15.S50 | January IAP 2015 | Graduate

Poker Theory and Analytics

Pages

Course Overview

This page focuses on the course 15.S50 Poker Theory and Analytics as it was taught by Kevin Desmond during IAP 2015 .

This course takes a broad-based look at poker theory and applications of poker analytics to investment management and trading.

MIT Sloan News published an article on how Kevin Desmond links poker strategy to risk management in this course: Game Theory: What Poker and Finance Have in Common.

Course Outcomes

Course Goals for Students

  • Create an environment for study of poker theory without the need for real-money wagering.
  • Develop the basic foundation for decision-making in poker.
  • Allow students to assess their own level of play and have a framework for improvement.
  • Provide an understanding of the current poker environment and how students might leverage talent for poker in the future.

Possibilities for Further Study/Careers

Management and leadership positions, finance, trade, global markets. The skills learned in this course are especially good for careers with high-pressure decision-making.

Instructor Interview

"I think there are a lot of ways for students to extract the benefits from poker without the need to risk money…"
—Kevin Desmond

In the following pages, Kevin Desmond describes various aspects of how he taught 15.S50 Poker Theory and Analytics.

Curriculum Information

Prerequisites

Permission of the instructor

Requirements Satisfied

Graduate subject credit

Offered

Offered during IAP when an instructor is available to teach the course

Assessment

The students’ grades were based on the following activities:

  • 1/3 Gameplay: cash-in 1+ or play 10+ Daily Turbos
  • 1/3 Casework: pass all three cases
  • 1/3 Attendance: attend at least 8 of 11 lectures

Instructor Insights on Assesment

Assessment in this course was an interesting puzzle. Poker is inherently a game where the results are not perfectly correlated with skill, but it is not in line with MIT’s academic standards for course credit to be determined probabilistically. However, gameplay was an important learning tool and the students needed to be incentivized to play their best. The solution was to split into two leagues, one for credit, and one for prizes, where students would earn credit based on volume of play, but earn prizes based on success. This would ensure every student had the capacity to earn credit in reasonable time. In addition to gameplay, the homework assignments were key to assessing the students’ understanding and ability to implement concepts taught during lectures.

Student Information

Enrollment

About 150 students

Breakdown by Year

The breakdown of student in this course was about 60% graduate students, 40% undergrads.

Breakdown by Major

The graduate students were about 50% Management, 50% Engineering/Math/Science. The undergraduates were about evenly split among years and primarily from Engineering. A small number of students were visitors from other universities.

Typical Student Background

Students primarily had an interest in competitive games and practical applications of statistics. Some students intended to work in areas where poker knowledge was critical (trading or gaming industry), whereas other students had a more academic or recreational interest in poker. Most students had a strong background in statistics/math.

Ideal Class Size

The ideal class size is around 100-300 students. Since the class is heavily focused on gameplay, the goal is for students to be playing against a mix of familiar players and new opponents. By the end of the course, most students had a general level of familiarity of the different playing styles of students in the class.

How Student Time Was Spent

During an average week, students were expected to spend between 19.5 and 24.5 hours on the course, roughly divided as follows:

In Class

The majority of in-class time was spent reviewing core concepts and solving example questions. Lecture attendance was mandatory for enrolled students and optional for listeners.

Out of Class

Outside of the classroom, a significant amount of time was spent playing poker in the online league, hosted by PokerStars. The average player played approximately 5000 hands of poker (the equivalent of about a year’s worth of live play). In addition, each of the three homework assignments required about 1-2 hours of work. For some students, about 2 hours per week was spent in review sessions.

Course Team Roles

Instructor

Usually a graduate student with a background in professional poker.

Faculty Advisor

Faculty sponsor who generally oversees the course, but doesn’t necessarily have an active role in teaching sessions.

The following set of lecture notes were organized specifically for the OCW course site, and correspond to the Lecture Videos. You will also find each linked on the corresponding Lecture Notes tab in the video gallery.

