Not the Grail, just a regular one - Bablokos!!! - page 63

 
Demi:


1. EURUSD and GBPUSD are not cointegrated. You open the cross chart and everything is obvious.

2. if these series were cointegrated, no regression would be necessary. Co-integrated instruments are traded on convergence from channel boundaries. It is only a question of finding the optimal weights of the instruments in the portfolio.

3. I demand that you immediately stop mocking econometrics! This is already taking on a systemic character! Enough!

Whatever you say, Your Majesty.
 
Avals:


The basis of paired trading is cointegration, and we cannot use correlation. Co-integration can be estimated even visually - it's flatness. I.e. the tendency of the kotyr to return to the average for example. Right now eurusd and usdchf are cointegrated. It can be seen in the eurchf cross. But the flatness is in a very narrow range.

Co-integration is based on the property that the greater the deviation, the more likely the return. The economic sense is that some participants have reasons to trade for convergence. Trying to jump in before them. So we need to understand the reasons why the instruments are now cointegrated, rather than fitting everything into a flat.

Good correlation - reliable return. But too good - small divergence - small profit.
 
Avals:


Co-integration can even be assessed visually - it's flatness

Completely wrong. Co-integration is the 'similarity' of the movement of the rows. The cohesion of their movement is always expressed as a return to some average. However, we must never forget that in quotients, unlike in geology, there is always a deterministic component, which must be removed from the quotient before applying statistical methods. Otherwise we are commensurate with these deterministic components. Hence my 40 pips and your fear of the spread eating away at the deviation from the mean.

 
Demi:


1. EURUSD and GBPUSD are not cointegrated. You open the cross chart and everything is obvious - there is no "eternal" flat.

2!


For a clearer discussion, take not currencies. Much more suitable correlated (and possibly cointegrated) commodities - BRENT oil and fuel oil (tickers BRN and HO, e.g. in mt4 ForexClub)

The fluctuation amplitude of the synthetic instrument(BRN-HO) in the last few months does not exceed $450-500 per lot.

 
leonid553:

For a better discussion, do not take currencies. Much more suitable correlated (and possibly cointegrated) commodities - BRENT oil and fuel oil (tickers BRN and HO, e.g. in mt4 Forex Club)

The fluctuation amplitude of the synthetic instrument(BRN-HO) in the last few months does not exceed $400 per lot.

Possibly.
and what does that tell you?
 
Demi:
Possible.
and what does that tell you?


The question is not for me. I didn't come here to argue, but to watch and listen to further discussion and reasoned opinions of "interested parties".

(In my chart above on the big history the index may incorrectly display the synthetic (spread) line BRN-HO 2 times a month, because on the 15th-16th of each month there is an expiry of the BRN contract and on the 29th-30th - expiry of the H0 fuel oil contract. On those days there may be a small artificial gap on the gluts of those instruments in mt4 FC).

 
leonid553:

For a better discussion, do not take currencies. Much more suitable correlated (and possibly cointegrated) commodities - BRENT oil and fuel oil (tickers BRN and HO, e.g. in mt4 ForexClub)

The fluctuation amplitude of the synthetic instrument(BRN-HO) in the last few months does not exceed $450-500 per lot.

1. Why take these particular quotes?

2. How is the synthetic instrument calculated?

3. Where is the entrance of the pose?

4. Where is the exit from the pose?

 

=============

I happen to find that the correlation of these instruments is much greater than the correlation of any other commodity!

The synthetic instrument BRN-HO is calculated and drawn as the difference (in dollars) of these component symbols for the position size ratio 1:1 (download the indicator and see the description at http://www.procapital.ru/showpost.php?p=813139&postcount=264 )

Entry/Exit (as Demi wrote above) - "...Co-integrated instruments are traded on convergence from channel boundaries. It's just a question of finding the optimal weights of instruments in the portfolio. " (c) We set up the channel borders so that the synthetic tool "catches" them at its local extrema.

I.e. in general case we trade SELL BRN - BUY HO on a bounce from the upper limit of the channel, and close positions at reaching the lower one. Then it is vice versa!

There is a forum thread where I often post such entries on the spread ( synthetic) silver-gold. See last such entry/exit https://www.mql5.com/ru/forum/122468/page211 - post leonid553 21.08.2012 15:04

 
faa1947:
Avals:


Cointegration can be estimated even visually - it's a flat

Completely wrong. Co-integration is the 'similarity' of the movement of the rows. The cohesion of their movement is always expressed as a return to some average. However, we must never forget that in quotients, unlike in geology, there is always a deterministic component, which must be removed from the quotient before applying statistical methods. Otherwise we are commensurate with these deterministic components. Hence my 40 pips and your fear of the spread eating away at the deviation from the average.


Well, a return to the average would look like a flat. A flat is not necessarily a bounce around a single zero level.

And as for my "fear of the spread")) - Don't be afraid to trade those 40 pips ;)

 
leonid553:

I accidentally discovered that the correlation of these instruments is much greater than the correlation of any other commodity!

The synthetic instrument is calculated and drawn as the difference (in dollars) of these component symbols for a 1:1 position size ratio (download the indicator and see the description at http://www.procapital.ru/showpost.php?p=813139&postcount=264 )

Entry/Exit (as Demi wrote above) - "...Co-integrated instruments are traded on convergence from channel boundaries. It's just a question of finding the optimal weights of instruments in the portfolio. " (c) We set up the channel bounds so that the synthetic trader "catches" them at its local extremums.

I.e. in general case we trade SELL BRN - BUY HO on a bounce from the upper limit of the channel, and close positions on reaching the lower one. Then it is vice versa! There is a forum thread where I often post such entries on the silver-gold spread. See last such entry/exit https://www.mql5.com/ru/forum/122468/page211 - post leonid553 21.08.2012 15:04

I have seen it, I just do not understand how synthetics is calculated.

The main question on synthetics. Where is the proof that it will return from the channel boundaries back?

Reason: