February 1999 Newsletter


If you’ve been thinking about purchasing another one of our products, or even thinking about upgrading your old release to the latest software, now
is the time to do it. Make some money on artificial intelligence and buy your Valentine a REAL present. During this sale period we are offering all
of our products at 15% OFF THE RETAIL PRICE! For the first time, this sale includes upgrades!

Go ahead and upgrade that ancient copy of NeuroShell 2 you have been trying to keep alive. If you are a Trader, it’s time for you to break down and get
the NeuroShell Trader to save yourself tons of time walk forward testing and building trading strategies. (You ARE walk forward testing and building
trading strategies, aren’t you?)

If you are in science, business, or academia, are you really serious about neural nets? What is your time worth, and shouldn’t you have the best
technology available? Maybe you owe yourself the NeuroShell Predictor and Classifier.

You MUST mention special offer #14 to receive this sale price.


Some of our customers are still not aware of our tech support web site, www.ward.net. Here is where we are posting the latest tips, documentation
changes, examples, bugs, and release news for all of our products. We modify this site at least once a week, and there are valuable tips and techniques
there for many products that will save your time and effort. You should check this site every couple of weeks for the products you own. You need a
valid product serial number to log on.


We had hoped that by now the beta of release 2.0 of the NeuroShell Trader would be available for download free from our tech support web site. It has
been delayed a week or two (what else is new in the software world?). Keep watching www.ward.net for it to be available for beta test – soon. That web
site already has posted the major enhancements in the release.

The Professional version of the NeuroShell Trader is also expected to be available in about a month (+ or -). It will contain genetic algorithms to
optimize indicators, inputs, and trading strategies, as well as new indicators and many other things. There will be an upgrade price for owners
of the NeuroShell Trader, as yet undetermined.

Work continues on a real time, intraday version of the Trader. However, the Trader can already be used effectively for intraday trading. Read how on


The February 1999 issue of Futures Magazine contains an excellent article by Murray Ruggiero, Jr. entitled “Born Again Neural Nets”. See page 52. This is
the first of a two part article in which Murray explains a powerful way to utilize neural networks: enhance traditional trading systems with neural
nets. Murray is a recognized neural network expert and trader. He was featured in Business Week, November 1992, for his work in neural networks in
finance. He has told us he welcomes calls for more information on his techniques at 203-469-0880.


While we are on the subject of articles, there is another excellent one by Desmond MacRae in the January 1999 issue of MAR (Managed Account Reports,
Inc.) It discusses neural nets in trading in general, and the NeuroShell Trader in particular. It also briefly profiles George Clarke, who you can
read about under the Real Stories section of www.neuroshell.com. We especially liked the last sentence: “In many ways, Ward may be almost solely
responsible for the ever widening use of neural nets because he has made them so accessible. Traders who now ignore them do so at their own peril.”
We hope to put this article on our web site if we obtain permission.

Managed Account Reports, Inc. (212) 213-6202


Here’s something different. A Boulder, Colorado company (not affiliated with Ward Systems) is building a robot kitten called Robokitten. It uses neural
networks and genetic algorithms, apparently to evolve and learn kitten behavior. The huge project is a research prototype with obvious hopes of
more practical projects to follow. But here’s what interested us:

1. The project is funded by a consortium of Japanese businesses.
2. The Washington Post published this article on the front page of its Science Section, sub heading Artificial Intelligence.

So the next time someone tells you that you shouldn’t use neural networks and genetic algorithms for a project, you will be seeing first hand why the
Japanese have been getting ahead us in so many areas!

We’ll try to get permission to put this article on our web site too, but if your library has past copies of the Post, look on Page A3 in the January 18,
1999 issue.


Ward Systems Group President Steve Ward is one of this year’s speakers at the OmegaWorld ’99 Conference sponsored by Omega Research. The NeuroShell Trader interfaces with Omega’s SuperCharts and TradeStation, and NeuroShell 2 has a TradeStation interface as well. This is the second OmegaWorld conference, and this year it will be in Las Vegas May 21-23, 1999. Steve’s topic is Evolutionary Techniques for Market Timing and Stock Picking. Ward Systems Group will also have a vendor’s booth at this conference. For more information about Omegaworld, see Omega Research’s web site


Maybe you thought that the human resources department would be the last place you’d find neural networks, but think again. Two of our customers have
commercially available systems using neural networks to help with hiring and placement decisions. Watch for a complete article in the May/June PCAI
magazine about these two systems, but here are excerpts:

Wendell Williams of Emergenetics Consulting Group, LLC, in Acworth, Ga., reports a success rate of 82 to 100 percent accuracy in determining the
probability of whether or not an applicant has the social qualifications to perform the job. (The nets do not determine the best candidate for the job,
only the applicant’s expected performance level in the job.) Williams has taken 25 years of personality and performance research to determine his
inputs to the neural nets.

He begins by giving the ESP test to current employees of the company in the same job family (i.e., engineers, managers, sales, clerical, professional,
technical, etc., would all constitute different families). He also asks their supervisors to rate each subordinate in 23 different performance areas
using a scale of “not enough”, “just right” or “too much”.

Based on this information, he creates 23 runtime net files and uploads them to the client’s site. When the client tests a new applicant, he enters
scores onto the web site and the program reports back the applicant’s expected performance level in each of the 23 areas.

“Using the NeuroShell Classifier has cut my analysis time in half. The Classifier combines three features that I always thought were mutually
exclusive: 1) ease of use, 2) a robust application, 3) and low price. Don’t underestimate the power because it looks simple,” says Williams.

