Discussing the article: "Pair trading" - page 2
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
...There is no practical application of this approach in a non-stationary environment....
)) how many times?
It makes no sense to consider the correlation of two symbols with each other at all, it can give an opportunity to determine the moment of positions opening, but in no way gives an opportunity to determine the directions of positions.
Therefore, we consider the correlation with sloping lines. It is necessary to count for two periods - for a long and for a short one. Correlation for the long period will give an understanding of whether the symbols are moving in one direction or in different directions.
Correlation for a short period will allow you to detect moments of divergence from the main trend.
For example, for a long period the correlation on both symbols is greater than 0, that is, they go in one direction - up.
If on the short period the correlation diverged, for example, on the first symbol is greater than 0, on the second symbol - less than 0 - enter.
Sell on the first one, buy on the second one. And so on in the same style ...
---
I would like to ask the author a question about the Olkin-Pratt correction. What is it and why? Is it really needed here?
What would happen without it?
Therefore, we consider the correlation with sloping lines. It is necessary to count for two periods - for a long and for a short one. The correlation for the long period will give an understanding of whether the symbols are moving in one direction or in different directions.
Correlation for a short period will allow to detect moments of divergence from the main trend.
For example, for a long period, the correlation on both symbols is greater than 0, that is, they go in one direction - upwards.
If on the short period the correlation has diverged, for example, on the first symbol is greater than 0, on the second symbol - less than 0 - enter.
Sell on the first one, buy on the second one. And so on in the same style....
Classic)))) but the same lag problem, it's the same way, and how to calculate the right ATR))))
Therefore, we consider the correlation with sloping lines. It is necessary to count for two periods - for a long and for a short one. The correlation for the long period will give an understanding of whether the symbols are moving in one direction or in different directions.
Correlation for a short period will allow to detect moments of divergence from the main trend.
For example, for a long period, the correlation on both symbols is greater than 0, that is, they go in one direction - upwards.
If on the short period the correlation has diverged, for example, on the first symbol is greater than 0, on the second symbol - less than 0 - enter.
Sell on the first one, buy on the second one. And so on in the same style....
---
I'd like to ask the author about the Olkin-Pratt amendment. What is it and why? Is it really necessary here?
What will happen without it?
the direction of symbol movement can be learnt using the first derivative...
the correction is needed if we use opening/closing positions by correlation value.... For example, -0.95 - open, -0.3 - close.... It can be dispensed with, but we should remember that the edges of the chart will be a bit compressed
Theoretically. But how do you practically calculate this APR?
If the question is about the ATR period, you can take the average period of holding a position on the TS and multiply it by N>1 (for the reserve). It is important that the ATR should be taken into account, because without it, the drain is assured. Practically. The only exception is coincidence when ATR1 is approximately equal to ATR2.
Therefore, we consider the correlation with the sloping lines. It is necessary to count for two periods - long and short. The correlation for the long period will give an understanding, whether the symbols go in one direction or in different directions.
Suddenly - correlation of a long uptrend with a downward sloping line can easily give a positive value greater than 0.5....
For example, the trend was formed by impulses, and the main time the course was corrected downwards.
suddenly - correlation of a long uptrend with downward sloping lines, can quietly give a positive value greater than 0.5.
For example, the trend was formed by impulses, and the main time the course was corrected downwards.
That's sensational! Show me a numerical series that gives such an effect.
https://www.mql5.com/en/signals/2020078?source=Site+Signals+Favourites
this signal seems to be based on arbitrage (divergence) of eurusd and usdchf pairs.
P/S not an advertisement, just caught my eye and judging by the history of trades it is so.
if you get a "Filling" Order add
request.type_filling=ORDER_FILLING_IOC;with every request you find