Cointegration of currency pairs | Page 3 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Currency Pairs - Same Currency, Different Direction? 7 replies

Synthetic hedges, cointegration, mean reversion and similar stuff 419 replies

Trading the cointegration 31 replies

Pairs of Currency Pairs 4 replies

Cointegration, Synthetic hedges, mean reversion in R, Tech Thread 0 replies

Attachments: Cointegration of currency pairs
Exit Attachments

Cointegration of currency pairs

  • #41
  • Mar 16, 2009 2:25pm Mar 16, 2009 2:25pm
Quoting c0d3
My reseach shows, hedge funds do global analysis in about 4hrs.

I'm curious how you came to this conclusion? Since this is a task which is easy to run in parallel, given enough computer power, it could be done in seconds.
  • #42
  • Jul 16, 2009 7:38am Jul 16, 2009 7:38am
Hi c0d3 & knocks420,

You stay in the co-integration stuff on Forex ?

You have a nice return ?

Knocks420 you auto trade with matlab ? how you send order of matlab to (ts or mt) ?

Regards.
Ludo.
  • #43
  • Feb 8, 2010 2:22pm Feb 8, 2010 2:22pm
Been testing cointegration possibilities on major currency pairs but doesn't seem to yield any great results. Anyone else been testing this relationship these days? Would any think its viable still as I'm close to giving up on this idea working out in the forex market at the moment.

SP
  • #44
  • Mar 13, 2010 7:52pm Mar 13, 2010 7:52pm
Thank you. I just wanted to say that. I wish more posts were like this one right here.

Quoting fxid10t
Using the definitions of the concept from here and here, the attached indicator for MT4 attempts to display cointegration data from the pairs under consideration similar to this.

[code]...
Time flies like an arrow; fruit flies like a banana.
  • #45
  • Mar 25, 2010 3:54pm Mar 25, 2010 3:54pm
c0d3 pm please.
  • #46
  • Jun 3, 2011 3:15am Jun 3, 2011 3:15am
Hi all, how far did everyone get with testing currency pairs for cointegration?

I previously coded a script in R which would read my CSV file for a list of US stocks and then run every pair in the list through a co-integration test using data automatically downloaded from Yahoo Finance (daily closing prices), also managed to spew out a beta co-efficient for current position sizing.

I never ended up pushing forward with this because I did not have the resources or skills required to continue such a project - there was too much work to do designing a system for equities. Things like survivorship bias for example were obstacles that I didn't know how to overcome.

But FX seems like a far better avenue to explore for a number of reasons:
- reduced margin requirements
- 24/5 market
- very high liquidity
Just wondering if people have found that opportunities do in fact exist for co-integration based stat arb in the forex markets?
  • #47
  • Jun 5, 2011 1:24pm Jun 5, 2011 1:24pm
Pairtradefinder now uses cointegration rather than correlation and they say their software works well with currencies. I use PTF for equities but doubt very much it works with FX. The big boys will have sorted that out by now if it did.
He was looking for the card so high and wild hed never need to deal another
  • #48
  • Jun 6, 2011 5:12am Jun 6, 2011 5:12am
Yes I would like to see the new version of PTF but don't really want to pay to upgrade as I don't believe the software is that great.
I just compiled a list of currently cointegrated FX pairs. Not even sure what kind of strategy to test on them. Becomes even harder when you have to account for the exchange rates with position sizing.
  • #49
  • Sep 1, 2011 11:47am Sep 1, 2011 11:47am
Perhaps you should look to this. If somebody is interested in cooperation
with cointegration analysis usong MATLAB and toolbox form http://www.spatial-econometrics.com/ Let me know.

