ABSTRACT
We present trading algorithms based on statistical analyses of market data. These analyses are also guided by established principles in financial economics. The market is assumed to consist of informed and noise traders. It is also postulated that the market is efficient, and the informed traders generally gain at the expense of noise traders. Any information about a stock is quickly impounded in price by the Efficient Market Hypothesis (EMH). If there are anomalies, statistical arbitrage techniques can be employed to successfully detect and exploit them for profitable opportunities. We present some commonly used strategies based on the statistical methods described in Chapters 2–4. Although the literature is replete with numerous strategies, most of them are some variations or combinations of basic strategies described here. As price-based strategies have been more or less fully exploited, it is important to consider strategies that use other information such as volume, market volatility, related stock performance, etc. Also presented are improved strategies based on machine learning methods.
