That's interesting - page 15

New comment
 
Yurixx:


I had already encountered this restriction some time ago. Obviously, I used a variable rather than a constant to organize the output.

I just recently had this very question and made a script to test it. There is definitely no limit of 255. Where it really is, those who want to find it can look it up themselves, I didn't.

int start() {
  string Out = "Write";
  int i = 1;
  while (i<=100) {
    Out = Out + "String" + i;
    i++;
//    Print(Out);
  }
  int h = FileOpen("TestString.txt",FILE_CSV|FILE_WRITE);
  FileWrite(h,Out);
  FileClose(h);
  return(0);
}
 
Farnsworth: Need a table for EURUSD, CLOSE, MINUTES:
  • Columns: minutes, from Monday to Friday
  • Rows: weeks

As much history as possible, the rows should be numbered in order of weeks, i.e. as if the time series should be "recorded" with a new row after the trading week ends.

Seryoga, well, it's kind of easy at all. But keep in mind, that each line of the Excel file has a different number of columns (the number of minutes in a day is different). Accordingly, the numbers in a given column will not necessarily correspond to the same time of day.

I'll try to sketch a script. You can choose the separator for the excel, I will put it into external variables. And the problem of quality history, without holes, is up to you.

 
Farnsworth:

It seems to me that you simply did not get beyond simple truths. When you get an estimate of the distributions for a series - each is valid, and differ by fractions. What a fucking misunderstanding. And selecting will lead to completely different results.

Farnsworth:

I'm not interested in what you mean by that.

Farnsworth:

Don't upset me with your stupidity of misunderstanding. Tired of reading your assumptions about me. Otherwise I'll start making assumptions for you. Do you think that's right? I get the feeling you're very, very far from practice.

. That was beautiful!

Farnsworth:

Sorry, but Schnoll has done more for science than you and no less than a "real academic". He writes clearly and precisely what his work is, where it came from and why he continued all his life. His main activity is different. At least the country owes (among other things) to him the emergence of a new and necessary department at Moscow State University. Why are you sneering here?

. Why can't I be snide? Especially if there's a reason there? Again, many times I have pointed out that the professor's reasoning about proteins is probably correct. But when he tried to make a great discovery in another field it turned out to be sad.

Farnsworth:

Yeah, Schnoll left crumbs, didn't wipe his hands, ate herring, and then measured up afterwards. What the fuck else are you gonna write?

. You must have a grudge in your eye. So I'll write again. The professor made a mistake in some of his experiments. He measured not only the parameters of interest, but also the effect of cosmic rays. That is why he has this dependence of results on eclipses and other things. Do you have any objections to this?

Farnsworth:

You're so stubborn. Am I looking for cosmic patterns on fox? No. I'm interested in Schnoll's work in a different context.

. Nivaros! If you're not looking for cosmic regularities on fore, and if that's not what you're interested in from the brilliant work of the Great Schnoll - then what are we talking about? Comparing histograms? You quoted in the topstart - amazing stuff.

Farnsworth:

You think you're stomping on Schnol like that? You're showing yourself.

. Shame on me some more!

Farnsworth:

You're exaggerating. It's easier for you, you don't have to go into it and understand it - and you want to write, you're a writer by nature (I don't know where to put the accent). Although, you know, I think you're more likely a teacher. Bad teachers have qualities inherent in you.

. Yeah?! And bad students have qualities inherent in you. So?

Farnsworth:

What are you talking about? Should I order the services of a psychic to understand what thoughts you have in your head, and what is left of them until your hands reach the keyboard?

. I say be more critical, all sorts of nonsense.


. And there's one more thing that surprises me. Usually, by the good old tradition, any scientific work is accompanied by a review. Even Fomenko's work has a review on "mathematical part" where reviewer Shiryaev writes that the mathematical method applied is quite a sound method. (About the "historical part Shiryaev does not write anything approvingly).

Schnoll does not have such a review. Either he didn't let the mathematicians review it, or they turned him down on various pretexts. All this should be alarming!

 
Candid:
Sextants:)
It's funny, I won't argue.
 
Farnsworth:
Get in.

I haven't read Schnoll, but the judge ... :) some insights from a cursory glance have developed.

Since he deliberately works with "untenable" histograms, he has to use specially devised methods of analysis. I did not understand them in detail, but my indirect impression is that they are, alas, already unquoted.

