Shame on PRT

Viewing 14 posts - 1 through 14 (of 14 total)
  • #12172

    Hello, here JohnScher.

    This little code i wrote with a little help

     

     

    What will i say

    I copy  the running code one to one from real trade to backtesting.

    And I did running the code real. And same time I did running the code simultaneous in backtesting.

    In backtesting there was no trade, in real yes!

     

    So i checked 4 other codes for this.

    Trades i did  –  i checked back with backtesting.  I saw diffirences between real trade und backtesting.

    Differences in points, differences in win or loose and differences in start or not start the trade.

    Shame on PRT, shame on IG.

    LET YOUR FINGER THEREOF.

    #12186

    Did u backtest with spreads on ?

    #12187

    sure, 2 points

    #12191

    Same timeframe? What instrument please?

    Why did you use 2 trailing stop system at the same time? The one from the platform and the MFE one from the site? Do you have any particular intention to do it? Thanks.

     

    #12193

    the problem is not 2 trailing stops, the problem is:  backtesting trades and realtime trades with the same code in to the same time  have different results !

    #12194

    Same timeframe, same instrument !

    #12195

    1:1 for my control, all is indentical! but not results of the trades!

    #12201

    What is the timeframe used already please?

    About the fact that ProBacktest don’t give same results as real live trading, this is a subject already vastly debated and discussed in all different languages here. In backtest, there is no inside bar look of what is the price touched firstly: takeprofit or stoploss, so it’s always takeprofit which is first tested.

    The “tick per tick” backtest is coming in the 10.3 version, which is still in Beta. I think it will be ready in the weeks/months to come, but this year for sure.

    #12252

    Nicolas.

    to do a correct work together

     

    check this code in backtesting and give the results here

    check from 24.aug.16 14:48:08 to yet.

    timeframe 1 hour, spread 2 , no brokerage fee

    i did running this code in real.

    So we the results in backtesting and i give the results by copy from real performance.

    try?

     

     

    //————————————————————————-
    // Hauptcode : test full Dax Long Trail8
    //————————————————————————-
    //————————————————————————-
    // Hauptcode : Test full Dax Short Trail8
    //————————————————————————-
    Set Stop Trailing 8
    trailingstop = 8

    if not onmarket then
    MAXPRICE = close
    priceexit = 0
    endif

    // Bedingungen zum Einstieg in Long-Positionen
    IF Time>080000 And Time<220000 AND Exponentialaverage [1] CROSSES OVER ExponentialAverage [6] THEN
    BUY 1 CONTRACTS AT MARKET
    ENDIF

    //case LONG order
    if longonmarket then
    MAXPRICE = MAX(MAXPRICE,close) //saving the MFE of the current trade
    if MAXPRICE-tradeprice(1)>=trailingstop*pointsize then //if the MFE is higher than the trailingstop then
    priceexit = MAXPRICE-trailingstop*pointsize //set the exit price at the MFE – trailing stop price level
    endif
    endif

    if onmarket and priceexit>0 then
    EXITSHORT AT priceexit STOP
    SELL AT priceexit STOP
    endif

     

    #12255

    I don’t need to test to see where is the problem now that I know the timeframe tested. As I told you, the 8 points trailing stop is triggered in backtest because it does test the takeprofit before the stoploss. In backtest, there is no tick per tick test of the price movement, while of course it is not the same at all in real life.

    This problem is a well known issue with the platform and you are not the first trader to encounter this behaviour 🙂

    #12271

    wellknown issue? not the first trader? and you are happy and satisfied about this?  – so  i am quite now, no problem… i leave, i am not longer interested in, thanks

    #12273

    Please keep in mind that I’m not a PRT employee, I’m a customer just like you.
    I’m definitely not satisfied with this and I’m looking forward to the tick per tick module that will come soon hopefully.

    #12312

    I believe this drawback with PRT is probably one of the first short-comings that each active client encounters. The problem lies perhaps not so much in the actual shortcoming, but in the insufficient PRT-documentation of the shortcoming, such that hours of work can be spent on something that ends up useless.

    That said we are all looking forward to a future version with look-inside-bar (also calles tick-walk-forward) back-testing functionality.

    #12313

    I think so, too. But the fact that the documentation should be improved bought me here and that’s a good thing.

    Let’s wait for the day when the PRT documentation just links here.

    By the way, maybe a FAQ will be helpfull.

Viewing 14 posts - 1 through 14 (of 14 total)

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