Use computerized evolution to create trading system?

Discussion in 'Automated Trading' started by lighnintrade, Nov 18, 2011.

  1. ronblack


    They have no idea of what they are talking about on that blog. One pig-ignorant guy is attacking another pig-ignorant guy.

    Trading systems created by evolutiononary processes are basically curve-fitting systems unless a number of conditions are met, which I am not going to discuss here because they are the basis for coming up with an edge.
  2. Half of that article is junk, but i could not find the original article i read on the subject some years ago.
    I see what you're saying about the curve fitting and this usually only occurs if you would set the parameters for success of the program to be show a profit by entering and exiting at certain times or with certain signals(usually TA). When programming the parameters for success for the evolutionary algorithm you should be able to set it to look for patterns that emerge in the stocks or indices. Patterns that you might not be able to find yourself but the program should be able to if done properly.
  3. Evolution itself is a curve-fitting process.
  4. I believe TSL is doing exactly what you propose. PAL is more of a data mining program with application specific taxonomy and meronomy, as MAESTRO would say. It is limited to price patterns:
  5. These parameters which you do not mention do not change with time, form the basis of market dynamics, and are an artifact of human nature.........

    A good system, however, will be optimized to harness them.

  6. This sounds like a good start.
  7. Yes, there are many ideas how this (evolutionary ATS development) can be accomplished, but they're not easily shared. It's up to you to figure that out on your own. There are some hints if you search around, but the core logic of the process is something one needs to work out himself.
  8. Putting 2 and 2 togehter is very difficult.

    As wide talesz mentioned the facts are available.

    The article you cited also specifies the other ingrediate indirectly. that is this consideration: A person must be able to use the information that he correctly determined he needs.

    When I was in college I had to sign an intellectual property agreement for the college and for Bell Telephone Labs. They were in conflict.

    So I used my BTL laboratory journal at college. I did study antenna design in an UHF lab, as a coincidental, matter of fact. When I graduated (and my last retake of a course in UHF was complete) the college refused to return my lab journal. The grounds were that as a student, I granted my intellectual rights to the college. what had happened, in fact, was that I did extra lab work with my free time in lab. The BTL work I was doing was on the Nike I and Nike Ajax and in particular the two types of antennae used in that system. So both organizations lost out since I returned to a civilian orientation upon finishing grad school.

    To explain the lack of understanding in this thread about the two requirements, what happens is most people cannot think through the opportunity the market has provided. By reading the history of the market operations and discoveries, you can easily find how historical personages solved the problem that is being considered by members of this emerging thread.

    When the work was done, the motables of history DID understand their challenge and they DID use the information they NEEDED to be successful in the problem solving required.

    The fact that the posts so far indicate what they do, simply means the posters do not understand how to go about the previously completed process.

    Work slowly to discover what is required of you and to dicover what you need to do the work. If you are incorrect in your present view, you have an additional problem; you have to work very hard to rid yourself of your mistaken beliefs.
  9. what?
    #10     Nov 19, 2011