One of the most pervasive mistakes in financial research is to take some data, run it through an ML algorithm, backtest the predictions, and repeat the sequence until a nice-looking backtest shows up. Academic journals are filled with such pseudo-discoveries, and even large hedge funds constantly fall into this trap. It does not matter if the backtest is a walk-forward out-of-sample. The fact that we are repeating a test over and over on the same data will likely lead to a false discovery. This methodological error is so notorious among statisticians that they consider it scientific fraud, and the American Statistical Association warns against it in its ethical guidelines (American Statistical Association , Discussion #4). It typically takes about 20 such iterations to discover a (false) investment strategy subject to the standard significance level (false positive rate) of 5%. In this chapter we will explore why such an approach is a waste of time and money, and how feature importance offers an alternative.
8.2 The Importance of Feature Importance
A striking facet of the financial industry is that so many very seasoned portfolio managers (including many with a quantitative background) do not realize how easy it is to overfit a backtest. How to backtest properly is not the subject of this chapter; we will address that extremely important topic in Chapters 11–15. The goal of this chapter is to explain one of the analyses that ...