Absolute courses - 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
In short, the idea is as simple as three rubles. That's why the author was so adamant that the rows are only fair for his area. That's the whole point. Even if we take into account the fitting, well, let's suppose, it has a right to life. The bottom line is this. We take two rows EU and UY for example all of February. On February 1 we assign D=1 and the rest according to the rates on February 1. Since the system of equations is overdetermined, we need a solution criterion.
Second part of solution: We assign D=1 to the last bar M5 already on February 28 and further all by analogy but in reverse order, i.e. in the past by February 1, and obtain absolute convergence of these six series in pairs, i.e. we practically have a proof of uniqueness of solution. But this solution will not work on the other interval, it means that with new data we will get more and more divergence. Actually it is what we trade. I do not see any other solution to this problem.
You are not Yusuf.
A different style of writing, different knowledge of terve/statistics - etc.
In short, the idea is as simple as three rubles. That's why the author was so adamant that the rows are only fair for his area. That's the whole point. Even if we take into account the fitting, well, let's suppose, it has a right to life. The bottom line is this. We take two rows EU and UY for example all of February. On February 1 we assign D=1 and the rest according to the rates on February 1. Since the system of equations is overdetermined, we need a solution criterion.
Second part of solution: We assign D=1 to the last bar M5 already on February 28 and further all by analogy but in reverse order, i.e. in the past by February 1, and obtain absolute convergence of these six series in pairs, i.e. we practically have a proof of uniqueness of solution. But this solution will not work on the other interval, it means that with new data we will get more and more divergence. Actually it is what we trade. I do not see any other solution to the problem.
Colleague grell, what are you talking about? When did I say that the series is valid only for my area? Is the system overdetermined? God be with you, young man. I wish I could define it, the underdetermined one. Even a fool could handle an overdetermined one. The second part's even funnier. Point the finger where I have D=1 at the right end of the interval. Absolute convergence? Series? What are you talking about? Do you even know what a series is (infinite sum) and what absolute convergence is? I can't even understand what your post is about.
My idea is really very simple. Originally (at the beginning of the thread) I presented it differently, but now (starting from page 20) I feel comfortable formulating it this way:
changes in exchange rates relative to real (not changing their value over time) benchmarks (absolute rates) are due to two mechanisms: joint movement (a common change in value, the same for all currencies within a closed triangle, for example) and separate, individual movement. It is postulated that it is the separate movement that forms the differences in the forms of RELATIONShips as you all see in the metatrader, and it is fundamental to know a common similar form for all currencies in the triangle REALLY (relative to an unchanging benchmark) for them. Then a linear regression can be performed and a levelling operation can be performed, i.e. the straight line corresponding to the evolution of the average shape can be subtracted. The dry residue will be left with three straight lines with slight slopes: the evolution of the DIFFERENT absolute rates E, D, Y (or any other) from their MEDIUM EVOLUTION (due to the change in the value of money (paper trash) in general, all at once), from which the PRIORITY of the change in the ED, EY, DY RELATIONS will be seen. Knowing this PRIORITY, we will make decisions about trading ED, EY, DY pairs, opening transactions only when the movement of the RELATIONSHIP is conditioned by the evolution of MULTIPLE absolute rates that comprise it.
P.S. Above was a request to name the theory. Let it be Absolute Rates Theory.
P.P.S. Once again, for those who do not understand. At the heart of the idea is simple and straightforward physics. The value of the paper rubbish that is thrown by all central banks in the world is divided into two components: an agreed-upon change (cheapening) according to the fact that if the Fed prints a billion dollars, the ECB, the Bank of Japan and other RF CBs will throw in their rubbish (so that the rate ratio does not change much - it is a question of trade advantages - and asymmetries are called currency wars), and INDIVIDUAL, which is a SMALL correction to this agreed-upon cheapening against time-varying units. And it is the individual peculiarities that determine the CONNECTIONS of the currency pairs we trade.
P.P.S.2 For the completely stupid: the reasoning about indices above is irrelevant. This is where I started. An index is essentially the same as any currency pair. Only the rate of one currency is not calculated against another currency, but against a basket of currencies. The important thing is that the value of this currency basket, gold, oil, whatever, CHANGES WITH TIME. It cannot be a ETALON. That is why we can select the ETALON in an arbitrary bar, relative to which we are going to build everything. Its size is not important. Only its STABILITY is important, i.e. UNCHANGIBILITY in time, which is achieved simply by DEFINITION.
The title is super.
