Adaptive digital filters - page 27

New comment
 
hrenfx:
Modification of the FRAMA indicator.

So what's in it?

Sorry, if anything, for the direct question. I hate to poke around in the code on the one hand, but knowing you I'm curious on the other.

 
A modification of John Eulers' idea is described very lucidly in the link above.
 

Man, you have no idea how approximate the dimensional formula given in the link above is. If you calculate the Dh by ticks, then we'll discuss the results.

But there are a million adaptive mash-ups out there, and they will all give similar results in terms of trading.

 
hrenfx:
The modification of John Eulers idea is very lucidly described in the link above.

I'm lazy. Creepy.

I don't want to read.

I don't want to think.

I don't even...

I want to make money. I'm about to resign from a serious company. Just quit.

What do I care whose modification it is? I don't give a shit about John's ideas, much less Euler's ideas.

I'd like your opinion on the matter.

 

let me explain.

We do not know in advance either the optimal dip depth (attractor reconstruction) or the optimal delay. Therefore, we cannot correctly estimate at which dimensionality of the phase space R.X. saturation occurs. More specifically - why does it suddenly say period=10 in the indicator parameters? Where did 10 come from? And why is the calculation step equal to exactly 1 bar, and not 2 or half bars, for example?

 
alsu:

Geez, you have no idea how approximate the dimensionality formula in the link above is. Calculate the Dh by ticks, then we'll discuss the result.

There are a million adaptive mash-ups to choose from, all with similar results in terms of trading.

Man, I don't give a shit about approximation. I haven't done any research or analysis, as you can see. I read the article, added 5 extra lines of code to the old indicator and posted it. And I don't give a shit about the simplified formula for calculating Hausdorff dimensionality and the rest.

If you want more or less reasonable work, take it:

Starchenko N. "Fractality index and local analysis of chaotic time series".

And it is clear that only morons calculate Hausdorff dimensionality on price VR quantized by astronomical time.

 
joo:

I would like to know exactly your opinion on the question.

The late Mandelbrot and his predecessor Hearst have confused everyone by emphasizing the study of self-similarity through the fractal dimension, i.e. through the Hausdorff-Bizikovich dimension. There is no timeframe self-similarity in the markets. We should use another aspect of this dimension.

What is commonly understood as fractals and how they try to apply it to the market is IMHO bullshit, even more bullshit than the well-known portfolio theories of Markowitz and Co. We should not apply the Mandelbrodian fractals or the theory of the so-called chaos or the so-called fractal analysis, which does not exist as such, and certainly not to the days, but simply the Hausdorff dimension as a numerical measure of the behavior of the price series.

 
hrenfx:

The late Mandelbrot and his predecessor Hearst have confused everyone by emphasising the study of self-similarities through the fractal dimension, i.e. through the Hausdorff-Bizikovich dimension. There is no timeframe self-similarity in the markets. We should use another aspect of this dimension.

What is commonly understood as fractals and how they try to apply it to the market is IMHO bullshit, even more bullshit than the well-known portfolio theories of Markowitz and Co. We should not apply the Mandelbrodian fractals or the theory of the so-called chaos or the so-called fractal analysis, which does not exist as such, and certainly not to the days, but simply the Hausdorff dimension as a numerical measure of the behavior of the price series.

I agree with the conclusion, but not with the message. By the self-similarity of a fractal Mandelbrot meant not identity on different scales, but rather "similarity" - remember, he observed just natural fractals where identity in the most cases is out of the question. And the guys who have made up "correct" fractals such as Koch curve and Sierpinsky carpet, which do not exist in the nature, have confused everybody's mind. But it should be taken into account that they did it with only one purpose - to show how fractional dimensions are calculated analytically. And then, as usual, newspaper "popularizers" took hold of it, and the above mentioned transfer of notions took place. The result was exactly what we are talking about, i.e. Nonsense)
 

Anyway, my nerdy opinion has been voiced.

The very idea of modifying the exponential average coefficient seems correct, rather than without it.

 
Thank you very much hrenfx
1...202122232425262728293031323334...41
New comment