Neill Carson and Associates of Houston, Tx. uses a variety of neural network configurations in NeuroShell in its employee selection products. Carson’s
“Pro” Series, is used by Southwest Bank of Texas in the initial hiring and placement of candidates. “The beauty of neural networks from a psychological
perspective is that they allow local norms and patterns to be used in prediction rather than relying on studies done elsewhere that may or may not
apply to the local circumstance,” according to Carson. Carson’s data strongly suggest what most managers intuitively know: customer service
characteristics may be different in Buffalo than in Dallas. Neural networks allow for “local” validation and prediction studies that would be cost
prohibitive (as well as less effective) using traditional methods.

“Neural nets are especially attractive for use with human behavioral and psychological data because they are insensitive to the often non-linear
nature of such data,” according to Carson. Another Carson product, “Sales Path”, uses neural networks to identify patterns of success in sales
employees. Use of the neural net to classify sales people led to the discovery that successful sales people fall into two categories that are
mirror images of one another: One pattern of high performance was characterized by high scores in the “traditional” sales traits of
sociability, empathy, conformity, and warmth. The second pattern was not especially sociable, low empathy, strong individualism, and aloofness with
people. These were the technically oriented sales people, who probably achieved their high performance by being a strong technical resource to the
customers. Carson noted that companies could hire underachievers by selecting an average of these two patterns.

For more information, contact:

Wendell Williams
Emergenetics Consulting Group, LLC
Acworth, GA 30101
Phone: (770) 792-6857

Neill Carson
Neill M. Carson & Associates
Houston, TX 77098
Phone: (713) 526-8132


Lamr Investment Technologies, LLC offers a pretrained net system for intraday trading. It also includes an order entry system. Their software is
called Profit Shares, and you can download it free! Well, ok, there’s a $19 registration fee. They apparently make their money on the trades you place,
but that isn’t expensive either. Privately they have told us it is very, very accurate, but that doesn’t surprise us. Their web site is www.lamr-it.com.


When you think about investment management, what comes to mind? Stocks, commodities, S&P 500, portfolios, predicting the market, etc. As an
Investment Manager for QInsight Group, Bill Carter muses on all of these facets of the discipline — except one. Most investors who use neural nets
are probably trying to predict the market, but not Carter. He actually uses neural nets to classify the economy. He then applies that knowledge to
predicting the market.

Carter’s method uses the NeuroShell Classifier to categorize the economy into one of five phases. Next, he analyzes each phase, and then picks
stocks depending upon which phase the economy is currently in. Bill spent six years working with linear and multiple regression, but his results were
never stable enough to use. Finally, with his first attempt at neural nets, he had a usable, working solution in less than four weeks!

Like many of NeuroShell Predictor or Classifier users, Carter did not want the world of neural net technology to weigh him down. He was attracted to
the Classifier because it was so easy to use, as opposed to other software tools. “The nicest thing is that I don’t have to concentrate on the
pre-processing form of the data,” according to Carter. He was not interested in becoming a neural net expert, but he was very interested in
the extremely accurate results he was getting.

So, how does Carter classify phases of the economy? It took years of research with multiple correlation tests to determine thirty-five financial
macroeconomic variables. The thirty-five variables may be grouped into five general groups, such as price indicators, interest rate indicators, “real
variables” (industrial production), monetary variables, and sector variables (location).

When we asked Carter about his training set, he replied, “it’s quick and easy!” He uses the historical data, of his variables. The Classifier model
reduced his previous set of thirty-five input variables to seven. “All seven variables the Classifier picked were determined to be critical, and I was
very impressed!” according to Carter. The new model went on to classify ALL five phases of the economy with 73 percent accuracy! Carter had previously
spent six months trying to make a working model with other methods, but he was never successful. Now, in less than four weeks with the NeuroShell
Classifier he had a “powerful, usable, working solution”.

Following classification, Carter predicts the next phase of the economy. From this information, he determines which industry groups will out perform
other groups. Finally, Carter advises individuals, as well as companies, on which stocks to purchase. He bases his advice on the current phase of the
economy. Because he is so successful with the Classifier, he has never overridden the nets.

Carter is impressed with the ease of use and power of the Classifier. He states, “it is excellent with economic and complex data. As a pure
classification model, it is the perfect software for me.”

You may contact Bill Carter at the QInsight Group, (619)-295-9292


Hans Wild forecasts daily share prices. Wild does not even remember how he encountered the Ward Systems neural net products, but they have made a
lasting impression in his short range trading. He believes the NeuroShell Predictor is “unquestionably the most user-friendly neural network program
in the marketplace!”

“One day a new release of NeuroShell Predictor was delivered onto my doorstep . . . Now I was able to achieve some absolutely fantastic forecasting results!” according to Wild.

Aside from being user-friendly, two features that attracted him included not having to worry about the number of hidden nodes or when to stop training.
Wild said he was able to see results after a few minutes, as opposed to several hours.

He experimented a lot in order to determine a useful combination of inputs, including open, high, low, close, and other various indicators. After
testing other neural net products, Wild exclusively uses Ward Systems Group software, including NeuroShell 2, NeuroShell Predictor, and NeuroShell

He jokes about his own neural net programming. “What awful performance my nets had!” he said, but this was before commercial neural net software was
available. Wild is not interested in the neural net strategies, only the results. His main concern is applying his results to trading, not what particular net is used. When we asked Wild if he ever overrode the nets, he replied “No, that doesn’t work. The neural net results are almost always better than my ‘market feeling’.”

The neural net software has saved Wild a lot of time and he achieves superior results. He credits this to the fact that he can run though more simulations in a shorter period of time. Therefore, his results are between 65 percent and 85 percent accurate. His advice to a first time NeuroShell Predictor user was, “Use it! Don’t despair if initially you don’t achieve good results. It takes some experimenting to find good input sets.”

Was this article helpful?

Related Articles