37
of the system. We have, though, already started the description process—the description
created for the Money Document will serve as a starting point. Over the iterative research
process, planning each loop will require team members to define goals and set boundaries
for that research.3.7.2. Research Quantitative Methods (Chapter 9)
Very rarely do we dream up completely new trading/investment ideas. Rather, we build
on ideas of the past and add new twists. Many times new trading strategies are essentially
copies of old ones that proved to be successful.
The next step is to research and benchmark the relevant mathematical and logical
models, which may include deriving proprietary algorithms and/or applying publicly
available research from journals, books, the Internet, or white papers. The research proc-
ess may also include gaining an understanding of the trade selection and execution meth-
odologies of other successful systems. The goal of research is to speed and refine our
path to the best algorithms, which will form the basis of a Business Rules Catalog for the
system. Building and maintaining a proprietary library of unique quantitative methods is
key to the long-term success of a firm.3.7.3. Prototype in Modeling Software (Chapter 10)
As we have said, despite its shortcomings, Excel as well as other modeling software (such
as Resolver, MATLAB, Mathematica, or Zack ’ s) are rapid development environments for
testing trade selection and execution algorithms. The goals of prototyping are to quickly
build several generations in order to evaluate whether a particular idea warrants further
investigation and to promote risk-based iterative development, where the hardest pieces
of the project are investigated first.
Currently, many trading/investment ideas are immediately built as working systems if
confidence is high that the trader and/or the system will make money. This type of devel-
opment often proceeds without proper definition prior to implementation despite under-
stood software risks. Prototyping algorithms enables us to clearly define algorithms, GUI
and data requirements, and develop a baseline application for regression testing. These
prototypes will form a foundation of the Technology Requirements Specification docu-
ment in later stages. Baseline regression with prototypes enhances defect detection and
removal more than any other strategy.3.7.4. Check Performance (Chapter 11)
We will use prototypes as a starting point in the discussion of how we are going to meas-
ure the performance of the system. Without a clear plan on how to measure performance
and define success and the variability of the trading system, we cannot proceed to the
next stage.
We define three types of trading/investment systems and each type will require a
unique mechanism for performance testing. The three foundation types are:1. Trigger systems. A good example of trigger trading/investment systems are ones
built on technical or valuation indicators.3.7. DESIGN AND DOCUMENT STRATEGY