LEC # TOPICS
1 Introduction (PDF - 1.3MB)
2 Analytical Techniques (PDF - 3.0MB) (Courtesy of Joel Fried. Used with permission.)
3 Basic Strategy (PDF - 3.3MB)
4 Pre-flop Analysis (PDF - 3.6MB)
5 Tournaments (PDF - 4.9MB)
6 Poker Economics (PDF - 2.6MB) (Courtesy of Aaron Brown. Used with permission.)
7 Game Theory (PDF) (Courtesy of Bill Chen. Used with permission.)
8 Decision Making (PDF - 1.8MB) (Courtesy of Matt Hawrilenko. Used with permission.)
  • PokerStars: Largest free-to-play online poker cardroom. Used for course home league.
  • Poker Tracker: Poker analytical software. Used for data analysis.
  • Universal Replayer: Java-based hand-history animator.
  • Poker Stove: Range vs Range equity calculator. Alternative to Poker Tracker for equity calculations.
  • ICM Calculator: Calculates payoff equity for tournament players. Alternative to Poker Tracker for ICM calculations.

Course Meeting Times

Lectures: 2–3 sessions / week for 4 weeks, 1.5 hours / session

Prerequisites

Permission of instructor.

Goals of the Course

  • Create an environment for the study of poker theory without the need for real-money wagering.
  • Develop the basic foundation for decision-making in poker.
  • Allow students to assess their own level of play and have a framework for improvement.
  • Provide an understanding of the current poker environment and how students might leverage talent for poker in the future.

Topics

  • Basic Strategy
  • Analysis Techniques and Applications
  • Preflop Analysis
  • Tournament Play
  • Poker Economics
  • Game Theory
  • Decision Making

Grading

Requirements

  • Gameplay - Cash in at least one or play at least 10 (Beginners Only) Daily Turbos on PokerStars
  • Casework - Pass all three cases
  • Attendance - Attend at least 6 of 8 lectures

Policy on Collaborations

You may interact with fellow students when preparing your homework solutions. However, you must write up solutions on your own. Duplicating a solution that someone else has written or providing solutions to be copied is not acceptable. If you do collaborate on homework, you must cite, in your written solution, your collaborators. If you use sources beyond the course materials in one of your solutions, you must also cite such sources.

Calendar

LEC # TOPICS KEY DATES
1 Introduction  
2 Analytical Techniques (Guest: Joel Fried)  
3 Basic Strategy Case 1 Out
4 Preflop Analysis Case 1 Due Case 2 Out 
5 Tournament Play Case 2 Due Case 3 Out
6 Poker Economics (Guest: Aaron Brown) Case 3 Due
7 Game Theory (Guest: Bill Chen)  
8 Decision Making (Guest: Matt Hawrilenko)  

PokerStars Home League

Pokerstars has created a play-money Home League for this MIT course for registered students. Each student must begin playing Beginners Only Turbos before moving on to Daily Turbos. Beginners Only Turbos pay out the top 20% and cashing in at least one of these tournaments (or alternatively playing in ten) is a requirement for passing the course.

Tournament Leaderboard Points

League rankings are based on tournament leaderboard points, which are paid out in a way similar to tournament prizes. With the exception of the Beginners Only Turbos, TLB points are paid out to the top 34% of players, according to the following formula:

\(   Points = if \:\: [ k > n * 0.34 ] \:\: then \:\: [ 0 ] \: \: else \:\: n * { { \sqrt{n} \over \sqrt{k} } \over { \sum_{x=1}^{int(.34n)} {\sqrt{n} \over \sqrt{x}} } } \)

n is the number of entrants

k is the place of finish (k=1 for the first-place finisher, and so on)

So first place in a 10-person turbo would earn you 4.38 TLB points, second would earn 3.10, and third would earn 2.52 for a total of 10.0 points awarded.

Daily Tournament Schedule

Note that the information listed below about the schedule is for archival purposes. The events in this schedule occurred in January 2015.

Starting Week 1

  • 6:00 PM - (Beginners Only) Daily Turbo
  • 8:00 PM - (Beginners Only) Daily Turbo
  • 8:00 PM - Daily Turbo
  • 10:00 PM - (Beginners Only) Daily Turbo

Starting Week 2

  • Sundays 2:00 PM - Weekly Major (Real Prizes)
  • 6:00 PM - Daily Turbo
  • 7:00 PM - Daily Turbo
  • 8:00 PM - Daily Turbo
  • 9:00 PM - Daily Hyper-Turbo

Course Info

Instructor
As Taught In
January IAP 2015
Level
Learning Resource Types
Lecture Videos
Problem Sets with Solutions
Lecture Notes
Instructor Insights