Zero sample correlation does not necessarily mean there is no linear relationship - page 29
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
I wrote above that there is a vector of weights which is adjusted when the sliding window is shifted.
Of course, we can take a year window and create a synthetic that has the same and maximum possible correlation with all the majors in that year.
But let's not forget what QC is, and that it depends on the sample and its size.
Come on.
I welcome any meaningful near-mathematical shamanism in forex. As long as the profits go up.
In this case, if your synthetic has a correlation much higher than 0.5, it probably points in favour of the dollar being more influential (dominant) relative to other currencies. And that same "maximum possible" correlation can serve as a measure of the dollar's "status". If in the same way synthetics are built from a set of pairs with the Yen (Euro, Pound) as the base (common currency), it may be possible to estimate the "status" of the Yen (Euro, Pound) on a given time interval.
If the interval is a moving one... ...it would make a good indicator, eh. You can share it when you do it... ;)
There's that shit.
Yeah, I'm aware of that. That's why I put a smiley face.
;-)
In this case, if your synthetic has a correlation significantly higher than 0.5, it probably points in favour of the dollar's greater influence (dominance) over other currencies. And that same "maximum possible" correlation can serve as a measure of the dollar's "status". If in the same way synthetics are built from a set of pairs with the Yen (Euro, Pound) as a basis (common currency), it is possible to estimate the "status" of the Yen (Euro, Pound) in a given time interval.
The task with common correlation is taken here. I have not managed to grasp the sense of solving such a task. That's why I asked a question.
About the measure of "status" of the base currency - I don't agree at all. Why depend on parrots at all? KK just depends on parrots. That's why I don't see much point in it.
And if we make the interval a sliding one... that would make a pretty good indicator, yeah. Feel free to share when you've done it... ;)
I was wrong about many variations of such curves. There is only one such curve (on the right):
I'm afraid to be wrong too, but one is not always)
Finding a curve that has the same correlation coefficient with the selected series is mathematically equivalent to finding in N-dimensional Euclidean space a point equidistant (corrected for a given vector of coefficients) from a given set of points . I note that such a problem may be degenerate, i.e. may have an infinite number of solutions - a curve, a plane or even several dimensional subspaces - depending on the degree of degeneracy. When applied to the market this means that the problem certainly may not become degenerate, but poorly conditioned it may become so, as the pairs under consideration may well include those strongly correlated with each other. We know very well what bad conditionality leads to. The solution of the problem starts to "feverish" at small fluctuations of initial conditions.
Therefore, without any doubt about the mathematical correctness of your calculations, I recommend first of all to get rid of strong correlations in the initial data - e.g. beyond +/-0.85 or choice.
The general correlation problem is taken here. I did not manage to grasp the meaning of solving such a problem. That's why I asked the question.
About the measure of the "status" of the base currency - I don't agree at all. Why depend on parrots at all? QC just depends on parrots. That's why I don't see much point in it.
I once posted a pretty good indicator. I saw the reaction of the "public"...The public is not the same here.
But it would still be nice to see the "fans"... ;)
Therefore, without in any way doubting the mathematical correctness of your calculations, I recommend getting rid of strong correlations in the raw data first of all - e.g. beyond +/-0.85 or choice.
The major majors in FOREX + silver and gold are poorly correlated.
Then, cards in hand, look for the centre of the described sphere))
I was simply pointing out the limitations of the task.
Finding a curve that has the same correlation coefficient with the selected series is mathematically equivalent to finding a point in N-dimensional Euclidean space that is equidistant (corrected for a given vector of coefficients) from a given set of points .
I thought it was a vector at the same angle to all the other vectors. The correlation coefficient is the cosine of the angle between the two vectors...