Stochastic resonance - page 23

 
Yurixx:

This is great! I've seen a lot of newbies or just horny people trying to measure their pips, and here - grandfathers trying to measure their beards! I want in too. :-)))

I've been working on my system for two and a half years now. But I have never changed the direction of my research. And probably I won't. If nothing comes of it - I'll have to retrain as a manager.

I think grasn has something to boast about, too.


What a word - grandfathers. There are no grandchildren, of course, but age is measured differently in forex. I don't mean forex, but in general, it is measured differently. I think this is the first time he has been called a grandfather in forex.
 
Mathemat:
Vinin wrote (a): It's just that I'm a bit stuck after the contest preparations. It's still a long way off. It took me three weeks to do the advisor. My family cried during that time. And mistakes are always made during a race.
So what, three weeks... I've been working with my Fibo-system for almost a year, not knowing what I will get (I cannot do it manually, until the algorithm of calculations is debugged - vicious circle). Just as with the eye: Darwinists argue that the eye came about as a result of evolution, while their opponents argue that it was immediately made such as it is now. Same thing with my Fibo-system: you work for a year, you work, you neglect your wife, you are struggling with your son 14 years old, what he can still do - but the system is still not done. You have wasted a year and a half on the National Assembly, haven't you?

So it turns out that the process is the highest, absolute value ... .
Hello, lads!
Since we're having such a binge on ale, I'd also like to give my opinion.
In my experience, it's hard the first five years, then you get used to it. And it's no fucking joke. (In my opinion). I started my first system in Omega in 1999, trying to trade at the same time. In two years I lost several deposits (there was no small lots back then). I am not sure where I got it from. I can't remember where I read it - man is formed as a trader in 5-7 years. Now I'm doing this shit again (what can I do?), but I try not to hurry (no matter how fast life occurs), because as soon as I "hurry", I make a mistake or fail, or in extreme case I lose the same time. On the forum, unfortunately, only half a year, but I caught myself thinking that I can no longer communicate without communication and would be happy to work in a team.

I also wanted to comment on the resonance:
Have read all the material from this thread, very interesting, but loaded not in a childish way. I've got too many windows(windows(c)) open in my head, and the stochastic resonance steals huge amounts of RAM. But never mind.
I spent many hours with soldering iron and oscilloscope (like many others) during 80's, and this phenomenon is well-known to me on intuitive level, and I think, that it can be and should be used. So far I don't have my own thoughts about its application, I just digest it, but grasn's approach seems to me rational:

The important thing is that it will not actually have any predictive qualities. And therefore I think that practical application of this direction is in development of a tool, probably more adequate and precise than many calculation methods of volatility I do not like so much. Most probably this tool should according to a "tedious" algorithm search for the side channels (flat), hold them for some time and control the noise level as soon as the level reaches the optimum value, then... it does not actually mean anything, except that one of the channel development limitations indicates a high degree of the local trend appearance, not just a trend, but comparable in size to the channel borders. .......

Regards,
Vladimir
 
Vinin:
Yurixx:

That's the spirit! I've seen a lot of newbies or just horny people trying to measure their pips, and here - grandfathers trying to measure their beards! I want in too. :-)))

I've been working on my system for two and a half years now. But I have never changed the direction of my research. And probably I won't. If nothing comes of it - I'll have to retrain as a manager.

I think that grasn has something to boast about as well.


What a word - grandfathers. There are no grandchildren, of course, but age is measured differently in forex. I don't mean forex, but in general, it is measured differently. But Niroba has been poking around for a year and a half with his demo results. He is the granddaddy of demo forex for sure.


Precisely because age in forex is measured by other things, I say "grandfathers". There are grandfathers in the army too, although they are all under 20. The presence of grandchildren or even children is not necessary.

In my opinion, grandfathers are those who have grown out of a fussy and irresistible desire to make a quick buck. The eye can see... There it is, right there.

Grandfathers are those who have understood their experience and have benefited from it. That is why they approach the question thoughtfully, patiently, calmly and purposefully. He or she does not chase some grail, does not show off in pips, does not count on miracles, does not ask others for an Expert Advisor or an indicator. But he is counting on himself, doesn't bother programmers (MQL is not such a problem), and digs, digs, digs .....

In line with this, Niroba can hardly be called a grandfather. :-)

 

to Yurixx

I think grasn has something to brag about as well.

I can, and have already bragged about it, including preliminary results, so far in MathCAD. Let's see what happens in MT, but knowing how the model works, I'm sure the results will be good here too. The model came out thick-skinned. I don't think bragging is anything disgusting and thank you among others for the valuable advice, after publishing the first conceptual test of the truncated draft system on methacvotes.

.... Thanks to this discussion I realised that firstly, the sum of values in a muwing can rightly be considered the sum of any values in a series, rather than the sum of consecutive values. Reason: the evaluation of the limits of the area of change is the evaluation of the BEFITS, not the current values. In addition, the minimum (maximum) of a moving average is obtained when the X value passes its minimum (maximum) - almost all elements of the moving average are near the range boundary - quite a realistic situation. This is also true for the price.

Second, due to the above, the integral equation, the solution of which may yield values Ymax and Ymin, isS(p(x)dx) = M/N. Here S(...) is a definite integral, p(x) is a function of the probability density function of series X. To determine Ymin, the integral is taken from 0 to some X1. As the result we obtain an analytic equation (if the integral is taken in analytic form) with respect to X1. Then by calculating the average value of X over this interval [0,X1] we obtain Ymin. .....

I don't quite understand how this helps to get a result, and how the probability convolution described in any, however serious, book on probability theory can help to find the universal curve we were talking about earlier? And S(p(x)dx) = M/N - is that right? But never mind, judging by a confident post it probably can be easily deduced, apparently it is the shortcomings of my purely pragmatic, engineering education - not a theorist at all, I was taught that if anything is to be done, so that it would really fly, not fall down and not kill people.

A story with a beard, so to speak. I decided to share more practical results, namely methods I've been using for a long time and now have given up. But recently, I discovered an interesting pattern. The notion of "sharing" is probably not very appropriate, if you read to the end, rather a "hint". Let me remind you a little bit of the basic idea. A low-pass filter is applied to the price range, designed for a certain specification (passband/slope frequency limits, nonuniformity coefficients, and so on). The first step gives a picture like this:


The better the filter is chosen, the closer the price range "on average" is to the resulting low-frequency signal (then simply the signal). The signal itself has several undeniable advantages:

  • a maximum is always followed by a minimum and vice versa
  • Knowing what the current local extremum was, we can say for sure what level the next local extremum will not go higher/lower than. Of course, this only applies to the signal itself but may be very useful when combined with estimation of "noise".

But there are some unpleasant characteristics

  • the main one being a significant delay in fixing the current local extremum

These are the qualities that I was going to exploit once, building my strategy on them. It's simple, we buy at the minimum and sell at the maximum, just like Schwager. The next step is to transfer the local extremums to the price series. After this operation, a sort of zig-zag appears. At the same time, the signal takes into account "energy structure" of the initial data and is in some sense better than a classical zigzag whose validity of extremums is very doubtful to me (and which I dislike as much as volatility). Having obtained such a zig-zag on a price series, one can already pass over to the notion of waves:


I will skip all artful trading schemes that I invented, but tell you about phenomenological dependencies for searching of which I collected statistics on numerous parameters of channels with the purpose of finding some relation between the parameters of the current wave and the most interesting ones, important for trading: the length and the spread of the future wave. So, I have not found any correlation. I also paid no attention to the classic zigzag - all in vain, there were no clues. But just recently, working on my model, I decided to try another very simple parameter "X" that does not belong to any wave, but nevertheless is a connecting element. The chart shows objects and links for some specification of the LFL (a lot of other parameters were removed as being of no importance for a certain analysis):

The number indicates the correlation between the objects. It can be seen that there is no direct relationship between the wave/channel parameters, but there is a significant relationship with this parameter "X" - the correlation value is 0.7 for the whole data set. If we carry out classification, the correlation rises to 0.9(which is actually a law). In other words, if we know a certain class of the current wave (trend/correction), we can determine the parameters of the expected wave - the spread and the RMS according to the found correlation (formula with the confidence interval estimation) very accurately. These same conclusions are confirmed by the Data Mining analysis. The calculation path is actually shown in the diagram. It is also possible to duplicate/verify the results obtained by different routes.

I have been working on my system for two and a half years now. But I have never changed the direction of my research. And probably I won't. If it doesn't work out, I'll have to retrain as a manager.

Yury, it is curious, you have got something similar, otherwise how to explain your determination and devotion to this direction, if I haven't confused anything. :о)

PS: what you said is also true for more complex structures:

  • a trend based on a previous trend
  • correction on the basis of a previous correction

In addition, a curious article http://alpari.ru/ru/ew_article/25780.html. But I could not find any proof of the statistics referred to by the author, although I have been dealing with wave theory for quite a long time and even use this structural approach in my model. Yes, one cannot understand waves by brain - one must believe in them. :о) For those who want to try to look for patterns, let me remind you of the basic rule: a tuple must contain all parameters of the object under study or a set of them.

