Indicators: SATL
It looks very promising... It is not only about this indicator, but rather about a set of indicators from the series of Vladimir Kravchuk's publications in "Currency Speculator". Digital filtering speaks for itself, having an impressive theoretical basis. Unfortunately, I didn't have a chance to analyse the maximum entropy method itself and dig into the sources in detail, so I would like to ask, where did you take the FIR coefficients for this indicator? Did you design them or did you take ready-made ones?
Z.Ы. imho, indicators of this kind and TS built on their basis deserve the closest attention... thank you very much for posting them!)
It looks very promising... It is not only about this indicator, but rather about a set of indicators from the series of Vladimir Kravchuk's publications in "Currency Speculator". Digital filtering speaks for itself, having an impressive theoretical basis. Unfortunately, I didn't have a chance to analyse the maximum entropy method itself and dig into the sources in detail, so I would like to ask, where did you take the FIR coefficients for this indicator? Did you design them or did you take ready-made ones?
Z.Ы. imho, indicators of this kind and TS built on their basis deserve the closest attention... thank you very much for posting them!)
It looks very promising... It is not only about this indicator, but rather about a set of indicators from the series of Vladimir Kravchuk's publications in "Currency Speculator". Digital filtering speaks for itself, having an impressive theoretical basis. Unfortunately, I didn't have a chance to analyse the maximum entropy method itself and dig into the sources in detail, so I would like to ask, where did you take the FIR coefficients for this indicator? Did you design them or did you take ready-made ones?
Z.Ы. imho, indicators of this kind and TS built on their basis deserve the closest attention... thank you very much for posting them!)
Actually, reading Kravchuk's article raises a lot of questions. For example, at the very beginning of the article he describes the Fourier transform in a strange way, it seems that the man does not know the question. Then follows a strange formulation of what is a digital filter. The wording is really strange, as if with a hint on some magic system. Further not better - "Discrete signals have a number of properties known only to a narrow circle of specialists . . ." and the Doppler effect.
In general, the article gives the impression of a hastily created advertising brochure rather than a technical description of the method. Moreover, there is no method at all. In simple words, Kravchuk made a spectral analysis of some fragment of EURUSD quotes, and according to the results he found out the presence of the main cycles and their harmonics. Based on the obtained result, he calculated coefficients for filters, which are designed to highlight these main (or the main) cycles.
The filters are ordinary, standard, there are many ready-made programmes for calculating their coefficients, I think I have seen something similar in MQL. You just set the cutoff frequency, the width of the transient response and the amount of attenuation in the passband and delay band.
Now let's assume that Kravchuk did everything as written and made no mistakes or falsifications. Even in this case, his so-called method can only be of some theoretical interest, because if you now make a spectral analysis of the latest EURUSD quotes, you will get a different result, other main frequencies and will have to recalculate filters for them. Therefore, it makes no sense to use the filters proposed by Kravchuk, they have no value at the moment.
If you want to repeat Kravchuk's feat on your own at this point in time, there is really nothing stopping you. Borrow a spectrum analyser from https://www.mql5.com/en/articles/292, search for a program written in MQL to calculate filter coefficients (or find third-party programs) and create them anew, even with each arrival of a new bar. I can assure you that you will not be able to repeat Kravchuk's results.
By the way, keep in mind that if you calculate SPM, then whatever method you use, you should get approximately the same estimates of SPM. The maximum entropy method is not an exception. You can find the SPM in any way that is most appealing to you.
You should not publish filters with incomprehensibly generated frequency response at all, they are of no value, you can generate a dozen of such filters in a few minutes.
Kravchuk's example is infectious. In the description from the published indicator we read:
"SATL (Slow Adaptive Trend Line) - a "slow" adaptive trend line is obtained with the help of a digital low-pass filter FNF-2."
Why adaptive? What is being adapted and by what criterion is the adaptation taking place?
"The LLF-2 serves to suppress noise and market cycles with longer oscillation periods."
Must be a typo. The LLF suppresses frequencies with shorter periods.
"Low-pass filters VLF-1 and VLF-2 provide attenuation A in the delay band not less than 40 dB and absolutely do not distort the amplitude and phase of the input discrete series of closing prices in the passband."
This is not true! Both amplitude and phase will be distorted.
These properties of digital filters provide much better (in comparison with simple moving averaging) noise suppression, which, in turn, allows you to dramatically reduce the probability of "false" buy or sell signals.
It turns out that moving average is not a digital filter, but what is it - an analogue one?
There is no analogue of SATL among widely known technical instruments. It is not a moving "average", but an adaptive estimation of the long-term trend line.
We have already said about adaptation - it does not exist!
Unlike moving averages, SATL has no phase lag relative to current prices.
That's not true! Well, not true at all!
Of course, the fact that a great number of people are raving about Kravchuk's work makes me doubt the correctness of my conclusions. So try to replicate the results obtained in Kravchuk's article. Maybe you will succeed. After all, agree that if the methodology is correct, it should provide repeatability of results.
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
You agree to website policy and terms of use
SATL:
Slow Adaptive Trend Line is used for suppressing market noises and market cycles with longer oscillation periods.
Slow Adaptive Trend Line (SATL) is formed with the digital filter of the low frequency FLF-2. Filter FLF-2 serves to suppress noises and market cycles with longer periods of oscillation. Filters of low frequency FLF-1 and FLF-2 provide attenuation in the stop band with no less than 40 dB and absolutely don’t distort the amplitude and phase of entry discontinuous price series in the pass band (bandwidth).
These properties of the digital filters provide significantly improved (in comparison with simple moving average) noise suppression that in its turn allows reducing sharply the probability of appearance of "false" signals for purchase and sell. There are no analogues to SATL among widely known technical instruments. This is not a moving "average", but just the adaptive line estimate of a long-term trend. Unlike moving average, SATL has no any phase delay with regard to current prices.
Author: Nikolay Kositsin