TSC Technical Forum On the Comp, a Bounce Then a Leg Down? By Gary B. Smith RealMoney.com Contributor 8/9/2004 8:32 AM EDT URL: http://www.thestreet.com/p/rmoney/techforumrm/10177100.html

Technical Analysis Meeting smart traders is a benefit of writing. Also, charts on Netflix, the Comp and Cabelas. The Comp seems oversold, but a bounce might only be followed by a down move. Netflix is at support, and might be a buy.

One of the privileges I get from writing this column and my newsletter is running across people who are bright, talented and have a lot to offer to the trading world. In the past I've profiled the work from my friend Wesson (amongst other things, he wrote my backtester program), Bob Drach and Jim Rohrbach.

I graduated from the University of California at Berkeley with degrees in Mathematics and Physics, and later did extensive graduate work in Mathematics and Geophysics at the University of Houston. I joined the Apollo group at the MIT Instrumentation Laboratory in 1966, and learned to program the Lunar Module onboard computer in assembly code. This culminated in my programming the terminal descent logic, both abort modes, and the lunar ascent for the Apollo 11 mission. Thus began my lifelong involvement both with space flight, and with the development of mathematical and logical algorithms.

I formed a small software company in 1981. Being self-employed provides some time and freedom to develop my own software products, in addition to working on a series of long-term contracts to support Space Shuttle software development, and to design and develop large-scale database applications for a major oil company. Most of my own software products are highly mathematical in nature: an "expert" system for nonlinear mathematical modeling (which includes neural network capability), factorization software, and tools for the numerical integration of sparse systems of stiff differential equations.

My interest in the stock market began in the late 1980's when I tried to apply my non-linear modeling tool, and its neural network methods, to the commodities market. This was not very successful, but it fired up my interest in the financial markets when I realized that, after all, they are all mathematical systems.

In 2001 I began to experiment with algorithms for predicting the short-term performance of stocks and mutual funds. Given the great availability today of reliable market data, company fundamental data and fast computers, this second assault on the financial world turned out to be very much more productive.

The first major step was to develop a complete Mechanical Trading System. This tool examines over 8,000 stocks daily and generates a series of trading portfolios of increasing size, and optimized for maximum annualized gains. The MTS constructs 10 Long portfolios and 10 Short portfolios, each containing from 1 to 10 equities. In the process of developing this program I discovered some new technical indicators: a "waxless" candle generalization of Japanese candlesticks, and a "nibble" indicator which can often detect insider interest in a stock just before it actually takes off. This program also employs most standard technical indicators and oscillators, as well as all available fundamental data. I felt that it was important, however, to fully try out my own ideas before the wholesale adoption of all the standard technical methods.