Home Forums General Discussion How far back to optimize an EA?

Trusted Trading Brokers banner
Viewing 5 reply threads
  • Author
    Posts
    • #7171
      User Avatarjacpin2002
      Participant

      Hi everyone!
      I have been autotrading with EA Studio live since the end of October. My current closed profit is 412.53USD. 80% of this profit came from one out of the 3 EAs that are trading. At what point should I optimize my EAs and how far back in historical data do I use for optimization? All of my EAs are at 70% or high winning percentage.

    • #7187

      I am glad to hear that you are on profit already so quickly. Most of the students give feedback to me that they succeed to make profits after the 3-4th month. However everyone is different, and some do it faster. Good job!

      It is not bad that most of the profit came from one EA. That is the idea of placing many EAs in one account, because the moment we place them on Meta Trader, we do not know from which one the profit will come.

      The optimization depends really on the market and the performance. If you have made optimization and you are doing profits, and you do again optimization most probably you will not see different results as parameters in the indicators.

    • #7195
      User Avatarjacpin2002
      Participant

      Thanks so much, Petko! If it wasn’t for you, I would have never had this kind of progress. Much appreciated to you and the developers of EA Studio.

      Ok, that makes me feel better. I believe that I go through a pretty rigorous process in generating the strategy, filtering to the one that I would like to demo, demo test, and then go live. To generate 96 EAs, it took me almost 2 months, but I feel the time is worth it.

      I’m not really sure I even want to do optimization now. I feel like if the EA passes the acceptance criteria in generation, the demo testing criteria to go live, and then maintains the same success rate in live trading, there is no need to optimize. If it starts to perform below the expectation, then I would just replace it with a new EA created from my process.

      So, I guess I’m trying to figure out maybe when should optimization be utilized and if so, is it worth it?

    • #7224

      You are very welcome! It makes me extremely happy when I hear that people who watched my courses start to make profit.

      Recently I received many questions, why I do not provide signals, or some kind of trade copier. The answer is simply – I want to build traders and not copiers.

      It is easy to do it, but when I am gone one day, the traders will be frustrated, and they will need another place to copy from.

      But when you are build as a trader, have the skills and the knowledge, you will succeed to continue trading, building EAs and be successful. That is what I want to achieve with the students.

      Yes, that is exactly what I wanted to say. If it is working, simply do not touch it. When it starts to lose, than you will look for solution. But with EA Studio the solution is easy – replace with new EA that you have already tested.

      Now the optimization normally is needed when you use Expert Advisors from 3rd place with different broker.

    • #7230
      User Avatarjacpin2002
      Participant

      @Petko-Ok, that makes sense. So, I’m only going to use the optimizer with the EAs that I get from you in the courses. Got it. Otherwise, I will stick with my original process of generating new EAs, when one starts to lose. Thanks a bunch.

    • #7233

      Glad to hear it makes sense. Of course I always suggest the traders to test and test, and test. What I mean is that you can give it a try optimizing them more often. But specially if you see current profits, the optimizer will not show any better results from the last optimization or the original generation process.

      Also, keep in mind that the optimization is a bit risky. What the optimize simply does is to find the best parameters for your strategy BUT only for the tested period. Not outside of it. There is no guarantee that these parameters would work as expected in the future. That is where the robustness tests come in hand.

      Just try to make is more simple, and you will do better.

      Cheers

Viewing 5 reply threads
  • You must be logged in to reply to this topic.
Shopping Cart