the non-computer crowd, but their reaction surprised me. They liked the system; they liked us. We took some suggestions from the crowd for ways to make our screens better. Different color schemes, different groupings, lines, and cosmetic things like this. The roving brokers liked our screens also. These are the floor brokers or market makers who work more than one pit, usually going from one pit to the other to fill orders or to stay where the action is. Their response was overly positive. They liked to look at screens that were always up-to-date, without having to yell out.

Tony Saliba and the Saliba system

We were not the first trading system in the pit and we were not the last. Tony Saliba gave it a try, and we would watch him daily. Burt Beckman was the programmer, systems manager for Tony. Their system came for a while, traded for a while, and then went without explanation. About six months after their monitors were taken down, I got a call from one of the marketing people of Tonys organization. He explained to me their now defunct system was for sale and asked if I was interested. I told him, I had one already, but that I would take a look at it. I tried to be humble and not rub in the fact that our system was still running.

Some of the other Fossett people joked with me about the value of an unsuccessful trading system. I was looking for system parts that might be useful. For example, our modem infrastructure would be outdated in just a few years and might be replaced with something faster. I went to their office and looked at their system. I was astounded! Only years later did I see a system as impressive, when I sat in the cockpit of a Boeing 767. It had a bewildering assortment of dials, meters, and other instruments filling 4 or more screens. I did not think I or anyone else I knew was smart enough to operate such a system. Jim Sauser said it early on, and it became a motto for the Nimble team, "Walk before you run, and crawl before you walk." This is a rule I have taken to heart and use constantly in my life. I do not know why the Saliba system failed, but I have a strong hunch it was all those dials.

Lessons Learned in Developing This System

• Walk before you run, and crawl before you walk.

• Trade without paying brokerage or bid/offer if you can.

• Systems never are as easy to implement as you think they will be.

• Sometimes technology leads you to a dead end, but being able to switch technologies will often get you past snags.

• Never pay fair value, just pay what you need to pay to get the trade accomplished.

Appendix D

Understand why a system should work logically. What is your edge, and why do you have it?

Always be introspective about your trading. Ease yourself into the water slowly. Practice continuous, stepwise improvement.

If an idea comes up, evaluate it fully, regardless of whose idea it is. If you want the freshest, get it from the source. Time and circumstance happen to us all.

About the Editor and Contributors

The Editor

Mark Jurik is the founder of Jurik Research and Consulting, a firm specializing in data modeling and time series forecasting. His greatest talent is integrating various disciplines as disparate as electrical engineering, psychology, and physics into unified theories useful for data pattern analysis. For his rapid problem-solving skills, he received a Letter of Appreciation from the U.S. Strategic Air Command in 1981. He has taught for several years at the University of California Extension at Santa Cruz and lectured at many conferences on aspects of neural net applications. Jurik has written for numerous publications related to financial data processing and was contributing author to the book Virtual Trading (Probus, 1995). He received a BA in Psychology and a BS in Electrical Engineering, both in 1978 from Cornell University.

Jurik Research and Consulting 408-688-5893

443 Los Altos Drive 408-688-8947 fax

Aptos, CA 95003 office@jurikres.com office e-mail

mgj@jurikres.com personal e-mail

www.jurikres.com Web site

Jurik Research and Consulting specializes in data modeling and time series forecasting. Technologies include advanced signal processing methods, neural networks, fuzzy logic, genetic algorithms, and data mining. Our mentoring program guides you through each stage of financial system building. Jurik Research also produces advanced indicators and special signal processing filters as add-in modules for both financial and scientific software applications.

The Contributors

Joseph Alotta (Appendix D) is principle with Open Door Investment Advisors, Inc., a fee-only manager of portfolios using no-load mutual funds. Formerly, he was