Machine learning in trading: theory, models, practice and algo-trading - page 796

 
By the way, it would be interesting to try to use MO for options. So that the TS would determine what model should be set now, either straight or strangle. Classification could do the job here...
 
Alexander_K2:
I AM A_K2. Please do not confuse me with anyone else! And the pseudo-Mark process will show itself, oh, how it will show. But, there are people here who are cooler than I am. We wait with the Grail. To each his own!

In 2003 and earlier, is it possible to look at the minutes?

When I accidentally downloaded the history of M1, I was astonished.

Well, it's heaven and earth.

And the process seems to be the same - for exchange...

Of course I remembered at once what "from theory to practice" led to, it is clear that to the quotation of the 3rd year, no other way
 
Mihail Marchukajtes:

Michael, I understand that you see the market from the inside, but you just don't know anything about programming and can't make a robot. But, so to speak, you want to show everyone how to make money. Right? Have you tried freelancing?

 
Alexander_K2:

Michael, I understand that you see the market from the inside, but you just don't know anything about programming and can't make a robot. But, so to speak, you want to show everyone how to make money. Right? Have you ever tried freelancing?

This is absolutely wrong. My trading robot and indicators are all written by myself. I don't want to show how to make money, but I can't look at people who erred in some points very significantly, so I try to help guide them in the right direction, so to speak. I am a teacher I do not give a damn let someone to teach, because I have an inquisitive mind and much that is interesting. I'm an inquisitive mind and much is of interest to me. But when I see a man who truly believes in his approach and does not see critical errors in it, I can not go away.

 
Mihail Marchukajtes:


Hmm... Then why do you need the team of algorithms and coders that you want to create?

 
Uladzimir Izerski:

If that's a statistic, you've already made progress this year. +0.01% is not a bad start. We all know how to teach, but we just don't want to learn.

Bravo!!! Well done, but everyone here has seen it more than once. You know what makes me different from everyone else? I do not hide it, unlike some who can not provide anything at all. Not to mention the fact that I cracked a nut since 05.03.2018 and it looks like this

Well...what do you say now? Interesting to hear....

 
Mihail Marchukajtes:

Bravo!!! Well done, but that's something everyone here has already seen, and more than once. And you know what makes me different from everyone else. I do not hide it, unlike some who can not provide anything at all. Not to mention the fact that I cracked a nut since 05.03.2018 and it looks like this

Well...what do you say now? Interesting to hear....

I can only commend). Wishing you good luck. It's honest.

 
Alexander_K2:

Hmm... Then why do you need a team of algorithms and coders, which you want to create?

Programmers are needed in R not to waste time and quickly create a system on any chosen package and prove or disprove the essence of the approach.

 
Renat Akhtyamov:

there is

That's why it is not real.

The good brokers on ECN accounts for a liquid instrument already for a long time have no difference between demo and real. I personally trade in that way, copying trades from demo to several real accounts. The result is the same, slippage in different directions compensate each other. Volumes, of course, are not 0.01 lots.
I completely agree with Yuri Asaulenko. If the testing methodology is correct, if we do not look into the future and correctly simulate spreads, swaps, commissions, and slippages, then there is no point in waiting for a long test on the real. Of course, we are speaking about the model for ourselves, because only the real will convince a third-party. But the lack of significant difference between the adequate testing, demo and real - is a fact.
 

After yesterday's verbal battles came to the conclusion that to make public my research in the field of regression, and especially to share experiences and disclose secrets as that at once did not want. Solely because of some individuals who themselves have not made any contribution to the development of the branch, but are able to call names solely because they can not understand what we are talking about here. I will continue to work in the direction of regression, but not as quickly as I would like, because I will have to deal with many things myself. Here I will publish only the results, both interim and final. And ask for help. Not without this. We are a community that should help each other.

If you are skilled in one of the R packages in the field of machine learning and this package is the most flexible, and you feel an unquenchable thirst for research and get a complete final product in the form of a specific algorithm, which results in good working regression models, confirmed by numerous tests and satisfying you as a trader. I am waiting for you in my personal area, and we will proceed there.

From you the skill to quickly write code in R for your chosen package. I will share my experience and disclosure of important nuances in the formation of a set of predictors, the organization of the AI system, as well as the method for selecting the model, which is most adequate for the output during repeated optimization of one and the same training file. Then the advisor and the transition to practice, that is, to earning dough, rather than babbling about the theory of distribution of Markovian or non-Markovian laws of theories of world earnings and aliens.

Ask yourself the question, "What are you here for?" To find and prove the right to the existence of a law discovered and known only to you. Or just to make money. And not just dough, but BABLY. If you're in the first category, then do not even bother to knock on my face. Because I'm here only to rake a shovel, and even with a bunch of tugriks. Preferably the amount with one and a lot of zeros.