Python for Algorithmic Trading

Chapter 4. Mastering Vectorized Backtesting

[T]hey were silly enough to think you can look at the past to predict the future.¹

The Economist

Developing ideas and hypotheses for an algorithmic trading program is generally the more creative and sometimes even fun part in the preparation stage. Thoroughly testing them is generally the more technical and time consuming part. This chapter is about the vectorized backtesting of different algorithmic trading strategies. It covers the following types of strategies (refer also to “Trading Strategies”):

Simple moving averages (SMA) based strategies

The basic idea of SMA usage for buy and sell signal generation is already decades old. SMAs are a major tool in the so-called technical analysis of stock prices. A signal is derived, for example, when an SMA defined on a shorter time window—say 42 days—crosses an SMA defined on a longer time window—say 252 days.

Momentum strategies

These are strategies that are based on the hypothesis that recent performance will persist for some additional time. For example, a stock that is downward trending is assumed to do so for longer, which is why such a stock is to be shorted.

Mean-reversion strategies

The reasoning behind mean-reversion strategies is that stock prices or prices of other financial instruments tend to revert to some mean level or to some trend level when they have deviated too much from such levels.

The chapter proceeds as follows. “Making Use of Vectorization” introduces vectorization ...