Viewing 9 posts - 1 through 9 (of 9 total)
  • Author
    Posts
  • #51210
    suraj.bhawan
    Participant

    I’ve taken the following courses:

    1. Forward Walk Optimiztion

    2. Forex Strategy Course: 12 EA’s

    3. Forex Trading Strategies from a Professional Trader: 5 EA’s

    4. Automated Forex Trading Course: 99 EA’s

    I was considering running the EA’s we receive from above in the validator with my brokers data, under acceptance criteria out of sample (20% OSS) minimum profit factor of 1.1

    Using the out of sample profit factor would act like it was a strategy run on demo.

    For robutsness I would select Monte Carlo, count tests 20-30 with validated tests at 80%.

    Thoughts on above?

     

     

     

     

     

    #51254
    Admin bar avatarAndi
    Participant

    Not a bad idea. I would only remove the OOS because these are ready strategies. You just want to see how they performed with your broker right? So you would like to see especially in recent months.

    If you are generating EAs, then yes, OOS is awesome.

    #51262
    suraj.bhawan
    Participant

    <p style=”text-align: left;”>Yes, ideally I want to see how the strategies would perform in recent times. The ones that perform well I’d take to my live account.
    The idea here was to initially try this method of application on a small account, if the method proves successful eventually apply in on a proper account.</p>
    Another thought that came to mind was that, once a week, use forward walk to optimise the strategies in action, more replicating the idea of continuous forward walk optimisation. Although I know you prefer not to touch what’s working well.

    #51263
    suraj.bhawan
    Participant

    The intervals of optimisation could be done once a week.

    #51278

    Hey Suraj,

    What Andi says is right. If you use my EAs on your broker’s data, just check if they do well. After that place them on a Demo and see which are the top performers. The OOS is useful when we generate new EAs.

    Walk forward is a fantastic robustness tool but again when generating the EAs.

    Simply, when you have ready EAs, all you need to do is to test them and decide which ones to place on a live account.

    #51311
    suraj.bhawan
    Participant

    I was looking at the EAs on EA Studio and noticed that there is a vast difference in performance from the MT4 Demo data and my broker data

     

    Stategies that are profitable over a period are actually loss making on the same period with broker.

     

    Wouldn’t optimisation using broker data make those strategies a better fit. I understand that OSS and Forward Walk are best used when generating own EA’s.

    #51315

    Hey Suraj,

    yes, it could be a difference with the broker’s data. That is why there is the option in EA Studio to import your own data because the brokers offer a way to different quotes, which results in different Historical data.

    I would suggest you generate EAs with your own data, instead of optimizing the EAs.

    Also, you can use the Historical data from the Free App which we have on the website, and there is a video in there where I explain how to fit that data for your broker.

    Cheers,

    #52551
    suraj.bhawan
    Participant

    Hi,

     

    I just had an idea that I wanted to run by you. Since I am using the EA’s generated by you, would it be a good idea to keep changing the portfolio based on the best performing EA in the previous week. For example, all th EA’s are first taken to the demo. Every Saturday, the best performing EA’s are taken to the live account, the same process is repeated in the following week. If 1 week is a short period the same concept could be applied either every two weeks or monthly.

     

    #52566

    Hey Suraj,

    I think one week would be great to use.

    Just test it is first on two demo accounts simulating one is a real account.

    Do not hurry risking on a live account before you found the right period and method of testing.

Viewing 9 posts - 1 through 9 (of 9 total)
  • You must be logged in to reply to this topic.