Python for Finance, 2nd Edition

Chapter 16. Automated Trading

People worry that computers will get too smart and take over the world, but the real problem is that they’re too stupid and they’ve already taken over the world.

Pedro Domingos

“Now what?” one might think. A trading platform is available that allows one to retrieve historical data and streaming data, to place buy and sell orders, and to check the account status. A number of different methods have been introduced to derive algorithmic trading strategies by predicting the direction of market price movements. How can this all be put together to work in automated fashion? This question cannot be answered in any generality. However, this chapter addresses a number of topics that are important in this context. The chapter assumes that a single automated algorithmic trading strategy only shall be deployed. This simplifies, among others, aspects like capital and risk management.

The chapter covers the following topics:

“Capital Management”

As this section demonstrates, depending on the strategy characteristics and the trading capital available, the Kelly criterion helps with sizing the trades.

“ML-Based Trading Strategy”

To gain confidence in an algorithmic trading strategy, the strategy needs to be backtested thoroughly both with regard to performance and risk characteristics; the example strategy used is based on a classification algorithm from machine learning as introduced in Chapter 15.

“Online Algorithm”

To deploy the algorithmic trading strategy ...