Zigzag indicator and neural nets - page 5

New comment
 

SK.

In general, when I often see your answers, I notice that we have the same idea about a lot of things. I'd really like to help klot(in appreciation of his libraries and work), he asked a very valid question on page 1 of this thread. "Works pretty tolerably well. I mean, the neural network, whose inputs have readings of this indicator, is practically sunk. But that is not the main thing. The main thing is to find out the algorithm to calculate the probability of Peak or Trough.

Any idea how it can be calculated?"

A neural network is unsinkable if its output is a zigzag, not its input. I was going to give him this idea. And he is perfectly capable of creating and tuning the NS on his own.

 
eugenk: However, Elliot was rather referring to a wavelet decomposition, for his wave patterns are localised in time.
...
The impression is that Elliot knew about golden ratio and Fibonacci numbers, and quite deliberately chose such numbers of waves in the law of combination.
...
So I don't believe in it, for where it comes from is not explained in any way at all.
eugenk, I will answer you point by point. The most important thing: Elliott was an accountant. Apparently he did not know about any wavelets or even Fibs when he wrote his first work on EWP. Probably, number 5 and 3 are chosen just from observations as minimal numbers which describe the structure of the waves. Especially since even numbers characterize an unfinished swing (4 is "up, down, up, down"; where is the final up movement?). The idea of Fibs - and rather not in the context of the number of waves in the larger ones, but in the context of the magnitude of corrections - was not his idea, but that of his friend (Collins, I think), a very strong analyst at the time.

And the explanations... Well, what explanations could there be? It was later, when Elliott decided to explain not only the fundamental movements, but all of the market movements (he worked with the Dow), he began to think up complex corrections, X links between them, failures of the fifth, irregular corrections and other distortions. And he called the wave principle itself almost the most basic law of nature. In short, he was carried away, and his students picked him up and developed him (the strongest among them being R. Bolton).
 
Prival:

SK.

In general, when I often see your answers, I notice that we have the same idea about a lot of things. I'd really like to help klot(in appreciation of his libraries and work), he asked a very valid question on page 1 of this thread. "It works pretty tolerably well. I mean, the neural network, whose inputs have readings of this indicator, is practically sunk. But that is not the main thing. The main thing is to find out the algorithm to calculate the probability of Peak or Trough.

Any idea how it can be calculated?"

A neural network is unsinkable if its output is a zigzag, not its input. I was going to give him this idea. And he's perfectly capable of creating and tuning it himself.


I think the input should be all the available market history in its raw form. And the output should be a cloud of event probability (similar to those shown in your last pictures in this thread). And depending on the shape, size and density (or height in 3D) of the cloud to form trade criteria and calculation levels of SL and TP. To my mind, the zigzag has nothing to do with it.

I hope to be in time for the next championship:)

 
SK. писал (а):

I do not agree that the original series cannot be predicted.

In order to forecast, you need to find parameters that characterise the market and exploit them. For example, repeatability, cyclicality, periodicity, fractals. In other words, it's necessary to reveal the dependencies to which the market is subject to some extent. And when this is successful, then, at the stage of strategy implementation into a program code, one or another tool or a combination of them can be used, and not necessarily from the list of standard ones. Not all of the laws that the market is subject to can be described by simple tools.


For prognosis, we do not need parameters, but an adequate theory. And my statement about the impossibility in principle to forecast should be understood as saying that there isn't a single model (let alone a theory) that would allow it at the moment. Accordingly, all existing tools are equally useless. Just like glasses to a monkey.

And my statement makes perfect sense. You, separating the grains from the chaff, have resolutely thrown ZigZag into the bin. I believe this determination should then either be extended to all other TA tools, or set aside for the time being. Time will tell whether ZigZag is needed or not. Including newcomers.

Mathemat:
"It can't be, because it can never be". And what will respected SK. and Yurixx respond to the existence of respected nen, who, judging by his published balance charts, is trading extremely successfully on the real, using some ZZ patterns (Gartley, Pessavento) and Fibo tools to predict levels of suppo/resistance - without any NS?


Too bad you didn't realise from my posts that I'm just defending ZigZag. I don't know how useful it is, but I think it's premature to give up on it. That's what I wrote.

eugenk:
The problems with all this begin when, firstly, the universality of these very patterns is postulated and, secondly, the 5-3 law is introduced and leads to the Fibonacci sequence. The first is plausible. In fact all the charts except for the minute one look so similar that they are nearly indistinguishable by eye. Though it should be justified in some way. The 5-3 law, on the other hand, is something completely incomprehensible. Why not 4-2 and 7-5? The impression is that Elliot knew about golden ratio and Fibonacci numbers, and chose such wave numbers in the law of combination quite deliberately. So I don't believe in it, for where it comes from is not explained in any way at all.

