Random Flow Theory and FOREX - page 82
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
Bandpass filters should not only divide by period, but also take into account the equality of areas between two LF filters or m/f of different periods of LF and HF.
Here's another outline. We take retranslators or modules, for example, of Open Clause minutes, and sum them up. For each new count we look for the point in the past, when the sum equals 1024, 512, etc., for example, points or percents. Then we calculate average claws from the main series with the trend. We obtain the period dynamic scales, or average, the number of which changes depending on the change of the path. But that's half the trouble, we should apply it to all timeframes.
Of course we can shift filters, but they can not be extrapolated by absolute values, we can extrapolate fluctuations that are in the traveled path function. And then it can be added to the general picture.
MA calculated by values from the left edge to the right one, and another MA vice versa by values from the right edge to the left one. An integral function between these two lines should be analyzed. The essence here is slightly different, the MA has a straight line in the impulse response, so the analysis will not be correct (analysis of the difference between these 2 mirrored MAs). You need other filters that have the impulse response as a 2nd order damped oscillating link, rather than the flapper.
Or try the following. On each new data set take the positive increments and divide them by the number of positive values in these last 1024 cells. The same is true for the negative values. Then add them together and then separately add all positive and negative last 1024 bars and divide them by 1024. The first and second series compare and analyse.
The essence of the matter is that the integral function on simple machines of that kind is not correct when detecting internal market fluctuations. Since maccha's coefficients are static. We must either, as I wrote above with dummies, or use filters with different with impulse characteristics of the oscillating link type of the 2nd order, then the specularity will not be such as on the first page,
The next step is to choose such a system of filters, for example, that their period increases in progression, but the distance covered by them will be approximately equal to each other. I wrote about it above.
We will continue our attack in the direction of analysis not of price direction, but of the size of movements, increasing the discretization at different timeframes from the smallest one, for example for 1 minute (how to insert the tick volume into this mix - later). The task may be considered from two points of view.
1 - Build a system of filters of different periods matching the price, equalizing phase changes between them, then extrapolate not the filters themselves but integral (area, or may be called incremental modulo) function between such filters with equalized phase.
2- Set parameters (as needed) for integral functions in advance, and build filters from them so that the sum of all coefficients is minimal (in fact, search for the minimum of the function).
It's a bit of the same thing, I think, as you talk about in your last posts about integral functions and their analysis and extrapolation.
...............
The bottom line is that an integral function on such simple machines would not be correct for detecting internal market fluctuations. Since the coefficients of the dummies are static. We must either, as I wrote above with dummies, or use filters with different with impulse characteristics of 2nd order oscillatory link kind, then the specularity will not be such as on the first page,
The next step is to choose such a system of filters, for example, that their period increases in progression, but the distance covered by them will be approximately equal to each other. I wrote about it above
..................
Or in the physical sense of understanding we can build an integral function not from filters calculated from right to left and vice versa, but calculated by nominal prices after which recalculated by prices reversed mirrorwise. The same in this integral function between these filters will be analogous to the market cycle density . I already mentioned here the TS that works both ways, or TS that works on a direct and inverted chart at the same time. It is approximately the same. When applied to coins it would look like taking profits through Parondo's paradox by applying strategies to different sides of a coin).
intermediate picture, to understand that it's not the direction of the sign that needs to be extrapolated but the size.
system of coinciding filters. Then from a low frequency (although you can probably do the opposite from the high frequency) fit the compression coefficients vertically for all of these lines, so that the sum of all equal to each other.
Next, we build a function of compression coefficients, and it is these functions that need to be extrapolated, then from them goes extrapolation of the original lines themselves from the picture.
At the stage of constructing indices from these values, we just use coefficients without extrapolation, and then they are extrapolated separately from the indices. (of course, we can include the equal volume, which will change the weighting of these functions in the overall calculation of the indices)
the picture here http://forum.alpari.ru/showthread.php?p=3075907#post3075907 unfortunately the specifics of that forum are such that it is not possible to discuss strategies without registration, so I have posted the picture, but I will have to register on their forum to see it. Such is the way things are. And here I have it not attached. (Do not take it as an advertisement)
to be continued...
You are annoying. May I answer?
Probably got it because the method is different, you are, as I understand it, replacing geometry by extrapolation of integral coefficients? i.e. via tangents from above to below,But.
that's to be considered later in terms of fractal geometry as well. It is just there to adjust differently what is extrapolated above needs to be done in spectral analysis. And one should not say that analysis by means of geometry does not predict anything. It is the same extrapolation but it has another name.
It is also possible not through the fractal geometry but through tangents to the price, from above and below, the price chase))) too. Except you can't put geometric methods into a cluster. And in Forex the useful signal is not only in the pair, but also in the cluster, because the capital is deployed in the group by the instrument, in particular in Forex, and how can we estimate the general picture of the market sector potential using geometry? The geometrical decomposition will be non-linear from the very beginning, it won't work to gather it back.
And spectral analysis, of course, requires a lot of resources (especially when it comes to reduction to the result), but it shows the true face of the market, what and how and where happens. For example, how do you extract useful information from the cluster? Or do you choose the tool with the help of your geometry? Although it may be done with the geometry. But it is easier to follow the currency pair, rather than the increasing range of instruments from this nest of currencies. And then trade the most profitable ones that have a higher potential. With the geometry you can say unambiguously that the target levels of this instrument are preferable to those of this one, and how much more sensible estimate it will be (of course and take into account the time of position holding).
spectrum analysis
Please make a picture like this: a point in a Cartesian coordinate system is running on a plane, the coordinates are the first and second strongest harmonics (estimates of instantaneous frequency values). This is called the phase portrait of the system. I want to look at the attractor, what it will be like. I would do it myself a long time ago, but working out my ideas works on the principle of FIFO, and there is a big queue of them)).
Please make a picture like this: a point in a Cartesian coordinate system is running on a plane, the coordinates are the first and second strongest harmonics (estimates of instantaneous frequency values). This is called the phase portrait of the system. I want to look at the attractor, what it will be like. I would do it myself long time ago, but I have FIFO method of working out my ideas, and there is a big queue of them)).
But in general, yes. I was puzzled by something similar. First I tried to build it not only from harmonics but also from hormonics of different characters, hormonics of signs in increments and hormonics of sizes in increments. Only to build it it is necessary to normalise them correctly. The axes are of a different nature.) (let's say so). To get rid of the absolute values, it takes a harmonic, because it is in fact the rate of change, but it needs to be normalized. If we look at your version, we will get that it is not the vector orientation in the phase space that is important, but the modulus size of the vector. Roughly speaking, we try to analyse and extrapolate its size.
Actually, yes. I was puzzled by something similar. Only I first tried to build it not just from harmonics but from hormones to different characters, hormonics of signs of increments and hormonics of sizes of increments. Only to build it it is necessary to normalise them correctly. The axes are of a different nature.) (let's put it this way). To get rid of the absolute values, it takes a harmonic, because it is in fact the rate of change, but it needs to be normalized. If we look at your version, we will get that it is not the vector orientation in the phase space that is important, but the size modulo of the vector. Roughly speaking, we try to analyze and extrapolate its size.
What is important is the trajectory of the vector's end, not its size or scale, but its form. The attraction type of the system is often determined visually