Quantitative Trading Systems by Howard Bandy is the type of book that begs the question: Just how serious do you want to get about this? If the answer is “really serious,” this is a book that needs to be on your shelf. Howard manages to compress a great deal of practical knowledge and experimentation into just over 300 pages of densely packed text. If you don’t already have a basic grasp of systems and terminology, you probably find yourself referring to the glossary and/or an outside source for the explanation; if you do have that basic grasp, prepare to have your horizons broadened. Howard sent me a copy to review, and I thoroughly enjoyed it.

There are two gripes that always pop up when dealing with books of this type, and that they appear in this review isn’t a sign of any shortcoming in Howard’s work; they’re unavoidable. I have some semantic quibbles and minor points of disagreement, but knowing me, that’s also probably unavoidable.

Quantitative Trading Systems sometimes comes across as an AmiBroker user’s guide. This is the software used to describe the trading algorithms and to test the results, so some level of live examples is to be expected. I suppose that a book like this could be written entirely in pseudo code, but whom would that help? Like it or not, every book on trading systems is held hostage by the software used to test, to whatever extent the book digs into the results.

The book focuses exclusively on technical analysis. I use “technical analysis” in the “old school” sense, meaning that it is confined to data derived from price and volume movement, including indicators build on the same. Unfortunately, these books are difficult to write for fundatechnical or fundamental strategies, for a variety of reasons. Aside from a lack of cheap and broadly available historical data, the historical data is subject to revision, which opens up another can of worms. Indeed, Howard consciously chooses to list reasons for avoiding fundatechnical or fundamental strategies, as opposed to the typical pattern of these books’ avoiding those methods because they’re following the “path of least resistance.” Because of the focus on technical and statistically analyzable strategies, there is a general slant towards systems that trade frequently.

The good news, and there is plenty of it, comes from the completeness and execution of the analysis and presentation. After laying out the five questions of any trading system over several chapters, Howard dives into explanations of the different types of systems. Covered briefly, he includes trend-following, mean-reversion, seasonal, pattern, sector, and rotation systems. He also includes a discussion of portfolios, as well as filters and timing systems.

It should be noted that the author is not trying to provide fish, but is, instead, describing how to fish, starting with selecting the cane to make a pole from. Some of the example systems strike me as “straw men” used to illustrate the principle behind the type of system without providing an actual, working system. The “pattern” chapter comes to mind, here, as many visual chart patterns can be described in code and several simple patterns can be profitably traded, but a single, simple, and non-profitable, example is used by Howard to describe the genre. This is a starting point for designing your own systems, not a book of systems to be bought and used. It also is not a taxonomy of systems inside each category: for that, you have to do your own research.

The remainder of the book fleshes out some other necessities of any well-rounded discussion of quant trading, including a discussion of testing and development. In-sample versus out-of-sample data is discussed, as are statistical tests for validity of the results. Walk-forward and Monte Carlo testing chapters are included.

This is very frankly the best book on trading that I’ve read this year, and one of the best I’ve ever read on the subject of quant trading systems. As a math nerd with an interest in system trading, I’m squarely in the book’s target audience, but if you have interest in the subject but lack the background in math and statistics, I would still strongly urge you to get this book. With time and a little effort, almost anyone can develop enough background knowledge to understand and apply the principles, even if they don’t use the same software used by the author. It all depends on how serious you want to get about this.