内部课程｜剑桥大学TheQuantHackers和北大量化交易协会内部培训项目全面启动

Internal courses｜Cambridge University TheQuantHackers and Beijing Quantitative Trading Association internal training projects are fully launched

<!-- @@@ -->

The internal training program jointly organized by Cambridge University TheQuantHackers and Peking University Quantitative Trading Association** has now started! This internal training is designed to help students who are interested in quantitative trading, machine learning, and derivatives learn and practice in depth. Whether you are new to quantitative trading or an experienced researcher, there is training content here for you!

About TheQuantHackers

TheQuantHackers originated from the University of Cambridge and is a cutting-edge community composed of a group of practitioners from the world's top universities and quantitative finance. It is committed to exploring the infinite possibilities in the fields of artificial intelligence (AI) and quantification (Quant). Our team members come from world-renowned universities such as the University of Cambridge, Oxford University, Harvard University, ETH Zurich, Peking University, and the University of Science and Technology of China. Our external consultants come from first-class companies such as Microsoft, HSBC, Jump Trading, Man Group, and top domestic private equity companies.

In-house training course partners

<!-- @@@ -->

The Peking University Quantitative Trading Association (QTA) is an academic society with the theme of quantitative trading. It is also one of the most influential quantitative societies in Peking University.

Provide support

<!-- @@@ -->

Cambridge University Algorithmic Trading Society (CUATS) is the most influential quantitative trading society in Cambridge, with partners including Jane Street, Citadel, Optiver, SIG, DRW, etc.

Project Highlights

We will deeply explore the key theories and applications in quantitative trading from multiple directions such as multi-factor models, machine learning, and derivatives. Each module will be led by a team of experienced tutors, combining theoretical learning with practical projects to help you achieve maximum improvement in the shortest time. At the end of the project, three practical projects from industry tutors will be provided to help participating students understand the daily work in quantitative work.

Multi-factor model direction

Introduction

Multi-factor models are the basis of quantitative trading research. The main goal of this module is to help community members establish an in-depth understanding of the market structure, become familiar with the process of multi-factor research, and design and optimize factors independently.

Course content

1. Theoretical Learning: Read "Factor Investment", the latest research reports and academic papers, and build an overall understanding of the multi-factor model.
2. Practical Research: Master the entire process of factor research, covering factor reproduction, construction and improvement. The tutor will provide specific research topics to help you discover high performance factors.

Whether you are a beginner or an advanced researcher, our team of mentors will help you grow quickly and delve deeper into the field of factor research.

Machine learning direction

Introduction

This module focuses on the application of machine learning in quantitative trading, combining the latest machine learning technology with financial data analysis to help you explore the potential of machine learning in unstructured data processing.

Course content

1. Machine Learning Pairs Trading Strategy

This strategy improves on traditional pairs trading by using machine learning to pair stocks with similar stock returns and company characteristics. By discovering these stock pairs and betting on mean reversion, the strategy profits when stocks deviate from historical relationships. The best stock pairs are found using agglomerative clustering and rebalanced monthly.

• Assets: NYSE, AMEX, NASDAQ listed common stocks
• Trading Logic: Based on a group of stocks with similar characteristics and past returns, go long underperforming stocks and short outperforming stocks.
• Fundamentals: Stocks that historically move together are expected to revert to the mean when they deviate. Incorporating company characteristics can improve stock pair selection and thus returns.
• Balancing Period: Monthly

2. Predicting stock returns through machine learning

This strategy does not rely on a single return estimate but rather predicts a stock's entire return distribution. Help traders make better decisions by estimating the likelihood that a company will beat or miss analyst expectations. Make accurate predictions using non-parametric machine learning models.

• Assets: US listed stocks
• Trading logic: Go long on companies that are expected to have a high probability of exceeding expected returns, and short on companies with low expected returns.
• Rationale: Forecasts of earnings distributions provide better insight into future stock performance, especially in the event of earnings surprises.
• Equilibrium Period: Quarterly (around earnings announcements)

3. Use machine learning to predict option returns

This strategy uses machine learning to predict option returns, identify mispriced options through large data sets that include factors such as liquidity, volatility, and more, and create long-short portfolios based on predicted returns.

• Asset: US Stock Options
• Trading Logic: Long options with strong predicted returns and short options with weak predicted returns based on a broad set of characteristics including implied volatility and liquidity.
• Basic Principle: By analyzing large amounts of option and stock characteristic data, machine learning can discover option pricing errors and arbitrage opportunities.
• Balance cycle: monthly (dynamically adjusted according to market conditions)

4. Use machine learning to construct factor portfolios

The strategy uses machine learning to build a portfolio that outperforms traditional factor models by analyzing hundreds of stock characteristics. A key advantage is the ability to dynamically adapt to changes in market conditions, particularly at different stages of the credit cycle.

• Assets: US listed stocks
• Trading Logic: Based on the machine learning model, identify the dominant factors that affect returns, the winners of long predictions, and the losers of short predictions.
• Rationale: The drivers of stock returns change over time, and machine learning helps capture these changes, producing better portfolio performance than static factor models.
• Balancing Period: Monthly

5. Use machine learning to assess company quality

This strategy uses machine learning to more accurately assess a company's quality, helping investors and stakeholders make better decisions by analyzing financial data and discovering quality companies that are likely to perform well in the future.

