Inside the Black Box: A Simple Guide to Quantitative and High Frequency Trading, 2nd Edition

CHAPTER 3

Alpha Models: How Quants Make Money

Prediction is very difficult, especially about the future.

—Niels Bohr

Having surveyed it from the outside, we begin our journey through the black box by understanding the heart of the actual trading systems that quants use. This first piece of a quant trading system is its alpha model, which is the part of the model that is looking to make money and is where much of the research process is focused. Alpha, the spelled-out version of the Greek letter α, generally is used as a way to quantify the skill of an investor or the return she delivers independently of the moves in the broader market. By conventional definition, alpha is the portion of the investor's return not due to the market benchmark, or, in other words, the value added (or lost) solely because of the manager. The portion of the return that can be attributed to market factors is then referred to as beta. For instance, if a manager is up 12 percent and her respective benchmark is up 10 percent, a quick back-of-the-envelope analysis would show that her alpha, or value added, is +2 percent (this assumes that the beta of her portfolio was exactly 1). The flaw with this approach to computing alpha is that it could be a result of luck, or it could be because of skill. Obviously, any trader will be interested in making skill the dominant driver of the difference between her returns and the benchmark's. Alpha models are merely a systematic approach to adding skill to the investment ...