Weekly Candle High (by candle ID) Bug

[hamedonline]

Hi there,
I believe there's a bug regarding weekly timeframe candle high by Candle ID (maybe for other values too).
I noticed this bug also effects calculation of Highest Price (Candles Period) for weekly timeframe.

The following EA is a simple demonstration of values. As you can see, all comment output values of candle high for different week candle IDs show the previous week's (candle ID = 1) value.
https://fxdreema.com/shared/F6I4KBxUb

@fxDreema Please do check this bug immediately since it seems to be a serious and urgent one.

[miro1360]

I think it is not bug, because candle ID 1 is previous candle ... current candle have value 0

[hamedonline]

@miro1360 Please check the shared EA and run it on your metatrader platform.
Why do you think the output values for weekly candles 1, 2 and 3 are the same?

[miro1360]

because tester starts in some moment and not all candles before this moment are used for calculation, you must forward tester and ignore few candles until your higher TF is covered by candles from lower TF ...

[hamedonline]

@miro1360 In some cases this could be true, but there's another point though that falsifies this assumption. No other timeframe (except weekly as we see) with the same (Candle High) condition, produces this kind of behavior. This is why I'm skeptical and somehow sure that it is caused by a bug...

By the way, I just updated the shared EA and you can see and confirm it for yourself (no duplicate values for Candle Low!).

I'm pretty sure this is a bug and it also effects calculation of highest price by candles period (only if you choose weekly timeframe in More Settings) .
=> Market Properties -> Highest Price (Candles period) :: More settings -> timeframe: Weekly

[miro1360]

it must check admin

[fxDreema]

I tried this EA and it shows correct results to me http://prntscr.com/hmo4m1

[hamedonline]

@fxdreema ok, here's the problem. When you add the indicator in forward testing, numbers are right. The problem exists in backtesting as I have attached the samples:
https://prnt.sc/hmxpis
http://prntscr.com/hmxq42

and since we need the numbers to fine tune the variables in backtests, ea never works as expected. I even had a division by zero error while backtesting (I tracked that and noticed ea returns zero for highest price by candles period)...
I'd appreciate if you shed some light on this issue.