System to hedge every major dollar pair for profit / low DD | Page 31 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Any tools to know what 1 pip is for every currency pair? 1 reply

Price info indicator needed - one pair's price x another pair 5 replies

Multi-pair trading or Few-Pair Trading. [poll] 15 replies

Trading 9 pair every day 143 replies

How nearly EVERY pair gap downward?! 8 replies

Attachments: System to hedge every major dollar pair for profit / low DD
Exit Attachments

System to hedge every major dollar pair for profit / low DD

  • #601
  • Jun 12, 2012 1:38pm Jun 12, 2012 1:38pm
My EA still needs more testing - it doesn't close trades properly. A new version is attached, I'll let you know when I'm confident that it works.
  • #602
  • Jun 12, 2012 2:25pm Jun 12, 2012 2:25pm
Quoting Dirtybrown
If you don't mind, could you please post your .set file for the EA? I'm thinking of trading it on th same pairs i've been using, usdcad nzdusd. What would be your recommendation based on this chart for the benefit of the thread? Thanks a million.
I really don't know what the values should be... I'd probably just eyeball the chart and enter my trigger and stoploss on the basis of where the standard deviation has gone in the past. I don't have a set file because I'm not confident that the EA works properly...

I uploaded a picture to my original post which explains the logic a bit better. You should be able to see what values go where:
http://www.forexfactory.com/attachme...1&d=1339528978
  • #603
  • Jun 12, 2012 2:36pm Jun 12, 2012 2:36pm
Yes this makes it very clear thanks a lot. Now anyone on the thread can also try to see if they see a problem in the code logic.

Quoting TheMaxx
I really don't know what the values should be... I'd probably just eyeball the chart and enter my trigger and stoploss on the basis of where the standard deviation has gone in the past. I don't have a set file because I'm not confident that the EA works properly...

I uploaded a picture to my original post which explains the logic a bit better. You should be able to see what values go where:
http://www.forexfactory.com/attachme...1&d=1339528978
Skype: heliosphan187
  • #604
  • Jun 12, 2012 2:53pm Jun 12, 2012 2:53pm
Quoting TheMaxx
I really don't know what the values should be... I'd probably just eyeball the chart and enter my trigger and stoploss on the basis of where the standard deviation has gone in the past. I don't have a set file because I'm not confident that the EA works properly...

I uploaded a picture to my original post which explains the logic a bit better. You should be able to see what values go where:
http://www.forexfactory.com/attachme...1&d=1339528978
Well the way I understand it it can't just sell above buy below thought it was dependant on if the numbers are + or - to what the direction of the trade should be ?
  • #605
  • Jun 12, 2012 4:03pm Jun 12, 2012 4:03pm
Quoting turboguru
Well the way I understand it it can't just sell above buy below thought it was dependant on if the numbers are + or - to what the direction of the trade should be ?
When I say buy or sell I mean the EA buys or sells the hedge ratios shown. So in the picture it would sell 1.00 NZDUSD and sell -0.77 USDCAD (which means buy 0.77 USDCAD)
  • #606
  • Jun 12, 2012 4:38pm Jun 12, 2012 4:38pm
Quoting TheMaxx
When I say buy or sell I mean the EA buys or sells the hedge ratios shown. So in the picture it would sell 1.00 NZDUSD and sell -0.77 USDCAD (which means buy 0.77 USDCAD)
K just makin sure thanks
  • #607
  • Jun 12, 2012 8:52pm Jun 12, 2012 8:52pm
Just found this thread, it's surprising because I never thought that anyone would manage to develop anything this useful for retail trading...

Trying out the ea.

Edit:

I also forgot to ask if anyone tried doing the cointegration search that one of you mentioned?
  • #608
  • Jun 12, 2012 10:22pm Jun 12, 2012 10:22pm
A whole new world of chart analysis unfolds when you look at these R charts. Metatrader's role can be properly minimized to that of an order entry api. You can have R and a database control when to trigger trades and send the results back to metatrader using python.


Here's a cigar to former meat traders!

Quoting saussiche
Just found this thread, it's surprising because I never thought that anyone would manage to develop anything this useful for retail trading...

Trying out the ea.

Edit:

I also forgot to ask if anyone tried doing the cointegration search that one of you mentioned?
Skype: heliosphan187
  • #609
  • Jun 12, 2012 11:04pm Jun 12, 2012 11:04pm
A few things,

What has everyone set for the diff parameter?

Also, I have this all running on the amd64 version of R, it seems to work fine (someone mentioned it wouldn't).

And the latest version of the ea isn't adjusting the sizes correctly, it attempts to open full lots on a small account.

Has anyone developed tools for multicointegration testing?

Are there any plans to change the code to work with crosses?

Statistics isn't my strong suite so I'm hoping you guys figure all of this stuff out.
  • #610
  • Jun 12, 2012 11:32pm Jun 12, 2012 11:32pm
I use the arbomat_demo, with defaults. Others will need to say what they use for the final release arbomat version.

I believe you are running the basket EA. You should first run arbomat manually to get a feel for how it works. The "Basket" version is still under testing.

I tried to get a non usd cross working based on pip values, couldn't get it to work. Maybe someone can figure it out or use my code as a requirement. 7bit mentioned early that the hedge rations "didn't work" for non usd pairs. Maybe a flaw in the code, not sure. Search the thread for fxez's posts, he showed some tests in R. I think I have some matlab code that can calculate hedge ratios I will try to find it.



Quoting saussiche
A few things,

What has everyone set for the diff parameter?

Also, I have this all running on the amd64 version of R, it seems to work fine (someone mentioned it wouldn't).

And the latest version of the ea isn't adjusting the sizes correctly, it attempts to open full lots on a small account.

