It is difficult to imagine a singular rules-based trading system that can accurately capture what expert level discretionary traders do because the expert trader is primarily forward thinking and speculating.
Yet, even the expert discretionary trader faces two significant problems. The first problem is that even if one achieves a high accuracy in speculation, it is difficult to know how much to bet because the precision is much more difficult to pin down. The second problem is that one cannot know that any specific trades or ways of speculating were even likely to have made money in the past. A good example is when a trader needs a given stop loss for a trade to work but uses a smaller stop loss.
One solution is to go fully systematic. The idea is to use the cognitive process not as the real-time basis for risking money but as the starting off point, the jumping off point, for a trading system. This can work but is a difficult, time consuming, and involved process. An alternative solution is needed for the discretionary trader who wishes to remain active but still have, at least, a historical basis for their decisions.
Imagine I present to you three coins that you can bet on. Each coin is biased in your favor but the bias varies. Now, it doesn’t matter which coin you want to bet on because you have the positive expectation. As long as the coin doesn’t have a memory, as long as the bias isn’t serially or anti-serially correlated then it doesn’t really matter how you decide to bet or skip bets.
Now imagine various trading systems or set of systems as the biased coin. If we have a trading system that can generate several short and long trades every day then we can bias the systems with our discretionary cognition. However unlike the random coins, we are able to take advantage of the active cognition aspects that active trading offers while, still, taking advantage of the historical basis, specification, and knowledge that trading systems benefit from.
However, there is a catch. We need a trading system or systems that both has an edge and can generate a lot of trades or else our idea is more likely to come and go without any trade to bias. And, it is quite difficult to build good trading systems that trade very frequently.
A compromise solution might be to create a quasi-advantaged trading system which may have an edge but only under specific scenarios. These quasi-advantaged systems can be built on simulation or selective data. These types of systems, for example, might allow us to trade optimal rules under given precepts. Our historical basis is not as strong as if we were biasing working systems but there is still a more systematic component over purely discretionary trading.
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.
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.