There is a substantive suggestion. Here we want to identify, as you say, real currency movements. Let's ask the question - why? I suppose, correct me, that we want to see which pair changes, e.g. ED, are due to the euro and which are due to the quid (individual movements in your terminology). And probably get some more nuggets, maybe such rates will lend themselves to better prediction, I do not know. So, how do we do it? We postulate that we should select rows E and D in such a way that they are as similar to each other as possible! It's kind of illogical: we wanted to identify the differences, but got the opposite, similar to the point of almost matching. My point is, it would be much more logical to look for exactly ortagonal series, i.e. those which are statistically NOT correlated with each other, i.e. their KK is not maximum, but minimum (tends to 0).
Once again, in short: I argue that your theory is based on an incorrect postulate, which needs to be replaced by the exact opposite.
And, at the end of the day, put the calculations out there already. We will at least check them for correctness.
And yes, P.P.S.3.
The absolute exchange rate charts will show the CORRELATION of currencies with each other. The real one, as it were. And that it is very close to 1 for any pair of currencies is undeniable. For all that paper rubbish is generally getting cheaper (and very fast - the dollar won 50 times in 50 years!) in a coherent way.
And what can be found by calculating the correlation between say EURUSD and GBPUSD is not at all relevant to the question of correlation between EUR and GBP, for it is FULLY determined by USD. Take any other quote currency instead of USD, even JPY, and you will get completely different results.
The title is super.
There is a substantive suggestion. Here we want to identify, as you say, real currency movements. Let's ask the question - why? I suppose, correct me, that we want to see which pair changes, e.g. ED, are due to the euro and which are due to the quid (individual movements in your terminology). And probably get some more nuggets, maybe such rates will lend themselves to better prediction, I do not know. So, how do we do it? We postulate that we should select rows E and D in such a way that they are as similar to each other as possible! It's kind of illogical: we wanted to identify the differences, but got the opposite, similar to the point of almost matching. My point is, it would be much more logical to look for exactly ortagonal series, i.e. those which are not statistically correlated with each other, i.e. their CC is not maximal, but on the contrary minimal (tends to 0).
Once again briefly: I assert that your theory is based on an incorrect postulate, which should be replaced with the exact opposite.
ZS And finally, post the calculations already. We will at least check them for correctness.
Colleague. Why-then you are absolutely right. Got it. Further, everything is logical. Since the correlations between E, D, Y cannot be arbitrary the sum of correlations corr(E,D)+corr(E,Y)+corr(D,Y) knowingly cannot reach a three. But the sum of correlations corr(ED,EDx)+corr(EY,EYx)+corr(DE,DEx), where EDx, EYx, DYx are respectively E/D, E/Y, D/Y, can easily reach exactly 3. Butcorr(E,D)+corr(E,Y)+corr(D,Y) cannot. The presence of some limit value less than 3 corresponds precisely to the fact that there are limits to the ratios ED, EY, DY. It is this subtle difference from 3 that we are interested in.
P.S. And yes, before speaking about orthogonality of basis vectors of whatever, learn to write the word orthogonality itself without mistakes.
P.P.S2. And yes, the question of correlation and the question of orthogonality, dependence (functional, stochastic, or otherwise) are in no way related. Typical example: sine and cosine. Correlation is zero. Dependence is strictly functional. Learn the math.
P.P.S.3 Am I correct in assuming that you personally will check my calculations?
And yes, P.P.S.3.
The absolute exchange rate charts will show the CORRELATION of currencies with each other. The real one, as it were. And that it is very close to 1 for any pair of currencies is undeniable. For all this paper rubbish is by and large getting cheaper (and very fast - the dollar won 50 times in 50 years!) in a coherent way.
Very much so: inflation is a regular factor, which can generally be deducted from both E and D without compromising the correctness of the E/D ratio, especially if one considers that it is roughly the same in different parts of the capitalist world. In addition, within a period of 150 5-minute bars, inflation is not visible, because it occurs at a much slower rate, a fraction of a percent per month or even per year. What we will see is small fluctuations that have nothing to do with long term global movements.
P.P.S.3 Am I correct in assuming that you will personally check my calculations?
Very much so: inflation is a regular factor which can be deducted from both E and D without affecting the validity of the E/D ratio.
Congratulations to you. You've got it. That's exactly what I'm going to do by subtracting from the INDIVIDUAL forms of E, D, Y the MEDIUM FORM, and thus finding the DIFFERENCES that solely form the forms of the ED, EY, DY RELATIONSHIPS in particular.
it is much slower - fractions of a percent a month, or even a year.