Market phenomena - page 37
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
as a memento
or better still, a screenshot, and print in triplicate...
I see different phenomena are considered by the letters of the Greek alphabet. Here are the phenomena for half the letters of the alphabet.
For the sentinel at 78,000 bars.
Same sentinel at the last 240 bar.
Same watchmaker, but 240 bar shifted by 1 bar.
Not very noticeable here, same picture, but shifted a little - one bar forward forecast is possible.
Same frequently 240 bar, but shifted by 10 bar or 11 in relation to the right-hand end.
A prediction 10 bars ahead is not possible.
It remains to assign letters of the Greek alphabet to the peaks. and there will be a phenomenon.
You will promote your and each other's brands on your own forum.
Dear Alex, don't jump to conclusions.
I see different phenomena are considered by the letters of the Greek alphabet. Here are the phenomena for half the letters of the alphabet.
For the sentinel at 78,000 bars.
Same sentinel at the last 240 bar.
Same watchmaker, but 240 bar shifted by 1 bar.
Not very noticeable here, same picture, but shifted a little - one bar forward forecast is possible.
Same frequently 240 bar, but shifted by 10 bar or 11 in relation to the right-hand end.
A 10 bar forward forecast is not possible.
All that's left is to assign letters of the Greek alphabet to the peaks and it will be a phenomenon.
Reread it a few times. Didn't understand anything. What letters, what the alphabet has to do with it, where here and who considers phenomena by the letters of the Greek alphabet, what "shifted ends". At least give me the citations you were referring to.
PS: good program, used before. well, you estimated the spectrum, if I am not mistaken on the basis of the criterion of maximum entropy. So what? I told you that it is useless, the regression model underlying this method has nothing to do with the market.
What do you want me to say? That there is non-stationarity? Yes there is, you should have asked before.
Reread it a few times. I don't understand anything. What letters, what the alphabet has to do with it, where here and who considers the phenomena by the letters of the Greek alphabet, what "shifted ends". At least give me the citations you were referring to.
PS: good program, used before. well, you estimated the spectrum, if I am not mistaken on the basis of the criterion of maximum entropy. So what? I told you that it is useless, the regression model underlying this method has nothing to do with the market.
What do you want me to say? That there is non-stationarity? Yes there is, you should have asked earlier.
As far as I understand the topic, the idea is to cut out what forms heavy tails and leave a stationary series. If I understand correctly, which will be the mainstream and we will use it.
The mainstream is perfectly searchable with maximum entropy and the fact that the peaks are floating is not the biggest nuisance. The trouble, including your methods, is that,
1) That we need stable characteristics at the right edge of the quotient and the less lags involved in determining these characteristics, the better.
2) the identified characteristics have to be extrapolated out of the sample, and we have to have some consideration of the credibility of that extrapolation.
Either we subordinate all our exercises to this goal, or all empty talk.
As far as I understand the topic, the idea is to cut out what forms heavy tails and leave a stationary series. If I understand correctly, which will be the mainstream and we will use it.
Yes, that's one area of research
The mainstream is perfectly searchable by maximum entropy and the fact that the peaks are floating is not the biggest nuisance.
This is a very controversial point.
The trouble, including your methods, is that,
1) that we need stable characteristics at the right edge of the quotient and the less lags involved in determining these characteristics, the better.
2) The identified characteristics must be extrapolated out of the sample and we must have some considerations about the credibility of that extrapolation.
Either we subordinate all our exercises to this objective, or all empty talk.
How can you predict without understanding the process being predicted. If we look from this point of view, the process "alpha" is well described (well, let's say as well :), for example by Ornstein-Uhlenbeck process, respectively by components that define displacement and diffusion (or in other words - volatility).
So, this process (betta) is really a big mystery, the fact is that the quote is not a random process, it is complex, non-linear, but not random - it is proved. It can 't be that the process betta, so randomly pulling the process alpha creates a completely non-random final series. There must be some sense here, it is necessary to understand what it is. And the appearance of a betta flow, (with different classifications) suggests some thoughts...
to joo
:о)
created a completely non-random final series
The market movement along some smooth curve is usually called a trend movement and I don't care how we got this smooth curve, because even in case of shocks its smoothness does not change, but the fitting error increases. In this sense all reasoning about SB with drift seems to me farfetched, at least for a forecast where more important is the presence of some additional market properties not considered during the formation of this smooth curve, which will change its direction on the next bar outside the sample.
From here the detrending algorithm comes to the fore, and then accounting for the residual from detrending and its effect on the trend itself. There are plenty of already developed methods, tools, tests and experience for this path.
created a completely non-random final series
The market movement along some smooth curve is usually called a trend movement and I don't care how we got this smooth curve, because even in case of shocks its smoothness does not change, but the fitting error increases. In this sense all speculations about SB with drift seem to me farfetched, at least for the forecast, in which more important is the presence of some additional market properties not considered at the formation of this smooth curve, which will change its direction on the next bar outside the sample.
What's that got to do with anything? I was writing about something else entirely. We are talking about co-organising stochastic processes with a random structure.
From here, we focus on the algorithm of detrending, and then we take into account the residue from detrending and its impact on the trend itself. There are plenty of already developed methods, tools, tests and experiences for this way.
It has nothing to do with trends and detrending. This is a dead-end way, you cannot remove trends, and there are no trends as such. Quite simply, there are several subsystems "competing" with each other. One sub-process takes over the control of forming a quote from the other.
What's that got to do with anything? That's not what I wrote about. It's about co-organising stochastic processes with a random structure.
It has nothing to do with trends and dendering. It is a dead end, you cannot remove trends, and there are no trends as such. Quite simply, there are several subsystems "competing" with each other. One sub-process takes over the control of forming a quote from the other.
The deaf cannot understand the sighted