Dukascopy
 
 
Wiki JStore Search Login

Something wrong with historical tester
 Post subject: Something wrong with historical tester Post rating: 0   New post Posted: Mon 30 Sep, 2013, 22:20 

User rating: 1
Joined: Tue 10 Jul, 2012, 11:51
Posts: 25
Location: Czech RepublicCzech Republic
Hi, when testing with last week data, it wrongly returns value of shifted indicator. I tried with CCI and RSI. Look at attached picture. When you shift indicator by 1, it actually returns value of it shifted by 2 (simply it multiplies the original shift by 2). Is there a bug or am I doing something wrong? Thanks.
Image


Attachments:
temp.vfs [21.76 KiB]
Downloaded 176 times
cci_shift1.jpg [382.59 KiB]
Downloaded 445 times
DISCLAIMER: Dukascopy Bank SA's waiver of responsability - Documents, data or information available on this webpage may be posted by third parties without Dukascopy Bank SA being obliged to make any control on their content. Anyone accessing this webpage and downloading or otherwise making use of any document, data or information found on this webpage shall do it on his/her own risks without any recourse against Dukascopy Bank SA in relation thereto or for any consequences arising to him/her or any third party from the use and/or reliance on any document, data or information found on this webpage.
 
 Post subject: Re: Something wrong with historical tester Post rating: 0   New post Posted: Mon 30 Sep, 2013, 23:48 
User avatar

User rating: 3
Joined: Tue 09 Oct, 2012, 11:59
Posts: 20
Location: Finland, Turku
Does it with MA's also 20 sift is 40 etc, makes it a pain in the... been just halving the value for visual and trying to remember it for normal.


 
 Post subject: Re: Something wrong with historical tester Post rating: 0   New post Posted: Mon 07 Oct, 2013, 20:00 

User rating: 1
Joined: Tue 10 Jul, 2012, 11:51
Posts: 25
Location: Czech RepublicCzech Republic
Hi vjf support,
Can you please give us an estimate when this will be fixed? Thank you very much.


 
 Post subject: Re: Something wrong with historical tester Post rating: 0   New post Posted: Tue 08 Oct, 2013, 07:21 
Visual JForex expert at Dukascopy
User avatar

User rating: 27
Joined: Fri 03 May, 2013, 18:25
Posts: 366
Location: GermanyGermany
Greetings!

I do not see any errors on that! Please bear in mind that "shift 0" means the current candle (in that expample), "shift 1" last candle and so on.
Hence it follows that in your picture the marked candles are wrong - it has to be every last candle.
When i tested the strategy i did not see any errors or bugs :)


 
 Post subject: Re: Something wrong with historical tester Post rating: 0   New post Posted: Tue 08 Oct, 2013, 11:25 

User rating: 1
Joined: Tue 10 Jul, 2012, 11:51
Posts: 25
Location: Czech RepublicCzech Republic
Hi,
thanks for your reply. I must do something wrong or miss something. Or maybe a bug is there.. :?
Please try to run the strategy again and put there a breakpoint in the end. When you get first 5min candle, the indicator get some values:
shift1=value a, shift2=value b, shift3=value c, etc..
when you get 2nd 5min candle, values should be shifted by 1:
shift2=value a, shift3=value b, shift4=value c

but it doesn't work this way. you get totally different numbers. Right after 3rd candle, you get the values that you're supposed to get at 2nd candle.

It seems to me that shift(x) actually returns value of shift(x*2)
Hope it's understandable.

PS: tested by historical tester in vjf (last week data).


 
 Post subject: Re: Something wrong with historical tester Post rating: 0   New post Posted: Tue 08 Oct, 2013, 14:08 
User avatar

User rating: 3
Joined: Tue 09 Oct, 2012, 11:59
Posts: 20
Location: Finland, Turku
Maybe I have bug in my knowledge about sift.

With MA sift 30 i get the MA price of sift 60.