But of course this is not Petrik or even Fomenko. The mere fact that he honestly describes both the failures and how people left the subject commands respect.

Apparently he has reasons for the truly amazing persistence with which he continues to search for patterns after so many failures.

 

Let me explain a little bit and still draw a diagram. For those who are left in the tank and can't fit through the hatch due to fullness. Sort of a joke :o) But rather out of a desire to be simply understood. And I promised some time ago to tell my colleagues about my system.

Market model

After a long search, as a working version of market model, I've adopted this thing "control systems with random structure". In my opinion (though not mathematics) - this model adequately describes the quoting process with all its subtleties.

Its essence is very simple. There is a finite number of structures that describe the transformation of input into output. Each such structure implies some model according to which the transformation takes place. The observed process is formed by a transition (switch) between the structures. All this is shown in the picture below:

Each model has a set of parameters, which can still change with each switch.

Adaptation to practice

Everything is great - but it is impossible to accurately identify such a system. Therefore, I introduce a "combined model": A=W(1)MODEL1(parameters)+ W(2)MODEL2(parameters)+....+ W(n)MODELn(parameters). Where W(n) are some weights of participation of these models in the prediction.

What do I work with

I don't work directly with quotes - it's an extremely complicated process. I use all sorts of cunning transformations, but what I've said also applies to them. The complexity is not going anywhere - it is inherited. It is impossible to simplify the process. And if you do simplify it, you can lose the process itself.

Analysis of time series evolution

Basic stage. At this stage, all possible structures are identified by some criteria. Statistics of switching between these structures is estimated. Transition frequency matrix for the structures is determined. In the future, I am thinking to use the so-called impulse neural networks (or wave networks). It is a very promising direction.

Algorithm

  • (1) by making some assumptions about behavior, a probabilistic estimation of the future state of the system at the given moment on the planning horizon is performed. The neural network crawls through the resulting initial state probability assessment matrix and in turn makes an assumption about the presence of an entry/exit point. It can very accurately tell whether or not there will be an entry/exit on the planning horizon. All that remains is to find it.
  • (2) A command is given to build an accurate forecast. It is performed:
  • - identification of the "on" current structure
  • - assessment of the choice of the most probable future structures
  • - identification of parameters of future models
  • (3) Next, the neural network estimates the coefficients of the combined model.

The results

Not to brag, I have nothing to brag about - I wrote here https://www.mql5.com/ru/forum/128060/page27. It is too early to learn MQL - I will improve the system up to the level of my requirements.

Conclusions:

Strange as it may seem, but the quoting process as a "whole" does not exist. Following this model and philosophy it makes no sense to build a distribution process and calculate statistics the same way. It will not help in identification at all. And this is where we need to turn to the so-called "fine structures" of which Schnoll speaks.

So there you go.

 
Mathemat:

Seryoga, well, it's not hard at all. But keep in mind that each line of the Excel file will have a different number of columns (the minutes in a day are different). Accordingly, the numbers in a given column will not necessarily correspond to the same time of day.

I'll try to sketch a script. You can choose the separator for the excel, I will put it into external variables. And the problem of quality history, without holes, is up to you.

I worked with MQL a long time ago. Then, Yuri promised to help as much as he can.
 
OK. If there's a problem, we'll sort it out. The main thing is to identify the last minute of the week. Or rather, not it, but the first minute of the next one.
 
Candid:

I haven't read Schnoll, but the judge ... :) some perceptions from a quick glance.

...

you misunderstood the word infiltrate. I didn't actually intend to pick up the falling banner from the hands of an academic.

Apparently he has reasons for the truly amazing persistence with which he continues to search for patterns after so many failures.

You're right about that.

Because he deliberately works with "untenable" histograms, he has to use specially devised methods of analysis. I have not dealt with them in detail, but my indirect impression is that they are, alas, already unquoted.

The question itself is a moot point and believe me, I don't want to get into quality control at all. Many stats have validity from 10 measurements onwards. Schnoll worked from 100-300 measurements. You can get a good estimate on 25 measurements and meet all the criteria. But from a practical point of view it would be bullshit. That's not the point. I've only just written what the subtlety is.

 
Farnsworth:

The market model

The more complex the system, the harder it is to criticise it... Some healthy criticism:

Why does it talk about the market model, but apply it not to the market, but to its individual parts - financial instruments? Why does it not consider the totality of financial instruments?

1...8910111213141516171819202122...28
New comment