Fractal theory - page 3

New comment
 

Generally speaking, if you follow this principle, it is possible to create some kind of pattern. I once built binary patterns by separating parts from each other with a zigzag.

You can make such a slicing on each timeframe, and then select the ones that repeat on all of them. Identify the probability of their formation.

And then we will dance from the current image in the market and compare it with possible patterns in the accumulated database.

Sharpening for each symbol pair.

 
Silent:

Advice can also lead astray.

A fractal is a multiple of some basis, one way or another.

Why H1 and M1, can you explain?

This is just as an example. M1 is the minimum available chart in mt4. You may also take a tick chart to search for the structure and trade on larger timeframes.
 
elugovoy:

Generally speaking, if you follow this principle, it is possible to create some kind of pattern. I once built binary patterns by separating parts from each other with a zigzag.

You can make such a slicing on each timeframe, and then select the ones that repeat on all of them. Identify the probability of their formation.

And then we will dance from the current image in the market and compare it with possible patterns in the accumulated database.

Sharpening for each symbol pair.

I don't understand the idea at all)), rephrase it please.

 

223231:

I do not understand the idea at all)), please rephrase it.

To extract "basic structures" of fractals (i.e. patterns) from the history, we take a zigzag indicator (you can also select the moving averages crossing and any combination of indicators). It divides the quotes history into these parts.

Then we define the characteristics of these parts. The methodology can be different here. It can be binary recodings of bars, indicators' relative indexes, in general, there is a field for imagination. Let us call such data quasi-cell.

We select and calculate quasi-cells for all (necessary) timeframes.

Then, we compare the resulting quasi-cells by timeframes, gathering statistics for each type of quasi-cell.

A deeper analysis can identify the transition of one type of quasi-cell to another (morphism, growth, development).

For analysis, it is easier to identify the probability of occurrence for each quasi-cell species.

This is the preparation and analysis. Then, when trading, a similar real-time analysis of the market is made and compared with the existing quasi-cells for possible formation.

During the appearance of new bars we can calculate what kind of a cell is forming at the moment (according to different timeframes), and with a reasonable probability we can say how the formation of this cell will be finished (according to the available types of cells).

Is that a little clearer?

 
elugovoy:

To identify the "basic structures" of fractals (i.e. patterns) from the history, we take a zigzag indicator (you can choose the moving averages crossover and any combination of indicators). It divides the quotes history into these parts.

Then we define the characteristics of these parts. The methodology can be different here. It can be binary recodings of bars, indicators' relative indexes, in general, there is a field for imagination. Let us call such data quasi-cell.

We select and calculate quasi-cells for all (necessary) timeframes.

Then, we compare the resulting quasi-cells by timeframes, gathering statistics for each type of quasi-cell.

A deeper analysis can identify the transition of one type of quasi-cell to another (morphism, growth, development).

For analysis, it is easier to identify the probability of occurrence for each quasi-cell species.

This is the preparation and analysis. Then, when trading, a similar real-time analysis of the market is made and compared with the existing quasi-cells for possible formation.

During the appearance of new bars we can calculate what kind of a cell is forming at the moment (according to different timeframes), and with a reasonable probability we can say how the formation of this cell will be finished (according to the available types of cells).

Is that a little clearer?

Oh, thanks, now I get the gist of it. It might come in handy a little later.
 
elugovoy:

To distinguish the "basic structures" of fractals (i.e. patterns) from the history, we take a zigzag indicator (you can choose the moving averages crossover and any combination of indicators). It divides the quotes history into these parts.

Then we define the characteristics of these parts. The methodology can be different here. It can be binary recodings of bars, indicators' relative indexes, in general, there is a field for imagination. Let us call such data quasi-cell.

We select and calculate quasi-cells for all (necessary) timeframes.

Then, we compare the resulting quasi-cells by timeframes, gathering statistics for each type of quasi-cell.

A deeper analysis can identify the transition of one type of quasi-cell to another (morphism, growth, development).

For analysis, it is easier to identify the probability of occurrence for each quasi-cell species.

This is the preparation and analysis. Then, when trading, a similar real-time analysis of the market is made and compared with the existing quasi-cells for possible formation.

During the appearance of new bars we can calculate what kind of a cell is forming at the moment (according to different timeframes), and with a reasonable probability we can say how the formation of this cell will be finished (according to the available types of cells).

Is that a little clearer?

Timeframes set the scaling coefficients, why on Earth should the market follow these very coefficients? Imho, it would be more correct to speak not about timeframes, but about time horizons, in general they can be arbitrary. Whether the market is divided into discrete horizons is a separate question. If so, their separation is also a separate question :).
At one time I looked closer to such a task and even made a special zigzag. But then it got stalled. And I put the zigzag on the market, for a great price :).

 
Candid:

Timeframes set the scaling coefficients, why should the market follow these coefficients? Imho, it is more correct to speak not about timeframes, but about time horizons, in the general case they can be arbitrary. Whether the market is divided into discrete horizons is a separate question. If so, their separation is also a separate question :).
At one time I looked closer to such a task and even made a special zigzag. But then it got stalled. And I put the zigzag on the market, for a great price :).

There is such a thing as normalization. If the patterns are normalized by the Maximum-Minimum value, the coefficients will be the same, with an acceptable error.
 
elugovoy:
There is such a thing as normalisation. If the patterns are normalised to a Maximum-Minimum value, the coefficients will be the same, with a margin of error.
Then the scaling will be set by the zigzag parameters. If you have in mind a continuous enumeration by the parameters of the zigzag (in addition to cycles in time), then the task turns out to be quite cumbersome to say the least. My idea was to try to build a hierarchy of specific horizons in a more economical way. That is, to catch the scaling factors which are objectively inherent to the market. If, of course, they are inherent in it :).
 
I realised what the problem is. The fractal structure has to be searched for on more than 1 pair. I am now developing a fractal model of 7 pairs. Since most of the money circulates in the main pairs, I will build a fractal linking all of them together, it should clearly show where funds are flowing. I will write when I have something useful. I think that in the end it will be possible to determine the current market situation with this structure.
 
then look for repeating fragments, with a given % deviation.
12345678910
New comment
Reason: