Non-fitting system - main features - page 5

 
Sorento писал(а) >>

Can you define dissimilarity?

Similarity or dissimilarity is the same thing in terms of value to TC. The Hurst coefficient given is a fairly recognised formal sign, but I don't think it is constructive.

We are not the first to encounter unsteady systems and it is not a matter of definition. These definitions have been given in the background, but the question has always been how to turn this non-stationarity into stationarity, without ever asking "how much trust can be placed in this stationary process, which we have cleverly derived from the stationary one".

 
faa1947 >> :

Similarity or dissimilarity is the same thing in terms of value to the TC. The Hurst coefficient given is a fairly recognised formal sign, but I don't think it is constructive.

We are not the first to encounter unsteady systems and it is not a matter of definition. These definitions have been given in the background, but the question has always been how to turn this non-stationarity into stationarity, without ever asking the question "how much trust can be placed in this stationary process we have cleverly derived from the stationary one".

Unless we consider a perfect sphere horse in a vacuum, everything we deal with in life is local in time. Including stationarity. How about starting from that?

I.e. consider not from within the process, and from outside - the criteria of loc. stationarity plots.

 

The systems under discussion should have one, very simple attribute - a complete lack of optimisable parameters. In my system it does, all parameters are taken from the price itself, you just have to ask it, and it knows a lot about itself. If the system has input optimizable parameters, and especially if it is based on any methods and combinations of TA, then the system will definitely start to fail. Absolutely, because TA (and waves in particular, etc.) has nothing to do with the market, and there is no basis for holding these optimal parameters after optimization. This is the deepest self-deception and a waste of your own, invaluable time (remember, time is a non-renewable resource).


Let me explain, a sign of a stable, confident business (if you treat the topic as a business) is its "duplicability". TA is the greatest deception, as there is an abundance of parameters, prototypes and peculiarities of psychology in the sandbox lined up for the players, all of which together will not just "go out" and realise the nonsense of these approaches. There will always be someone fickle, but always successful, say, building levels or arcs or other bullshit better than others. But of course this is mine, imho, came to this about three years ago, as I've written about here sometimes. :о)

OY!!!! What am I???? Everything works, of course it works ... in combination with experience and intuition. They should be fed into the optimizer's input :o)))

 
Mathemat писал(а) >>

Do you have any real ideas on how to account for non-stationarity in the tester?

Your complaints on vile supporters of stationarity, and also your crown phrases - "dynamical system", "chaos theory", "BP is non-stationary", "sequence of digits of number Pi" I already learned by heart.

I don't know, I haven't thought about it. The tester gives inputs-outputs in position some final numbers and that's enough for me. But I'm against talking about developing methods of increasing confidence in test results. Actually there is an optimization theory and genetic algorithm is not the only one. I haven't seen any assessment of this genetic algorithm on the forum compared to others, what type of BP it is suitable for. And is it even necessary?

I have more important questions unresolved. We take SPM and see one resonant frequency, then it falls apart into several resonant frequencies and then it's already difficult to identify the resonant frequency. Everything happens gradually. I posted it once, I can post it again for you. What are some ways to catch this change in AFC and especially the phase?

 
grasn >> :

The systems under discussion should have one, very simple attribute - a complete lack of optimisable parameters. In my system it is, all parameters are taken from the price itself, you just have to ask it, and it knows a lot about itself. If the system has input optimizable parameters, and especially if it is based on any methods and combinations of TA, then this system will definitely start to fail. Absolutely, because TA (and waves in particular, etc.) has nothing to do with the market, and there is no basis for holding these optimal parameters after optimization. This is the deepest self-deception and a waste of your own, invaluable time (remember, time is a non-renewable resource).

Let me explain, a sign of a stable, confident business (if you treat the topic as a business) is its "duplicability". TA is the greatest deception, as there is an abundance of parameters, prototypes and peculiarities of psychology in the sandbox lined up for the players, all of which together will not just "go out" and realise the nonsense of these approaches. There will always be someone fickle, but always successful, say, building levels or arcs or other bullshit better than others. But of course this is mine, imho, came to this about three years ago, as I've written about here sometimes. :о)

OY!!!! What am I???? Everything works, of course it works ... in combination with experience and intuition. You have to feed them into the optimizer's input :o)))

))) What you call "asking the price" is the TA. At least shoot yourself! (God forbid - live 200 years.)

Anything that analyses the content of the quote stream is TA.

 
Sorento писал(а) >>

Dear!

Give a definition. however crude.

Just above gave about Hearst, is there anything you're not happy with specifically?

 
Svinozavr >> :

))) What you call "asking the price" is TA. Shoot yourselves in the face! (God forbid - live 200 years.)

Anything that analyses the content of the quote stream is TA.

Then what do you call "fractal analysis" (I don't mean that crap from one .... figure), what do you call "self-organising stochastic systems", what do you call .... just mathematics and geometry (predominantly it) after all? I take it these are some kind of TA fields?

 

Maybe it should be taken into account what is used by the vehicle, i.e. let's assume the script searched for something on the history and found that from such and such a date of such and such a month of this year

it found a stable dependence up to now, we can only assume that it has something to do with a large batch of published economic data

on this basis, TS is built and optimization will give something, but the discovered property suddenly appeared and suddenly may disappear without noticing it or without knowing it

You may continue optimizing till you are blue in the face and it is possible the cp will not fail.

Maybe one should keep an eye on the basis, if it can be monitored

 
grasn >> :

Then what do you call "fractal analysis" (not that crap from one .... figure), what do you call "self-organising stochastic systems", what do you call .... just mathematics and geometry after all? I take it these are some fields of TA?

Can't add anything to that.

Anything that analyses the content of the quote stream is TA.

 
Svinozavr >> :

There is nothing I can add.

Anything that analyses the content of the quote stream is TA.

eh.... I wish you would have stopped at the thesis: "Can't add anything"... but you didn't - you added, ....