Has anyone developed tools for multicointegration testing?

Are there any plans to change the code to work with crosses?

Statistics isn't my strong suite so I'm hoping you guys figure all of this stuff out.
Skype: heliosphan187
  • #611
  • Jun 13, 2012 12:32am Jun 13, 2012 12:32am
Quoting Dirtybrown
A whole new world of chart analysis unfolds when you look at these R charts.
130% agreed ! And I can tell as a reguler user that R has nearly no limit: you can do model fitting, simulations, forecasting (ie: just google "Automatic Time Series Forecasting: The forecast Package for R" for instance). Output of the best of statistical modeling technics now can go directly to your MT4 to place orders smartly... And if it s not enough yet (which I doubt), you can write your own R programs

Quoting Dirtybrown
You can have R and a database control when to trigger trades and send the results back to metatrader using python.
I don't get it... I mean, the mt4R.mqh already allow getting results from R... So what Python is here for ?
  • #612
  • Jun 13, 2012 1:01am Jun 13, 2012 1:01am
Thanks for the suggestions will have a look. It's true, arbomat just scratches the surface of what is possible. Shortly I will be posting some test results based on FXEZ's sample code.

We should brainstorm some new R functions, will get back to you once I have the examples done.

And I think you are right about python, I'm half asleep its 2 am here

Quoting Fx Juju
130% agreed ! And I can tell as a reguler user that R has nearly no limit: you can do model fitting, simulations, forecasting (ie: just google "Automatic Time Series Forecasting: The forecast Package for R" for instance). Output of the best of statistical modeling technics now can go directly to your MT4 to place orders smartly... And if it s not enough yet (which I doubt), you can write your own R programs



I don't get it... I mean, the mt4R.mqh already allow getting results from R... So what Python is here for ?
Skype: heliosphan187
  • #613
  • Jun 13, 2012 2:18am Jun 13, 2012 2:18am
I thought the python code that 7bit wrote was magically retrieving data for mt4?

Something like that?

Haha I probably shouldn't be messing with all of this stuff.
  • #614
  • Jun 13, 2012 3:07am Jun 13, 2012 3:07am
Quoting Dirtybrown
I think I have some matlab code that can calculate hedge ratios I will try to find it.
See attached.
Attached File(s)
File Type: txt example3_6.txt   3 KB | 1,101 downloads
  • #615
  • Jun 13, 2012 12:28pm Jun 13, 2012 12:28pm
Quoting naser007
Hi All

i test this system on 5 baskets
you can for see my result go on http://www.myfxbook.com/members/nase...hedge-r/322867

good luck
Hi All
Today profit is : +2.41%
Today pips : 127 Pips

http://www.myfxbook.com/members/nase...hedge-r/322867
  • #616
  • Jun 13, 2012 12:38pm Jun 13, 2012 12:38pm
Nice work. It will also be helpful if you can take a screenshot when you enter or exit the basket so we can learn from your experience. Thanks for your contribution Naser, look forward to hear more.

Quoting naser007
Hi All
Today profit is : +2.41%
Today pips : 127 Pips

http://www.myfxbook.com/members/nase...hedge-r/322867
Skype: heliosphan187
  • #617
  • Jun 13, 2012 3:43pm Jun 13, 2012 3:43pm
Quoting Dirtybrown
Nice work. It will also be helpful if you can take a screenshot when you enter or exit the basket so we can learn from your experience. Thanks for your contribution Naser, look forward to hear more.
under his open trades, it shows a 133 pip loss.

Also, I opened a 1H NZDUSD / USDCAD hedge yesterday and left it over night. It hasn't budged unfortunately.

It's too bad we can't have a quantitative strategy that actually outperforms a typical working ta system. I assume they exist...
  • #618
  • Edited 11:04am Jun 14, 2012 10:03am | Edited 11:04am
Quoting naser007
Hi All
Today profit is : +2.41%
Today pips : 127 Pips

http://www.myfxbook.com/members/nase...hedge-r/322867
hi All
Up to this time of today profit is : 1.92%
Today Pips : 112.70 Pips

i think this System is HolyGrail, but before your must Find Best basket and TP

more information's Please see http://www.myfxbook.com/members/nase...hedge-r/322867



Update : Today profit up to : +2.64% and 160 pips
  • #619
  • Jun 14, 2012 12:09pm Jun 14, 2012 12:09pm
Quoting naser007
i think this System is HolyGrail
Oh no. Now you've jinxed it.
You are in a maze of twisty little passages, all alike.
  • #620
  • Edited 3:25pm Jun 14, 2012 1:08pm | Edited 3:25pm
As far as I'm concerned, my pair trades posted in usdcad nzdusd is stable and profitable. I'm using it with real money. It's quite easy to trade and analyze the price pattern. There is some shifting, similar to "repainting" in metatrader. The key is to find cointegrated pairs or baskets.

Though typically the augmented dickey fuller tests are used on several dozen individual pairs, usdcad nzdusd pass this test. They do not qualify under Granger cointegration however -> http://en.wikipedia.org/wiki/Cointegration

In my experience on this pair (usdcad, nzdusd), ADF is sufficient to prove cointegration. Additionally, the hedge ratios barely budge using 1 hr data on the Arbomat demo version with default settings. This is another confirmation of cointegration, specifically covariance. Higher covariance = less coefficient shifting = less or no need to reblanace your hedge proportions.

Quoting saussiche
under his open trades, it shows a 133 pip loss.

Also, I opened a 1H NZDUSD / USDCAD hedge yesterday and left it over night. It hasn't budged unfortunately.

It's too bad we can't have a quantitative strategy that actually outperforms a typical working ta system. I assume they exist...
Skype: heliosphan187
0 traders viewing now