From theory to practice - page 455

 
Evgeniy Chumakov:
Welcome back!

Hello, Eugene!

When will you open a signal, even on demo?

All the same, we have to admit - only a trading signal allows to judge whether the trader is moving in the right direction or not.

My profit =0% and I understand that it is impossible to do without unambiguous accounting of ACF and/or entropy (non-entropy) in the algorithm. It makes you work and move forward, which I wish you to do as well.

 

I've come across a situation where I need to dynamically change the sample size by some criterion, but I don't understand how to implement it.

For example, my sample size was 200 and I was just a little short of the signal .... But when the sample size was 300 there was a signal.

 
Evgeniy Chumakov:

I've come across a situation where I need to dynamically change the sample size by some criterion, but I don't understand how to implement it.

For example, my sample size was 200 and I was just a little short of the signal .... At the same time I had a signal with the sample size of 300.

IMHO - sample size should be = const.

"Floating" is the quantile of the confidence level - you and I have already discussed this.

The normal distribution for the sum of the increments in the sliding window is obtained in the measurement limit. And at this current moment the distribution sort of changes, but not much - from a gamma distribution of a certain order to a normal distribution.

Something like this:

https://ru.wikipedia.org/wiki/Уравнение_Фоккера_-_Планка, only slightly more complicated - forming skewed distributions.

I was given a task by Sorcerer to make the same cartoon, and I haven't made it yet... He apparently got angry and left the forum... That's a shame.

So - it's important to see the current probability distribution and dynamically change the confidence level quantile.

Уравнение Фоккера — Планка — Википедия
Уравнение Фоккера — Планка — Википедия
  • ru.wikipedia.org
Впервые уравнение было использовано для статистического описания броуновского движения частиц в воде. Хотя броуновское движение описывается уравнениями Ланжевена, которые могут быть решены численно методом Монте-Карло или методами молекулярной динамики, задачу в такой постановке часто трудно решить аналитически. И, вместо сложных численных...
 

Question for respected radiophysicists:

Is the ACF for a discrete signal calculated at signal time shift deltaT = const or can deltaT still be a variable?

 

Question for traders with more experience in data collection and processing:

If we take a sliding window = 24 hours and collect tick volumes = the sum of all ticks in that time window - would that sliding sum represent a Poisson distribution?

 

I really need answers to these questions, my friends, in order to save time for research.

Please help with the answers - the Grail is closer than ever and, with impatience and excitement, my hands and feet are shaking - I can't work...

 
Alexander_K2: The Grail is closer than ever and, with impatience and excitement, my hands and feet are shaking - I can't work...


Imho. The grail is when the price unfolds at values Density = 0, distribution function = 0/1 . Then only space is higher.

 
Evgeniy Chumakov:


Imho. The grail is when price reverses at Density = 0, distribution function = 0/1 . Then only cosmos is higher.


Here's an example. Like a bang on the wall, the price stopped and went back. (Although I put the profit out of the light, rather there closing ~ 30 pips lower from the opening price).


 
Evgeniy Chumakov:


(True, I put the profit from the background, rather there the close is ~ 30 pips lower from the opening price)



TP/SL ratio = 3 to 1.


 
By the way Alexander! Maybe the reason why your profits are circling zero is because you don't use MM. Losses eat up profits and it's good to be at zero.