That's interesting - page 12
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
is the constancy of the ratios of the golden ratio. It's a manifestation
of the now widely studied property of self-similarity." p.359
I'm not referring to Schnoll, just perception. Someone is very critical of everything.
How do you take it, then? Schnoll says, you should have it like this, because my squirrels have it like this, I say, no way, it's not like this, it's the way it's supposed to be.
All right. Forget it. You don't get me. There's nothing more to talk about.
All right. Forget it. You don't understand me. There's nothing more to say.
So, as they say, "closer to the body".
To be honest, I'm starting to regret having started this topic. Almost everyone is trying to kick, not even xyz do not understand what it is about and for what purpose.
But since I've started, decided to put my suggestion now, and do my business. Will not have time at the weekend, and what I wanted to tell more about belongs to my system, along with the charts. I will have to explain some of the results, I will have to talk about the system, and that will require even more time and desire, which has almost disappeared. Why do it? Many people don't burden themselves with it at all, drawing completely unquestioning conclusions. Therefore, I will be brief, or rather very brief.
What Shnol has discovered is genius, and even if it is not confirmed in any way - just for thinking outside the pattern, the academician can be given a monument even now.
There is a huge problem - model identification. Of the many methods, which are more or less universal, the method of maximum likelihood stands out. This method has a simple and clear concept: parameters of the model are estimated by maximizing the likelihood function. The parameters that maximize the value of the likelihood function are interpreted as the most likely ones. The function itself is a joint probability density function.
Those who have tried to identify models on the basis of this and other methods, I think they will agree that it is not a worthwhile task. For complex models it is practically impossible, even for time series with normal distribution (the form of distribution is not so important here).
Here, for example, a lot is shown:
Schnoll, in fact, drew attention to what is really objective - histograms. Introduced the term "fine structure", emphasising this. Not some mythical distribution (let me remind you, this is statistics, experiment, work with frequencies). But completely objective, real shit. The actual algorithms by which the experiment is processed "iron out" this data, removing all "knowledge", and making much fundamentally impossible. By the way, one of the "reliable" methods of determining the type of distribution is again maximum likelihood.
And I completely disagree with HideYourRichess.
Это распространённое заблуждение. Уложить что то в матстатистику, так просто - нельзя.
It's not that simple and unambiguous. Not at all. In many cases you can. Very many cases. Matsatistics, exactly, is one of the, so to speak, "near-scientific sciences", if you can put it that way literarily. I've been dealing with trend and chance statistics for a very long time - it's a "mess" there in general, and a well-grounded and scientifically proven one at that.
So, starting from the second part, the academician's memoirs contain a very good vision of actually a new approach to model identification. But colleagues - don't speak about stars, believe me, they have nothing to do with it yet, and I do not understand them at all. So, I have some "vibrational phenomena" that may increase the effect, although the "Influencing Factors" (the terms have not been settled yet) are a separate subject.
I will add a little more - when for myself I introduced the concept of "dead statistics". It happens so, everything seems to be correct, and the process is simple, and the model is theoretical and the distribution, I know exactly what it is - but it doesn't work. But it is necessary to make some adjustments, and that's it.
Well it is impossible, in principle impossible to identify the process by (say) 10 prediction points and the theoretical probability density function. You have to be more subtle here.
to HideYourRichess
What can I say regarding your categorical statements. Put out the histograms. The academic has laid them out, if you do not lay them out it is pointless to talk. I'm not interested in talking like this, in such a dialogue.
You have read a couple of paragraphs, quite innocuous about Fibo numbers in his work and just "moved away". And I think it's bullshit, but hone your snide attitude on others.
THE ACADEMIC IS NOT SUGGESTING THAT YOU TAKE YOUR MONEY AND TRADE ON FIBO AND HIS THEORY. AND I'M NOT SUGGESTING THAT EITHER.
