Question on pair trading. - page 5

New comment
 
...But unfortunately, there is no cointegration in forex. Is there a rationale?(or method)(Do not send to the wiki!)
 
YOUNGA:
...But unfortunately, there is no cointegration in forex. Is there a rationale?(or method)(Do not send to wiki!)


To put it in simple terms - cross of two cointegrated instruments should constantly "hang out" in a fairly narrow horizontal range. Have you ever seen this in forex?

And for a mathematically correct method, google it. There were a couple of research articles there

 
YOUNGA:
...But unfortunately there is no cointegration in forex. Is there a rationale?(or method)(Do not send to wiki!)

This is where I've laid it all out in full accordance with the theory. Everything DEMI writes is the principle in action of "heard the bell".

Granger got a Nobel for cointegration. The concept has been widely used in the market since the 80s. There is software support in all econometric packages.

 
faa1947:

This is where I've laid it all out in full accordance with the theory. Everything DEMI writes is the principle in action of "heard the bell".

Granger got a Nobel for cointegration. The concept has been widely used in the market since the 80s. There is software support in all econometric packages.


Two eternal questions for you personally:

1. Example of two cointegrated tools in the forex. Just please don't make weird graphs of unknown regressions..... Just two cointegrated instruments and we'll see for ourselves on the cross.

2. where is the money? Obviously trading on cointegrated instruments is arbitrage. Where?

 
Demi:


Two perennial questions for you personally:

1. an example of two cointegrated instruments in forex. Just please don't make weird charts of unknown regressions..... Just two cointegrated instruments and we'll see for ourselves on the cross.

2. where is the money? Obviously trading on cointegrated instruments is arbitrage. Where?

The concept of "cointegration" assumes that the student has some level of education, at least knowledge of regression analysis (2nd year of university). There is nothing I can explain to you personally, given your previous posts.
 
faa1947:
The concept of "cointegration" assumes a certain level of education, at least knowledge of regression analysis (2nd year of college). Personally, in view of your previous posts, I can't explain anything.


No need to explain! I beg you - please do not explain! We will see everything ourselves, like in the nineteenth century! Otherwise your explanations....

Simple and blunt - two cointegrated tools in forex. Courage!

 
YOUNGA:
...But unfortunately there is no cointegration in forex.

Well, that also needs to be proved :) If we can't find it, it doesn't mean it doesn't exist. It's like a gopher :)

I just think that the problem is that everyone is looking for cointegration in the short term. But the real cointegration should be in the long term. For the economic ties between countries are not so rigid as to return equilibrium (parity) of currencies within a few days/weeks. Economic distortions due to an unfavourable exchange rate often last for months or years. Take Switzerland or Japan.

 
Meat:

Well, that also needs to be proved :) Just because we can't find it doesn't mean it doesn't exist. It's like a gopher :)

I just think that the problem is that everyone is looking for cointegration in the short term. But the real cointegration should be in the long term. Because economic ties between countries are not so rigid as to return to equilibrium (parity) within a few days/weeks. Economic distortions due to an unfavourable exchange rate often last for months or years. Take Switzerland or Japan.

I thought that if we take 2 instruments, for example EURUSD and USDJPY, which are strongly correlated, and we look at their cointegration by constantly monitoring EURJPY, maybe we will see the cointegration on shorter timeframes?
 
Meat:

Well, that also needs to be proved :) If we can't find it, it doesn't mean it doesn't exist. It's like a gopher :)

I just think that the problem is that everyone is looking for cointegration in the short term. But the real cointegration should be in the long term. Because economic ties between countries are not so rigid as to return to equilibrium (parity) within a few days/weeks. Economic distortions due to an unfavourable exchange rate often last for months or years. Take Switzerland or Japan.

Determining cointegration is a purely mathematical problem with known algorithms, tests, constraints and unsolved problems. Of course, as in all statistics, an intuitive understanding of the reasons for the equilibrium of the two series, e.g. seasonality, is very important. But this understanding should always be supplemented by very specific calculations. And if this intuitive assumption is confirmed by calculation, then cointegration can be used in trading.
 
Meat:

Well, that also needs to be proved :) Just because we can't find it doesn't mean it doesn't exist. It's like a gopher :)

I just think that the problem is that everyone is looking for cointegration in the short term. But the real cointegration should be in the long term. Because economic ties between countries are not so rigid as to return to equilibrium (parity) within a few days/weeks. Economic distortions due to an unfavourable exchange rate often last for months or years. Take Switzerland or Japan.

That is why I have doubts about the advantage of trading in pairs. In the case of an erroneous entry it may take a long time to wait for the currencies to converge (wait for outlasting losses) or to close on a stop-loss.
1234567
New comment