How to minimise index correlation - page 5
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
and what is the use of such orthogonality (exactly in this form). and in general I want to move away from the word orthogonality, now they will give an anathema ato.
I do not understand what you mean, brain paralysis begins, then you talk about orthogonality and calculate indices using a commonly accepted formula, then you write that you have your own method of calculating the indices, which of course is a secret.
There's nothing else to explain...
At the point where the vertical line is, all currencies are stupidly equated to one USD=EUR=GBP=JPY=CHF=CAD=AUD=NZD=1.0 (i.e. 0.0% change). On each bar we calculate the percentage change of each currency and draw a point. I cannot show you how it is calculated. We calculate it basing on the orthogonality requirement for the currencies.
I personally see anticipation in phase changes.
First of all, I messed up a bit. It should be like this
for 2 currencies (on the EURUSD pair) EUR=MathPow(1.1,1.0/2.0)=1.049 USD=MathPow(1.0/1.1,1.0/2.0)=0.953
for 3 currencies(EURUSD,GBPUSD,EURGBP) EUR=MathPow(1.1*1.1,1.0/3.0)=1.067 USD=MathPow(1.0/(1.1*1.0),1.0/3.0)=0.969
The advantage of this orthogonality is the independence of currencies and the possibility to analyse each currency independently, while a currency pair cannot be analysed separately from the others.
I consider the well-known dollar index formula as a commonly accepted calculation formula for indices, and I have never used it.
I also think that I should stop using the word "anticipation", because they call anything with this word. I want to do a coincident decomposition, i.e. without lagging and without leading(or pre-empting).
What kind of phase change and what kind of anticipation can be seen in them, I can't understand. And how it all relates to decomposition into indices, I'm sure I'll never understand.
First of all, I messed up a bit. It should be like this
for 2 currencies(EURUSD) EUR=MathPow(1.1,1.0/2.0)=1.049 USD=MathPow(1.0/1.1,1.0/2.0)=0.953
for 3 currencies(EURUSD,GBPUSD,EURGBP) EUR=MathPow(1.1*1.1,1.0/3.0)=1.067 USD=MathPow(1.0/(1.1*1.0),1.0/3.0)=0.969
The advantage of this orthogonality is the independence of currencies and the possibility to analyse each currency independently, while a currency pair cannot be analysed separately from the others.
I consider the well-known dollar index formula as a commonly accepted calculation formula for indices, and I have never used it.
I also think that I should stop using the word "anticipation", because this word means anything. I want to do a coincident decomposition, i.e. without lagging and without leading(or pre-empting).
What kind of phase change and what kind of anticipation can be seen in them, I can't understand. And how it all relates to decomposition into indexes I'm sure I'll never understand.
I understand what a prejudgement is (or so I think), it does not mean that we can get the price before it appears on the monitor)) just that we can count the indices so that they are equal, then the prejudgement will look like a prejudgement on the coefficients. and we can take this prejudgement immediately in calculating the indices, then we get different synthetic and natural pair, what is wrong in synthetic - NOT equal to the natural pair, if this synthetic runs ahead? (its phase runs ahead)
what's wrong with synthetics NOT being equal to a natural pair if those synthetics run ahead? (its phase runs ahead).
I don't know... I haven't seen that yet. I doubt it's even possible. I don't see what would make that happen.
Again we have to go back to the question of orthogonality. There are orthogonal indices. You propose to abandon them and go to some other ones. They will necessarily be less orthogonal, there will be their mutual influence on each other and it will be impossible to analyse them separately strictly speaking. Are you sure you are willing to pay that price for dubious leading synthetics?
I don't know... I haven't seen such a thing yet. I doubt it's even possible. I don't see what would make it happen.
Again we have to go back to the question of orthogonality. There are orthogonal indices. You propose to abandon them and go to some other ones. They will necessarily be less orthogonal, there will be their mutual influence on each other and it will be impossible to analyse them separately strictly speaking. Are you sure you are ready to pay such a price for dubious leading synthetics?
It's the same as calculating the future filter coefficients. I don't have an idiot proof in the form of calculations, it's difficult, especially in Excel, to reconstruct the synthetics from the anticipatory coefficients.
By the way, have you tried to use your filters to calculate indices? I wrote in the branch on tendential planimetry. cf has no problem with vertical compression during smoothing.
The idea was this: take fans from LF (digital) filters and build indexes from each individual filter. get something like a fan for each individual index.
It's the same as calculating future filter coefficients. I don't have an idiotic confirmation in the form of calculations, it's difficult, especially in Excel, to directly synthetically reconstruct from pre-emptive coefficients.
By the way, have you tried to count indexes using your filters? I wrote in the branch on tendential planimetry. cf has no problem with vertical compression during smoothing.
The idea was this: take beraeys from LF (digital) filters and build indexes from each individual filter. we get something like a fan for each individual index.
