Machine learning in trading: theory, models, practice and algo-trading - page 1347

 
Alexander_K:

:))) Alexius! You, I see - you know. However, I hasten to report - the market cannot be taken by theorists alone. For one reason - it does not take into account such a concept as "time". Yes, yes, ordinary time - seconds, hours, ... Without understanding the role of time and its consideration, all the power of TViMS is also powerless.

All of us are looking and looking and still not finding it.)

Taking time into account is the introduction of non-stationarity into the price model.

 
Aleksey Nikolayev:

All of us have been looking and looking, but we still haven't found it.)

Time is the introduction of non-stationarity into the price model.

Don't worry, Alexei. If you pay due attention to the behavior of the price wave pack(probability density of tick increments, its nonparametric kurtosis, asymmetry) within a certain time sliding window (and not just some sample volume), then you will reach the results that I have (+30-40% per month). But, I want more. So, when you've gone through all of this, write an article - come back to the branch of the TA. Together we'll continue the way to the coveted Grail.

 
Aleksey Nikolayev:

All of us have been looking and looking, but we haven't found it yet.)

Time is the introduction of non-stationarity into the price model.

... And, as a consequence, loss of foundation for TV application ? And, including spectral analysis is inapplicable, because, as you argue, in this case there is not even the concept of spectrum. Am I expressing your position correctly?

 
Oleg avtomat:


Alexis reminds me of me 1.5 years ago, when I thought that one theorist could crush the market to dust. I have tears in my eyes...

 
Alexander_K:

Alexis reminds me of me 1.5 years ago, when I thought that one theorist could crush the market to dust. I have tears in my eyes...

.

https://www.mql5.com/ru/forum/70676

 
Oleg avtomat:

... and, as a consequence, the loss of the foundation for the application of TV ? And, among other things, spectral analysis is not applicable, because, as you claim, in this case there is not even the concept of spectrum. Am I stating your position correctly?

If TV only considered stationary processes in the broad sense (and close, reducible to them), that would be correct. Read at least the table of contents of Korolyuk's handbook - there are other classes of random processes.

 
Alexander_K:

Alexis reminds me of me 1.5 years ago, when I thought that one theorist could crush the market to dust. I have tears in my eyes...

And you look like a failing student who failed to calculate the ACF SB.

 

No, well, I use TViMS, too. But at least now I understand that you can't do without it. Time, you can't do without it. Time cycles, some subtle temporal structure that I don't fully understand. And inside of time periods - yes, I use a theorizer to change the market.

Without taking into account the time theorizer does not work.

 
Aleksey Nikolayev:

If TV considered only stationary processes in the broad sense (and close, reducible to them), it would be correct. At least read the table of contents of Korolyuk's handbook -- there are other classes of random processes.

"at least the table of contents" -- good ;)))

But have you forgotten (?) this :

Forum on trading, automated trading systems and testing trading strategies

Machine Learning in Trading: Theory and Practice (trading and not only)

Aleksey Nikolayev, 2019.02.18 20:44

The notion of a spectrum is defined only for a stationary process. Price is not, if only because of the growth of dispersion over time.


Do you continue to insist that there is no concept of spectrum for a non-stationary process?

 
Aleksey Nikolayev:

And you look like a student who failed to calculate the ACF SB.

Alexei, dear... It's a Heaviside function, why count it? Or do you also think you're the smartest one here? Show me your diploma, then. Brag.