a trading strategy based on Elliott Wave Theory - page 253
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
I dare only note that your friend, expanding his consciousness, has not managed to go beyond the implicit assumption about eternity of order. That's how it is: a person knows what should be done, and it seems to him that he does exactly that, but in fact remains in the frames of an unconscious delusion. Alas! Expansion of consciousness is a tricky thing. If a man saw where to expand it, he would have expanded it long ago. So, he is turning his head, staring into his eyes, but nobody knows when it will happen and what will be the reason.
I would only dare to note that your friend, expanding his consciousness, cannot go beyond the implicit assumption about the eternity of order. That's how it is: a person knows what should be done, and it seems to him that he does exactly that, but in fact he remains within the limits of an unconscious delusion. Alas! Expansion of consciousness is a tricky thing. If a man saw where to expand it, he would have expanded it long ago. So, he is turning his head, staring into his eyes, but nobody knows when it will happen and what will be the reason.
Yuri, it was just a lyric digression in two sentences, and you make such far-reaching conclusions. This man does not suffer from these ailments (his head is straight, his eyes do not dart around, he is aware of other points of view), apart from that he is perfectly all right in every way. By the way, he is also a physicist in the past.
You intrigue me by turning from mini forecasts to the structure of the Universe. Really you have decided to work out your own field theory. :о)))
All right, if you do not like philosophy, I shall try not to speak about it any more. It is better to explain, if it is not a commercial secret, what you meant by your use of mini prediction:
It is interesting for me, though not for the price, but for the indicator.
I admittedly don't really get it.
Sergey, you have misunderstood me. About your friend in that paragraph only the first sentence. All the rest is about people in general and me in particular. Everything I have written is first of all about me, so don't attribute it to criticism, arrogance or anything else. And these generalizations are not made from your post, but from my personal experience.
The fact is that broadening of consciousness is something I have been consciously striving for in recent years. Therefore, the topic is close to me and therefore the philosophy is also close to me. I would be happy to talk to you in that direction, although, I repeat, the format of the forum is not conducive: too much banging on the keys (for everyone) to explain my point of view. Would you like to do it here, in this thread ?
Regarding field theory you are mistaken, I have long been not interested in private theories, and even unified field theory, as it is also a private theory. Why be limited to physics when you can grasp the whole universe without tearing it apart and turning it into a simple sum of primitive models? :-)) However I liked theory of Evans, but exactly due to its conclusion on ontological and epistemological questions of the universe (sorry for language, I do not know how to say it simply).
As for the indicator, everything is simple: you look for the local price extremum, I look for the indicator. Since it changes quite rapidly, I have a very small sample.
Yuri, I understood that very well, but I couldn't help myself in the sense of a joke. Forgive me. :о)
The fact is that broadening of consciousness is something that I have been consciously striving for in recent years. Therefore, the topic is close to me and therefore the philosophy is also close to me. I would be happy to communicate with you in that direction, although, I repeat, the format of the forum is not conducive: too much banging on the keys (for everyone) to explain my point of view. Would you like to do it here, in this thread ?
Regarding field theory you are mistaken, I have long been not interested in private theories, and even unified field theory, as it is also a private theory. Why be limited to physics when you can grasp the whole universe without tearing it apart and turning it into a simple sum of primitive models? :-)) However I liked theory of Evans, but exactly due to its exit ontological and epistemological questions of the universe (sorry for a language, I do not know how to say it simply).
I don't think any philosophical reasoning would be appropriate here, and besides, I'm afraid I may not be a worthy interlocutor because of my "practical limitations". You can do some keyboard tapping, but it will take time and inspiration :o)
Concerning the indicator it is simple: you look for the local price extremum, I look for the indicator. Since it changes quite rapidly, I have a very small sample.
I understand and I also remembered, not so long ago we were discussing the topic of local extrema. In some sense the results of my investigations of indicator prediction (in my case it is a low pass filter) were disappointing. I haven't yet worked out a sufficiently accurate method for predicting local extrema, though much depends on the indicator
The fact that the resulting curve resembles a parabola is not surprising. MNC calculates coefficients in such a way that the function which most resembles the data source "wins".
A couple of remarks.
Any smoothing procedure causes the FAC of the time series to become more positive. Indeed, the spread of smoothed points decreases and there is an opportunity to "wait" for the next point in the "right" place. This gives the illusion of winning over a random process. In reality, there is no advantage, because it is the Bid price that is traded, not its average value (there is an inevitable phase delay). All I am saying is that LOC is essentially a digital smoothing procedure with all the problems that come with it.
And secondly, if there are two time series with zero FAC each (random) and positive correlation between them, then an honourable addition of the series is equivalent to a smoothing procedure (arithmetic average). This is about working with the (High+Low)/2 series. Yes, each of the terms of that expression is weakly predictable. Yes, the derived series is noticeably predictable. But what good does it do? That fact alone doesn't allow forecasting the initial time series.
A moving average is an example from the same area. There is no statistical advantage in using it for a random process. The reasonability of using smoothing procedures is possible only for time series with positive FAC. For the majority of instruments in the Forex market and for different TFs the FAC is negative!
Any smoothing procedure causes the FAC of the time series to become more positive. Indeed, the scatter of smoothed points is reduced and it is possible to "wait" for the next point in the "right" place. This creates the illusion of winning over a random process. In reality, there is no advantage, because it is the Bid price that is traded, not its average value (there is an inevitable phase delay). All I am saying is that LOC is essentially a digital smoothing procedure with all the problems that come with it.
And secondly, if there are two time series with zero FAC each (random) and positive correlation between them, then an honorary addition of the series is equivalent to a smoothing procedure (arithmetic average). This is about working with the (High+Low)/2 series. Yes, each of the terms of that expression is weakly predictable. Yes, the derived series is noticeably predictable. But what good does it do? That fact alone doesn't allow forecasting the initial time series.
A moving average is an example from the same area. There is no statistical advantage in using it for a random process. The reasonability of using smoothing procedures is possible only for time series with positive FAC. For the majority of instruments in the Forex market and for different TFs the FAC is negative!
Hi Sergey!
I do not have any illusions concerning my very weak (at the moment) forecasting model. So I offered it for discussion on the Forum. Besides, I wrote about the ultimate goal of the forecast, namely, prediction of a local extremum in the reversal zone, but not the price itself. In this respect I have simplified the formulation of the problem as much as possible. Let me remind you:
A forecast is considered correct if all the bars lay (or just touched) the channel borders in the forecast area. Everything else will be done by an additional logic based on the formed price series, the current price, the current position in the reversal area and its borders.
This simplification is not accidental. I wrote about how I tried quite a few professional forecasting tools in relation to price without much success.
I don't think the moving average should be used as a smoothing. Rather, the original purpose is to gain "knowledge" of whether the current trade is above or below the average price. That fact in itself is already valuable.
There was a great science-fiction writer, I think it was Henry Kuttner. The main character in one of his series was a professor who loved drinking. He did all his inventions exclusively when he was drunk. The fun part starts in the morning. So, one day, working on his mini-prediction, he decided to expand his mind with beer. I woke up in the morning (the morning began in my headache) with a vague feeling that I had invented something (apart from the feeling that comes from a headache). I looked curiously at the algorithm I wrote in MathCAD, and what it was doing.
So, second version of mini prediction or how to apply the stupidest trick (apparently the next version will be about applying the stupidest trick). In principle it's the same (from what I've understood in the code I've written), in addition, if anyone is interested, you can watch the start here "grasn 21.02.07 19:05". The prediction is done on the clock, and the primary series is taken as (H+L)/2.
Step 1. Determination of sample length
I assume that the main "mini" movement will be limited to a confidence interval by a linear regression, sample length N, and I use it as a reference. For selected current sample I iterate to history with step +1. For each obtained sample I define a number of parameters. Sampling length N is fixed based on several conditions triggering at the same time.
Step 2: Prepare input data for calculation
Next, I get the residuals from the subtraction of a linear regression from the price series. Those are the residuals that I will be forecasting.
Stage 3: Modelling the movement
There are some changes, but all within ANC. The matrix of basis functions has the dimension |N x N| and looks as follows (it's a good idea to keep it that way):
There were several decomposition functions in the first version. As a result, I left only one function, but it has added "weight" and got the name "wave function" as a gift from me. The wave function has seven parameters that are calculated solely on the basis of the resulting series. There is no arbitrariness. "Different" wavefunctions are obtained by fixing some parameters from each other. The seventh parameter is time samples. I should say right away that I don't use the Fourier transform, it's just senseless for this case.
Step 4: Forecast
Here we have a simple prediction function as follows
where, lambda is the skyline coefficient, derived from "square worms" (grasn 21.02.07 22:59).
Legend:
-the vertical red line -limits the sample to the right, it was taken for the prediction
-grey lines - High and Low correspondingly
-Blue crosses - actual price series
-red squares - movement pattern, which is also a forecast function.
Note that I am already predicting the price itself. :о))))
And here's the forecast itself, I've removed the RMS channel boundaries and the confidence interval:
And some more forecasts ...
...
...
...
Ugh, "tired" of screenshots. Just kidding! Of course there are some wrong predictions, but only if the price goes immediately outside the confidence interval. Still working on the model, but I'll do a test and go through all the hourly historical readings (over 50,000 of them) as soon as possible. I just need to think of reasonable criteria to estimate the quality of the forecast. I have an idea to attach a channel to the (H+L)/2 forecast series: http://www.altertrader.com/publications01.html
Question!
My colleagues, I found out that MACD (without the signal line) is much better in forecasting than the price series. How can I go from MACD prediction to the price series? Any thoughts?
Colleagues, I found out that MACD (without signal line) is much better predictive than price series. How can I go from the MACD prediction to the price series? Any thoughts?
That will get you into the famous current of Mathematical Market Prediction - http://forum.alpari-idc.ru/thread24517.html
Read the review here - http://www.investo.ru/forum/viewtopic.php?t=127490
Есть мысль к прогнозному ряду (H+L)/2 прикрутить канал: http://www.altertrader.com/publications01.html
Read the review here - http://www.investo.ru/forum/viewtopic.php?t=127490
Thanks, I'll read it.
As for the reviews, I have read them. But I need to define quality control of my forecast, and this is exactly what I need. Having made a price forecast, why not build a channel using this link? I think this channel will give more accuracy than confidence intervals or RMS.