Vizualisation Backtest

New comment
 

Hello,


does anyone have a problem with vizualsation on MT5 Backtest?

and there is no difference, if the vizualisation in on or not.


let me explain, 

i've tested an simple function for inside Bars


bool InsideBarShort(const string sym,const ENUM_TIMEFRAMES tf,const int shift)
  {
   bool insidebar=false;
   double High1=iHigh(sym,tf,shift);
   double Low1=iLow(sym,tf,shift);

   double High2 = iHigh(sym,tf,shift+1);
   double Low2  =iLow(sym,tf,shift+1);

   double Spanne2;

   Spanne2=High2-Low2;

   if(//High1<(High2-((Spanne2/100)*ProzentHoeheBarvorInsideBar)) && 
      High2>= High1 &&
      Low2 <= Low1)
     {
      Print("-----------Low1: ",Low1," Low2: ",Low2, "High: ",High2-High1," Low: ",Low1-Low2);
      insidebar=true;
     }

   return(insidebar);
  }

this is written in an #include,


in the onTick i use a very simple function

 if(Hour()>=Start && Hour()<=End && 
         InsideBarShort(_Symbol,PERIOD_M5,1) == true 

         )
        {
         OrderEinstig = High(_Symbol,PERIOD_M5,1);
         OrderStopLoss = High(_Symbol,PERIOD_M5,2)+SL*_Point;
         OrderTakeProfit = (OrderEinstig - TP*_Point);
         
         PendingOpen(_Symbol,ORDER_TYPE_SELL_LIMIT,Lots, OrderEinstig, OrderStopLoss, OrderTakeProfit,MN,Text,TimeCurrent()+60*60);

        }


when i start the backtest, i get following results




when i look to the time in the report, everything is right, only the vizualisation failed.


does anyone else have this problem?


amando

 

I am afraid I have no idea what is the problem you are trying to show ?

Please provide better (and with no time/prices scales cut off). It's so small I can't read them.

 
amando:

Hello,


does anyone have a problem with vizualsation on MT5 Backtest?

and there is no difference, if the vizualisation in on or not.


let me explain, 

i've tested an simple function for inside Bars


this is written in an #include,


in the onTick i use a very simple function


when i start the backtest, i get following results




when i look to the time in the report, everything is right, only the vizualisation failed.


does anyone else have this problem?


amando

I believe we have the same problem. Arrows at vizualisation are placed incorrectly. It seems to be a bug, but it could be also dessincronized data. I just turned off plotting arrows and do tape reading check.

But as Alain said, please post full screen, erasing only sensitive account information and without resizing.

 
Alain Verleyen:

I am afraid I have no idea what is the problem you are trying to show ?

Please provide better (and with no time/prices scales cut off). It's so small I can't read them.

I believe I have the same problem. Arrows are being plot somewhere at different prices. Sometimes I think it's a bug, sometimes I realized it's database quality, such as dessincronization or invalid Ask/Last/Bid values.

 
Arthur Albano:I believe I have the same problem. Arrows are being plot somewhere at different prices. Sometimes I think it's a bug, sometimes I realized it's database quality, such as dessincronization or invalid Ask/Last/Bid values.

Ever consider that there is nothing wrong, but that you are forgetting about the spread?

OHLC Bar/Candle prices are all Bid prices only, but Buy Opens(In) and Sell Closes(Out) at the Ask Price (not Bid) and those may seem quite off from the bars/candles on the chart, especially on M1 - M5 time-frames.

 
Arthur Albano:

I believe I have the same problem. Arrows are being plot somewhere at different prices. Sometimes I think it's a bug, sometimes I realized it's database quality, such as dessincronization or invalid Ask/Last/Bid values.

Well I don't know from the information provided.

 
Alain Verleyen:

Well I don't know from the information provided.

Indeed. I will investigate itm more deeply, but AFAIK, Last could be outside Ask/Bid, and Ask/Bid moves and there is no closed deal (no aggression). On the simulator, if it uses Ask/Bid prices and not last, results are unrealistic and arrows are placed outside the candle. I posted a topic where I asked a different simulation, "every last tick" (or every price changing tick) mode. But I had no answer. My answer to it was placing limit orders after data-filtering and validation, such as:

Ask>Bid; (after market hours and before market open there are bid/ask ticks, but there are no price changing ticks)

MqlDate.hour>=market open hour;

(Last >= Ask || Last =< Bid)

And so on...

He is simulating using: "Every tick based on real ticks".

But as you say, there is need for full print-screens and also, IMHO, a print screen on tick data surrounding the position/deal. "Read the source Luke!" :)

 
Arthur Albano:

Indeed. I will investigate itm more deeply, but AFAIK, Last could be outside Ask/Bid, and Ask/Bid moves and there is no closed deal (no aggression). On the simulator, if it uses Ask/Bid prices and not last, results are unrealistic and arrows are placed outside the candle. I posted a topic where I asked a different simulation, "every last tick" (or every price changing tick) mode. But I had no answer. My answer to it was placing limit orders after data-filtering and validation, such as:

Ask>Bid; (after market hours and before market open there are bid/ask ticks, but there are no price changing ticks)

MqlDate.hour>=market open hour;

(Last >= Ask || Last =< Bid)

And so on...

He is simulating using: "Every tick based on real ticks".

But as you say, there is need for full print-screens and also, IMHO, a print screen on tick data surrounding the position/deal. "Read the source Luke!" :)

Arthur this topic is about Forex (EURUSD), there is no "Last" price implied here (on Forex last price is always equal to bid and is of no use). Please try to not confuse people ;-) I would suggest to open you own topic to report your observations.
New comment
Reason: