Machine learning in trading: theory, models, practice and algo-trading - page 994

 
Yuriy Asaulenko:
Admittedly, I have not seen - I was not interested in anyone's signals or even reports from the real. I do not see the point. I do not understand this universal desire.
If your hourglass is about to run out, but you do not want to tell people about the strategy and make a contribution - that's your problem. But why go to such lengths to grovel in this thread, grovel in another. Who said that no one earns, hold a candle or what. This thread is about the MO, do not forget. But I do not see with whom I can communicate. Sanych and Perervenko deflated and have not adapted the MO to the market, the rest are enthusiastic amateurs, like me. There is nothing to communicate with you at all, you do not know anything on the subject. So I think it's just a shame, don't be as embarrassed as some of the others. But people of average intellectual wealth apparently do not give this literacy in any way.
 
Alexander_K2:

Yuri, I'm telling you again, the principle is right.

But, you see, this is the thing - even cool ideas do not resonate with people if there is no signal (like a passport) or it is negative like mine. I see it in my own example - well, no positive balance, it would seem - pick up the banner, bring the work to completion, please people. No - no one is interested.

So in this case, too.

Well, no one has "equity in the sky," or even any equity at all, and that's it - the theme immediately becomes passé and uninteresting.

Conclusion: Each subject should simply have a "man with a signal". A positive one! Then life begins.

We are waiting for such a man. We hope and believe.

We hope and believe. 150% true.

+

 
Maxim Dmitrievsky:
Sanych and Perervenko deflated and did not adapt the MO to the market, the rest of the amateur enthusiasts like me ...... But people of average intellectual prosperity apparently do not give this literacy in any way.

I will copy my post of six months ago, testifying to the fact that it is extremely difficult to find a pattern in the market:

I get the impression that there is nothing to teach NS at all, because there are almost no typical situations in the market. Either specially trained robots with big money are breaking patterns on purpose.
My pattern finder couldn't determine a typical double top because in the last year in 20 most similar cases the price worked out a double top pattern and flew further up.
The forecast is represented by the bold red (high) and white (low) lines, gray and dark red variants from the history.

And the most similar options are not so similar.

Sometimes it predicts in one direction and the price goes in the opposite direction.


So it turns out 50/50, like with a coin.
 
Maxim Dmitrievsky:
If your hourglass is about to run out, but you don't want to tell people about strategy and make some contribution, that's your problem. But why so humiliate yourself and fall in this thread, fall in another. Who said that no one earns, hold a candle or what. This thread is about the MO, do not forget. But I do not see with whom I can communicate. Sanych and Perervenko deflated and did not adapt the MO to the market, the rest are enthusiastic amateurs, like me. There is nothing to communicate with you at all, you do not know anything on the subject. So I think it's just a shame, don't be as embarrassed as some of the others. But people of average intellectual wealth apparently do not give this literacy in any way.

If my ongoing series of articles is not an attempt to apply IO to the market, then we have a different understanding of this concept.

It's time to get rid of adolescent maximalism and the desire to evaluate everything. You still know little, know little, be humble...

Participation in this branch is inefficient because of the conflict of interest: my goal to give part of the knowledge in the examples that I understand, the purpose of most participants in this branch - show, prove, then we may be to follow. Those who want to learn study the articles and follow them. I know by experience that the best way to learn something is by doing examples.

Good luck

 

In order not to be hopelessly sad, I once again attach Kolmogorov's work on BP forecasting.

From it, it is simply obvious that we should work with the first and second price differences. And if the expectation of returnees is always strictly = 0, then with the second differences we need to work - to do such a tricky (according to Aleshenka - exponential) sampling, so that the ACF took some tricky form (I was not familiar with it).

And voila - the Grail from Aleshenka crying in tears is taken and distributed to the suffering people.

 
Vladimir Perervenko:

If my ongoing series of articles is not an attempt to apply IO to the market, then we have a different understanding of this concept.

It is time to get rid of adolescent maximalism and the desire to evaluate everything. You still know little, know little, be humble...

Participation in this branch is inefficient because of the conflict of interest: my goal to give part of the knowledge in the examples that I understand, the purpose of most participants in this branch - show, prove, then we may be to follow. Those who want to learn study the articles and follow them. I know by experience that the best way to learn something is by doing examples.

Good luck

What a young max God forgive me. I do not understand the strategy in your articles - some zigzags, digital filters. Why torture neural networks so much, when the same xgboost will be enough in the vast majority of cases. The main thing is the strategy.
 
elibrarius:


I get the impression that there is nothing to teach NS at all, because there are almost no typical situations in the market. Either specially trained robots with big money specifically break patterns.

You overgeneralized the situation: typical situations are not in your predictors, and the market has nothing to do with it. You should be concerned with predictors, or more precisely, the predictive ability of predictors:

  • Either your predictors have predictive power for your target, and that predictive power changes little as the window moves, in which case it makes sense to look for a model for that set,
  • or your predictors do NOT have predictive power for your target, in which case there is nothing to judge.

 
Vladimir Perervenko:

If my ongoing series of articles is not an attempt to apply IO to the market, then we have a different understanding of this concept.

It is time to get rid of adolescent maximalism and the desire to evaluate everything. You still know little, know little, be humble...

Participation in this branch is inefficient because of the conflict of interest: my goal to give part of the knowledge in the examples that I understand, the purpose of most participants in this branch - show, prove, then we may be to follow. Those who want to learn study the articles and follow them. I know by experience that the best way to learn something is by doing examples.

Good luck

Vladimir, first of all there must be a sense.

It's not there

 
Renat Akhtyamov:

Vladimir, first of all there has to be a point

It's not there.

This is where I don't understand the point. Explain, please.

 
Vladimir Perervenko:

This is where I don't understand the point. Explain, please.

I already wrote that doing any strategy takes a decent amount of time.

Probably even more neuronics.

No one has any profit

So what's the point?