August 2014 Newsletter – John Ehlers’ Early Onset Trend Indicator / ChaosHunter Templates

August 2014 

John Ehlers’ Early Onset Trend Indicator Tells You When to Get In the Market and When to Get Out Before the Trend Is Over

by Marge Sherald, CEO

In the August 2014 issue of Technical Analysis of STOCKS & COMMODITIES magazine John Ehlers introduced his Quotient Transform indicator as a means of “detecting a trend early and knowing how long to stick with it.”  Most analysts use a variation of moving averages to determine a trend, but at the risk of the indicator introducing so much lag that you miss some of the opportunity to profit.

When we created the chart for the Traders’ Tip that appears in the magazine (thanks to some indicator programming by Denham Ward), we only displayed two different variations of the indicator on the chart, as per the editor’s guidance for the tip.  However, if you read the entire article, you discover that John Ehlers included details for creating trading systems using his quotient transform indicator.

He says the quotient transform indicator may be applied to oscillator indicators as long as they are transformed into an indicator that produces output values between -1 and 1.  The article includes his suggestions for making this transformation.

For example, if you want to use the RSI indicator which normally produces values between zero and 100, he suggests subtracting 50 from the value of the RSI indicator and dividing the result by 50.  We did this and created the Scaled RSI indicator that is displayed on the chart below.

The chart displays the Scaled RSI indicator and four different versions of the Quotient Transform indicator that substitute the Scaled RSI in order to create the trading rules.
 

The Trading Rules

Long Entry:  Quotient Transform RSI 0.8 > 0

Long Exit:  Quotient Transform RSI 0.4 < 0

Short Entry:  Quotient Transform RSI – 0.8 < 0

Short Exit:  Quotient Transform RSI -0.4 > 0

The values 0.8 and 0.4 are constant values used to compute the Quotient Transform indicators.  Positive values are used on the long side, and negative values are used for the short rules.  (See the complete article for an understanding of his methods.)  The smaller constant of 0.4 is used to exit the trade before the trend has run its course.  We optimized only the constant values and the results looked promising.  SPY and energy stocks Duke Energy, Consolidated Edison, and Southern all produced positive equity curves.

Click here to download the chart we created and the required Quotient Transform.DLL.  Once you extract the files, copy the .DLL to your c:\NeuroShell Trader 6\template directory.  When you open the chart you will have access to the Quotient Transform indicator for use in other charts.

ChaosHunter Templates Save Time OPTIMIZING Models
by Marge Sherald, CEO

After we ran the article in last month’s newsletter on how to create templates for ChaosHunter, we had a few tech support calls that reminded me of something we forgot to mention.  The callers generally asked what we had done to make ChaosHunter optimization so much faster in Release 4.  While we would like to claim some credit, the truth is that nothing was done to increase optimization speed in Release 4.   However, the users asking the question were building trading models and had discovered the Steve Ward Trading Models templates (both daily and intraday) that come with ChaosHunter.  Since he used the program for 10 years before it came on the market, we figured he probably knew what he was doing when he created the templates.  When I build trading models I often start with his template and adjust according to the instrument and the type of model.  When I find something I like, I save a copy as a template for later use.

ChaosHunter Release 4

Since we didn’t speed up optimization, you might be wondering what we did add in Release 4.

First of all we added the ability to edit the formula found by ChaosHunter and immediately apply the modified formula to your data file and view the results. This is another time saver because you don’t have to export the formula to another program before you can tweak it and see the results.

We also added a graph to view the results on the out-of-sample data set WHILE THE MODEL IS BEING OPTIMIZED.   Prior to release 4 you could view the statistics for the out-of-sample data, but the graph makes it much easier to recognize a good model.

We also added a complete new example that addressed the problem of predicting processing load by the hour when you only have a limited number of inputs.

If you’ve been enjoying summer like many of us, it’s not always easy to keep up with the changes in the software programs you rely on.  Below is a quick round up of how the latest program updates from Ward Systems Group can increase your productivity this fall.

New Options for Data and

Brokers for NeuroShell Trader
FXCM

If you have an FXCM account, you can use data directly from that account in NeuroShell Trader 6.4 or higher. You can also send trades directly from NeuroShell Trader to your FXCM account, so you have the option of automatically executing trades if you choose.  For information on opening an FXCM account, check out FXCM.COM  FXCM is an independent legal entity and is not affiliated with Ward Systems Group Inc. Trading foreign exchange (forex) on margin carries a high level of risk, and may not be suitable for all investors.

TradeStation

If you have a TradeStation account, you can use data from TradeStation even if the TradeStation program is not running in the background. You can also send trades directly from NeuroShell Trader to your TradeStation account, so you have the option of automatically executing trades if you choose.

To download the FREE software that connects NeuroShell Trader and TradeStation, log intoWWW.WARD.NET with your NeuroShell Trader serial number and go to the Release News section.

For information on opening a TradeStation account, visit TRADESTATION.COM.

Optimized Renko Bars

Ward Systems Group has released a set of renko bar indicators that may be optimized based on price change, volume, or a combination of price & volume.  Developed by Richey Enterprises for NeuroShell Trader, the InterChart Tools Renko Bars include parameters that allow the user to specify the number of ticks used to calculate both the up and down part of the renko bar.  Since rising price jitter is often different from falling price jitter, and since the function of the bars is to absorb noise, the IctRenko Bars permit an asymmetrical definition to accommodate this difference. The Ict Renko Bars are unique to NeuroShell Trader.

