Discussion of article "Grokking market "memory" through differentiation and entropy analysis" - page 5
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
Plus for the article, interesting. But false axioms cut the eye, for example:
исследователям необходимо работать с приращениями цен... Эти преобразования делают временной ряд стационарным, удаляя всю память из ценовых последовательностей.
The problem is that price increments are stationary but contain no memory of the past, whereas the price series contains the entire amount of available memory but is non-stationary.
Switching to increments not only does not remove the memory, but even manifests it better, since the interfering strong autocorrelation of values inherent in the integrated series is eliminated.
Example of a stationary series with memory: y[i] = -y[i-1], realisation: 1, -1, 1, -1, 1, -1, 1, -1 ...
The autocorrelation of the values of an integrated series is apparent memory; it carries no information, and appears only because each successive value is plotted relative to the previous one.
Real memory is memory in increments. Examples:
"if it has been growing, it will continue to grow" - is trending (persistence).
"if it was growing, it will continue to fall" is reversibility (antipersistence).
For the same reasons, calculating persistence for a differentiated series is probably meaningless.
In short, mathematical science has already studied everything here and laid it out long ago.
Plus for the article, interesting. But false axioms cut the eye, for example:
The transition to increments not only does not remove memory, but even manifests it better, since the interfering strong autocorrelation of values inherent in the integrated series is eliminated.
Example of a stationary series with memory: y[i] = -y[i-1], realisation: 1, -1, 1, -1, 1, -1, 1, -1 ...
The autocorrelation of the values of an integrated series is apparent memory, it carries no information, and appears only because each successive value is plotted relative to the previous one.
Real memory is memory in increments. Examples:
"if it has been growing, it will continue to grow" - is trendiness (persistence).
"if it was growing, it will continue to fall" is reversibility (antipersistence).
For the same reasons, calculating persistence for a differentiated series is probably meaningless.
In short, mathematical science has already studied everything here and laid it out long ago.
Well, it describes two kinds of "memory" - one is related to non-stationarity and mean shift, the other is related to regularity of increments. This is a free interpretation of "memory" by the author of the modification of fractional diff., so I didn't change it.
e.g. trends, seasonal trends, cycles... only then increments as remnants of former glory, the last stage of differentiation.Thank you to the author for the excellent work!
This is an example of a serious scientific approach to the study of such a complex, non-stationary process as the movement of market prices.
Maxim, you first wrote about the contribution of players to the overall movement.
Could you please tell me which players' actions you consider as a signal and which as noise?
With longer horizons, the largest number of participants with the same views, the "memory" of whose actions?
Interesting your opinion, thanks.
Interesting article!
Maxim, at the beginning you wrote about the players' contribution to the general movement.
Please tell me which players' actions you consider as signal and which as noise.
With longer horizons, the largest number of participants with the same views, the "memory" of whose actions?
Interested in your opinion, thank you.
Thank you. In the context of random processes if we consider, "memory" is also a random phenomenon. I.e. cumulative reaction of participants to some events, which is expressed in the repetition of their actions, i.e. cyclically. For example, the reaction to economic news is usually expressed by bursts of trading activity, and the news comes out with a certain periodicity, it is also a kind of memory. That is, it is impossible to say which participants' actions will have a consequence, rather it is a self-organising process under the influence of external information.
Or a simple example from nature: in spring or summer everything starts to bloom not because plants have agreed among themselves, but because natural conditions or cycles favour it. But these cycles cannot be called a global regularity, rather, it is also an accident that entered the attractor for some time (until a meteorite falls on the planet, for example, or the Earth collides with the celestial axis).
Well, in the context of such definition, there is a possibility to estimate these cycles somehow, for example, with the help of entropy of the process and to make a forecast, if these cycles really exist at the moment. That is, it is not a "memory" of someone's specific actions (answering the question), but rather a memory of the system as a whole about external influencing conditions that lead to spontaneous self-organisation.
не потому что растения договорились между собой
Yes, I know what you are talking about, I am close to such thoughts.
The huge masses of the universe, the ones we can see, are almost all in rotational motion. And a photon, which is like a blob in a confined space i.e. a particle, manages to fly through two holes at the same time and shows us that it is a wave. And we too, when competing for resources, make multidirectional (chaotic) movements and create a wave. Trying to be aware of all this, of course, is how generalisations are pushed. Although generalisation is an insidious tool:)
As I understood, in our case memory is the inertia of a significant movement in our sensations. I.e. the process that started the movement has already ended, but the price is still flying?
As I understand it, in our case memory is the inertia of a significant movement in our sensations. I.e. the process that started the movement has already ended, but the price is still flying?
Yes, as a variant of interpretation. External information enters the market, there is a point of bifurcation and instability, then this information settles down due to self-organisation of participants. Within this process, recognisable patterns of behaviour, cycles, are formed. It is like throwing a piece of meat to the dogs, they do not eat it immediately, although the hypothesis of efficient markets says that immediately, while it is not even meat but a running hare (which sounds a bit absurd, but there is such a hypothesis).
They can overlap or repeat after a certain time, if there are periodicities of information arrival. If the market stops receiving organising information, at such moments it looks more like a SB, apparently, as there is no structuring component whatsoever
In game theory, eating a thrown piece of meat is called an auction, the noise arises because of the large number of participants, as everyone eats it differently and with different appetite. Someone cures it and then sells it, and someone pecks it in pieces or takes it from others, someone spits it out.
Yes, as a variant of interpretation. External information enters the market, a point of bifurcation and instability arises, then this information settles down due to self-organisation of participants. Within this process, recognisable patterns of behaviour, cycles, are formed. It is like throwing a piece of meat to dogs, they do not eat it immediately, although the hypothesis of efficient markets says that they eat it immediately, while it is not even meat but a running hare (which sounds a bit absurd, but there is such a hypothesis).
They can overlap or repeat after a certain time, if there are periodicities of information arrival. If the market stops receiving organising information, then at such moments it looks more like an SB, apparently, as there is no structuring component at all
About cycles...
Intraday they are clearly visible. See the GBPUSD charts for this month:
On the bottom indicator of the Sorcerer (currently warming up, probably on the edge of the forest) this periodicity is present in the process variance (red and blue lines) - practically a sine wave.
It is important to be able to predict the behaviour of dispersion (volatility) and make deals only when the price (or the sum of increments) goes beyond the dispersion, when the inflection point is passed (marked with green circles on the charts).
The waves go round and round ;) :)
Here is the information provoked outrage and the price went. Is it possible that when the "memory" is gradually stopped, the price stops at any arbitrary place? Or these places are still limited areas, which we consider as levels. When one person thinks, "I'll get to this point and take it." And the other person thinks: "I can't take it anymore, I'll take another 20 pips and close".
I.e. price increments are different at different parts of the price scale.