a completely random process and FOREX. - page 6

 
Korey:



I actually meant the following, I assume that k is the number of cycles:



i1 = fix(rand*N)+1

k=fix(rand*100000)+1

for j=1:1:k

rand;

end

i2 = fix(rand*N)+1

c=r(i1);

r(i1)=r(i2);

r(i2)=c;

end;

No difference.

Just to be sure. another version.

close all;

N=1000;
r=NORMRND(0,0.0077,1,N);

r1=r;

for i=1:1:10000
i1 = fix(rand*N)+1
for j=1:1:fix(rand*100000)+1;
rand;
end
i2 = fix(rand*N)+1
c=r(i1);
r(i1)=r(i2);
r(i2)=c;
end;

figure;
%r=r-05;
for i=2:1:length(r)
r(i)=r(i)+r(i-1);
r1(i)=r1(i)+r1(i-1);
end

grid on;

plot(r);
figure;
plot(r1)




That's it. There's nothing to add.
 
I forgot:
"y(n+1)=a0*y(n)+a1*y(n-1) .... a5*y(n-5) + b.noise we get a linear neuron + noise. what good is this?"
I don't really understand what this line means. What else is a linear neuron? A neuron is an adder and an initialisation function, but a linear what? You add white noise and you get a neuron? That's ridiculous.
 
Continuation about the noise. Here I decided to finish this topic.

Thermal noise is good noise!!!

I think so.

Let's take it from the sound card.

coolEdit + Matlab

do a mono 32float recording. and there's nothing there.


we record silence, but in fact the spectrum is not zero. -97dB. Very quiet

let's zoom in and get





Save to wav.

let's see dense distribution in Matlab


normal distribution. But it is smooth =). of course 1755129 reports.

and now. main thing
sum up with the same program from the first page. already good. only spikes are strong.


shuffle it to the chart below. This is not what forex looks like.



Now in order to get rid of the jumps associated ( strong outliers I think not of thermal nature ) let's take the derivative and count the additive noise again.

Although it is not very fair.

We get what we wanted in the beginning. total additive noise = 0 - white.



If we do a re-stop as above.



we get the same thing. probably at the limit of n re-stops-> infinity.


I don't really understand what this line means. What else is a linear neuron? A neuron is an adder and an initialization function, but a linear what? Add white noise grash-> and you get a neuron??? That's silly.

sorry. correct.

about neuron and linear regression equation
not to be confused, but they are very similar. note that 1st order. when there are no terms above y(i)^1.

linear neuron is where activation function is f(x)=a*x;

weighted sum.
x=w1*x1+w2+x2....+...wn*xn;

1st order regression equation

y(n+1)=a0+a1*y(n)+a1*y(n-1) .... a(m-1)*y(n-m)+e(n)

e(n) - noise with specified parameters of m.o. and dispersion.

very many models for forecasting are built like this.

for example like this.

y(n+1)=a0*y(n)+a1*y(n-1)+s0*e(n)+s1*e(n-1)

where e with dispersion parameters and mo. and generated by the computer. damn it. =)


neural network due to redundancy of connections is not so sensitive to noise.

in some cases neural network can work better than AR.

Have you Grasn used AR?
grasn-> To start with, just come up with a condition that the "price" is guaranteed not to be negative under any initial condition - you'll realise that it's not that simple. And investigate it, and what you are doing now is bullshit in the literal sense. There are a lot of processes, both natural and technical, resembling quotes. You can easily get a series from PI that resembles quotes with Fibo, levels and other attributes.
it's not a market model ;)) I'm not going to build it. EKLMN. making the price not go to "-" is not difficult.
the point is different.



If you grash understand everything about random processes and say that someone is taking a bullshit.
try to build support and resistance lines on the chart below for PI, everything will be very convincing =))))

show me on this chart unnatural behavior for the market. then it will be an argument. except that it went to "-".

I can generate data for 20 years ahead using PI. =).

regarding Pi is an irrational number. Decimal numbers form a pseudo-random sequence.
you will get the same idea. see the picture below.


Histogram.
data for up to 1700000 digits.









the figures I observe in all these pseudo-random sequences are very similar to the real market.
decided to share this information. it may not be new to many people.

for me it is more than surprising. the figures are so similar.

understand me correctly, but I thought that FIBO is a fundamental one.