For details, visit WWW.NEUROSHELL.COM.

GeneHunter Adds Evolution Strategy, 64-bit, and Excel 2013

 

Ward Systems Group launched Release 4 of GeneHunter, a powerful software solution for many types of optimization problems. GeneHunter is actually two products in one:  an Excel Add-in and a library of genetic optimization functions that you can call from your programs.

New Optimization Method

Release 4 adds the Evolution Strategy optimization method to the Excel interface .  The classic GeneHunter optimization method remains in the program.  Evolution Strategy is a variant of genetic algorithms that uses real numbers instead of integers in chromosomes, and therefore does not cross segments of a chromosome, but instead crosses whole chromosomes. The individuals represent potential solutions to a problem. The individuals are tested by a fitness function and the results are used to determine if the individual will be included in the next generation of potential solutions. The Evolution Strategy method is the default optimization method used in NeuroShell Trader.

Excel Version Can Run as 64-Bit and Works with Excel 2013

The Excel Add-in is updated to work in both 64-bit as well as 32-bit versions of Excel and Windows, up to and including Excel 2013. The GeneHunter Excel Add-in may now be run as either a 64-bit or 32-bit application in Excel 2013 and 2010.  The GeneHunter Add-in continues to run as a 32-bit application in Excel 2003 and Excel 2007.  GeneHunter runs under Windows XP, 7, 8, and 8.1 operating systems.

New .NET Programming Examples

Release 4 adds examples for Microsoft .NET versions of C* and VB.   The GeneHunter programming toolkit also includes examples written in VB6, MSVC6, and Delphi6.

For more details on how GeneHunter can solve your optimization problems, visit WWW.WARDSYSTEMS.COM/GENEHUNTER.ASP

 

 

ChaosHunter Release 4
ChaosHunter® is a stand-alone software tool designed to produce readable formulas to model your numeric data.   Release 4 has added several features to save time in finding the best model for your data.

 

A new formula editor lets you change the formula found by ChaosHunter and immediately apply the modified formula to your data file and view the results. This saves time because you don’t have to export the formula to another program before you can tweak it and see the results.

A new graph lets you view the results on the out-of-sample data set WHILE THE MODEL IS BEING OPTIMIZED.   Prior to release 4 you could view the statistics for the out-of-sample data, but the graph makes it much easier to recognize a good model.

A new example shows you how to predict a value in a time series when you only have a limited number of inputs.

Technical support for any ChaosHunter release prior to Release 4 will terminate on December 31, 2014. 

Time to Examine Your Models to Determine if They’re Right for Today’s Market Conditions

by Marge Sherald, CEO

The national and world news is full of unrest.  There are so many different stories from around the world the news coverage can’t decide which story deserves the top focus.  This constantly shifting focus should carry over to your trading models.  Which model should you currently be using?  Which model should you have in your tool box to immediately run when the markets change.  (The batch mode of the Trader Power User versions is a great way to quickly run through your model options with the most recent data.)

This need for a quick change might even affect how you structure your model.  For example, you can use the Adaptive TurboProp2 add-on neural net for predictions and take advantage of it’s ability to retrain itself every so many bars.  The same may be said for Adaptive Net Indicators, which can use the optimizer to determine the optimal training set size and the number of ahead bars to predict.  (This also means that the oldest bars in the dataset are dropped off as new bars are added and the model is retrained with each new bar.)  Adaptive Net Indicators include a set of indicators designed for classification signals, such as buy/sell/hold.  These nets give you a probability that the current input/output pattern matches one from the training data so you can have more confidence in the classification.

Click here for more information on adaptive neural nets from Ward Systems Group.

Partner Add-ons

The MESA 91 indicators from John Ehlers include Adaptive versions of the RSI, CCI, and Stochastic indicators, another way to keep up with changing markets.  Bowfort Technologies Adaptive Indicators add-on includes more than 20 adaptive indicators, some traditional and others unique to this add-on.

Click here for more details on adaptive indicators from John Ehlers and Bowfort Technologies.

Instead of using adaptive indicators, another approach is to change the bars used in the indicator calculations.  See the InterChart Tools Renko Bars article below for some additional information on this approach.

One key to trading success is to keep your models flexible so they can handle changing market conditions.  Another solution is to have a variety of models that you can call on as the market shifts. NeuroShell Trader gives you plenty of flexibility to do both.

The InterChart Tools Renko Bars Work with Daily Data
by Marge Sherald, CEO

We’ve had several tech support requests asking if the InterChart Tools Renko bars would work on daily data and the answer is yes.  One concern that was raised was whether you might stay in a trade too long when the market took a turn but another Renko bar was not yet generated.  One method you can use is to enter and exit trades based on the Renko bars, but to use a trailing stop based on a daily bar to exit a trade if the market is not going your way.  The same can be said for any size bars; simply base your trailing stops on the bar size of the base chart rather than the Renko bar.

In this example a long exit is triggered by the daily trailing stop even though a Renko bar was not present. 

 

The example chart above is available for download from www.ward.net.  Be sure to log in with your InterChart Tools Renko Bars serial number to access the chart.

No matter what size bar is used for the base chart, you always have the option of optimizing the Renko bars based on a different number of ticks for the up and down portion of the bar. Since any bar’s function is to absorb noise and rising price jitter is often different from falling price jitter, the IctRenko bars permit an asymmetrical definition to accommodate this.   See how optimizable Renko Bars can improve your trading signals.

Was this article helpful?

Yes No

Related Articles