• Assets: US listed stocks
• Trading Logic: Go long on undervalued high-quality companies based on company quality scores predicted by machine learning.
• Rationale: Machine learning can more accurately assess company quality by analyzing granular financial data to better predict future performance.
• Balancing Period: Every year

6. Use machine learning to predict intraday returns

This strategy aims to predict the direction of intraday stock returns using an LSTM model. By looking at past intraday prices, the model can make profitable trades during the trading day, but recent backtests suggest improvements are still needed.

• Asset: S&P 500 Stocks
• Trading logic: Use LSTM to predict, go long on the top 10 stocks that may have the best performance during the day, and short the last 10 stocks that may have the worst performance.
• Fundamentals: Past intraday stock price movements help predict future intraday returns, providing profitable short-term trading opportunities.
• Balance Period: Intraday

This direction is suitable for students who are interested in machine learning, especially those who want to learn in depth through programming practice.

Derivatives direction

Introduction

If you are interested in options and other derivatives and want to apply them to your personal trading strategy, this module is for you. We will take you through an in-depth understanding of option pricing models and hedging strategies.

Course content

1. Theoretical Learning:

• Option basics: BS model and its applications
• Dynamic analysis of Greek letters
• Volatility surface and stochastic volatility model
• Pricing implementation of exotic options

2. Practice Project: Combine numerical methods and derivatives strategies to build a backtesting system and conduct optimization testing.

This direction is especially suitable for students who are interested in derivatives trading and complex financial instruments. It is better to have a certain programming foundation.

---

TheQuantHackers practice project mini-Project

1. Option Pricing

Project introduction

This project provides an in-depth look at option pricing models and their application in hedging strategies. The project includes programming and mathematical exercises, aiming to help Cambridge TheQuantHackers and Peking University QTA students understand different methods and practical applications of option pricing.

Main content

1. Black-Scholes Option Pricing Model

• Includes Itô calculus exercises

2. Discretization of stochastic differential equations and Monte Carlo option pricing

• Use numerical methods to simulate option prices

3. Dynamic hedging strategy based on market sensitivity

• Applications including delta, gamma, vega

4. Dupire Model and Heston Model

• Two important volatility models

5. Option pricing model based on fast Fourier transform

• Carr and Madan models

6. Introduction to Complex Options

• Such as barrier options, etc.

7. Practical Examples of Algorithmic Trading Strategies

2. Limit Order Books

Project introduction

This project focuses on limit order books and their operation within the market microstructure, focusing on how exchanges match orders and the implementation of various algorithms and order types. This project focuses on programming exercises and less mathematics content.

Main content

1. Concept of Limit Order Book

• Covers limit orders and market orders
• Implement order book based on price-time algorithm
• Discuss order cancellation and market inducement strategies (such as market deception)

2. Matching Algorithm

• Implemented proportional matching algorithm and expanded to include top order and liquidity providers (LMM)
• Add other order types (market order, stop loss order, iceberg order, etc.)

3. Current fair price of the asset

• Calculate mid-price and weighted versions
• Micro prices based on Markov models
• Heuristics using secondary market data

4. Extension Suggestions

• Extended research using deep learning to predict future order book states and conduct trading

3. Implied Volatility and Volatility Surfaces project (Implied Volatility and Volatility Surfaces)

Project introduction

This project focuses on how to calculate implied volatility, draw volatility surfaces, and apply them to trading strategies. The project contains a lot of programming and mathematics exercises. Cambridge TheQuantHackers-Peking University QTA students will learn how to process actual market data and analyze volatility.

Main content

1. Volatility Surface Concept -Introduction to the Black-Scholes pricing model

• The definition and existence conditions of implied volatility

2. Get option chain data

• Use the yfinance library to obtain option chain data
• Organize option chain information into pandas data frame

3. Calculate Implied Volatility

• Analytical solution of implied volatility under special circumstances
• Calculate implied volatility using bisection method and Newton-Raphson method
• Scaling using the Brent method

4. Interpolation and drawing volatility surface

• Ensure that the volatility surface is arbitrage-free
• Draw surfaces using linear interpolation and cubic spline interpolation

5. Characteristics of Volatility Surface

• Discussion of Volatility Skew, Smile and Term Structure
• Typical shapes of volatility surfaces in various markets

6. Extensions to Practical Applications

• Straddle options trading in the Forex market
• Predicting option returns through volatility term structure

---

Registration method

This internal training is open to members of the Cambridge TheQuantHackers community and the Peking University Quantitative Trading Association. No extensive quantitative experience is required. The TheQuantHackers community is currently open to the public. To participate in future internal courses and internal lectures, you can scan the following QR code to apply to become an TheQuantHackers member. Attach your resume to get the application results faster~

Scan the QR code to sign up now!

---

北京大学量化交易协会(QTA)是一个以量化交易为主题的学术型社团，也是北京大学校内最有影响力的量化社团之一。

提供支持

剑桥大学TheQuantHackers和北京大学量化交易协会联合举办的内部培训项目现已启动！本次内培旨在帮助对量化交易、机器学习、以及衍生品感兴趣的同学深入学习与实践。无论你是刚接触量化交易的新手，还是有丰富经验的研究者，这里都有适合你的培训内容！

关于TheQuantHackers

TheQuantHackers起源于剑桥大学校内，是由一群来自世界顶尖高校和量化金融从业人员组成的前沿社区，致力于探索人工智能（AI）与量化（Quant）领域的无限可能。我们的团队成员来自剑桥大学、牛津大学、哈佛大学、苏黎世联邦理工学院、北京大学、中科大等世界知名高校，外部顾问来自Microsoft、HSBC、Jump Trading、Man Group、国内顶尖私募等一流企业。

内培课程合作伙伴