According to this, Niroba can hardly be called a grandfather. :-)

I was enlightened here, it turns out that Alex is a whole interesting phenomenon, already attracting the attention of experienced psychologists, with its own name:

https://ru.wikipedia.org/wiki/%D0%A2%D1%80%D0%BE%D0%BB%D0%BB%D0%B8%D0%BD%D0%B3

Anyway, he's a godsend for any DC. When I'm sad, I always look at his posts and my mood immediately lifts :o)))

to eugenk

Although not asked:

Alas, the relevant one in the parallel forum is clearly unreadable.

It's hardly possible to write clever things all the time, sometimes people need to relax, for example Candid has recently published a cat - a cool cat, I liked it, he was so fat. And he added a photo of another cat - a millionaire who's going to (what a funny thing) go to space. Alex immediately calculated how much time he would need to trade, to pay for an excursion to the same destination and get ahead of the cat. By my reckoning he should blow all the sponsors of the real championship to the wind. :о)

to eugenk, Candid

Colleagues, the support/resistance levels by definition cannot be potential pits. The price is almost never at those levels, while the object is in a potential pit most of the time. Then already talk about levels, at which the price "sits" for a long time. But I assure you that serious disappointment awaits you with the potential. Be surprised that the process under discussion has nothing to do with fictitious potential pits, or fictitious support and resistance levels. :о)

to VBAG

In my days in the 80's I spent much time with a soldering iron and oscilloscope (like many people) and this phenomenon is intuitively familiar to me, so I think it can and should be applied.

Greetings! My logic is simple - since you sat with a soldering iron and oscilloscope, it means you know DSP more or less, and if so, you probably know how adaptive filters are designed. Perhaps some questions on them will come up in the future, can I get in touch? :о))))))

 
grasn:

to VBAG

Greetings! My logic is simple - since you sat with a soldering iron and an oscilloscope, it means you more or less know DSP, and if so, you probably know how adaptive filters are designed. Perhaps some questions on them will come up in the future, can I get in touch? :о))))))

Sergey!
Your email is not in my profile. Write to me, I'll be glad to chat with you.

P.S. You're right on the money, speaking about DSP - I know exactly more or less! I'm more of a practitioner and looked more into specifications of operational units and microcircuits than into Zalmanzon. I've only recently gone back to it.
 
grasn:

to Yurixx

I think grasn has something to brag about as well.

I can, and have boasted before, including of the preliminary results, ...

I do not think bragging is a disgusting thing and I am grateful to .... among others.

Well, Sergey, what difference do the results have, including preliminary, when aksakals measure their beards? And since you have nothing against bragging, show us your beard, pls. Show how long it is. :-))

I have been working on my system for two and a half years now. But I have never changed the direction of my research. And I probably won't. If nothing comes of it - I'll have to retrain as a manager.

Yury, it is curious, you have got something similar, otherwise how to explain your determination and devotion to this direction, if I haven't confused anything. :о)

My system is totally different from yours. And the purposefulness and dedication seem to have two reasons: 1) its potential has not yet been exhausted or even realised and 2) I simply do not have any more worthwhile ideas. That is, there are individual interesting ideas, of course. But there are some that would form a system - alas.
 
grasn:

After this operation a kind of zig-zag is obtained. In this case, the signal takes into account the "energy structure" of initial data and is in some sense better than a classical zigzag, the validity of which for me is rather doubtful (and which I dislike as much as volatility). Having obtained such a zig-zag on a price series, one can already pass over to the notion of waves:

I will skip all artful trading schemes that I invented, but tell you about phenomenological dependencies for searching of which I collected statistics on numerous parameters of channels with the purpose of finding some relation between the parameters of the current wave and the most interesting ones, important for trading: the length and the spread of the future wave. So, I have not found any correlation. I also paid attention to the classic zigzag - all in vain, there were no clues. But recently, working on my model, I decided to try one more very simple indicator "X".

Personally I don't consider a zigzag as an indicator at all, to me it is just a style of chart layout :). So really it's all about how the extremes are defined. I had other ways, but, characteristically, I still haven't found any connection between the characteristics of the current ... er ... wave (I do not like this word :) and characteristics of the future one. That is, parameter X is such a find, which can only be envied. And congratulations, of course :).

grasn:

Colleagues, support/resistance levels by definition cannot be potential pits. The price is almost never at these levels, while the object is in a potential pit more often. Then already talk about levels, at which the price "sits" for a long time. But I assure you that serious disappointment awaits you with the potential. Be surprised that the process under discussion has nothing to do with fictitious potential pits, just as it has nothing to do with fictitious support and resistance levels. :о)


Let these statements remain on your conscience, grasn:)
 

to VBAG.

Email sent. Really, no email in my profile, but I've added it.

to Yurixx

Sergei, what do the results matter, including preliminary, when veterans measure their beards? And as you have nothing against showing off, let's see your beard, pls. Show how long it is. :-))

On the 'current' model, which is now being transferred to MT - five years (bragging right there, without leaving the till: 'Stochastic Resonance';; (14.10.2007 12:49). I hope we won't go any lower than the beard? Otherwise there are a lot of all sorts of offshoots there - they'll get banned, won't they? :о))))

My system is absolutely nothing like yours. And purposefulness and devotion seem to have two reasons: 1) its potential has not yet been exhausted or even realized and 2) I simply have no worthwhile ideas. That is, there are individual interesting ideas, of course. But alas, I don't have anything that would form a system.

I didn't expect it to be absolutely similar, it's just that we had a long discussion about extrema (started here: https://www.mql5.com/ru/forum/51428), including via mail, and I understood then, that our approaches were very similar. But it's not important at all, though I would like to know your approach conceptually, being very curious :o)).

to Candid

Let these statements remain on your conscience, grasn:)

grasn has a conscience, (here again Yuri will say - bragging) if there is any use of potential energy, it's in determining the "stability" of the channel, but not in calculating the future "pit". Although, technology is marching by leaps and bounds, they say some nano technologies have appeared. You're not using it, by any chance? :о))))

I personally do not consider the zigzag as an indicator, to me it's just a style of marking up a chart :). So really it's all about how the extremes are defined. I had other ways, but, characteristically, I still haven't found any connection between the characteristics of the current ... er ... wave (I don't like this word :) and characteristics of the future one.

Right! Just in case - I will highlight and take on my conscience this modest statement, which in fact is nothing else but my own opinion: the statistics based on wave theory and Fibo levels is, to put it mildly, but roughly speaking, wrong. This wave is three times more frequent than this one - nonsense, it does not exist as a class. And by definition it cannot be so when two "wave-patterners" may mark one and the same area differently. Statistics is an exact science, you get as much as you pay for.

So, parameter X is a lucky find and one can only envy it and congratulate you, of course :).

Special thanks and good luck hunting for your parameter, except that I've abandoned the approach described long ago, and now I'm thinking of returning to it as a supplement. :о)))

 
grasn:

Right! Just in case - I will highlight and take on my conscience also this humble statement, which in fact is nothing other than an opinion of my own: the statistics put under wave theory and fibo levels are lying to put it mildly but crudely.

It depends on which levels and from which swings. If from the swings of one wave order, as Svaney does, then it is really rubbish: distributions of retracements are very smooth, without sharp spikes on fibs. But nen is somehow thinly (hehehe...) on Fibo instruments, right?

P.S. Should I read DiNapoli?
 
Mathemat:
grasn:

Right! Just in case - I will highlight and take on my conscience and this humble assertion, which in fact is nothing but my own opinion: the statistics put under the wave theory and Fibo levels to put it mildly, but roughly speaking lies.

It depends on what levels and from what swings. If from swings of the same order, as Svaney does, then it's really rubbish: distributions of retracements are very smooth, without sharp spikes on Fibs. Butnen is somehow thinly or poorly (hehehe...) working with Fibo tools, right?

My "righteous anger" is mainly directed at waves, with fibs it is a bit more complicated, there is indeed something there, but a bit different. The parameters that the authorities suggest to use for fibos do not work, there are no patterns - Chaos. Maybe, heh heh - looked in the wrong place and in the wrong way - not the point, the important thing is that I was interested in evaluating "predictability" based on correlation analysis and Data Mining and found some features, shared some of them.

And there is indeed a lot of nonsense, for example a certain Gunn assures that 45 degree movement is meaningful - fuck it, it means nothing at all.

PS: Poorly, hehehe, everybody works somehow, besides, I just don't know how exactly nen works, maybe you do?

PPS: Reread the post, you'll probably get it slightly wrong. Fibo does work, but in a slightly different way. And ... never mind - it doesn't matter.

P.S. Have you read DiNapoli?

If you're asking me, can you tell me how to use a smiley face to indicate a shrug? It's a joke - read it if you don't have anything more interesting at hand. :о)