How do you achieve a qualitative leap in market analysis? There is an option: - page 7

New comment
 
Do you have results showing that the market is not random? I've done a lot of different analyses of quotes, then substituted random, and the difference was minimal. I don't know methods of proving time series randomness, nor do I know if they even exist. In fact, I have never encountered evidence that any time series in nature is random (ant population, heartbeat, etc.), nor the reverse.
 
Actually from a statistical point of view - the market is almost random, with little trend component. But there is enough of that...
 
In general, does randomness have a place in the world? - It's more of a philosophical question. About life: why 4 weighting factors? It seems to me that one can only speak about the effectiveness of this type of neural network if one refuses to use stoplosses during optimization.
 
Reshetov:

As for wavelets, this is quite a scam. If we take any function, decompose it into a Fourier series and reconstruct it relative to the zero harmonic level, it falls under the definition of a wavelet, because the integral of the function histogram at this very level is 0. The wavelet operators only invent that their "inventions" supposedly contain more information than the Fourier transform. Fucking lobbyists are lying.

Astonishing knowledge of the subject - I mean, wavelet analysis. Actually, in wavelet analysis.
decomposition is not in a basis of infinite-time sinusoids, but in a basis of short
"wavelets. This makes it possible to analyze non-stationary
series. Information display in wavelet analysis is made in contrast to the Fourier analysis,
on a two-dimensional plane. Due to these features, wavelet analysis has received the widest
It is used in a huge number of areas - seismic, radar, compression
and information security, medicine, etc. By applying wavelet analysis to the input signal, the learning curve of neural networks will increase by orders of magnitude.

It would be interesting to know how a connoisseur of arbitrage, analytical geometry, neural networks and Fourier analysis, can build a Fourier decomposition and then extrapolate the simplest almost
table analytic function y=A0*sin(x**2) given on the interval say from 0 to
10*pi. Within wavelet analysis, this is not difficult to do.
 
Itso:
Actually from a statistical point of view - the market is almost random, with little trend component. But there is enough of that...
Many people also confuse random and equally likely events. If you flip the wrong coin, heads and tails will fall at random, but with different probabilities. Knowing the difference in probability can be used to gain an advantage in this eagle toss. Similarly with ticks, if a pips up and a pips down have different probabilities, then it is a sin not to pocket that very difference. And why the fuck would #GM stock move randomly, say, if the company makes $x profit per share on sales. Or the price of corn futures isn't random either if it's been gnawed by locusts. All markets are not random, but correlated with different factors which already determine supply or demand.
 
New:
Reshetov:

As far as wavelets are concerned, this is quite a scam. If we take any function, decompose it into a Fourier series and restore it, it falls under the definition of a wavelet with respect to the zero harmonic level, since the integral of the function histogram at this level is 0. The wavelet operators only invent that their "inventions" supposedly contain more information than the Fourier transform. Fucking lobbyists are lying.

By applying wavelet analysis to the input signal, the learning speed of neural networks will increase by orders of magnitude.


Why the hell would it increase by orders of magnitude? I'd rather lie about 20 percent. But by orders of magnitude. Entropy - a measure of the amount of information cannot grow by orders of magnitude, just because some function is called a wavelet, which is fashionable among suckers.
 
getch:
Do you have results showing that the market is not random? I've done a lot of different analyses of quotes, then substituted random, and the difference was minimal. I don't know the methods of proving time series randomness, nor do I know if they even exist. In fact, I've never encountered evidence that any time series in nature is random (ant population, heartbeat, etc.), nor the reverse.
And who can blame you if you took quotes and let's say random straying from Bernoulli's scheme, obtained one way or another and couldn't see the difference between the two. Go and see an oculist at your leisure, maybe he can help?
 
I didn't use Bernoulli's scheme, I compared it with a pseudo-random sequence obtained from the built-in Random function in MathCad. You can blame the function, but I'm sure there is no correlation between it and the time series of quotes. Let's be more specific, since you don't need the help of an oculist, show me where the differences are. Since you are so confident, why not back it up with explicit evidence.
 
getch:
I didn't use Bernoulli's scheme, I compared it with a pseudo-random sequence obtained from the built-in Random function in MathCad. You can blame the function, but I'm sure there is no correlation between it and the time series of quotes. Let's be more specific, since you don't need the help of an oculist, show me where the differences are. Since you are so confident, why not back it up with explicit evidence.
If the correlation coefficient between some time function and any other time function is close to 0, then they are independent of each other. But this is not to say that the lack of correlation with a random process can be an indication of the randomness of the second function. It would be surprising if a random process correlated with some other random or non-random process.

You'd better read a maths book at your leisure. Maybe you'd find some familiar letters there. You're banging on your chest, like you've been researching quotes for randomness. I thought that you must have really examined the probability distributions in quotes, calculated all sorts of dispersions and derivative functions, defended several dissertations and published several scientific works. But as a result it turns out that getch is an ordinary amateur, who has signed up for his own incompetence, or, to put it simply, lameness.
 
I don't know what you have more of, an inferiority complex or a superiority complex. But the frequent manifestations of that in you do not move me anyway. I can't disrespect my interlocutor, such notions. Now about casualness. The first thing each person familiar with mathematics does with quotes is apply probability theory to them. It was done by many people. The results are silent. Suffice it to say that these results are practically not used in writing automated systems (the statement is unfounded, of course). Now imagine that the quotes became random for some time, will the systems written by many people produce a different result? My unsubstantiated opinion is that they will produce the same result. To check this, take any Expert Advisor and run it through any pseudo-random sequence. And then compare. All this, of course, says nothing. Did I ask you to show the non-randomness of the time series of quotes, or do you just not want to waste time on such "nonsense"?
12345678910
New comment