Question on pair trading. - page 4

 
Meat:
You must have made a typo, referring to buying EURUSD rather than EURGBP? If you take lots 1:1, it would not be identical to EURGBP. To be identical you need to balance the volumes of positions in terms of money, not the number of lots
Yes, of course EURUSD. Corrected.
 
Another question for connoisseurs. Is it possible to determine the moments of convergence and divergence of pairs by the curve of the correlation indicator?
 
khorosh:
One more question for the experts. Can we use the curve of the correlation indicator to determine the moments of convergence and divergence of pairs?

You can top EURUSD at the bottom GBPUSD + correlation between them. Rather more promising than a divergence convergence indicator.

 
Strange reasoning - if the signals are correlated then you can make money on it, perhaps you can make money if the signals are correlated "in general", but sometimes there is a discrepancy between them
 
ivandurak:

You can top EURUSD at the bottom GBPUSD + correlation between them. Rather more promising than a divergence divergence indicator.


Why then at the place marked by arrows, with maximum spread, the correlation is close to one? Or am I misunderstanding something? What is the correct way to use correlation for this purpose?

 
khorosh:


Why then is the correlation close to unity at the place indicated by the arrows, when the spread is at its maximum? Or am I misunderstanding something? What is the correct way to use correlation for this purpose?


Correlation is not a "at the moment" indicator. Correlation is calculated over some past period - the average temperature in the ward.
 
khorosh:


Why then is the correlation close to unity at the place indicated by the arrows, when the spread is at its maximum? Or am I misunderstanding something? What is the correct way to use correlation for this purpose?

Currencies can go in the same direction but at different speeds. Get correlation close to 1 and maximum spread spread spread. If you look at the cross at this point, there will probably be a trend. And the signal to open positions in this case is likely to be against the trend.
 
ivandurak:
Currencies may go in the same direction, but at different speeds. You will get the correlation close to 1 and maximum spread spread spread spread. If you look at the cross at this point, there will probably be a trend. And the signal to open positions in this case is likely to be against the trend.

Let one be open against the trend and the other against it - the important thing is that the "spread" then "collapses".

 
ivandurak:
Currencies can go in the same direction, but at different speeds. Get a correlation close to 1 and a maximum spread spread spread. If you look at the cross at this point, there is likely to be a trend there. And the signal to open a position in this case is likely to be against the trend.
.
After the maximum spread, it is time to enter the market, but the correlation indicator does not give any signals. Why do you think that "... it is more promising than the divergence indicator "?
 
khorosh:
After the maximum of the spread it is time to enter the market, but the correlation indicator does not give any signals. Why do you think that this is more promising than the divergence indicator?


The correlation indicator and the divergence indicator are the same thing. What you mean is not correlation but cointegration. That's what faa wrote about.

The point is that if, for example, these two pairs were cointegrated, then their spreads would hang in a channel and they would always converge. Then when the spread reaches the border of the channel, you could enter the trade and make a guaranteed profit - arbitrage. But, unfortunately, there is no cointegration in forex.

The correlation in the markets "wiggles" in time very strongly. And for some instruments it even changes its sign in the opposite direction with time.

Reason: