Author's dialogue. Alexander Smirnov. - page 7
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 the article there is a FIGURE1 (EURUSD Daily), if you sketch it out and compare it - there are similarities, although there are also differences.
And to top it all off, a comparison by muving Hal.
blue - SmirnoffMA
red - HullMa
period =12.
I will elaborate on my algorithm, which so far many people have not understood.
Open "VS" No. 01(75) 2006, p. 33. 33.
Synthetic Moving Averages (SMAs) are implemented here using an algorithm that has inner and outer loops. The outer cycle is typical and implements a "clowes" sliding average with a time window of 4 bars. The value of this window is constant and does not change during the use of any CCC. The internal cycle is the 'highlight' of the CCC. The value of alpha equals 2/(m+1)=0.6667 at m=2. It is programmed into the program and does not change under any circumstances. The scheme of internal cycle implementation is shown in Table 1. There are two branches here: the upper one is the averaging of the "claws" using EMA in the "backward" window; the lower one is the averaging of the resulting averaging products using EMA, but "forward". Actions must be performed strictly according to table 1 with the same accuracy in both branches (otherwise group delay compensation will not happen!) At averaging "backwards "+"forwards" the total alpha value will already be 2/(4+1)=0.4, i.e. m'=2+2=4. This is an algorithm of realization of the least effective of CCC family at m'=4.
It is possible to obtain an SSS algorithm for m'=8. To do this, we need to assign C1 to Q5, C2 to Q6, C3 to Q7, and C4 to Q8. The "backwards" and "forwards" cycle is repeated, but with new data. Thus, in CCC we are only averaging the "claws" themselves on the first pass backwards, and then averaging with changing directions of averaging products repeatedly. The inner loop can be "raced" to infinity by changing m' (but the multiplicity of m' must be kept equal to 4!).
For other m' you will, hopefully, construct the CCC by yourself.
And now look at Fig.1 and Fig.2 of this article. Do they compare with yours? They were obtained with "TS 2000i" and are very similar to Djuric's fig.
Is it clear at least now? My undergraduate students have defended a bunch of diplomas in the last 4 years, a PhD student is completing her dissertation. And they've never had a problem with this algorithm.
I recommend a good book to the "backwoods rude" and "Moscow bamboos", who have no idea about the course "Digital Signal Processing", but express their "critical authoritative opinion": A.B. Sergienko Digital Signal Processing. - SPb: Peter, 2002. - 608 с. (note page 216).
There are other books, but they just "don't pull" them. I will not dwell on theoretical aspects. I will only note that all the qualitative declarations in this article have real quantitative estimates and have been published.
Well here is a comparison of JMA (red) and SmirnoffMA (blue). On the screenshot both MAs have period of 12. JMA is faster and smoother. And at periods of 10 and 14 SmirnoffMA disappears at all - it must be some kind of a glitch. I don't see that SmirnoffMA is better yet....
Why compare? We should not compare, but use. As one of the options: consider the Djuric MA as a fast MA, Smirnov MA as a slow MA. The signal is their intersection. Further as in the textbook.
The author is really missing... You get a dialogue of the author without the author
Alexander, can you attach your version of TS 2000i indicator here? Some EA developers will be able to translate it in MQL4. You may compress it with ZIP archiver first, otherwise it won't stick.
....I only note that all the qualitative declarations in this article have real quantitative estimates and are published.
Perhaps you have not noticed my 1st post in this thread. I would like to suggest once again that you at least post pictures. Where Jurik filters together with your filter, and test signals are fed (preferably several pictures which show all properties). Then at least you will have a visual evaluation. As a scientist you should know methods of quantitative evaluation, maybe I missed something but I did not see them in "VS" №01(75) 2006. Comparison of Djuric (and not without him) with your algorithm.
1. Post a grad student in a thread (it's the first time we hear about her and we don't even know if she's pretty?)
2. If the author can't post the post-graduate girl's personal, let him announce the date of the public defence (of the post-graduate girl) and we will come, for sure we will come.
3. Next, discuss Smirnov's behaviour from a trader's perspective, and even from an automatic trader's perspective (rather than whether Jurikkk is right).
4. The posts are real: if you have not seen - write not seen, and if you have not understood - write not understood.
5. Instruct the real Smirnov to refer to and explain in the same post, otherwise we are waiting here.