A Discussion of Trading System Development using the Scientific Method

Warning: Declaration of ElementorPro\Modules\Posts\Skins\Skin_Content_Base::register_controls(Elementor\Widget_Base $widget) should be compatible with Elementor\Controls_Stack::register_controls() in /home/customer/www/blueowlpress.com/public_html/wp-content/plugins/elementor-pro/modules/theme-builder/widgets/post-content.php on line 30

The Blue Owl Press hosts a series of articles that outline the trading system development procedures I recommend.  The website also has links to my books, my videos, auxiliary papers and articles, and resources (including books, videos, and courses) related to the development of trading systems for stocks, commodities, futures, and ETFs.

The underlying approach is the scientific method.  Trading systems are combinations of models and data.  The data includes historical price and volume of traded securities.  The models are fit to the data in the development phase where patterns that precede profitable trades are identified, then validated using the walk forward process and forward testing of out-of-sample data to ensure that the trading systems are robust, profitable, and safe to trade.  

Traditional trading system development platforms, such as AmiBroker, TradeStation, and NinjaTrader, where the model is adjusted using backtesting and optimization are discussed.  Machine learning and artificial intelligence techniques using Python and the scikit-learn libraries are also described.

Measurements of goodness-of-fit, such as accuracy, expectancy, profitability, compound rate of growth, and drawdown are explained.  The advantages of state signals with mark-to-market evaluation over impulse signals are explained.  Several original and unique metrics, including safe-f and CAR25, are used to guide development, assess risk, and normalize results to enable comparison of alternative uses of funds. 

Techniques for management of trading systems that have successfully passed validation and are being actively traded are introduced.  These include correct use of position sizing as a management tool, and enable the trader to determine whether the system is healthy and working, or is broken and should be taken offline. 

The focus is to develop systems in which the trader has confidence — systems that are built using the evidence in the data — that can be measured, monitored, and managed — that recognize the risk tolerance of the trader and manage trading by managing risk.

Thoughtful discussion is welcome.

Dr. Howard Bandy

Leave a Comment