(hope you pay attention)
Prove, argue, if you are serious about something, or do not waste my time. Otherwise "someone saw something". Do you think that only you know statistics? Sorry, but you're like a student who was taught to do something and just fucks around strictly according to the algorithm, and everything that comes out is "bullshit".
Suppose you don't agree, you think it's nonsense - you have every right, but never mind - no one is going to risk their lives in testing. You have expressed your opinion, why are you chattering like a woodpecker "it can't be", "it can't be". I personally heard it.
Above, in blue, is the change in tick volumes averaged over the days of the week for 2009. Below is a picture from Schnoll's work.
No robot will find a pattern, but a human brain will immediately have ideas. Uncle Elliott's 8 wave formation? ))))
Farnsworth:
It's not that simple and straightforward. Not at all. In many cases you can. Very many cases. Matsatistics is one of the, so to speak, "near-scientific sciences", if you can put it that way literarily. I've been involved in trend and randomness statistics for a very long time - it's a "mess" there in general, and a reasonable and scientifically proven one at that.
. There is no "mess" in matstatistics, but there are plenty of people who try to use its methods thoughtlessly. So, the "mess" is in their heads. The applicability and possibilities of matstatistics have long been studied and known. Matstatistics is not omnipotent and is not the mother of all sciences, it has its own corner among scientific methods and does not pretend to be more.
Farnsworth:
One cannot, as a matter of principle, identify the process by (say) 10 prediction points and the theoretical probability density function. You have to be more subtle than that.
. I'm just sure, just by the fact that you have already referred to it somewhere, you have read Shiryaev's lecture, about Renko and Kagi volatility. Why do you repeat the mistakes that Shiryaev explicitly says, "What is your process?" You see? First you identify the process, and only then can you talk about "10 prediction points and probability densities". But not the other way around.
Farnsworth:
So, starting from the second part, the academician's memoir contains a very good vision of actually a new approach to model identification.
. Just in case, not to repeat myself, I have read the second part carefully enough. In fact, it suggests comparing histograms "by eye". It also describes a sad story of how 5-7 people in different years tried to formalize this process - in general, there is no reasonable criterion, except for the fact that "I see so". There's a lot more out there, surprisingly.
Farnsworth:
What can I say about your categorical statements. Put the histograms out there. The academic has posted them, if you don't post them there is no point in talking. I'm not interested in talking like this, in a dialogue like this.
You have read a couple of paragraphs, quite innocuous about Fibo numbers in his work and just "moved away". And I think it's bullshit, but sharpen your snide attitude on others.
. I see a terrible misunderstanding here. We're criticising the amazing results of an Academician from the RAEN (Academy of Charlatans) here. Not my work. My work, these are commercial results, the property of the company. I cannot, I have no right to publish them. And I have already voiced the conclusions - there is nothing of the kind in my data.
Farnsworth:
I AM NOT SUGGESTING TO TAKE YOUR MONEY AND TRADE ON FIBO AND HIS THEORY. NOR AM I SUGGESTING THAT.
(hope you pay attention)
. Did I ask the sacramental question "Where's the money, zine?" or did I ask looking intently into his eyes "Pucker up the stairs"? No, I don't care about that, in this case.
Farnsworth:
Prove it, argue it, if you are serious about something, or don't waste my time. Otherwise "someone saw something". Do you think you are the only one who knows statistics? I'm sorry, but you're like a student who was taught to do something and just fucks around strictly according to the algorithm, and everything that comes out is "bullshit".
Suppose you don't agree, you think it's nonsense - you have every right, but never mind - no one is going to risk their lives in testing. You have expressed your opinion, why are you chattering like a woodpecker "it can't be", "it can't be". I personally have heard.
. Somebody here has already said it, I'll just repeat it - I'm against sexting.
Above, in blue, is the change in tick volumes averaged over the days of the week for 2009. Below is a picture from Schnoll's work.
No robot will find a pattern, but a human brain will immediately have ideas. Uncle Elliott's 8 wave formation? ))))
Hello, Seryozh. I didn't immediately realise you were you. You changed your nickname. Then, like, everything is clear.)) Then... I decided to take a pause. I wonder... what they're going to agree on.
