What makes an unsteady graph unsteady or why oil is oil? - page 7

 
Prival писал(а) >>

The main thing is to determine the signal-to-noise ratio according to theory.



Followed your posts and glad to be back on the forum.
A large number of matte methods imply a signal in BP. And what is the signal in financial BP, and in the marketplace? It seems to me that we need to define it. Anyway the concept of signal in the market does not coincide with the concept of signal in the radio. For a trader it is a signal of a position, i.e. a signal produced by TS. More generally, it is a reversal of BP or more precisely, the probability of reversal. The change of the candle colour is a reversal, but the probability is zero; the marked trend is not zero, but we can fall under the reversal.

 
NTH писал(а) >>


I.e. we can say that: a non-stationary process is a process in which the result of a particular case is unknown. ?


Unsteady means no MO and no dispersion.
To the question of the topic. A common financial time series model is trend + wave and varying periodicity + noise. The wavelet decomposition of BP is curious. We get a matrix, 1st column is trend. The 2nd is the trend in the difference between trend and BP, i.e. noise, etc. It is argued that on no more than the 6th column of the matrix we have a stationary series. I was interested in another issue in this whole story. We have a trend that will end sooner or later. Somewhere in the next columns of the matrix a new trend is born. The emergence of this new trend is the beginning of the end of the existing trend. Could it be the emergence of stationarity earlier than column 6? Or something else?
 

From the point of view of practice. If I open a position a) "on the fly", b) by some super-duper analysis, c) by voodoo magic), the point remains - the result of this particular trade is unknown to me. As well as the next price move, bar colour, etc. is unknown to me.

 
NTH писал(а) >>

From the point of view of practice. If I open a position a) "on the fly", b) by some super-duper analysis, c) by voodoo magic), the point remains - the result of this particular trade is unknown to me. Just like the next price step, bar colour etc. is unknown to me.

You have to predict the price reversal. The whole TA is based on this.
 
thought.
any fitting is a search for patterns.
there are no other methods to find patterns.

In the optimization period - profit. it means that the patterns have been found.
Other periods - no profit. It means that other regularities are present.

It means that these time periods are somehow different.
Some other main indicators characterise them.

The system that has yielded a profit on the optimization period should be used on the same similar time periods.

The only question is how to classify the market and find the main indicators characterizing them.
 

Well, all TA is certainly not built on that. Predictions are probability, and with some logic it can be ruled out and a pattern can be detected.

 
Stanley1888 писал(а) >>
thought.
any fitting is a search for patterns.
there are no other methods to find patterns.

on the optimization period - profit. it means that the patterns have been found.
other periods - no profit. it means that other regularities are present.

It means that these time periods are somehow different.
Some other main indicators characterise them.

The system that has yielded a profit on the optimization period should be used on the same similar time periods.

The only question is how to classify the market and find the main indicators characterizing them.

Fitting is the deduction and transformation of the state of dynamically changing elements of a non-stationary process. I.e. - fantasy! Fitting neglects reality (static elements of the process). Optimisation says: "you will lose in the end! )

 
faa1947 >>:

Нестационарный - это значит, что нет МО и дисперсии.

God be with you, where did they go? In non-stationarity, simply put, at least one of the parameters is not a constant. Scientifically speaking, it's https ://ru.wikipedia.org/wiki/Стационарность.

 

What makes us think that financial markets are non-stationary.

Maybe probability patterns are constant over time and financial markets are stationary.

 
Stanley1888 писал(а) >>

financial markets are stationary.

An example, please.

faa1947 wrote >>


Been following your posts and glad to be back on the forum.
A lot of maths implies a signal in BP. And what is a signal in financial BP, and in the marketplace? It seems to me that we need to define it. Anyway the concept of signal in the market does not coincide with the concept of signal in the radio. For a trader it is a signal of a position, i.e. a signal produced by TS. More generally, it is a reversal of BP or more precisely, the probability of reversal. The change of the candle colour is a reversal but the probability is zero, the marked trend is not zero but we can fall under the reversal.

Signal ~ fair value?