Statistical Arbitrage

This two day workshop introduces delegates to statistical arbitrage strategies, including pairs trading, with particular reference to research, testing and implementation. Relevant software (MATLAB) will be used throughout the workshop to illustrate examples and to help students practice the essential steps in developing a stat arb strategy. No prior knowledge of MATLAB is required. (Note: Students will be able to apply the principles learnt during the workshop, regardless of which software they choose to use thereafter).

March 7 to March 8, 2016
Duration: Two days (9.00am to 5.00pm)
Location: The Tower Hotel – London, UK
Trainer: Ernest Chan
Course fee: £1890 + VAT – Register online

Course Outline


+ The different types of statistical arbitrage strategies
+ Stationarity, cointegration, mean reversion, and momentum

The Essentials of MATLAB

+ The pros and cons of using MATLAB
+ Quick survey of syntax
+ Exercises: building some useful utilities for trading research

Directional Trading

+ Concept of stationarity, and why it is useful
+ Statistical test for stationarity: adf
+ Exercise: Testing for stationarity
+ Testing for mean-reversion: computing half-life based on Ornstein-Uhlenbeck formula
+ Why is computing half-life better than computing average holding period?
+ Exercise: Computing the half-life of mean-reversion
+ Exercise: Backtesting a Bollinger Band strategy for AUDCAD and EURCHF

Pairs and Triplets Trading

+ Concept of cointegration and why it is useful
+ How is cointegration different from correlation?
+ Statistical tests for cointegration: cadf and Johansen
+ Exercise: Find out if GLD-GDX is cointegrating
+ Finding the best hedge ratio
+ Exercise: Backtesting a Bollinger Band pairs strategy
+ Trading cointegrated triplets
+ Exercise: Backtesting a Bollinger Band triplet strategy
+ What are the best markets to pairs trade?

Index Arbitrage

+ Trading an ETF against a basket of its component stocks
+ Two ways of constructing a basket
+ Exercise: Backtesting an index arbitrage trading model

Long-Short Portfolio

+ Ranking stocks in an index based on various simple returns criteria
+ How minor variations in strategies can produce big differences in returns
+ Important biases and pitfalls in backtesting long-short portfolio strategies
+ Exercise: Backtesting variations of a long-short portfolio strategy