You're Spending Too Much Time Developing a Trading System
Genetic System Search for Technical Analysis (GSSTA) employs the genetic programming form of evolutionary computation to search for technical analysis investment trading systems. It harnesses your computer to perform the mundane, time-consuming task of trading system development. Consider the following scenario.
The Old, Typical Path to Technical Analysis Systems
Developing a technical analysis system is often tedious. You might start with a 30-period simple moving average. That works okay so you add a trigger at an ADX level of 30. That's better. Now fold in a Relative Strength Indicator. It's a little worse so make an optimization run. The long buy system is profitable; move on to the long sell rules. Start by trying a few values in a Parabolic Stop and Reverse indicator.
Modifications and Dead Ends
Maybe you add a 10-period Exponential Moving Average to the mix. No, that's worse. Take it out. You read an article on the Stochastic Oscillator so you try that. Let's adjust the threshold number a bit. Then you notice that the drawdown is 80% and you know you'll never stay with that system. Once the long trading formulas are in place you might tackle the short side.
That's just for one stock—and just for now. When market conditions change you'll have to slog through the process again.
Tedium Sets In—Doubts Arise
The early phases of your trading system construction can be fun, even exciting, but this wearisome exercise continues until you get tired, bored or frustrated. It is unlikely that you will exhaust the search before it exhausts you. The system you settle for might be reasonable but you know it can be improved.
Time Flies
You also realize that the time spent developing a viable trading system is time not invested in the opportunities which regularly appear. The market is constantly in motion and with that movement comes the potential for financial gain. There must be a better way to find a worthwhile set of investment formulas.
Brute Force
How about a brute force attack? Let the computer do the work. The first method that comes to mind is to program the computer to step through every variation of every indicator along with all possible values of the indicator's parameters. For instance, run the triangular moving average of the open, the high, the low and the close from 2 to 200 periods. You will quickly see that, while the computer has no trouble with the assignment, it will take a very long time to complete the job. Modern computers are fast but it still amounts to an enormous task. You don't want to wait that long.
Now a Better, Easier Way to Find Technical Analysis Rules
The construction of an successful investment system is often time-consuming and must be performed periodically as market conditions change. Furthermore, a system that works well for one stock may not be profitable for another security. Whether you are new to technical analysis or an experienced chartist, you realize the importance and difficulty of finding rules that work well.
Genetic Programming
Evolutionary computation techniques such as genetic algorithms provide efficient solutions to complex problems. Use the power of genetic programming to reduce your workload. Genetic System Search for Technical Analysis is a long name for a program that will shorten your quest. Just provide a stock's price and volume data (such as Equis MetaStock) and start the search. In a matter of minutes the program produces a set of investment formulas.
Generations of Evolution
Learn more about the algorithms underlying GSSTA on the Details page.
Create a Custom Trading System for Each Stock
Technical analysis traders can spend a significant amount of time developing a trading system and staring at stock charts. The prospect of designing individual systems for each security they trade is only rarely considered. They settle for one-size-fits-all. In addition, they want to use that system as long as possible to reduce the effort expended in the creation of revised systems.
The benefit of the Genetic System Search for Technical Analysis program is that it is easy to create a unique set of buy and sell formulas for each stock. Traders are not limited to one system for a universe of securities. Furthermore, it is simple to conduct a system search as new data become available.
Previously traders wondered, "Can a set of rules developed for IBM be used to profitably trade Proctor and Gamble?" and "How often shall I adjust my trading rules?" With the GSSTA program those questions don't arise. The program allows a trader to construct a system for each stock and revise it often.
Benefits
Smart Optimization
The ideal investment system will provide maximum profit, minimum drawdown and minimum brokerage fees. Most technical analysis programs that optimize parameters merely try to maximize the profit. Genetic System Search for Technical Analysis takes volatility and drawdown into consideration too. You're not going to stay with a system in which your equity drops far into negative territory or signals ten losing trades in a row. The program searches for systems with less drawdown and a lower number of consecutive bad trades.
Tailor Your Investment Signals to Each Stock
You can run Genetic System Search for Technical Analysis searches on historical data for each stock that you follow. Create a distinct set of buy and sell formulas for each security.
Keep Your Investment Rules Up to Date
Perform the system search as often as new price and volume data arrives. It's easy to employ a revised investment system every week or even every day. Don't let your strategy go stale. Stay on top of changing market conditions.
An Open System
The rules produced by GSSTA are not hidden. This is not a black box. The program uses traditional indicators available in any stock charting package which results in transparent investment formulas. For a sample of the rules, see the Screenshots page.
An Alternative to the Brute Force Approach
A computer can be programmed to step through every combination of investment signals and variables to find the best system but the number of calculations is massive. Evolutionary computation and genetic programming are used to search quickly in promising areas thereby reducing the time required to find a solution.