Image
Image


Attachments:
B_1_30_30.vfs [14.21 KiB]
Downloaded 158 times
B_1_30_30_C.png [101.57 KiB]
Downloaded 405 times
B_1_30_30_B.png [116.56 KiB]
Downloaded 459 times
DISCLAIMER: Dukascopy Bank SA's waiver of responsability - Documents, data or information available on this webpage may be posted by third parties without Dukascopy Bank SA being obliged to make any control on their content. Anyone accessing this webpage and downloading or otherwise making use of any document, data or information found on this webpage shall do it on his/her own risks without any recourse against Dukascopy Bank SA in relation thereto or for any consequences arising to him/her or any third party from the use and/or reliance on any document, data or information found on this webpage.
 
 Post subject: Re: Something wrong with historical tester Post rating: 0   New post Posted: Mon 21 Oct, 2013, 21:03 

User rating: 1
Joined: Tue 10 Jul, 2012, 11:51
Posts: 25
Location: Czech RepublicCzech Republic
Hi gentlemen,
sorry for bothering again. Is there anything new about this? It seems the problem still persists.
Thx


 
 Post subject: Re: Something wrong with historical tester Post rating: 0   New post Posted: Tue 22 Oct, 2013, 08:56 
Visual JForex expert at Dukascopy
User avatar

User rating: 27
Joined: Fri 03 May, 2013, 18:25
Posts: 366
Location: GermanyGermany
Hi guys!

I did test it again and I created an investigation.

Unfortunately it seems that you are right!

Thank you for your that! :)


 
 Post subject: Re: Something wrong with historical tester Post rating: 0   New post Posted: Thu 31 Oct, 2013, 20:54 

User rating: 0
Joined: Sat 05 Oct, 2013, 13:54
Posts: 6
Location: SpainSpain
Same problem with MACD. Takes the values of shifted histograms multiplying by 2, and another thing. With every new candle, the histogram shifted by 1 is moved 1 position back. So, with the first candle, it's all ok. With the next candle, shifted 1 takes the value of the previous candle#2, shifted 2 candle#4, shifted 3 candle#6. With the next candle, shifted 1 now is previous candle#3, shifted 2 candle#5, etc.


 
 Post subject: Re: Something wrong with historical tester Post rating: 0   New post Posted: Sat 02 Nov, 2013, 14:17 
User avatar

User rating: 0
Joined: Tue 22 Oct, 2013, 16:46
Posts: 12
It seems to be general issue.
Bbands have similar error : viewtopic.php?f=141&t=50284


 
 Post subject: Re: Something wrong with historical tester Post rating: 0   New post Posted: Mon 18 Nov, 2013, 19:45 

User rating: 1
Joined: Tue 10 Jul, 2012, 11:51
Posts: 25
Location: Czech RepublicCzech Republic
Hello gentlemen,
any guess when we can expect this fixed? Many thanks.


 
 Post subject: Re: Something wrong with historical tester Post rating: 0   New post Posted: Wed 11 Dec, 2013, 12:22 
Visual JForex expert at Dukascopy
User avatar

User rating:
Joined: Mon 22 Apr, 2013, 11:30
Posts: 583
Location: UkraineUkraine
Greeting.

It will be fixed in the next version update.

Keep Calm and Carry On.


 
 Post subject: Re: Something wrong with historical tester Post rating: 0   New post Posted: Tue 17 Dec, 2013, 17:32 
User avatar

User rating: 0
Joined: Wed 20 Nov, 2013, 07:35
Posts: 7
Location: United Kingdom,
Hi, what is the ETA for the update?

Is there a workaround? ie somehow change the JFX code, once have built the strategy in VJFX?

thanks


 

Jump to:  

cron
  © 1998-2011 Dukascopy® Bank SA
On-line Currency forex trading with Swiss Forex Broker - ECN Forex Brokerage,
Managed Forex Accounts, introducing forex brokers, Currency Forex Data Feed and News
Currency Forex Trading Platform provided on-line by Dukascopy.com