No. Changing filter coefficients and indexes from filters is a nightmare.
I decompose the market into indexes, and indexes into frequency components by filters without perversions. That should be enough. All that's left is to do a proper extrapolation.
No. Changing filter coefficients and indexes from filters is some kind of nightmare.
I decompose the market into indices and the indices into frequency components by filters without perversion. That should be enough. What's left to do is a proper extrapolation.
1-This is the second time it all comes down not to the problem of building indices, but to the secret method of this very normal extrapolation.right?
2-but my assumption (on the contrary) is that if such extrapolation methods are available, it is not the method itself that matters, but what it can give, the anticipation by some coefficients.
So I'm looking for a method to identify and describe it, instead of initially looking for the method itself in order to get a pre-prediction. all my left hemisphere started eating the right one.))
The question should be asked: Why do we need it, what do we want to see in the index? And then about methods of interpreting index readings. And if you have methods that already work fine, then there is probably no need to start a vegetable garden.
I would like to hear from you the answer to the question .
because the purpose of constructing indices varies from person to person.
alternatively
1- to look at the level of divergence between the currencies, to assess the speed of their divergence relative to each other and to look for moments when two currencies are not going in the same direction, but when they are going in different directions from the horizon. you cannot see this on a pair.
2 - To calculate each currency relative to the whole basket in order to select the most promising instruments.
3 - in order to judge the group velocity of index instruments, the simultaneous movement of pairs, a certain centre of mass, relative to which the movement of currency instruments fluctuates.
well, that answers the question (briefly and perhaps not quite ideal, but somehow). it is possible to improve the criteria. so what about the centre of mass?
I would like to hear from you an answer to the question.
Because the purpose of building indices varies from person to person.
an option
1- to look at the level of divergence between currencies, to assess the speed of their divergence relative to each other and to look for moments when not two currencies go in the same direction, but when two currencies go in different directions from the horizon. on a pair this is not visible.
2 - To calculate each currency relative to the whole basket in order to select the most promising instruments.
3 - in order to judge the group velocity of index instruments, the simultaneous movement of pairs, a certain centre of mass, relative to which the movement of currency instruments fluctuates.
So, here is the answer to the question (briefly and probably not quite ideal, but something like that). you can improve the criteria.
You have pretty much described my goals for index construction.
I calculate the movement of a currency relative to a basket of currencies, in fact I have looped 8 currencies. Since I internally understand trend trading better, I watch when two currencies go in different directions from the horizon, i.e. I decide on the index of a given currency going up, down or flat. Of course, it is visible separately for each pair, but for me it is much easier and faster to analyse 8 indices, than to analyse 27 pairs. After having analysed the indices I select pairs for trading and start searching for entry points without fanaticism (it all happens though).
That's why I asked you: " Why do we need it, what do we want to see in this index?". To understand what the correlation has to do with it? Correlation of what with what? And why reduce it to a minimum?
You have pretty much described my index construction goals.
I calculate currency movements relative to a basket of currencies, in fact I have looped 8 currencies. As I understand trend trading more internally, I look when two currencies are going in different directions from the horizon, i.e. I decide on the index of a given currency going up, down or flat. Of course, it is visible separately for each pair, but for me it is much easier and faster to analyse 8 indices, than to analyse 27 pairs. After having analysed the indices I select pairs for trading and start searching for entry points without fanaticism (it all happens though).
That's why I asked you: " Why do we need it, what do we want to see in this index?". To understand what the correlation has to do with it? Correlation of what with what? And why keep it to a minimum?
Well, I can not say exactly, because I have not counted yet. But this is my version so far.
As long as there is correlation, what does that tell us? Perhaps the movement between correlated series is dependent on each other and has a relationship.
That is why we need to find the point when correlations are minimal (zero in the ideal case, but zero isn't zero) and recalculate this point with every new count.
You may use the same algorithm to calculate the minimum correlation point (for example, when you calculate the minimum correlation point you may not reach zero, but you may not see how, so I need to ask).
Why compare the correct trading method, extrapolation, which is not publicly available on the web, and the construction of indices.
it does not matter which method it is, if there is no indexes correctly indicating currency movements.
we can use them to find the centre of mass of the group,
and to construct the movement of this centre (please note: not the position of this centre in space, the position is not explicit,
namely the speed of its movement), do not receive a full picture. after all there should be a movement of the centre of mass
of a group of elements, and the motion of the mass of each individual element of the group, will be near the motion of the common centre.
And if that centre has inertia, then it will show up in the behaviour of the elements in the group.
In other words, the extrapolation method that Alexey has,
which allows you to get a prediction, it's not the merit of the method itself, it means that in the price
there are predictable elements.
The essence can be obtained by finding such a method of extrapolation, or by correctly
The effect should be in both cases, but the indices will give you the choice of the best instrument.