So far, we've gone over all the concepts and ideas involved in algorithmic trading. We went from introducing the different components and players of an algorithmic trading ecosystem to going over practical examples of trading signals, adding predictive analytics into algorithmic trading strategies, and actually building several commonly used basic, as well as sophisticated, trading strategies. We also developed ideas and a system to control risk and manage it over the evolution of a trading strategy. And finally, we went over the infrastructure components required to run these trading strategies as well as the simulator/backtesting research environment required to analyze trading strategy behavior. At this point in the book, you should be able to successfully develop a deep understanding of all the components and sophistication needed to build, improve, and safely deploy all components of an algorithmic trading strategy business stack.

The goal in this final section of the book is to begin to look beyond the deployment and operation of algorithmic trading strategies by considering things that can go wrong in live markets or slowly deteriorate as time passes, by trading signal edges vanish, and how new market participants are added, or more informed participants join the market and less informed participants leave. Financial markets and market participants are in a constant state of evolution, so algorithmic trading businesses that are able to evolve over time and in the face of changing market conditions, adapt to new conditions, and continue to be profitable, are the only ones that can survive long term. This is an extremely challenging problem to tackle, but in this chapter, we will go over the hurdles we typically encounter and offer some guidance on how to tackle them. We will discuss why strategies do not perform as expected when deployed in live trading markets – and show examples of how to address those issues in the strategies themselves or the underlying assumptions. We will also discuss why strategies that are performing well slowly deteriorate in performance, and then we'll look at some simple examples to explain how to address these.

In this chapter, we will cover the following topics: 

• Strategy performance in backtester versus live markets
• Continued profitability in algorithmic trading