Yoghurt systems and canned systems or The relationship between trading tactics and the reliability of historical test results - page 14

New comment
 

It is strange how easy it is for you, although the Russian language is rich, but I think it is impossible to treat concepts so loosely.

A law is a verbal and/or mathematically formulated statement that describes the relationships between different scientific concepts, proposed as an explanation of the facts and recognised by the scientific community to be consistent with the data at this stage. An untested scientific statement is called a hypothesis.

Regularity is a necessary, essential, constantly repeated interrelation of phenomena of the real world, defining stages and forms of formation process, development of phenomena.

God be with him with scientific recognition. You plan to find the law (regularity) of the curve that you see on the screen and exploit this constantly repeating interrelation.

First define the object of study. In those 5 points the object is wrong, you are investigating the wrong thing, there is no analysis (search) for a pattern in the curve. And there are studies of the behaviour of the signal line, histogram, etc. And these are other objects with other characteristics and interrelations.

Let me try to explain by example.

  1. Verbal formulation. At the beginning of exchange trading sessions, the price (curve) moves in the opposite direction to the trend of the trading day of the exchange. The physical reasoning - some large market participants have some unknown information that allows them to draw such conclusions. Suppose the curve should go down, consequently, these players (those who know) will sell the currency to those who want to buy it, until the rate book allows them to do this and they do not change their minds. Let's check this statement visually. For this purpose let's use the indicator of trading sessions KimIV.

Here is the picture

Red circles indicate the beginning of trading sessions or when one is left alone in the market. Turquoise arrows is direction of movement of the curve in this time interval. The crimson arrows is the global direction of movement in the middle of the session. Picture is taken from 2.06.2008

It seems to be visually similar. And it does not contradict our statement. Let us go further.

  1. Let's put hypothesis H0 - that our statement is true (statistically significant), against hypothesis H1 that is not true (no statistical relationship).
  2. Let's select a currency pair. Prepare a sample for the statistical analysis (an array of ticks, the more the better). We introduce censoring, i.e. we remove from the sample Saturdays, Sundays, Christmas holidays, Independence Day, etc. (when the exchange is closed). You can introduce a more stringent censorship (at the discretion of the researcher) remove days when important news are released (those that change the market, let's say interest rate changes in a selected pair).
  3. Now how to determine the direction of short-term (initial) movement. From my point of view, the most reliable way is to use Spearman's coefficient of rank correlation. The description is available here 'Spearman's Rank Correlation Coefficient' (thanks as always to Rosh). But it's not the indicator, but the idea behind it (although you can adapt the indicator as well). The global direction of the straight line can be determined in different ways, options H or L, which is greater from the point of the beginning of the time section in question or the magnitude of the slope of the straight line plotted by the ANC.
  4. Then we use the developed and well known theory of statistical hypothesis testing and draw conclusions. If yes - the hypothesis H0 is accepted, there it is a log and we start to examine it further for drawing TS. I want you to note that if the hypothesis H0 is rejected, it does not automatically mean that the converse hypothesis is true (H0 - short-term movement coincides with a global movement). This hypothesis must also be tested and by the same scheme.

Again I want to remind you to look for patterns in the curve, not in indicators or Expert Advisors (looking for patterns in indicators is the same rake, but from the side view, optimizing on the history that does not give any confidence for tomorrow). This curve has no SL, no TP, no profit or loss and it does not care what indicators you attach to it.

P.S. KimIV sorry, I asked in your thread to make a bubble sorting procedure, just to test the hypothesis stated here (needed for Spearman), but I didn't exactly formulate the ToR, I needed a procedure for sorting a multidimensional array by a given column, in this case 2-dimensional. But no longer needed, as I try to do another, from my point of view more important, on this research as always not enough time.

If suddenly someone is interested, and do the study and get something (a negative result - is also a result), if you do not mind, share. At least by post one phrase "the idea works", if you do not want to make the result light.

I enclose the procedure for calculating the Spearman rank correlation coefficient on Matcad. It may come in handy for someone.

Files:
spirmen.rar  42 kb
 

Bravo, Sergei!

Thirteen pages of juggling with the word "regularity" and who can do what. Well, if they do not want to learn, at least do not call their fantasies a pattern. You'll make a fool of yourselves!

 
Prival писал (а) >>

It is strange how easy everything is with you, although the Russian language is rich, but I think one should not treat notions so freely.

In my opinion, everything is correct, and there is such a pattern, and it has been known, if my memory does not change, for more than one year. And it is possible and necessary to use it.

By the way, while I was there :)) the moments of session openings are set incorrectly in Kim's work. I did not get into the essence, but the rectangles are not correct.

 
Korey писал (а) >>

Who's here for this and that, who's here for that :-))

That was great!

Korey wrote (a) >>

===!!! MTS is forced to work where no legitimacy can be traced!

What's that? Is that a joke? A joke? How can you program something that doesn't have a pattern?

In and out on MathRand() ?

:0)

 
Prival писал (а) >>

If anyone is suddenly interested and does the research and gets something (a negative result is also a result), if you don't mind sharing. At least post one phrase "the idea works", if you don't want to make the result light.

First column sum of MACD histogram from signal ( Main[2]<Main[1] and Main[2]<=Main[3] ) count from 1 to the previous opposite signal.

The others are named. The distribution is done according to the maximum profit (by High) from the point where we entered to the opposite signal.

As a result, distribution of profit interval fraction depending on the sum of histograms, there are small positive res....

This is an old work, so I don't remember some things from it, but if you wish it may be restored...

By the way only Buy was considered...

Files:
gbpusd60_macd_buy_nicfa.zip  185 kb
 
Yurixx писал (а) >>

What's that? Is it a joke? A joke? How can you program something that has no regularities?

In and out on MathRand() ?

:0)

Below H4 there are NO patterns or not yet as long as the market crowd is working mostly manually (laugh or cry?)))
I.e. the reliability of TC prediction is sharply reduced when moving to H1 and tends to zeros on M1.
I checked this maxim myself)) when trading manually in the strategy tester.

 
Korey писал (а) >>

Below H4 there is NO pattern or not yet while the market crowd works mostly manually (laugh or cry?)))
I.e. validity of TC prediction is sharply reduced when moving to H1 and tends to zeros on M1.
I checked this sentence personally)) when trading manually in strategy tester.

NO! It's different. At H4 there are not so many bars yet to understand that there is no pattern there either:)

 
Korey писал (а) >>

Below H4 there are NO patterns or not yet as long as the market crowd is working mostly manually (laugh or cry?)))
I.e. the validity of TC prediction sharply decreases when going to H1 and tends to zeros at M1.
I checked this sentence personally)) when trading manually in strategy tester.

Wait a minute, what about Batter? https://www.mql5.com/ru/users/Better/ Or is it not a pattern you think?

 
Let's be clear: telling a thread about the patterns used in TC or coaching a network is not the same thing ! - who highlighted the rules from the nn of betters?
The championship regularities ended in february this year. since then that instance of Bettera EA has been leaking. i.e. the "regularities" only worked for 5 months
 
Korey писал (а) >> "patterns" only worked for 5 months.

But that means they are there after all. And you say there isn't. Or rather that they are minimal. And 5 months isn't bad at all.....

1...7891011121314151617181920
New comment