There are a few good reasons to build graybox systems and some not-to-so good reasons. One not-so-good reason might be if you really haven’t developed the skills required to develop profitable systems. But, if you can develop profitable systematic systems then why should you still be interested in graybox systems?
First, if you have a strong market read and strong discretionary trading ability then if you going fully systematic means giving up a unique advance. On the other hand, a purely discretionary trader is taking a significant and unnecessary handicap. The problem is not if you should have some quantitative/systematic mix but what the proper blend is and how to achieve it.
One solution is to treat the enterprises as separate endeavors. And, in fact, for the most part my systematic strategies tend to trade infrequently and do not use similar strategies to my discretionary methods. It works. But, it results in a disconnect. They aren’t providing synergistic benefit. Traditional systems are not well suited to discretionary integration because of the lower trade frequency. Absolutely, if a profitable and proven system can be part of your arsenal but it is not a means toward self-improvement.
In order to achieve true synergistic benefit, we need a new type of system that can aide the discretionary trader on a day-to-day basis. Traditional systems fall flat in those regards. Beyond the infrequency, the best graybox systems should encapsulate our “best plays” and not simply be the result of quantitative methods.
One key problem is the difficulty in developing systems that can trade as frequently as we need. As such, we may need to alter our normal perspectives on what a system is and the performance requirements to meet. Indeed, there are a lot of possibilities, some which may work great or not at all, once we go beyond traditional backtesting. The possibilities are exciting. The canvas is wide open which presents the first serious problem because the structure doesn’t exist anywhere else and must be created.
As a graybox trader, you are paving the way and building something completely unique. We can imagine whole classes of graybox systems that work in very different sorts of ways. For example, automatons or tactical strategies might do extremely well in certain market conditions but not provide an “edge” as we traditionally define; instead relying on the discretion of the trader to turn them on and off. “Best play” systems on the other might hand might be composite systems that seek to encapsulate and systemize our best trades that make sense to us. Finally, predictive systems seek to enhance our market read by surfacing the most predictive information for utilization in our decision making process.
As you can see there are many types of “graybox systems” with very different designs and objective and each can provide unique source of advantage. The hardest step is the first step. The goal shouldn’t be to develop the perfect solution because no such solution exists and we need some experience to determine what works best for us. The goal should rather to just build something to start with and build on it and add too it based on actual experiences.
The question is: what type of graybox do you want to create? Do you trade a graybox system that you developed? How does it work for you? Let me know by direct contact using the link to the right or drop a comment below.
Curtis is passionate about markets. He has developed top ranked futures strategies. His core focus is (1) applying machine learning and developing systematic strategies, and (2) solving the toughest problems of discretionary trading by applying quantitative tools, machine learning, and performance discipline. You can contact him at email@example.com.
Please log in again. The login page will open in a new tab. After logging in you can close it and return to this page.