Prediction market is emerging as a new type of financial market, with total trading volume reaching $50.25 billion in 2025. By using financial incentives to turn the ‘wisdom of crowds’ into measurable forecasts, such market have shown distinctive value in areas such as sports, political elections, and economic indicators. As a result, prediction market now represents an important frontier in financial innovation and data-driven forecasting.
The purpose of this course is to introduce students to prediction markets and their role in aggregating information and forecasting future events. Students will learn to leverage quantitative methods to interpret market data, build models, and make informed predictions across a variety of domains. Topics include using option chain to estimate the duration of Iran war, modeling and forecasting sports teams and game outcomes, analyzing information diffusion in social networks for presidential election predictions, and predicting movie box office performance using social media signals and movie script analysis. By exploring these diverse applications, students gain hands-on experience in combining data science, finance, economics, and computational modeling to generate actionable forecasts from complex, real-world data.
There are mainly two sections of this course:
The first section introduces the fundamentals of prediction markets, including betting mechanisms, blockchain-based smart contracts as implemented in platforms like Polymarket, and the structure and function of order books. This section also examines prediction markets from a financial perspective, covering the concepts and pricing methods of binary options, portfolio optimization for binary options, and risk management strategies.
The second section explores a range of real-world prediction applications, including estimating the implied probability of the closure of the Hormuz Strait, sports game outcomes, movie box office performance, presidential election predictions, and key economic indicators. Students will apply a variety of quantitative and computational tools, such as prompt engineering for generative AI, complex network analysis, logistic regression, and machine learning algorithms, to build and evaluate predictive models.
Prerequisites
No prior background in mathematics or statistics is required.
Preferable math knowledge: Linear algebra (such as Math 216, 218, or 221), probability (such as Math/Stat 230, Math 340/Stat 231, or Math 238L/ECON 238L)
Programming Skills:
Preferred but not required.
Class assignments and projects will require students to execute their ideas and models into betting strategies, and these are accomplished through certain programming skills or prompt engineering using generative AI tools.
Learning Objectives
Upon successful completion of this course, students will be able to:
Objective 1: Understanding of the Prediction Market
- Develop an understanding of the mechanisms, underlying technology, and economic principles of prediction markets.
- Learn key concepts of risk management and how to apply portfolio theory in prediction markets.
Objective 2: Quantitative Analysis
- Design and execute quantitative betting strategies across several topics.
- Implement, compare and evaluate prediction approaches used in computational social science and other domains.
Objective 3: AI application
- Apply prompt engineering with generative AI tools for information gathering, analysis, and forecasting of specific events.
Course Materials
Primary Course Materials – The primary course materials will be based on lecture notes written by the instructor. There will be assigned reading materials for each lecture to help students dive deeper into specific topics and concepts.
Secondary Reference Materials — The instructor will provide research papers, research reports, websites and other materials for students to read. For example,
- Stacked ensemble model for NBA game outcome prediction analysis:
https://www.nature.com/articles/s41598-025-13657-1
Course Project
The course project is the primary component of students’ final grade and is designed to synthesize the knowledge and skills developed throughout the course. Students will work in groups of 3-5 to design, implement, and evaluate a quantitative betting strategy. Each group will progress through the full development cycle: drafting the idea, constructing quantitative models, coding/prompt engineering, and evaluating, refining their approach based on empirical results, and presenting a comprehensive report.
Evaluation will consider the rigor of the scientific logic, the appropriate application of quantitative tools, and the clarity of the analysis and presentation. While strong performance metrics are valued, the grading emphasizes methodological soundness, critical thinking, and the ability to justify decisions. Prediction market is a new and challenging domain. Even if outcomes do not match expectations, students can still achieve high scores through rigorous analysis, critical thinking, and thorough testing of models and strategies.
Performance Evaluation
Student grades are primarily based on projects and assignments. Projects assess students’ ability to design, implement, and empirically evaluate quantitative betting strategies, while assignments are to guide students for implementing the methods in given research papers.
| Course Element | Percentage of Grade |
|---|---|
| Project | 60% |
| Assignment | 30% |
| Attendance | 10% |
General Course Outline
Section 1 Introduction to Prediction Market
- Introduce prediction market and its basic structure, including smart contract mechanisms, order books, and pricing methods.
- This session will also cover key concepts such as the efficient market hypothesis, hedging, arbitrage opportunities, and the gambler’s ruin problem in prediction market.
Section 2 Hidden Information in Financial Market and Prediction Market: Iran War as an example
- A brief introduction to financial derivatives, including futures and options, along with the fundamental logic of their pricing.
- Using the Iran war as an example, explain how to estimate the expected duration of the conflict and the implied probability of closure of the Strait of Hormuz by extracting hidden information from oil futures and options markets.
Section 3 Binary Option Pricing and Cryptocurrency Binary Betting
Introduce the fundamentals of binary options and their pricing methods, including the Black–Scholes framework and tree models. This lecture will also explore quantitative approaches for short-term cryptocurrency price prediction.
Section 4 Risk Management and Binary Option Portfolio Optimization
Cover portfolio optimization and bet-sizing methods for binary outcomes and risk management, including the Kelly criterion, the mean–variance framework, and return–entropy approaches.
Section 5 Overview of Computational Social Science
Provide a brief theoretical overview of computational social science, covering core concepts from probability theory, complex networks, and game theory.
Section 6 Complex Network and Sport Team Fundamentals
Concepts of complex network, network analysis for Sport teams
Section 7 Sport Game Prediction
Game prediction models (Bayesian inference, logistic regression, AI)
Section 8 Prompt Engineering for Prediction Market
Provide a guide for using most popular AI tools such as ChatGPT to collect, analyze data for prediction market.
Section 9 Movie Box Office Prediction
- Hidden information in movie scripts for box office prediction
- Factors for box office prediction such as budget, genre, stars, social media, etc.
Section 10 Presidential Election Prediction
Epidemic and information spreading models (SIR model), network analysis of social media, presidential election prediction via social media.
Section 11 Economic Prediction
- Illustrate the concepts of inflation, CPI, PPI, M1, M2, etc.
- Analytical framework for inflation