Krzysztof
Attached File(s)
File Type: pdf cointegration.pdf   425 KB | 2,713 downloads
  • #50
  • Sep 2, 2011 2:20am Sep 2, 2011 2:20am
Here is a complete MATLAB strategy and pdf describing pairs trading
Attached File(s)
File Type: pdf cointegration-en.pdf   284 KB | 2,961 downloads
File Type: zip cointegration.zip   27 KB | 1,319 downloads
  • #51
  • Sep 2, 2011 10:53am Sep 2, 2011 10:53am
Quoting fajst_k
Perhaps you should look to this. If somebody is interested in cooperation
with cointegration analysis usong MATLAB and toolbox form http://www.spatial-econometrics.com/ Let me know.

Krzysztof
The attached pdf file cointegration.pdf appears to be corrupt and won't open with Adobe Reader.
You are in a maze of twisty little passages, all alike.
  • #52
  • Sep 2, 2011 11:09am Sep 2, 2011 11:09am
Quoting stevegee58
The attached pdf file cointegration.pdf appears to be corrupt and won't open with Adobe Reader.
I just tried to download it and its OK. Perhaps you should try to download it again
  • #53
  • Sep 2, 2011 12:08pm Sep 2, 2011 12:08pm
Quoting fajst_k
I just tried to download it and its OK. Perhaps you should try to download it again
I downloaded again and it's OK now.
You are in a maze of twisty little passages, all alike.
  • #54
  • Sep 4, 2011 12:19am Sep 4, 2011 12:19am
first page of this thread is hilarious..

for discussion how important is co integration when scalping of intraday time frames for stock/futures pairs?
Grabbing the bull by the horns!
  • #55
  • Sep 4, 2011 1:00am Sep 4, 2011 1:00am
Quoting whitegoodman
first page of this thread is hilarious..

for discussion how important is co integration when scalping of intraday time frames for stock/futures pairs?

TF does not matter here. For sure will be easier to scalp cointegrated pair as in this case there is guarantee that Technical Analysis based strategy will work due to stationarity of data.

Krzysztof
  • #56
  • Sep 4, 2011 1:45am Sep 4, 2011 1:45am
I think you are talking about Leading indicator. There is some kind of study that the differential of the bond rates 10 year maturity of the countries in question are a leading indicator of price future path. However you should make some backtesting before using it in a trading system.

Check the rates here http://markets.ft.com/RESEARCH/Markets/Data-Archive

Good Luck!
  • #57
  • May 26, 2012 12:07pm May 26, 2012 12:07pm
Is there any consensus on which forex pairs have high degree of cointegration, historically?
  • #58
  • May 26, 2012 1:55pm May 26, 2012 1:55pm
Hey mwduke.

I've read that EURUSD and EURCHF have, due to the linked economies. Since the CHF cap, that has most likely changed.

It will be interesting to know. If you find anything, put it on the thread.
  • #59
  • May 26, 2016 11:07am May 26, 2016 11:07am
I am working on an automatized quantitative strategy that use cointegration in Forex. I am backtesting this strategy in Python.

Please see below the python file:
https://drive.google.com/file/d/0B1A...ew?usp=sharing
See below the data that I used to backtest:
https://drive.google.com/file/d/0B1A...ew?usp=sharing

My Algorithm is:

  1. Read File
  2. Transform the data with the log return
  3. Treat Outliers
  4. Realize the linear regression
  5. Test for cointegration with ADF test
  6. If the spread is stationary then apply the best ARMA model
  7. Forecast and using beta calculation to control the risk

The equity curve of this strategy is attached below.

Could you please help to understand how can I improve this algo?

Thanks

Attached Image
David
  • #60
  • May 26, 2016 7:09pm May 26, 2016 7:09pm
Quoting sushiparlour
Been testing cointegration possibilities on major currency pairs but doesn't seem to yield any great results. Anyone else been testing this relationship these days? Would any think its viable still as I'm close to giving up on this idea working out in the forex market at the moment. SP
It works on stocks better rather on currency market. I think on forex its not about math, you're not building portfolio there, so trading algorithms in some cases can be completely different.
0 traders viewing now