Whether there is a process whose analysis of one part does not allow predicting the next part. - page 11

New comment
 
faa1947:

I am suspicious of the statements of loners, I prefer a crowd, and a large one at that.

In honour of the holiday, give me the answers to your questions. I'd love to read, but it's a shame to look it up myself.

Strange, though... You're going through a lot of books, you're providing links, you're counting something, you're doing some tests and so on.... that's a lot of activity! But to think about it a bit, that's where the hitch comes in...
 
avtomat:
Strange, though... You go through a lot of books, you give references, you do some calculations, some tests and so on.... that's a lot of activity! But to think about it for a while, that's where the hitch comes in...

Same thing.

A meaningful question .... and that's it... More moralizing. Answer your own question, not mine, at least once.

 
faa1947:

Same thing.

A meaningful question .... and that's it... more moralizing. Answer your own question, not mine, at least once.

Here's a clear example ...

You're suspicious of the statements of singles, you prefer the crowd and the masses - it's your word. Ridiculous, really...

 
avtomat:

Here's a prime example ...

You're suspicious of the statements of loners, you prefer a crowd, and a large one at that -- those are your words. Ridiculous, really...


Happy Holidays!
 
joo:

Hi.

I propose the community to invent a process, that can't be predicted (so that you can't make money on this prediction). At the same time, the process should not have stationary stat-characteristics in time.

"Is there a process, the analysis of one part of which does not allow predicting the next part" --- the question is incorrectly posed. The answer (positive or negative) depends substantially on the forecasting algorithm used. For one and the same process different forecasting algorithms will give different results.

 
avtomat:

faa, here's a function for you:

what do you say about the differentiability and predictability of this function? And tell me about its determinability!

Funny, judging by the logic of the indicators I'm working on now this function is at the same time:

  • A rigidly deterministic antipersistent one-lag memory row*;
  • A rigidly determined persistent near infinite memory.

I.e., if this indicator were a robot, it would define for itself two trading horizons at once, on one of which it would be an antitrend TS with a deal holding time of one lag and on the other one it would continuously hold a long position.

____________________

* Strictly speaking, the memory of such an antipersistent series can be larger than one lag and in the perspective also be infinite, but any other period larger than one will always be undefined due to the lack of data (simplicity of a function). For example, a function may draw a dash upwards based on the fact that two of the nearest three dashes were horizontal - but again, we will never know this due to lack of data. At the same time, one lag approximation is sufficient and indicators will select it as the nearest one, on the principle of working with what is, not what could be.

 
faa1947:

I prefer not to argue with you anymore, your delusions are too strong and cannot be overcome by reasonable arguments. I'd rather redirect my energy to something less useless...
 
C-4:

Funny, judging by the logic of the indicators

and the function --- it's the easiest thing in the world... ;)
 
joo:

Tried. Used a grid with one hidden layer. Was able to achieve a positive MO - predicted the direction of the next increment. So your version is no good.

->

What do you think, is it the merit of the grid or the learning algorithm?
 
joo:

Hello.

I propose the community to invent a process that cannot be predicted (so that on this prediction can not make money). At the same time, the process should not have stationary stat-characteristics in time.


Hello!

Life cannot be predicted in such a way as to make money on that prediction.

But: "Inspiration cannot be sold, but the manuscript can be sold" :)

If you are firmly planted on nets - look for a fast algorithm, which precedes a slow one. Imho, of course :)

1...4567891011121314151617
New comment