Viewing 4 posts - 1 through 4 (of 4 total)
  • Author
    Posts
  • #39143
    facundo
    Participant

    Hi,

    I would like to ask for some help to understand what am I doing wrong. I have been using in Demo account and the putting the profitable strategies in a real account and for the 3rd week it is losing in general.

    1)Once I generated a portfolio of strategies for EURUSD M15, H1, H4.

    2) I put them in a Demo account. I check then the strategies that had 3+ trades and Profit >1.2.

    3) I get those strategies and I put them through the validator of EA Studio to check if they pass the Montecarlo validation.

    4) If it passes i put it in real.

    For some reason in the majority of the cases they are still losing in Real account. I do this every 3 days and still they lose. Also I see that when I pass them through validation, in the Journal they open more trades that are actually opened in the Demo account.

    Maybe I am doing some typical mistake i shouldnt be doing?

     

    thanks

    Facundo

     

     

     

    #39150

    Hey Facudon,

    From what I read there are a few things that might be wrong:

    1. Your broker obviously has different live and demo price quotes. Why do you get a different number of trades? Did you try to backtest the EAs over the live Historical data? Does it match with the Demo?

    If the Demo server is different from the live server, the Demo testing is basically useless.

    2. Why do you run the Monte Carlo after testing on Demo? Why don’t you use it in the Reactor while generating the strategies?

    3. How do you generate the strategies? What are your acceptance criteria?

    4. How many counts of bars do you use when you are generating the strategies?

    Let me know more, so I can give you better advice.

    Cheers,

    #39168
    facundo
    Participant

    Hi Petko,

     

    Thanks for the fast response. I reply following your questions

    1. Your broker obviously has different live and demo price quotes. Why do you get a different number of trades? Did you try to backtest the EAs over the live Historical data? Does it match with the Demo?

    [Facundo] You might be right. I will try with JFD and see if that one is better.
    I don’t know why I get different number of trades. But in Demo account I select the strategies that are profitable then i put them in EA Validator and I see in Validator trades that the Demo account didn’t open.

    No. I didn’t try to back test with the live historical data. only the demo.

    If the Demo server is different from the live server, the Demo testing is basically useless.

    2. Why do you run the Monte Carlo after testing on Demo?

    [Facundo] Because I thought that way i can validate if the Strategy which is giving profitable trades on demo is passing montecarlo validation. I did it to double check. Maybe is not the right way.

    Why don’t you use it in the Reactor while generating the strategies?

    [Facundo] I didn’t understand this question. I did use the reactor to create initially the strategies. Is that what you mean?

    3. How do you generate the strategies? What are your acceptance criteria?

    [Facundo] with the reactor. I generated them 2 weeks ago and I creating new ones every week. Then I put the strategies in the Demo account and select for Real account only the trades that are profit more than 1.2

    My acceptance criteria is

    Min count trades : 300

    Profit factor : 1.2

    Min r square: 60.

    10% OOS

    And montecarlo validation

    4. How many counts of bars do you use when you are generating the strategies?

    I used the meta trader Demo. So as many as it gives me.

     

    thanks for the help

    F

    #39684

    Hello Facundo,

    Yes, maybe it is not a bad idea to test a different broker and see how it goes there. But you definitely need to look for a match between the backtest and actual trading, no matter on Demo or Live account.

    My question about the Reactor is: Do you use the Monte Carlo in the Reactor or just after testing on Demo?

    I would suggest you increase the min count of trades to 500.

    If you use MetaTrader-Demo, this is the data from JFD Demo server, so you better use it, rather then testing with the broker you have at the moment.

    Kind regards,

    Petko A

Viewing 4 posts - 1 through 4 (of 4 total)
  • You must be logged in to reply to this topic.