ERUUSD spectra - is this proof of non-stationarity? - page 9
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
For a start, stop confusing periodicity and cyclicity all the time.
The cyclicality is inherent in all repeating processes.
Somewhere on this forum (I don't remember where I was reading diagonally) it was said correctly that a cycle is the market's memory window of an event.
The market remembers some event and there is a cycle, and it naturally fades,
and of course we should start with the event.
Finally, this is an interesting idea. But how do you find the cycle? Most likely, we should calculate SPM in the size of this "memory window of the market". If it is smaller or larger, you will get a wrong spectrum. SPM calculated within the memory window will be cyclic and change gradually, and not as it is in my stats.
Finally, an interesting thought. How do you find the cycle? Most likely, the SPM should be calculated in the size of this "market memory window". If it is smaller or larger, you will get a wrong spectrum. SPM calculated within the memory window will be cyclic and change gradually, and not as it is in my stats.
If you don't have any other ideas about event detection, you can go by zigzag for example.
The spectrum should be searched until a new extremum of the zigzag is reached, parameters can be selected using the tester.
Once a new extremum has been set it means a new window and a new search. After all, the spectrum is evolving so what is the point of referring to the one that has already been cancelled?
I recommend to deduct the linear regression from the quotes with the same window from extremum to zero before looking for the spectrum.
Then you will bypass the Kotelnikov-Nyquist theorem.
Before searching for a spectrum I recommend to take a linear regression from the quotes with the same window from extremum to zero.
then you will bypass the Kotelnikov-Nyquist theorem(thanks to Prival.)
If there are no other ideas for detecting an event (that is the starting point of the event), you can use a zigzag, for example.
The spectrum will be searched until a new extremum of the zigzag is set, parameters can be selected by the tester.
Once a new extremum has been set, this will open a new window and initiate a new search. The spectrum will, after all, float so why cling to one that has already been cancelled.
I recommend that you subtract the linear regression from the quotes with the same window from the extremum to zero before looking for the spectrum.
Then you'll get around the Kotelnikov-Nyquist theorem.
There is something to the idea. But, ZZ is redrawn, and once drawn, you don't need it anymore. Besides, ZZ has such a parameter as period. We can introduce a restriction: 1. if the new reverse is at least x-pips from the previous one and 2. if it is at most y-pips from the previous one and 3. if the number of reversals on the z-bar length is approximately the same. Will this advance us?
Just for the record.
There seem to be two main paradigms for dealing with the price chart to date:
- adjusting to the existing order, in the hope that it will hold for a while longer (for which, of course, spectral analysis is very good)
- bringing the chart to the stationary form using some methods and working with the stationary function
I think the second method is more confident. In terms of future calculations. Although, it is more complex in terms of developing a profitable strategy.
Just for the record.
There seem to be two main paradigms for dealing with the price chart to date:
- adjusting to the existing order, in the hope that it will hold for a while longer (for which, of course, spectral analysis is very good)
- bringing the chart to the stationary form using some methods and working with the stationary function
I think the second method is more confident. In terms of future calculations. Although it is more complicated in terms of developing a profitable strategy.
There are a lot of discussions on this forum. I am not aware of the results. If we are talking about the stats from which the thread was started, in fact the SPM is constructed not by a chart but its derivative - an autoregressive function with a moving average with the maximum entropy compared to the noise.
How do you do that? And for what purpose?
And so that the period is not longer than the window.
There is something to the idea. But, ZZ is redrawn, and once drawn, according to you it is no longer needed. Besides, ZZ has such a parameter as period. We can introduce a restriction: 1. if the new reverse is at least x-pips from the previous one and 2. if it is at most y-pips from the previous one and 3. if the number of reversals on the z-bar length is approximately the same. Will this advance us?
So take it from the 2nd extremum to 0. (The second one will not be redrawn for sure)
Only the ZZ parameters should be adjusted, so that windows would not change every couple of bars, at least some statistics would be present.