PP Fractals // Error message

Forums ProRealTime English forum ProOrder support PP Fractals // Error message

Viewing 14 posts - 16 through 29 (of 29 total)
  • #189023

    Peter, As i mentioned above (#189007) the original code generates a profit VS the code modified by Nicolas generates a loss. So, there is something in this PP Fractals code that plays a role of filter beyond the graphing.

    #189024

    what @Nicolas is saying. So to me it now seems that he (from his own code ever back ?) tells that the arrays were only there for graphing, while you don’t graph anything. So just eliminate the code concerned (?).

    Which is exactly what he has done for you (I checked and understand it now). So the challenge is : why is the result not the same.
    Probably because the approach with the array failed ? or the arrays themselves fail ? (it really is too easy to make mistakes, the way PRT set them up).

    #189029

    I’m not putting in doubt what Nicolas has done. However, I see (as you see it in your backtest) there is a big difference between with and without.

    Do you mean that there might be something wrong in the calculation of the Arrays that generates “fake” Profits? It’s very likely. May be the solution/answer is with Vonasi the author of the original code 🙂

    #189030

    May be the solution/answer is with Vonasi the author of the original code

    For info … we’ve not seen Vonasi hereabouts for 9 months or so!

     

    1 user thanked author for this post.
    #189032

    I think the problem lies in this part :

    My “coder hunch” tells me that z is eliminated from the new code, while z is (was) necessary for this part. So it is not about a and b, but this z. I can’t see quickly how to change that – hopefully @Nicolas can.
    And maybe I am wrong. But both codes really are 100% the same, except for the part which is commented out (and the new usage of resistancevalue ans supportvalue).

     

    1 user thanked author for this post.
    #189033

    For info … we’ve not seen Vonasi hereabouts for 9 months or so!

    umd followed him. If you know what I mean … (and if not then too bad).

    #189034

    I see exactly what you mean 🙂 🙂 🙂

    1 user thanked author for this post.
    #189038

    Something else (not), we should not forget that something *is* wrong somewhere with the arrays approach. And avoiding their purpose (the graphing) is not really the very best idea.

    Without the arrays, this error still occurs for the 1M backtest. I will report it, then together with nontheless’s report (elsewhere) they may see commonalities.

     

    #189042

    Problem is that the array can increment 12 times per hour, so the 1 million column of  the array could be reach pretty fast, please try the new version below:

    #189046

    Thank you very much Nicolas!!! I left home, so I’ve no access to my PRT now. I’ll test it later today and let you know.

     

    Basically, what I’m trying to acheive is a Strategy that works without any time-based indicator for Entry Signals (no MA, no RSI, etc.), only price, price and price. I found the Swing Line that you shared in another post quite effective with (UPDATEONCLOSE) and TF>H1, but I need to add more filters like Breakout of Trendlines.

    May be we can do that without Arrays?

    #189051

    Thank you for your help, @Nicolas.
    The result is unchanged, but more importantly, this is not workable because of slowness and *that* tells me that something with the array-accesses is not correct (asking for an undefined element will trigger the error-trap under the hood and this is ultimately slow). The 100K takes 20 minutes to calculate. This was not so at all with the earlier version from Khaled (using arrays).

    I am only the messenger. 😐

     

     

    1 user thanked author for this post.
    #189054

    Dear Trading Companions 🙂

    First, please let me thank you for all the help and time you have spent trying to help. Obviously, this is more about the learning journey than the immediate end result in itself.

    Second, the 2nd code reworked by Nicolas didn’t work LIVE. I had again and again the same error of 1,000,000 rows.

    Third and happily, I managed to replace the PP Fractals code by another code from Vonasi called Fractals Average https://www.prorealcode.com/prorealtime-indicators/fractals-average/, where I replaced the two trigger variables with “MidAvg”. I got results not far away and it is live now as we speak. Let’s how it behaves over the afternoon.

    I’ll keep you posted.

    If anyone thinks of a piece of code somewhere on this forum which uses only price for Trend, Breakout, etc., please let me know. I defenitly want to build a system without any time-based indicators (MA, RSI, etc.).

    Again, thank you for your support. I feel blessed.

    #189064

    New try (I did not compare tjhe result with initial version)

     

    1 user thanked author for this post.
    #189075

    It is normally fast again.
    The results are not like the original (at all).

    1 user thanked author for this post.
Viewing 14 posts - 16 through 29 (of 29 total)

Create your free account now and post your request to benefit from the help of the community
Register or Login