pub shall I go?
5.5. Existence
5.6. Related models
5.7. References

Chapter 6: An epistemic foundation for iterated weak dominance

6.1. Iterated weak dominance
6.2. Example: To which pub shall I go?
6.3. Assuming versus believing
6.4. Assuming the opponent's rationality
6.5. k-th order assumption of rationality
6.6. Reference

Exercises to Lecture 4

Part III : Conditional beliefs in dynamic games

Lecture 5: Common initial belief in rationality

Chapter 7: Common initial belief in rationality

7.1. Example: Father and son
7.2. Dynamic games
7.3. Conditional beliefs
7.4. Optimal strategies
7.5. Epistemic model
7.6. Common initial belief in rationality
7.7. Algorithm: Ben-Porath procedure
7.8. Connection with lexicographic beliefs
7.9. Dynamic games with perfect information
7.10. Related models
7.11. References

Exercises to Lecture 5

Lecture 6: Common strong belief in rationality & Epistemic foundations for backward induction

Chapter 8: Common strong belief in rationality

8.1. Example: Father and son
8.2. Strong belief in rationality
8.3. Common strong belief in rationality
8.4. Example: Watching TV together
8.5. Algorithm: Pearce-Battigalli procedure
8.6. Proper belief revision
8.7. Related models
8.8. References

Chapter 9: Epistemic foundations for backward induction

9.1. Example: Kramer versus Kramer
9.2. Overview of epistemic foundations for backward induction
9.3. Backward induction
9.4. Forward belief in material rationality
9.5. Belief in rationality at future and parallel information sets
9.6. References