Microstructural Features for Market Making
Volatility, Imbalance, Liquidity and more!
“Garbage in - Garbage out”
You’ve probably heard this a lot in the context of model building.
This is the case not only for machine learning models / predictive models but also for market making models and others that don’t directly try to make a prediction.
When it comes to microstructural features I distinguish between a couple of different categories:
Real-time: Only uses the current state of the market.
Historical: Uses historical data.
Trades-based: Uses only trades data.
Order book-based: Uses only order book data.
Hybrid: Uses both trades and order book data.
This article serves as an overview of different metrics that fall into those categories.
You can come up with infinitely many such metrics. You should use what let’s you achieve the task at hand.
Table of Content
Real-time Trades-based
Historical Trades-based
Order book Data Structure
Real-time Order book-based
Historical Order book-based
Historical Hybrid
Microstructural Volatility
Final Remarks