Live 12-hour course covering the entire pipeline of algorithmic trading - strategies including both risk premia and advanced strategies, including research and development methodology, and the gritty details including data sources, databases, class structure from signals to strategies to portfolios to optimizers to back-testers.
Numerous examples of strategies in futures will be gone through in detail, while strategies in equities, fixed income, and FX will be described and referenced.
What am I going to get from this course?
- Understand sources of risk premia strategies, i.e., CTA and bank Quant investment strategy returns
- Understand more general quant strategies, the sources of risk, and basic trading strategies in each investment class.
- Assemble the necessary components from the data sources/APIs, to databases to back-testers to portfolio optimizers and risk management tools into a quant investment system
- Assemble appropriate class structure for algorithmic trading optimization and portfolio construction.
- Understand the place for ML methods and models, how to use them and steps to take to avoid over-fitting.
- Know how to do research into algo trading - the steps to take to continually research and develop new strategies and alpha sources.
- Know how to structure your code, create appropriate classes, sub/meta-classes, etc, to reduce the chance of spaghetti-code, help you organize the both the code-base and your thoughts on it.
Background - Types of Algo Trading - The Industry
Overview of some of the flows, some of the styles and some of the firms dominating this fast-growing area.
The Pipeline - From Beginning to End
Process, Flow and Architecture will be touched on here, and discussed in more detail in the later sections. We will try to discuss the many different variants of how strategies are formed and how we can accommodate most of the variations with the appropriate architecture.
Module 2: Signals and Strategies
More Advanced Signals
i) Volatility Risk Premium – Short gamma, long vega, forward vol
ii) Relative Value/Pairs/StatsArb – Mean Reversion
iv) Equities Risk Premia / Smart beta
v). Event Driven
vi) Other, exogenous features
Endless methods for combining simple features are always popping up. We will look into some simple but flexible strategies in some detail, allowing exogenous features or more complex times-series models to be used. We will discuss ML approaches and the extent to whether they may add anything to this process, depending on the asset class and the depth of the feature-space.
Module 3: Portfolio Strategies and Portfolio Optimization
Module 4: Overfitting & Robustness
Rule of Thumb vs More Exact Methods
We will discuss the prevalance of overfitting in both the Sciences and in Finance. Rules of thumb as a means of remaining parsimonious are altogther a reasonable approach to this area. We put these in context.
Multiple Hypothesis Testing Methods
i) Adjusted SRs
ii) Reality Test and similar bootstrap methods
Covariance Penalties (AIC/BIC/Mallows Cp) and Cross Validation
Process and Pipelines
i) Python setups – Venvs & Packages
(1) Data Sources/APIs
(2) Databases and DB technologies
iii) SecMaster & Ongoing Product Research
i) Creating Tradable Histories
ii) Structuring the Project : Instruments, Signals, Strategies, Portfolios and Optimization methods
ii) Signals, Transformations, Scalings. Signal Combination - Strategies
iii) Portfolio Optimization
iv) Performance Measures and Reporting
v) Risk & Attribution