Machine learning in trading: theory, models, practice and algo-trading - page 1569
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
And then the "physicist" starts talking this nonsense at a math forum, he is publicly put down and he sincerely doesn't understand why he has become a laughing stock...
In order to put someone down, it's a good idea to get a little higher than the plinth yourself :D
That's it, get your heels out of here, you're not interesting
Brownian motion as a non-Markovian random process
The theory of Brownian motion, well developed during the last century, is an approximation. Although in most practically important cases the existing theory gives satisfactory results, in some cases it may require refinement. Thus, experimental works carried out at the beginning of the 21st century at the Polytechnic University of Lausanne, the University of Texas and the European Molecular Biological Laboratory in Heidelberg (under the direction of S. Genay) showed the difference of the Brownian particle behavior from the theoretical predicted by the Einstein-Smoluchowski theory, which was especially noticeable when the particle size increased. The research also touched upon the analysis of motion of surrounding medium particles and showed an essential mutual influence of motion of a Brownian particle and a motion of medium particles caused by it on each other, i.e. a presence of "memory" of the Brownian particle or, in other words, dependence of its statistical characteristics in future on the whole prehistory of its behavior in the past. This fact was not considered in Einstein-Smoluchowski theory.
Uncles may have learned their lessons at the institute, but they're getting old and fucked up.
Uncles are still in themselves, but they did not deal with the markets in practice, at least they did not trade significant amounts of money for the family budget, and this is IMHO the main difference between a practicing trader and a theorist. I myself am not even a theorist yet, just at the stage of acquaintance, trying to figure out WHERE TO GET RELEVANT INFORMATION ON THE THEME. On speculations wellooooo much information noise and deliberate misinformation from all sides, even I did not expect. Just recently I wanted to invest in super-profitable PAMMs, stopped only by the past experience with MMM, and then those same people explained how these pams are organized and everything was in its place...
I am not even a theoretician yet, but I have not dealt with the markets in practice, at least they did not trade significant amounts of money for the family budget, and this is IMHO the main difference between a trader-practitioner and a theorist. I myself am not even a theorist yet, just at the stage of acquaintance, trying to figure out WHERE TO GET RELEVANT INFORMATION ON THE THEME. On speculations wellooo much information noise and deliberate misinformation from all sides, even I did not expect. More recently I wanted to invest in super profitable PAMM, stopped only the past experience with the "MMM", and then those same men explained how these pammas are organized and everything began to place ...
I have no personal experience anywhere :) here and there
There are several stochastic market models, among which is Random Wolf. But that doesn't tell you anything. It didn't tell me anything, for example.
The eternal unsolved topic is if a coin goes on tails 10 times, what is the probability of heads. In a perfect mathematical model, it's the same = 50. In the real world... if I'm not mistaken no one can prove or disprove whether the tosses depend on each other.
In the market, obviously by definition, movements are RELIABLE on prehistory, and market laws lie more in the realm of mass psychology, sociology, game theory, and only marginally in statistics, no more than any other processes.
That's why I conclude that if MO is to be used, it is certainly not to try to find regularities in price increments.
I wonder if someone made a neural network where inputs are not prices or formulaic indicators, but solutions of different strategies with logic unit and/or probabilistic model.
Maxim Dmitrievsky, you know what I mean?
Brownian motion as a non-Markovian random process
The theory of Brownian motion, well developed during the last century, is an approximation. Although in most practically important cases the existing theory gives satisfactory results, in some cases it may require refinement. Thus, experimental works carried out at the beginning of the 21st century at the Polytechnic University of Lausanne, the University of Texas and the European Molecular Biological Laboratory in Heidelberg (under the direction of S. Genay) showed the difference of the Brownian particle behavior from the theoretical predicted by the Einstein-Smoluchowski theory, which was especially noticeable when the particle size increased. The research also touched upon the analysis of motion of surrounding medium particles and showed an essential mutual influence of motion of a Brownian particle and a motion of medium particles caused by it on each other, i.e. a presence of "memory" of the Brownian particle or, in other words, dependence of its statistical characteristics in future on the whole prehistory of its behavior in the past. This fact was not considered in Einstein-Smoluchowski theory.
Uncles may have studied hard at the institute, but they are already old and crazy.
Admittedly, I am not familiar with the Einstein-Smoluchowski theory, but it is possible to assume the presence of a subtle memory, practically imperceptible in the conditions of the beginning of the big bang. If there is a reference point of motion of this particle, then memory is possible, if there is no starting point, memory can not be in principle. Well this is exclusively in my theory. What's AAA????
The eternal unsolved topic is if a coin goes on tails 10 times, what is the probability of heads. In a perfect mathematical model, it's the same = 50. In the real world... if I'm not mistaken no one can prove or disprove whether the tosses depend on each other.
In the market, obviously by definition, movements are RELIABLE on prehistory, and market laws lie more in the realm of mass psychology, sociology, game theory, and only marginally in statistics, no more than any other processes.
That's why I conclude that if MO is to be used, certainly not to try to find regularities in price increments.
I wonder if someone made a neural network where inputs are not prices or formulaic indicators, but solutions of different strategies with logic unit and/or probabilistic model.
Maxim Dmitrievsky, do you know what I mean?
I used to use just increments and the result was zero. I even made a few videos on the subject
I added time cycles (hours, days, months, etc.) to the increments and it started to work (so far without MO). I.e. dependences lie in time intervals, which interact with each other and each with itself, but with a lag.
Hours, days, etc. are categorical features that can't be compared to each other, but there are models in which they fit well, such as CatBoost. Plans are in the works to try it.
Basically, this is the logic block - the interaction of different time slots with each other (see my last article). I think MO should be able to handle it, too. I have no other thoughts on what other regularities there may be, except lag dependencies.
Admittedly I am not familiar with the theory of Einstein-Smoluchowski, but it is possible to assume the presence of a subtle memory, practically imperceptible in the conditions of the beginning of the big bang. If there is a reference point of motion of this particle, then memory is possible, if there is no starting point, memory can not be in principle. Well this is exclusively in my theory. What's AAA????
Admittedly, I am not familiar with it too, but I can also assume that SB will never be pure SB, always on generation will be influenced by something in periodic sense, at least the relic radiation :) but this is too thin things