For the new and beginning developer, one problem you might encounter is not having enough capital to run more then a few systems. This can lead one not to work on their system development as much as they could. It is my belief that all market movements are the result of market cognition. Market cognition can be used to explain the entire cognition of all participants in the market but can also be seen as an individual factor and is the greatest single determining factor, outside of capital, for your success as either a discretionary or quantitative trader.
Traders with high market cognition can predict the markets. Discretionary traders with high market cognition can use implicit learning or more commonly known as pattern recognition and conscious processing to make high probability trades. System traders can backtest and validate their best ideas to find the ones that have been historically profitable.
As a system developer, if you are not actively thinking about the market, coming up with new ideas, and developing systems then your market cognition can stagnate. But, recall that most small developers do not have enough capital to run more then a couple systems. This can lead to stagnation. A possible solution is the super system.
Imagine many systems generating signals across many markets. Out of all the signals there can only be one best signal. The liquidity of markets is often championed but without intelligence to make good trading decisions then the ability to trade has limited value.
The super system concept is rather simple. You feed all your individual or sub system trades into a super system instead of trading them outright. The super system ranks all signals based on (1) the quality of the trade as projected by the sub system (profit factor, winning percentage probability, risk) , (2) the performance of the system in the current market conditions, and finally (3) either a mixture of factor components or factor components that work best for the current market conditions. The super system sorts and ranks all the trades and only takes the best trades from your sub systems.
For example, at an individual system level the developer needs to balance or optimize the trade quality (profit factor) with the net profit. Only taking the best trades will tend to lower the net profit while taking more trades will increase the net profit but may decrease the quality of the trade. In the super system concept, the sub system doesn’t generate a binary signal but rather produces a prediction or forecast that includes the projected profit factor. This allows for the super system to optimize for highest net profit and trade quality across multiple systems with limited capital. As well, the super system can be programmed to trade a specific mixture of factor components. This can be static or dynamic. For example, the super system could be designed to trade at least 50% trend signals and 50% mean reversion signals over some trading window.
As for the advantages, your results can be better then any individual system as you can maximize your net profit and profit factor dynamically across all the systems. It also causes one to think about optimization in new ways because the super system can perform the optimization in real-time. The ability to ensure mix of given signals could also dampen individual factor components. Some of the disadvantages or challenges: your results will not replicate any individual system and you will not achieve a true diversification. There are also certain requirements for this to work well that your subsystems will need to adhere too: will require more active systems, trades should be independent events, must be able to produce a trade quality metric/value that offers value based on the signal and the performance of the system, and systems should not rely on outlier profits.
The super system concept can also be used as a graybox concept where either (1) the super system attempts to find the optimal trades for your predictions/forecasts based on its independent analysis or (2) the super system generates a forecast which either determines your stop loss levels and/or the direction you are allowed to take trades.
The author 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 the author at firstname.lastname@example.org.
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.