Bottom line.
I haven't been a scientist for six years, but I stick to methodology. In particular, the razor device of a medieval British monk: Don't create unnecessary entities unless absolutely necessary.
Scattered results? Within this methodology? Right, right... Doesn't this methodology involve washing your hands before an experiment, say? Or taking induced radioactivity out of the box?
Literally (about "washing your hands") Zhvanetsky comes to mind: You must be more careful, guys!
I happen to know a little bit about biophysics. I have been told that literally everything affects the experiment. Right down to whether the lab tech has her period.
As for the radiation... Hmm... You're sure you've heard of Peter Kapitsa? He was a genius of experimentation. Well, here goes. I don't know how and with what someone measured what, but it was clearly not Kapitsa. Removing side effects on measurement is the main problem of a physics experiment.
Figuratively, "you have to wash your hands."
Alas and nah.
===
))) Respectfully. Accept and other...
. There is no "mess" in matstatistics, but there are plenty of people who try to use its methods thoughtlessly. So, the "mess" is in the head. The applicability and possibilities of matstatistics have long been studied and known. Matstatistics is not omnipotent and is not the mother of all sciences, it has its own corner among scientific methods and does not pretend to be more.
Yes you are right of course, technically. And I went a bit overboard with the "mess", but there are quite a few problematic situations - purely practical. Fact.
I'm just sure, just by the fact that you've already referred to it somewhere, you read Shiryaev's lecture, about renko and kagi volatility. Why do you repeat the mistakes that Shiryaev explicitly says, "What is your process?" You see? First you identify the process, and only then can you talk about "10 prediction points and probability densities". But not the other way around.
You seem to be diligently failing to understand. First an assumption is always made about a process about which little is known a priori, a criterion is developed, then a series is taken and only then do you identify the process - test your assumption. That's the only way. Identification methods don't work without the process itself to be identified. (that's just silly). Whether it is predictive or from somewhere else is irrelevant in this case. The likelihood function is a concrete thing, not abstract, it needs a series (natural/forecast/...whatever) to work.
Assume that the process, with some distribution, is close to e.g. AR(p), or arima(,,,) or whatever. but what p to take? If you know Rasr, you will derive a likelihood function and use some optimal method to find p that corresponds to the max of this function .Shiryaev meant something else entirely, and nobody repeats mistakes.
Just in case, not to repeat, I have read the second part carefully. In fact, it suggests comparing histograms "by eye". It also describes a sad story of how 5-7 people in different years tried to formalize this process - in general, there is no reasonable criterion, except for the fact that "I see so". There's a lot more out there, surprisingly.
I'm not suggesting that we should pray for this work. I merely made a suggestion as to how identification methods could theoretically be made more effective. Did I write in a way that I have already proved something or have you made such concrete arguments after which you must ask the admins to delete everything :o) Why so categorical?
. I see a terrible misunderstanding here. We are here to criticize the amazing results of an Academician from RAEN (Academy of Charlotans). Not my work. My work, these are commercial results, the property of the company. I cannot, I have no right to publish them. And I have voiced the conclusions - there is nothing of the kind in my data.
You have a misunderstanding. I did not invite to criticize. I myself can criticize everything and everyone. I don't give a ... on these REANs and stuff. I invited you to read and discuss, giving your, so far, conceptual vision. And I don't really need your secret data.
. Did I ask the sacramental question "Where's the money, zine?" or did I ask staring into the eyes "Pucker State"? No, I don't care about that, in this case.
You're not alone, and I don't give a shit.
. Someone's already said it, I'll just say it again - I'm against sexting.
And I'm not a supporter of sexting. Where do you see sextant? Why are you imagining things? Maybe I was suggesting between the lines to take over the world as well?
The nature of changes in tick volumes has long been well known and described. There is no need to look it up, everything is clear.
I completely agree with you there. I even got bored :o(