Advantages and disadvantages of an automatic vehicle - page 3

 
bas:

ATS is the only way to effectively exploit a pattern because a pattern is an algorithm, by definition. In this sense, automatic trading has no disadvantages.

And the system, i.e. the pattern (any pattern), has only one disadvantage - anything can happen in the future, including things that you have not foreseen.

this is a brilliant idea! a pattern is an algorithm and in order to use an algorithm, you need an algorithm to use the pattern! best statement.

all you need next is an algorithm to find a pattern and the grail is ready. An ideal TS would consist of an algorithm for finding patterns and forming algorithms to use those patterns. That is, not only the work with regularities must be formalized, but also the search for and control over the availability of the used regularities.

 
Maxim Romanov:

that's a brilliant idea! a pattern is an algorithm and to use an algorithm, you need an algorithm to use the pattern! best statement.

all you need next is an algorithm for finding patterns and the grail is ready. An ideal TS would consist of an algorithm for finding patterns and forming algorithms to use those patterns. That is, not only the work with regularities must be formalized, but also the search and control of availability of the used regularities.

This is how the MoD deals with it).

 
Renat Akhtyamov:

Vladimir, the FA theory is there, the formulas are there, but (!!!) the input data for the calculation are either not there at all, or they are there, but not in full,

unfortunately

The advantage of TA over FA is that everything is at hand, namely the price and nothing else but formulas to twist it.

It does not work with FAs. It takes more than one or two years to accumulate knowledge. The breadth of the information field is immeasurable here.

May I unveil the secret about the formulas?

 
Yuriy Asaulenko:

That's how the MoD does it).

I know) but there will be no grail on machine learning anyway. Then there will be competitions whose pattern finding algorithm is better, faster, more adaptive and there will still be losers. It's like in war. If you let out an army of infantrymen with spears to attack the peasants, you win, but if the enemies have tanks.... But when everyone has tanks, it's a competition whose tank is better. But people have nothing to do in war with tanks, although there will be loopholes. There will be loners who will create diversions. So it is with trading. There is room for hand trading, but how long will you survive with only a spear on the battlefield with tanks, planes, guns and other equipment. And if you also replace soldiers with robots that do not feel pain, then there is not much room for a meat man, but there will always be one.
 
Maxim Romanov:
I know, but there is no such thing as a grail on machine learning. Then there will be a competition in whose algorithm for finding patterns is better, faster, more adaptive, and there will still be losers.

You are contradicting yourself. If there are regularities (which I completely agree with), then the machine doesn't need to be taught them. You have to "bluntly" work them out. And the system should be based on this.

I am not very good at expressing my thoughts. If you have found a pattern, you should work it off. After all, it repeats. You find another one, you work it off. These regularities you find yourself - if you look at the chart every day, you cannot miss them.

 
Uladzimir Izerski:

The advantage of TA over FA is that everything is at your fingertips, namely price and nothing else but formulas to twist it.

This does not work with FAs. It takes more than one or two years to accumulate knowledge. The breadth of the information field is immeasurable here.

Can you reveal the secret about the formulas?

no problem

Personally, I like this particular textbook

i hope you enjoy reading it too

Interested in Volume 2/Section 3 etc.

http://institutiones.com/download/books/1618-mezhdunarodnaya-ekonomika-kireev.html

Международная экономика - Киреев А.П. - Учебник в 2-х томах
  • Экономика
  • institutiones.com
Год выпуска: 1997, 2001 Автор: Киреев А.П. Жанр: Международная экономика Издательство: «Международные отношения» Количество страниц: 416, 488 Описание: Структура учебника «Международная экономика» обеспечивает последовательное рассмотрение основных закономерностей функционирования международной экономики. Он состоит из двух частей. Первая часть...
 
Mihail Marchukajtes:

It is precisely time in the market that means nothing. Earnings are made on the price axis, not on the time axis. So it is better to evaluate a level at a specific price rather than the number of bars that price has spent at that level. In other words, earnings are on the Y-axis and not on the X-axis where X is a time scale.

If a person doesn't read the market like a book (like me, for example) time doesn't matter, as well as price and bars.

 
Maxim Romanov:
I know) but the grail on machine learning won't come out anyway. Then there will be a competition in whose algorithm for finding patterns is better, faster, more adaptive, and there will still be losers. It's like in war. If you let out an army of infantrymen with spears to attack the peasants, you win, but if the enemies have tanks.... But when everyone has tanks, it's a competition whose tank is better. But people have nothing to do in war with tanks, although there will be loopholes. There will be loners who will create diversions. So it is with trading. There is room for hand trading, but how long will you survive with only a spear on the battlefield with tanks, planes, guns and other equipment. And if you also replace soldiers with robots that do not feel pain, then there is not much room for a meat man, but there will always be one.

However, if you change the paradigm itself, all these inferences become meaningless.

The other paradigm is that no one needs to win or even fight in this war. Any multiplicity of strategies inevitably generates some new market statistics and new patterns. And it is up to us to use them as long as they exist, whether with robots or with our hands.

To put it simply, as long as we are not a link in the food chain of large participants - we and our actions are of no interest to them.

 
Ibragim Dzhanaev:

You are contradicting yourself. If there are regularities (which I completely agree with), then the machine doesn't need to be taught them. You have to "bluntly" work them out. And the system should be based on this.

I am not very good at expressing my thoughts. If you have found a pattern, you should work it off. After all, it repeats. You find another one, you work it off. You find the rules yourself - if you look at the chart every day, you cannot miss them.

I am not contradicting, the patterns change, some are replaced by others.

 
Maxim Romanov:

I'm not contradicting, patterns change, some are replaced by others.

If a pattern changes so quickly. then it's not a pattern. it's something else. Of course things change, but not that fast.

If you find a pattern, it's enough for a lifetime, and you'll leave it to your children.

Reason: