ChaosHunter makes no a priori assumptions about the final form of its trading models. It does not assume any theory of market movement, e.g. Elliot wave, support-resistance, regression to mean, fractal, random walk - not even chaos theory. It is unlike modeling techniques such as regression and neural networks, which optimize coefficients of equations with relatively fixed structure to fit price related curves. (ChaosHunter can fit price related curves, but not in the preferred trading mode). ChaosHunter does not ask you to decide if or how it should utilize technical indicators, or which ones (although it does allow you to provide suggested ones).  

What ChaosHunter does is crunch away using natural selection techniques to evolve unique, profitable trading models using building blocks of basic, simple functions from arithmetic, algebra, trigonometry, neurons, logic, and technical indicators. Evolution selects which building blocks are used and how they should be fit together to form equations that produce buy/sell signals. ChaosHunter, in effect, evolves its own technical indicators. 

The users select historical price data that ChaosHunter will use to backtest its evolutionary creations. If desired, the users may also select or limit the building blocks or provide more technical indicators for ChaosHunter to consider. Then ChaosHunter starts crunching to find models with no further guidance. 

ChaosHunter shows you all models that it has found with no hidden secrets, so that these models can be placed into any of a number of trading platforms. ChaosHunter is not evolving computer programs, just equations, so for most trading platforms programming will not be required.