RC 106 Backtest problem with EURUSD M1 Strategy

Hi Mark,
the following strategy shows different backtests on SQ 3.8.2 and SQ 4.106
This Strategy is a EURUSD M1 winning Strategy on jfd-Demoaccount.

See the following graphic. (I attached this graphic this report too, if this graphic is not viewable below)

I have done the following tests (I attached the files too)
1) Backtest of EURUSD M1 35.31 with 3.8.2 and duka data (M1 precision), all backtests same time
2) Backtest result with M1 precision (This backtests fits the trades on Metatrader 4 account you can see myfxbook)
3) Backtest with 3.8.2 (Ticksimmulation)
4) Backtest results of ticksimmulation -> it fails, this strategy is loosing now, you can see attached pic
5) Backtest with SQ4.106 with sel timeframe precision
6) Backtest results of sel timeprecison with SQ4 -> it fail totals, the trades are totally different, trade # totaly different
7) Backtest with SQ4.106 ticksimmulation
8) Backtest results SQ.106 ticksimmulation SQ4 -> it fail totals, the trades are totally different, trade # totaly different
9) Results of myfxbook on jfd demoaccount. You can see the result is nearly the result of 3.8.2 (M1 precision backtest) -> this is a winning strategy
on jfd-demoaccount
(On jfd-Real account it is winng too, but the profit is not so high)

I attached all backtestresults and the strategy, you can doublecheck it.
I thing something is not 100% correct with the SW 4.106 backtester oder Indicators of this strategy ?

thomas
Attachments
21.rar
(910.19 KiB)
  • Votes 0
  • Project StrategyQuant X
  • Type Bug
  • Status Fixed
  • Priority Normal

History

t
#1

tnickel

16.07.2018 12:41

Task created

t
#2

tnickel

16.07.2018 12:55

You can look custom analysis for the correct ea. (There are two on this demoaccount)


https://www.myfxbook.com/members/tnickel/j76-jfd747-p79-eurusd-m1/2324201

MF
#3

Mark Fric

10.08.2018 11:36
there seems to be problem with Pivots indicator, it is also very slow in SQ X. I'll look at it further but in the next build.
MF
#4

Mark Fric

15.08.2018 08:34
zkus prilozenou strategii - jednak mysalim nesedi testy mezi SQ a MT4, ale taky pocitani Pivot indikatoru je extremne pomale
TB
#5

Tomas Brynda

31.08.2018 11:35

Status changed from New to In progress

MF
#6

Mark Fric

13.09.2018 07:54

Attachment Image3.png added

Attachment Image1.png added

Attachment Image2.png added

jak presne ty testy spoustis?

Ted jsem to testoval na jinem MT4, a prochazi mi to.


Jak jsem to testoval:

- do MT4 jsem nahral vsechny aktualni SQX indikatory z custom_indicaotrs

- spustil jsem SQX, nacetl jsem si .str strategii a ziskal jeji MQL kod

- v MT4 jsem si vytvoril novou strategii, skontorloval,ze mela vypnute vsechny trading options, spustil

- ulozil jsem si result z MT4, tu samou strategii jsem spustil jako test v TestStrategiesMT4 - tam to proslo

- tu samou strategii jsem spustil se stejnym nastavenim v Retesteru - tam jsou odchylky.


Takze v testu to prochazi, ale v Retesteru ne. Musi tam byt jeste nejaky rozdil v nastavenich, podivas se na to?


Ale asi to neni problem Keltner Channelu.




Dalsi vec - asi mas chybu v implementaci MTKC v SQ. Mas tam:

@Override

public double OnBlockEvaluate(int relativeShift) throws TradingException {


switch(OutputIndex){

case 0: return Indicators.MTKeltnerChannel(Input, Period, Deviation).Upper.get(relativeShift + Shift);

case 2: return Indicators.MTKeltnerChannel(Input, Period, Deviation).Lower.get(relativeShift + Shift);

default: throw new TradingException("Invalid output index: " + OutputIndex);

}

}


Ten druhy case ma byt case 1:, protoze Lower je druhy output buffer v poradi. Ted Lower buffer nebude vubec fungovat, asi se ani nikdy nevola.

MF
#7

Mark Fric

13.09.2018 07:55
jeste poznamka - submitnul jsem tu strategii P79 i vysledky MT testu do reposibory
TB
#8

Tomas Brynda

16.10.2018 18:03

Status changed from In progress to Fixed


Votes: 0

Drop files to upload

or

choose files

Max size: 5MB

Not allowed: exe, msi, application, reg, php, js, htaccess, htpasswd, gitignore

...
Wait please