From theory to practice - page 70
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Let's imagine a black box (inside the traders and the marketplace, the exchange) with no external influences, and living a life of its own - the output of the box: a flow of quotes. Even without external influences it will change somehow.
Now let this BS receive random (for the observer) delta-functions of different sign and intensity (news, for example). The NM starts to react somehow and we observe not the effects themselves, but the response of the NM to them + independent life of the NM itself.
The memory is there, but the output is a superposition of response to many events and the self life of the NM. Even in case of a simple control system (ACS) the problem of division of what is not very solvable.
Well in principle I have long ago constructed such model, the question is how to display projection of levels in increments?
Well in principle I have long ago built such a model, the question is how to display the projection of levels in the gradients?
No, you can't.
The ground itself, structurally, is nothing more than an integrator. By differentiating we obtain only the reaction of sets of traders to the impact. In terms of filtering, differentiation is lifting the HF component and suppressing the LF one, i.e. after differentiation we only get noise around zero.
SZZ we have to integrate again in order to isolate something).
Let's imagine a black box (inside the traders and the marketplace, the exchange) with no external influences, and living a life of its own - the output of the box: a flow of quotes. Even without external influences it will change somehow.
Now let this BS receive random (for the observer) delta-functions of different sign and intensity (news, for example). The NM starts to react somehow and we observe not the effects themselves, but the response of the NM to them + independent life of the NM itself.
The memory is there, but the output is a superposition of response to many events and the self life of the NM. Even in the case of a simple control system (ACS) the problem of division of what is not very solvable.
What about my 36 currency pairs? What kind of supercomputer would I need to trade on such streams at the same time? I'm disappointed...
Pity... It's a pity we never got to hear from the head of the transport department. (с)
Who deletes other people's posts and why? Alright, I don't mind doing it again:
Alexander_K2, what about the distribution of the modified row? It's a minute thing to calculate.
Have I calculated the RMS correctly?
Here's the formula:
Here is the result in Excel:
You have calculated the RMS correctly. However, see how much it would be if n=1. You'll wonder what kind of nonsense this is. The name "n - volume of statistical population" is very vague, usually they write that n is the number of elements in the sample. Then the RMS according to this formula can't be calculated if there is only one element. That's why the square of the RMS is called a "biased" variance estimate. There is also an unbiased one, where n is n1-1 instead of n in the denominator. The square root of the unbiased variance estimate is called the standard deviation.
The nature of this conflict is that one item has one degree of freedom. If many-many features are defined from a small number of data, they become dependent on each other. In this case the arithmetic mean is included in the RMS calculation. So to speak, one degree of freedom has already been used. The "strange" behaviour of the denominator of the standard deviation is just to say that both the mean and the spread cannot be determined from a single element. It can be seen that the standard deviation is always greater than the standard deviation by a factor of [n/(n-1)]^0.5. However, if the number of elements in the sample is large, you can forget about it, because it is not much. When n=100, it is (100/99)^0.5=1.005, which is half a percent. Moreover, if we know for sure that the RMS tends steadily to some value.
This is where the tricky part comes in. "RMS tends to", i.e. the laws of large numbers work. If the real phenomenon being measured actually has this stability. In other words, the basic hypothesis of probability theory is fulfilled - the relative frequency of an event tends to some value as the number of events increases. This is also called "statistical stability". If it does not exist, all classical probability theory is inapplicable to the phenomenon. This difference is discussed in the huge quotes from Oleg avtomat, which start fromhttps://www.mql5.com/ru/forum/221552/page58#comment_6191471 and onwards. They are hard to read. In my opinion, it is much more fun to view the presentation of Gorban's report with pictures and graphs. It will create a more optimistic and constructive mood, such as this phrase:
"It has been shown that ocean swell, traditionally regarded as a pronounced destabilizing factor, can improve the performance of hydroacoustic stations."
Even for exchange rates, the author has walked around looking for the phrase "Averaged over 16 decades, the statistical instability parameter (continuous curve) and the range of variation of this averaged parameter, defined by RMS, (dashed curves) for the Australian dollar (AUD) quote against the US dollar (USD) for 2001 (a) and 2002. (б)".
I attach the presentation, and for those who want more sources, here's a list of presentations, sometimes with file addresses, from the list "Archive of past "Image Computer" seminars http://irtc.org.ua/image/seminars/archive from 2002-2017. Gorban has up to a dozen monographs on developments in "hyperrandom" phenomena:
I.I. Gorban Theories of Hyperrandom Phenomena. Theory and Practice. Section 7. System Analysis.
I.I. HURBAN I HYPERRANDOMNESS KIEV NAUKOV DUMKA 2016. - 288 p. ISBN 978-966-00-1561-6