Tagged: 

Viewing 15 posts - 1 through 15 (of 28 total)
  • Author
    Posts
  • #7650

    Hey Andi,

    Glad to hear from you again.

    Yes, it is not bad to reduce the data. In EA Studio the maximum is 200 000 and that is for a reason.

    In FSB Pro I go maximum to 300 000 on M1 chart.

    Yes, I know FSB Pro is fantastic strategy builder…keep in mind that I tested it long long time before placing it on our website.

    #7617
    Admin bar avatarAndi
    Participant

    Thank you indeed, I missed to see your answer Petko.

    So should I reduce my Historical data? I guess this way I will put more weight on the recent market data.

    And by the way FSB is the best tool I have ever used in trading.

    #6805

    Andi,

    keep in mind that you can collect long time Historical data from one server from one broker, and at one moment if they move your account to another server, or if you open another live account it could be on another server.

    Normally the broker should notify you about that, but I wouldn’t trust it.

    Anyway, Thetrader is right, the more bars you have, the better, but your profit per day will reduce. Simply, the FSB Pro looks for such strategies that are working profitably over the whole period, which cause smaller daily profits.

    Kind regards,

    #6703
    Thetrader
    Participant

    Hey Andi, I think your data is a bit too much. With so many bars you will get Moreno robust strategies but their daily results will be not that satisfying. The more data you have, the more you limit the average profit per day.

    The only think you can test is if your demo account matches with the journal.

    #6682
    Admin bar avatarAndi
    Participant

    Hello guys!

    Interesting topic about the FSB here. I actually succeeded to download some historical data from my broker, they have it on their website. And on M1 I have more than 650k bars. I am not sure if so many bars would be good idea?3

    Also, how can I be sure that this Historical data is from the server I am using for the broker?

    #6452

    Thank you, Edu! Appreciate it. It is always about testing…

    Let me know if you have other questions about FSB.

    #6443
    edu
    Participant

    happy birthday, i have buyed your course. Now i have to test again 😉

    #6417

    Sure, Edu. Let me know if you have any other questions.

    Kind regards,

    #6416
    edu
    Participant

    thank you again

    i saw your new 100 Ea probably i am going to buy it with the new discount for you birthday 🙂

    #6415

    Hey Edu,

    for the first post:

    The Optimization results seems better, and I think the Monte Carlo looks just normal for the number of bars you have. Later when you have bigger number of bars, the strategies will get more and more robust.

    The USDJPY looks so much better. I see on your FSB back test that you have more than 300 count of trades which is a reliable number.

    Also, when you perform the Monte Carlo, you can select the last two options together. No need to do it separately.

    Cheers

    #6410
    edu
    Participant

    (Please see the previous message before.)

    New Update, after EURUSD-EURGBP-GBPUSD, i was able to collect more data for USDJPY

    This is the before:
    USDJPY_PRIM_1809

    USDJPY_PRIM_1810

    This is the after:

    USDJPY_DOPO_SCHEDA

    usdjpy_dopo_1

    usdjpy_dopo_2

    This IS Monte carlo:

    usdjpy_montec

    usdjpy_montec2

    If you can give me your opinion also about this and the other in the previous post would be highly appreciate

    Edu

    #6406
    edu
    Participant

    Me too as entrepreneur, i have no fixed day off.
    ————————————————————

    I understand the problem and i did it again here the results:
    nuovogbpyjpy_secondo_indiczioni<br />what is the boiling point for aluminum<br />

    it seems that it has huge drawdown (when i use an Ea in real i typicaly use 0,10 lots)

    I also did the montecarlo test (that i finally understand how it works)

    mont_229<br />what is the boiling point for aluminum<br />

    montecarlo_229_pt_2<br />what is the boiling point for aluminum<br />

    How it looks to you? My broker and this strategy work well together?

    —————————————————————————————————————————-

    Now i have 3 good “looking” EA (EURUSD-EURGBP-GBPUSD), 1 (USDJPY) With too little data (but in demo in 5 days it closed everyday in loss) and this.

    Probably i will buy the 99 ea course even if testing all of them would be difficoult considerinf FX blue doesn’t track very well the magic numbers

    Thank you again
    Edu

    #6404

    Hello Edu,

    you are missing something with the optimization.
    The first time you did it with step of 1, and the second time with step of 10.

    It is good to do optimization with round numbers as step. But make sure the parameters of the strategy before the optimization to be round as well.

    For example, make Stop Loss 80 and Take Profit 70, and then do optimization with step of 10. Same for the other parameters

    Your Monte Carlo is just fine. The over-optimized strategies will show total loss with all tests, which is not your case. And yes, you are correct.

    No weekend for me during the last years, mate 🙂

    Cheers

    #6401
    edu
    Participant

    Wow thank you i wasn’t aspecting a reply on saturday.

    So i tried as you said, here i removed the “acceptance criteria” and the equity changed

    test_senza_accepance_criteria

    after i try to put bigger criteria, but anything changed

    try_to_put_better_criteria

    So i try the montecarlo test. First parameter

    montecarlo1

    then the second (doesn’t look good)

    montecarlo_2

    What do you think? Should i remove this from the demo and try another ea?

    ————————————————————————————————————————————————————–
    Before i didn’t explain well, i was asking:

    If Monete carlo test:

    – Give me an equity similiar to my ea’s equity ——-> Good Strategy and not overoptimize
    – Give me a different equity ——-> overoptimize strategy

    is this correct?

    Thank you and again have a good weekend
    Edu

    #6391

    Hello Edu,

    this backtest looks bad. When you optimize in EA Studio, you have an acceptance criteria. Try to remove it, and do the optimization. This way you will see if the issue is there. If it passes, than try to make the acceptance criteria not so strict.

    The Monte Carlo itself is the robustness test for over optimization. If it fails, it means that the strategy is probably over-optimized.

    Another old method to see if a strategy is over optimized is to change manually some of the parameters in the indicators. If you see huge difference in the equity, this is also a sign that the strategy is over-optimized.

    Have a beautiful weekend too!
    Petko A

Viewing 15 posts - 1 through 15 (of 28 total)
  • You must be logged in to reply to this topic.