Machine learning in trading: theory, models, practice and algo-trading - page 2371
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 that says what? .... ))
Vova is an asset of this forum and should be protected
You, cool dude, better help the stompers.
You're a cool trader)), and they still haven't learned how to trade with quasi-positrons for 5 years.
You, cool dude, better help the stompers.
You're a cool trader)), and they still haven't learned how to trade with quasi-positrons for 5 years.
It says that you've been stomping on the same place for five years))).
And in five years you haven't trampled anything. And you've got a lot of attitude.
So far.
The fact that no one understands your philosophical nonsense, not even you - it shows that we're trampling in place for 5 years ))))
You're beautiful )))))
Diagnosis confirmed...
Bye!
I get banned for every sneeze
Of course they're afraid that you might say something clever.
Of course they're afraid you'll say something clever.
2) The simpler the space, the more repeatable it is, it is easier to look for patterns and they will not repeat every 2 years
4) Smart simplification removes noise
What are the useful properties (what you need from a model)
1) The model must be adequate to market movements
2) repeatability of data within the model
3) simplicity
As an option, based on observations:
There are probably a bunch of fast algorithms out there that you can't compete with if you look at things realistically.
There is an option to use automation to gather data from all available sources of information.
The point is to automate the data collection and analysis from all possible sources, including information such as the state of the company, the opinion of actual consumers about the company and products and so on (everything is important here, "other" is my abbreviation).
Next, to see if others have already had such cases and such background.
On the basis of this issue a decision on the further variants of events, the most likely.
In the short interval it does not give an advantage, unless you know in advance about the courts and other strong factors... But this is unlikely to roll, even if it is possible (insider).
But in the medium and long interval it might work and the position can be acquired gradually, not explicitly...
--
Or you can create your own information. Information - that affects all the markets at once or a certain asset. I understand that this may be not quite legal and so on. I did not get into the question so I can't say for sure.
Yes, there have been such things before, targeted and explicit. Nevertheless, there are many assets and new ones will emerge and the options to ride a strong new trend or something similar are there for sure. So is taking part in the creation of that trend.
---
Conclusion: Information and world events strongly influence prices. And it is probably possible to automate the process of gathering or producing information, for example, only on one "target" asset. Imho one prepared series of deals is enough, even once every 5-10 years (for example 2 years is not much), to obtain a significant profit, and not to struggle in terms of speed and "skill" with other algorithms daily (it is "expensive" for sure and not so effective and it is "noise"). IMHO.
As an option.....
The option is unaffordable for many reasons
I will say one - the antithesis of the third point
I say that the price should be reconstructed into a model that is better for the brain and algorithms to understand, it compresses the information qualitatively, discarding unnecessary information, and you say - and let's add millions more of the same garbage / noisy information and then ...
If you follow the postulate "the price takes into account everything" - you don't need to analyze the salaries of every clerk in a million companies to understand where the price is going right now
The option is unaffordable for many reasons.
I will say one - anti-compliance with the third point
I say that the price should be reconstructed into a better model for the brain and algorithms to understand, it compresses the information qualitatively, discarding unnecessary information, and you say - and let's add millions more of the same garbage/noisy information and then ...
If you follow the postulate "the price takes into account everything" then you do not need to analyze the salaries of every clerk in a million companies to understand where the price is going right now
The actual price (that has already happened) may take everything into account, but I'm not so sure about that. And it seems to me that the price may have changed a lot, in the past. Either separately under someone, or as a whole, so it's not so obvious. Also in the future.
Thousands of factors are not noise or garbage. They are prerequisites for future consequences. You don't have to cover all companies to begin with. You can start with just one. Either create one yourself and the whole "background" around it, or get involved in the process of creating or analyzing the "noise" around an existing company in order to bend it over, for example, or, on the contrary, to rock it to growth. It's quite crude and straightforward, but the idea seems to be clear for sure. And it completes my post above.
The actual price (already happened) may account for everything, but I'm not so sure about that.
How can that be?
And it seems to me that the price could be greatly altered, in the past. Either separately under someone, or in general roughly at all and directly stated, but it seems like the idea is clear for sure. And complements
How can that be?
Thousands of factors are not noise and garbage. It's the preconditions for future consequences
..........
.....
It's not all about the model, it's about collecting signs