We often get questions about using indicators in NeuroShell Trader charts that vary depending on market conditions.  I’ve created a prediction that uses an indicator from the Price Momentum indicator category in NeuroShell Trader and another indicator from the Advanced Indicator Set 2 add-on.   The results look promising so I thought I would share them.
The NeuroShell Trader has an indicator called the Relative Momentum Index, which measures price momentum. It is calculated in a manner similar to the more common Relative Strength Index (RSI) indicator, but with a lookback period that can be optimized as opposed to the fixed one period lookback of RSI. (Relative Momentum Index ranges between values of 0 and 100. The closer the index is to 100, the stronger the indication of an overbought market. The closer the index is to 0, the stronger the indication of an oversold market. In general, index values above 50 indicate a possible uptrend, while index values below 50 indicate a possible downtrend.)
I included two snapshots of the Relative Momentum Index as inputs to a prediction.  The first one used the 1 hour bars from the original chart. Using the NeuroShell DayTrader Power User’s ability to mix time frames, I added a copy of the same indicator that used two hour bars for a slightly longer term view of market action.

Advanced Indicator Set 2

Next, I added another input to the prediction called the Variable Length Moving Average indicator from Advanced Indicator Set 2, a collection of user requested indicators including some developed by Marc Chaikin and J. Welles Wilder.  (Click here for a complete list of indicators from Advanced Indicator Set 2.)
I applied the Variable Length Moving Average to the Average Directional Movement (ADX) indicator from the Trader’s Price Momentum Indicator category.  The average periods were varied by a simple calculation of (1 + 20) with an optimization range from 1 to 40 for the parameter that started out as 20.  (In fact both addition parameters could have been optimized.)  Any other calculations could have been used to vary the length of the average, including other indicators.

The Variable Length Moving Average indicator from Advanced Indicator Set 2 lets you vary the number of periods used to calculate the average.

The chart above displays the out-of-sample results for the prediction based on a variable length average used as an input to the neural net.

The system equity for a basket of stocks picked from different sectors looked promising for both the training data and the out-of-sample period.  This chart is available for download fromwww.ward.net in the Examples section.