BIG CALL FUNCTION PROBLEM

Forums › ProRealTime English forum › ProRealTime platform support › BIG CALL FUNCTION PROBLEM

Viewing 15 posts - 16 through 30 (of 33 total)
1 2 3
  • 06/06/2025 at 10:36 AM #248009
    Franro

    https://www.prorealcode.com/topic/htf-indicator-call-bug-in-backtest-code/#post-247994

    You can see here aswell!

    06/06/2025 at 10:42 AM #248011
    PeterSt

    Yes, that is the topic I meant.

    06/06/2025 at 11:12 AM #248012
    JS

    Hi,
    Just to clarify…
    You’re using “ignore1hourtimeframe” in the line “ignore1hourtimeframe, cb2 = CALL “indicator21″”…
    Are you sure you’re allowed or able to use this? Where does it come from?
    Also, cb2 = average[200](Close), but cb2 isn’t used in any specific condition.
    So you’re just calling an average with: ignore1hourtimeframe, cb2 = CALL “indicator21”.

    06/06/2025 at 2:54 PM #248014
    JS

    I think I know where the issue lies…

    The Call instruction is only executed on the close of the candle…

    When you use UpdateOnClose, both systems behave the same and there are no differences in the outcomes…


    If you don’t use UpdateOnClose, the system with the built-in indicator will check every five minutes (or     TimeFrame(Default)) whether the conditions are met, while the system using the Call instruction has to wait for the result of the condition from the hourly chart (TimeFrame(1 hour))…


    You can also see this in the backtest results: when a position is closed, the system with the built-in indicator enters a new position after five minutes, while the system with the Call instruction only re-enters after an hour…


    In live trading, the code is always executed on the close of the candle, so there is no difference there…

    In short: the Call instruction is not executed on every tick, but always on the close of the candle…

    06/06/2025 at 5:26 PM #248015
    Franro

    JS this has not been a problem for multiple years. I am testing the same code from 4 years back, which now has changed backtest results.

    It is a serverside issue.

    06/06/2025 at 5:34 PM #248016
    Franro

    You can do it with any indicator and try. I just came up with a random one.

    If you do not use updateonclose the backtest is shown wrongly. It does not matter what the indicator is.

    And once again this has not been a issue for over 6 years of using multi timeframe. I litterelly tested over 10 systems even systems imported with correct backtest, which now shown wrong.

    And there is nothing wrong with the calling I just did. I have been calling indicator for years even before on a version you could just have same ignore variable across different timeframes. After that the new version of PRT encourage you to rename it on a new timeframe.

    You can name it ignore1, ignore2 etc. It does not work on ANY CODE which it has done for mulitple years, something has changed over the last week/weeks. I just wrote that code to show a simple explanation of it.

    06/06/2025 at 5:38 PM #248017
    Franro

    Also ProRealTime seems to be aware of this issue and it will hopefully be fixed any time soon.

     

    06/06/2025 at 6:34 PM #248018
    phoentzs

    It’s usually the same problem… most people forget “updateonclose” in complex MTF systems.

    06/06/2025 at 6:55 PM #248019
    Franro

    phoentzs this is not about forgeting. This has never been a problem for several years. I noticed it 2 days ago when backtesting a system and I was shocked by the results. Then I started testing all my systems with calling indicators. All systems that have been totally fine over the years (but changed totally now). It is a serverside problem on PRT, not about forgetting updateonclose.

    It is only happening when CALL the indicator. If you have the same system without call function from indicator, you get another backtest result WITHOUT updateonclose.

    Also PRT seem to have commented it on being adressed, so how come you still point at the updateonclose?

    06/06/2025 at 7:02 PM #248020
    Franro

    https://www.prorealcode.com/topic/htf-indicator-call-bug-in-backtest-code/#post-248010 You can try yourself (:

    06/06/2025 at 7:05 PM #248021
    phoentzs

    I’m not familiar with the calling function of external indicators… I simply don’t use it. I only use indicators, or let’s say, trading decisions, within a system. And if you omit “updateonclose,” you get a completely different result. But it’s always been that way.

    06/06/2025 at 7:06 PM #248022
    Franro

    But why do you comment it is the same problem as usual, if you do not understand the problem or checked yourself?

    06/06/2025 at 7:08 PM #248023
    phoentzs

    Calm down…

    06/06/2025 at 7:10 PM #248024
    Franro

    I am calm, I justed asked. It is just confusing you saying it is the same problem like there is some other underlaying issue. I just wanted to ask incase you had a solution.

    Sorry if it came out wrong.

    06/06/2025 at 7:14 PM #248025
    phoentzs

    Because it often happened that people were surprised that their system didn’t do what they expected… the solution was this one little word.

  • Author
    Posts
Viewing 15 posts - 16 through 30 (of 33 total)
1 2 3

Find exclusive trading pro-tools on ProRealCode Marketplace

Similar topics:
“big difference between positions in backtesting and live trading.
2 pole oscillator by big beluga
big increase in price
help >> function close not working (it does not check price levels properly)
highest function in for loop
htf indicator call bug in backtest code
looking for a function, which can tell if there is an event happend in the last
margin call rates functions in prt
mystup, mystdown = call “orderflow”
nice setup for a big short

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