Once, in a favorite thread on a parallel forum, in a polemic with Niroba I wrote about the ZigZag and the origin of the Elliott ratio 5-3. Only odd numbers can be present in this pair of numbers - this fact has a perfectly clear origin that follows from the structure of the ZigZag. Here Alex-Bugalter has already fiddled with the ZigZag and I am sure he knows why this is so. And the actual values of 5 and 3 are to some extent (one must assume) based on statistics, i.e. they occur most often in real life. And that's all about Elliott's "theory".

To confirm (or disprove) it, one would simply have to analyse these statistics. Having a ZigZag indicator this is very easy to do. However, I, as I do not use EWT, do not need it. I would, on the other hand, advise some novice Eliot to do it as an exercise. The benefits would be many, but most importantly he would get a statistical picture of the validity of EWT. For someone willing to risk their hard-earned money, it makes sense in my opinion.

 
Yurixx писал (а):
.. at the moment there is no model (let alone theory) that allows this.
It would be more correct to say, not published in the open press :)
 
SK. писал (а):
Yurixx wrote (a):
... there is currently no model (let alone theory) that allows this.
It would be more correct to say, not published in the open press :)


I don't agree with you there. There are these works, only they should be applied to the Forex market.

"Machines won't learn to think until humans learn to think"

The first fundamental results in discrete-time filtering theory belong to Soviet scientist A.N. Kolmogorov [1] (1941), and in continuous-time - to American scientist N. Wiener [2] (1942). Scientists R.E.Kalman and R.S.Bucy [3] (1960, 1961) obtained final results on linear filtration theory of Gaussian processes in discrete and continuous time. Fundamental results in nonlinear filtering theory belong to Soviet scientist G.L.Stratonovich who worked out the theory of nonlinear filtering of Markov random processes since 1959 [4,5,6, etc.].

There are also works by Levin B.R. http://www. computer-museum.ru/connect/levin.htm and Tikhonov V.I.

I would give all Nobel prizes to Stratonovich for his GREAT equation. I just think one should be able to prepare it (the equation) for Forex. Which is what I am trying to do now.

  1. Kolmogorov A.N. Interpolation and extrapolation of stationary processes. -Proc. of Russian Academy of Sciences, Mathematical Series 1941, No.5, pp.3-14.
  2. Wiener N. Extrapolation, interpolation and soothing of stationary time senses. New-York: John Wiles.1949.
  3. Kalman R.E., Bucy R. New results in linear filtering and prediction theory, ASME trans, J.Basic Ehg, March, 1961, V-83D, p.95-108.
  4. Stratonovich R.L. Conditional Markov processes. -Moscow: Lomonosov Moscow State University, 1966.
  5. Stratonovich R.L. About a priori-conditioned quasi-optimal filters. - Radiotekhnika i elektronika. 1981.
  6. Stratonovich R.L. Principles of adaptive reception. -M: Sov. Radio, 1973.
 
Prival писал (а):
 There are these works, only they need to be applied to the Forex market.

I was talking about publications on the practical application of scientific research to forex problems.

And as far as mathematics is concerned, we should all be studying. That's not even up for discussion. What a zigzag...:)

 
Yurixx писал (а): I wish you had understood from my posts that I was just defending ZigZag. I don't know how useful it is, but I think it's premature to give up on it.
Yes, Yurixx, I see it now. But here's how to practically compile that very statistics for Elliottians, I have no idea yet (if we take into account the perversions of Elliott itself within EWP, which I noted in response to eugenk). Roughly speaking, an (Elliottian) wave does not necessarily end at the ZZ swing break point of the same scale.
 
Mathemat:
Yurixx wrote (a): I wish you understood from my posts that I'm just defending ZigZag. I don't know how useful it is, but I think it's premature to give up on it.
Yes, Yurixx, I see it now. But here's how to practically compile those very statistics for Elliottians, I have no idea yet (if we take into account the perversions of Elliott itself inside EWP, which I noted in reply to eugenk). Roughly speaking, an (Elliottian) wave does not necessarily end at the ZZ swing break point of the same scale.


So it is not the same scale.

I would simply find a classical figure 5-3 on a chart of a suitable price, select the ZigZag parameter so that it reproduces this figure, and then calculate, on the one hand, how many such figures on a specified interval of history and, on the other hand, how many trends of no less than a specified size (for example, the size of 5 waves in total) have happened over that time in total. It's not all statistics of course, but it's simple and accessible, and shows the total share of the 5-3 pattern in the sum of all market patterns in this piece of history.

This is just the simplest idea of calculation. A number of subtleties need to be added to it, but these are subtleties of a technical nature, everyone can solve them on their own.

 
Prival:


I don't agree with you there. There are these works, only they need to be applied to the Forex market.


I agree with SK. We were talking about the theory of the market, not about the mathematical methods that can be applied to it. I will not go into the details of it, I will not comment on it, I will not comment on the correctness of it.

But thanks for the books. It is very interesting.

By the way, what is "zinear filking" ? If it is "linear filtering", it would be good to correct it. And at the same time look through all references for errors. We have to look for these books. :-)

1234567891011
New comment
Reason: