Help with Fourier - page 6

New comment
 
ANG3110, you are going in the right direction, even right thinking, but you should forget about Fourier, it is not suitable for prediction. Here is a hint for further development of your theory, the smallest frequency in the frequency spectrum of your indicator (MinFreq) should be less than PI divided by the number of bars analyzed (BarsAnaliz). MinFreq = PI / BarsAnaliz. By Fourier MinFreq = 2*PI / BarsAnaliz, i.e. 2 times larger. And then think for yourself.
 
lsv писал (а):
ANG3110, you are going in the right direction, even right thinking, but you should forget about Fourier, it is not suitable for prediction. Here is a hint for further development of your theory, the smallest frequency in the frequency spectrum of your indicator (MinFreq) should be less than PI divided by the number of bars analyzed (BarsAnaliz). MinFreq = PI / BarsAnaliz. By Fourier MinFreq = 2*PI / BarsAnaliz, i.e. 2 times larger. And then think for yourself.

Thanks "great teacher" for the tip. Ha ha ha.
I hint that it's not an indicator, but a script, and I change the period by moving the mouse as I like and the number of harmonics, too. The rest, I repeat, one must be able to tune it, to see and understand the picture. It reminds me a little of tuning a receiver to a radio station. Which ratios to choose, i.e. how many bars (by the way, I don't use the concept of number of bars, everything is time-based and switching from timeframe to timeframe, the number of bars is calculated automatically, I've already started to forget what the number of bars is, for me it's time-based), it can be seen in the minimum of skewness relative to the Fourier lines and the maximum amplitude of the harmonic sum.

Forget about Fourier, well forget it, who's stopping you if it's not useful for you.
 
passing by...

The Fourier transform only works correctly with stationary series. Absolute quotes are not stationary and contain a lot of noise.
In practice, it is advisable to track non-stationarity - along with calculating the spectrum, calculate its variance.

If the ratio of spectrum peak height dispersion to the value of this height exceeds a certain threshold, the spectrum peak is formed by non-stationaries and is not reliable, i.e. cannot be used for forecasting of quotation movement.

Thus based on spectrum dispersion you can detect unreliable spectra that cannot predict and most probably will confuse the picture, and vice versa - reliable peaks.
Try it...

 
ANG3110 писал (а):

It is true that the process is non-stationary. But it can be made relatively stationary given the nested cycles. As for forecasts, I've tried many methods in my time starting from power regression, adaptive divergences of different types up to "caterpillar-like" methods. But I like Fourier most of all. Where the market will go tomorrow or the day after tomorrow? Let's carry out several trials and get rid of "guessing by coffee grounds". At least the cycles work in the vast majority of cases. On Friday one of my acquaintances decided to leave an open order for EUR on Monday. I looked and said that the pullback would not start before Monday. But it might not have been.
Once upon a time, many, many years ago, I was accused of being addicted to astrology by citing my own writings in a very reputable book, only they didn't know that I had written it. So there I would have put these smart guys on the high seas for a few days fishing without radar, radio or even a compass. I would see where they would sail without taking into account the angle of the Sun, star gauges. Better yet, put them on Forex, and of course on a real account, I think getting clever and "teaching others" would immediately discourage them.


I wonder. Can I get a link to "writing in a very authoritative book"?
About the fishermen - are we talking about astrology or astronomy?
(If..., the answers can be private)
 
ArtemRG:
passing by...

The Fourier transform only works correctly with stationary series. Absolute quotes are not stationary and contain a lot of noise.
In practice, it is advisable to track non-stationarity - along with calculating the spectrum, calculate its variance.

If the ratio of spectrum peak height dispersion to the value of this height exceeds a certain threshold, the spectrum peak is formed by non-stationaries and is not reliable, i.e. cannot be used for forecasting of quotations movement.

Thus based on spectrum dispersion you can detect unreliable spectra that cannot predict and most probably will confuse the picture, and vice versa - reliable peaks.
Try it...

Yes, thanks for the hints, maybe they will be useful if I resume work with the Fourier transform.
 

Interesting!

my time studying MUWINGS
I came up with this strange indicator, or rather a simple set of muwings
synthesizing them for all TFs

the set works perfectly in the flat

there was an idea to make it all through a fast Fourier transform
but I have not got to

 
YuraZ писал (а):

Interesting!

In my time, when I was studying muwings.
came up with this strange indicator or rather just a set of muwings
by synthesizing them over all TFs

the set works perfectly in a flat

I had an idea to process it all through a fast Fourier transform
but I haven't got around to it

That's a nice picture. At my time I built such "rainbows" on different muvings and on T3 and on AMA and on VIDYA and on MESA and on compensatory muvings and on regressions of different degree and on averages with angular momentum adaptation and on logarithmic muvings and on stepped onesand on proportional-time shifted with proportionally changing period and on High-Low time Price Channal and on amplitude step ones, and on some other types, I cannot even remember all of them now. I even built their integral functions, i.e. the average of the sum of a multitude of averages with different periods in the arithmetic and geometrical progression. I even built a synthesis of various types of functions. The whole problem is to choose the pair of curves that best fit the situation at the right moment. I think Fourier analysis can help with this.
 
SK. писал (а):
Interesting. Can I have a link to "writings in a very authoritative book"?
About the fishermen - are we talking about astrology or astronomy?
(if..., answers can be private)

I'll answer privately in my email, because we're talking about phenomena of extra-physical memory here... I'm not that old in terms of physical age.
 
YraZ, excuse me, can you explain what is actually shown on the picture? As for what works in the flat, to be honest, I always thought that it works perfectly on a perfectly phoned (say heads - buy, tails - sell) opening, with a stop greater than the spread of the flat. Yes, it may not be optimal in terms of drawdown, but it works...
 
ANG3110, I'm sorry, I've read all your posts and I get the impression that you're hiding something. Either you are seriously hiding something (I'm not judging, money is an intimate thing), or you have an idea, need help, but don't want to open up to the end (this is not good), or you are talking about times of the order of months. I myself started to master all this precisely with the idea of applying the Fourier transform. And pretty quickly became disillusioned.
12345678910111213...19
New comment