April 2014 Newsletter – Building Trading Models in ChaosHunter

April 2014 

Building Trading Models in ChaosHunter
Commentary by Marge Sherald, CEO

Building trading models in ChaosHunter requires a certain mind shift if you’re used to building models in NeuroShell Trader.  For one thing, you’re not predicting anything and that affects your choice of an output in ChaosHunter.  In ChaosHunter, the output for a trading model is the price time series on which trades will be made.  So, for example, you can’t use the percent change in open as the output in a ChaosHunter trading model because you can’t buy and sell on a percent change.

Another important distinction is that ChaosHunter presumes the trade will be made on the next bar’s open if you select open as the output.  If you want the model to fill on the open 5 bars into the future, you’ll have to include another column in your data file that represents the open 5 bars into the future.  You can easily manipulate a price data file in a program such as Excel simply by copying the open column and then shifting the cells 5 rows up.  Remove the cells at the end that don’t have an output value or include those rows in the “out-of-sample” period.

Technical Indicators from Data Files

If you export your data from a program such as the NeuroShell Trader, it will be easy to include some technical indicators in your data file.  However, Indicators imported from data files will be treated as VALUES rather than as indicators with parameters that may be optimized.  For example, if you’re working with the Relative Strength Indicator(RSI), ChaosHunter will look at the value for the RSI rather than trying to find different values for the number of periods used to calculate the RSI in order to make the most profit.  If that is what you want, you should use the technical indicators included in ChaosHunter.

Technical Indicators Created in ChaosHunter

The ChaosHunter Formula tab includes a list of technical indicators that may be applied to price data streams included in your data file.   Click on the check box beside any of the technical indicators that you want to be considered for use in your formula.  If you select any technical indicator, you must also choose a time series that the indicator may use for calculations.  Available time series from your data file are listed on the left side of the Formula tab.  You may select several and ChaosHunter will determine which one(s) will be included in the final formula.

Unlike the NeuroShell Trader, there is no need to set an optimization range for the indicators.  That is taken care of automatically in ChaosHunter.

If you are using technical indicators from inside of ChaosHunter, it is suggested that you DO NOT  select any of the time series such as open, high, low, close, or volume on the Inputs tab.

The ChaosHunter Formula tab allows you to select technical indicators and the corresponding time series used to calculate the indicator.

A Mini Berkshire Hathaway Portfolio with 3 Neural Nets

Commentary by Marge Sherald, CEO

We took a sample of the stocks traded in Warren Buffet’s Berkshire Hathaway fund and added SPY and DIA to create a model based on three different neural nets. (Trading rules could be substituted for the nets.)  We first added a neural network that used Average Directional Movement (ADX), Commodity Channel Index (CCI), Relative Strength Index (RSI), and Stochastic %K. Then we created another neural network with different indicators: Accumulation Distribution Sum, MACD, Relative Momentum Index (RMI), Williams %R, and Positive and Negative Money Flow. We added a third neural network which used the Neural Indicators add-on model described in the March 2014 newsletter.

This combination of nets bought and sold a basket of stocks whenever one of the three models generated a trading signal.   The models held up very favorably in out-of-sample testing.  The results for the entire portfolio and each individual stock are shown below the chart in the portfolio view.

NeuroShell Trader makes it easy to create models for multiple securities. Simply add a chart page for each security of interest and any model you create will be applied to each instrument. The models may be individually optimized for each security or you may choose to optimize a single model for all of the securities in the chart.

Was this article helpful?

Related Articles