Hull average cross over strategy – where I am going wrong?

Forums ProRealTime English forum ProOrder support Hull average cross over strategy – where I am going wrong?

Viewing 15 posts - 16 through 30 (of 53 total)
  • #167823

    working in ProOrder due to GRAPH

    Been working outside all day, think the sun has got to me … I haven’t put it in ProOrder yet! 🙂

    I cobbled 8ficusst code together while eating my tea (sorry Vonasi, dinner!) and was surprised at the results … hence I posted it.

    #167834

    GraHa: Anyway, why can’t / don’t you trade your backtest version … it contains Buy and Sellshort and you say it is the same as the Auto version??

    Hi again – thanks for all your efforts – they are much appreciated.

    I really wish I knew more about how the AUTO software runs a code, but the manual does not reveal this in detail. I have tried ELSE and many other instructions in the conditions part of the code, but to no avail. The manual is far too sparse when it comes to describing crossover models designed for AUTO TRADING. Maybe someone in these ProRealCode forums has already solved a similar problem, but I don’t know how to search through all these thousands of posts — maybe there is a way to do this that you know about?

    But just for clarity:  As far as my understanding goes, BackTest codes with multiple variables and multiple conditions cannot be used exactly as they are for AUTO trading, as the variables and conditions must be detailed individually so that the AUTO software can choose which elements to match to the price action at any specific time.

    Still trying!!

     

    #167835

    GraHal wrote:

    Did you try below …

     

    I did already try this combo, thanks.

    MODERATORS COMMENT:

    • Always use the ‘Insert PRT Code’ button when putting code in your posts to make it easier for others to read.

    Your post has been edited.

    #167856

    BackTest codes with multiple variables and multiple conditions cannot be used exactly as they are for AUTO trading

    Above is what we all do else what is the point in backtesting / optimising  if not to refine a strategy so as to make profit running in Live.

    Okay I know some use the backtest platform to watch forward running live then take manual trades when conditions are met etc.

    My hunch on why your Auto-System code is taking trades at every bar is that the conditions are so loose that conditions are met for an opposing trade / reversal at every bar?

    You could test my hunch by using below and you should get a trade every 10 bars?

    You would need to add code to get the 1st trade open so as to get the party started! 🙂

     

     

    #167858

    I have also tried this format, where ELSE is used. (In the conditions at the bottom of the page)

    But that never worked either………

     

     

    #167859

    So your question is …

    Why does the code in the post above take a trade at every bar?

     

    #167865

    In answer to the question above (which may be wrong?) …

    Your code does not take trades at every bar … see attached.

    So what is the question / problem we are trying to find? 🙂

    #167867

    I will try your code suggestion and will report back, thanks Gra.

    I understand that “looseness” could trigger too many trades, but the fact is this problem does not occur with BACKTEST using the exact same code.

    So maybe the BACKTEST system needs different coding expressions than the AUTO system??

    #167869

    Gra, do you know if the AUTO system uses the same methodology as the BackTest system uses at each bar as it progresses?

    It would help a lot if you could find out. I’m not asking for proprietary code/information or anything remotely like that, but it would be helpful to know how the AUTO system selects the options that I load into my codes. Otherwise we are battling in the dark here.

    #167871

    Leave it with me for the time being, so I can run one of the full codes on AUTO on my system and show you what I mean.

    Speak soon.

    #167911

    Both Backtest and ProOrder are in sync … see attached for a Long opened at 17:52 on both.

    That’s at UTC +2 so UK time = 16:52

    #167952

    Thanks for this GraHa.

    Of course I will have to wait until the live market starts up again to continue testing these AUTO systems……….

    #168039

    Can you let me know which code you tested? — IE; was it the first (shortened code) I posted?

    #168044

    I have also tried this format,

    I tested the version shown in the post above.

    #168067

    OK, I see the version shown now. I’ll test it again today and let you know my findings.

Viewing 15 posts - 16 through 30 (of 53 total)

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