Initially, I just decided to compare the patterns I found in the market. after getting them and comparing
with different sections of history, I decided, to compare with a pseudo-market (that's a good name).
I decided to see to what extent the patterns are present in the noise. I didn't decide to start trading or fitting with GA, but to look at
what I had dug up.


Well, the charts do turn out to be similar, for crying out loud. =))))

Conclusions:
1. Absolute random process (white noise) really doesn't look like FOREX.
2. The pseudo random number process is similar to FOREX.

3. Despite much, the question of where such a parody of the real market comes from is open ???






 
D.Will писал (а):

1. Absolute random process (white noise) is really not like FOREX.
2. The pseudo random number process is similar to FOREX.

3. Despite much, the question is how such a parody of the real market is open ???

1. it was interesting.
2 и 3. Evil market makers using GSH to disguise their intentions ? :).

By the way, I, for one, have yet to get any objective evidence that Fibo ratios work in the market. Not that this is my goal, but for each new automatic markup method I do a check and I don't see any Fibs.
 
lna01 писал (а): By the way, I, for one, have not yet been able to get objective evidence that Fibo ratios work in the market. Not that this is my goal, but for every new automatic markup method I do a check and I don't see any Fibs.
It depends on how you look for those Fibs. If in the same way as Swannell, i.e. analyzing only waves of the same order, then there really aren't any special "resonances" visible there: smooth p.d.f. without any prominent convexities. And if one searches through Fib clusters from waves of different orders, something might come out. I haven't found it yet :)
 
1 The fact that the applicant investigated the thermal noise
received from the open line input of a digital audio system
does not detract from the merit of the work, as special Motorola noise transistors are not available to all,
and Soviet pioneer and school children's palaces are no longer funded.
2. The paper reveals a paradoxical circumstance - a random permutation
in a stochastic sample is equivalent to an integral transformation.
Accordingly, the applicant did not bother to explain why the integration takes place,
and so we do not know by which formula. However, this also does not detract from the merits of the paper.
3.The main merit of the scientific material in question is that the honourable DeVille has disproved
H. a bunch (read: x-ton) of coryphaei and their books about the random market, and we believed them.
 

to D.Will

Few people really understand random processes, that's why they are random :o)))))). But that's not the point. For example, the behaviour of elementary particles in magnetic fields is exactly like the behaviour of quotes. The graphs are simply indistinguishable, besides you don't need to "move" anything, the model as such is there. Traders used to carry these charts and shouted that now we can predict everything. And what about the physicists? Do you think these "stupid" physicists began to build support and resistance levels and calculate the "Fibo" levels to analyse the behaviour of their elementary particles? :о))))

  • Support and resistancelevels, in essence, are zones of "concentration" of local extremums of one, prevailing type, i.e., in other words, the places with minimal attendance of the investigated value. Such places I will find you on ANY chart very much, especially where the noise has been added.
  • Fibo levels. What makes you think they are fundamental? Have you statistically verified their presence, or is it all emotion?

You are raving that your calculations are very similar to real quotes. Reread your posts, no one is denying that. I was just asking, they are similar - so what? You replied - like "just like that".

PS: For the third time I am writing that AR models work on such BPs, I will not write about it anymore.

 
I will certainly give my opinion:
The detected effect of ordering the noise structure by random permutations,
i.e. noise disorder measure reduction under any predetermined influence (still to be proven)))
is fundamental, on a par with Belousvois' chemical reaction,
I. Prigozhin's Nobel Prize-winning work Order out of Chaos (etc.),
In short, dear DeVille - I advise you to publish scientifically.
 

To D.Will

The figures I observe in all these p.random sequences are very similar to the real market. I decided to share the information. it may not be new to many people.

I wrote that the information is not new at all. If the term "p.random" means "completely random", then re-read carefully what you have written. They are not completely random sequences at all, they cannot be completely random by definition.

 
Korey:
I will certainly give my opinion:
The detected effect of ordering the noise structure with random permutations,
i.e. reduction of the measure of noise disorder under any predetermined influence (yet to be proved))
is fundamental, on a par with Belousvois' chemical reaction,
I. Prigozhin's Nobel Prize-winning work Order out of Chaos (etc.),
In short, dear DeVille - I advise you to publish scientifically.

I agree with my colleague - publish and I'm off to get popcorn. :о)))