From theory to practice - page 1476

 
khorosh:

Probably like this: HU or Z

)))

Either way the formula is ))))

Renat Akhtyamov:

ok.

Rotate the map by 90 degrees since the beginning is clear

turn off time ( X coordinate) and leave only the market mechanism

but the only unanswered question is what is around the edges of the price(?) the only thing we need to know is the Y-coordinate

 
Renat Akhtyamov:

it

as they say.

I don't care if you trade or not, you'll still get...

;)

That's it?!

44 years you've been deriving this formula, Rena? And you're ready to publish the preprint? I don't get the joke this morning...

 

My point is this.

I am very sorry that the topic has become a dumping ground. Comparing the posts posted here with my private messages from people really suffering, fighting like lions against the market, I come to the conclusion that we should either start a new thread, or go to the Smart Lab.

I'll think about it.

 

So that you can understand the level of people who do not want to participate in this circus, but genuinely want to beat the market, I will reply to one of them here. I think he will understand that the message is addressed to him.

So:

"What you are working on is difficult to formalise mathematically. However, if you imagine your fractal images as probability distributions(build histograms), you will obviously see stable unimodal distributions. By calculating the sample size on which they appear, you are likely to come to an understanding of periodic market structure as I did. Working within these limits, you can build a TS.

Describing your patterns with linear equations/functions, on the other hand, is not possible. But, I am still reading and thinking...".

Do you feel the difference in level?

 
Alexander_K:

So that you can understand the level of people who do not want to participate in this circus, but genuinely want to beat the market, I will reply to one of them here. I think he will understand that the message is addressed to him.

So:

"What you are working on is difficult to formalise mathematically. However, if you imagine your fractal images as probability distributions (build histograms), you will obviously see stable unimodal distributions. By calculating the sample size on which they appear, you are likely to come to an understanding of periodic market structure as I did. Working within these limits, you can build a TS.

Describing your patterns with linear equations/functions, on the other hand, is not possible. But, I am still reading and thinking...".

Do you grasp the level difference?

Do you ever grasp the difference between a histogram and a probability distribution? It's not even about terminology (the histogram approximates the density of the probability distribution). The point is that for any sample (for any set of numbers) a histogram can be drawn, but for it to approximate any distribution this sample must satisfy rather strict conditions (be a sample ofa sequence of i.i.d. values). Price increments are not - especially in a flat environment (which is what your system is mostly working with).

Independent and identically distributed random variables - Wikipedia
Independent and identically distributed random variables - Wikipedia
  • en.wikipedia.org
The i.i.d. assumption is important in the classical form of the central limit theorem, which states that the probability distribution of the sum (or average) of i.i.d. variables with finite variance approaches a normal distribution. Often the i.i.d. assumption arises in the context of sequences of random variables. Then "independent and...
 
Alexander_K:


You feel the difference in level?

Yeah, we've had tougher dudes than that.

What's the point?

Here just knowledge is not needed, you only need your own brain to figure it out without tips, because they are nowhere to be found.

But if you don't have the brains, then "ouch".

;)

 
Alexander_K:

That's it?!

44 years you've been deriving this formula, Rena? And you're ready to publish the preprint? I don't get the joke this morning...

well first of all understand this, i.e. it's impossible to make money because there's a grail on the other side that will move the kotir as it pleases.

 
Alexander_K:

So that you can understand the level of people who do not want to participate in this circus, but genuinely want to beat the market, I will reply to one of them here. I think he will understand that the message is addressed to him.

So:

"What you are working on is difficult to formalise mathematically. However, if you imagine your fractal images as probability distributions (build histograms), you will obviously see stable unimodal distributions. By calculating the sample size on which they appear, you are likely to come to an understanding of periodic market structure as I did. Working within these limits, you can build a TS.

Describing your patterns with linear equations/functions, on the other hand, is not possible. But, I am still reading and thinking...".

Do you feel the difference in level?

You can see the periodicity of the market by eye and without a level of education)

The problem is that once the working structure is defined it is more likely to collapse.

Even if you set a neural network on each working sample and assemble a committee of networks into one, it won't do much.

Start a new branch.

 
Alexander_K:

Thank you, Bass. Good to know.

Read the book and your 15 years of misery will end quickly.

I don't have any misery) and without any books by the way)

 
Aleksey Nikolayev:

For any sample (for any set of numbers) a histogram can be drawn, but for it to approximate any distribution this sample must satisfy rather strict conditions (be a sample of a sequence of i. i.d. values). Price gradients are not - especially in flat conditions (which your system mostly works with).

Do we need it (to approximate the distribution in strict accordance with theory)?

We are working with discrete data. And there is and will never be anything precise and constant in the market.

We have to somehow work with what we have.

For example, the histogram of increments looks more like Laplace than Gauss. That's why we take the least moduli method rather than the least squares method. And so on.