a trading strategy based on Elliott Wave Theory - page 284

 
to Neutron

And here is the first result of BP processing by an algorithm that has nothing to do with the wavelet transform (see post above)! For comparison, on the right, here is a picture of Andre69:



I'd say the match is satisfactory. By the way, the code in MathCad contains ONLY the recurrence formula for the LPF - 10 lines and that's all, and the counting time is 1 sec.
It is pleasant, that results received by absolutely different methods are similar.



I think I am familiar with what you showed on the left. I have seen similar pictures. Another variant of multiplescale analysis. In general, all roads lead to Rome. And that's great!
 
to Candid

This is the impression that emerged from the sum of the pictures: a regular structure exists in a certain range of frequencies. Clutter dominates both too high and too low. I wonder if this is a property of this section of BP or of the market in general. <br / translate="no">


Seems a bit wrong to me.



Here are two wavelet spectra (Morlet wavelet, normalized by 1, frequency increases from left to right, expressed in conditional units) for two far apart (several months) sections of the price series. They are different, of course, but not too different. It feels like the market has a sort of own frequency, which drifts back and forth over time, but smoothly. I think there is something to that.
 
Another picture to go with the theme:



Another wavelet (from the Gaussian family), a different way of representing the conversion coefficients. This is for the same chunk of the price series. Which one, unfortunately, I didn't write down (I did it long ago), so I can't quote its graph. The scale extends from the lower figure to the upper one.

Looking at all this already makes me want to think about skeletons and pr....
 
<br / translate="no"> Looking at all this already makes you want to think about skeletons and pr....


They (skeletons) are what you should be thinking about. It's the only useful thing you can use to get the dynamic characteristics of the system.

Everything else is just pretty pictures and nothing more.
 
to Andre69
Here are two wavelet spectra (Morlet wavelet, normalised by 1, frequency increases from left to right, expressed in notional units) for two distant (several months) segments of the price series. They are different, of course, but not too different. It feels like the market has a sort of own frequency, which drifts back and forth over time, but smoothly. I think there is something to this.


Now this is very interesting!
Andre69, can you cite a similar picture, but averaged over at least ten non-intersecting windows and with vertical whiskers plotted on the chart, whose spread corresponds to the standard deviation for a series of results?
That would be indicative.
 
to grasn


Глядя на все это уже хочется думать о скелетонах и пр....


They (skeletons) are the ones to think about. This is the only useful thing with which you can get the dynamic characteristics of the system.

Everything else is just pretty pictures and nothing more.



Not the only one!
The picture (matrix of coefficients) contains ALL the information that is in the skeleton. But they, dear ones, contain only a tiny fraction of what is in the full picture. Do not be fooled. Skeleton is just a convenient and illustrative method of analysis.
 
To Neutron

To Andre69
Here are two wavelet spectra (Morlet wavelet, normalised by 1, frequency increases from left to right, expressed in notional units) for two distant (several months) segments of the price series. They are different, of course, but not too different. It feels like the market has a sort of own frequency, which drifts back and forth over time, but smoothly. I think there's something to that.


Now this is very interesting!
Andre69, can you cite a similar picture, but averaged over at least ten non-intersecting windows and with vertical whiskers plotted, the spread of which corresponds to the standard deviation for the series of results?
That would be indicative.




So far there is only this one:


This is an averaged spectrum over seven consecutive, non-overlapping chunks of the price chart. Done on an hourly chart, i.e. averaging over about a year. You can see that the spectrum is smoothing out (which is expected), but far from all the way through (which I think is more interesting)
 
By the way...
Not about wavelets at all.

While having fun with price series charts I accidentally discovered a very simple and fully automatic algorithm for finding support/resistance levels.

Here is a picture:



Do a simple transformation over the price curve (you can change its parameters), then look for flat areas.

Don't mind if I'm too slow to open the door and it's all known and uninteresting for a long time.
If I'm wrong, I'll be glad to give you the details.

Good luck to you all and keep up with the trends!
 
to grasn


Looking at all this already makes you want to think about skeletons and pr....


They (skeletons) are what you should be thinking about. It's the only useful thing you can use to get the dynamic characteristics of the system.

Everything else is just pretty pictures and nothing more.


Not the only one!
The picture (matrix of coefficients) contains ALL the information that is in the skeleton. But they, dear fellow, contain only a fraction of what is in the full picture. Do not be fooled. Skeleton is just a convenient and illustrative method of analysis.


I am not enraptured, I am guided by practice. I can point out that the original signal also contains ALL the information. What is useful for forecasting is in the skeletons.
 
to Andre69
This is an average spectrum over seven consecutive, non-overlapping chunks of the price chart. Made on an hourly chart, i.e. averaging over about a year. You can see that the spectrum is smoothing out (which is expected), but far from all the way through (which I think is more interesting)

Andre69, what does the maximum with abscissa 1000 correspond to in the above graph? Is it the size or repetition rate of some regular perturbation on the BP of the Head-Shoulder type, or something more exotic? In short, how to interpret the result of a Morlet wavelet on BP?