Machine learning in trading: theory, models, practice and algo-trading - page 2857
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
Let's say some mammal (in the words of Katshchyik) can't explain something and whisper it, and calls it NOTHING in other words. It is doing science, according to you, and is worthy of the title of academician and other titles.
But from the point of view of an ordinary observer, it is just a crazy grandmother, because it is engaged in nonsense and confuses other mentally immature people.
No of course there are explainable and understandable connections before, and after, and in between a black box with a random outcome. Children under 3 - 7 years old and after 15 -21 years old. Everything seems to be the same, but the cuts are different... fates, I mean.
Not gone, but something she realised but couldn't bring it to the nobility))))
I would very much like to meet a person here who will say - thanks to MO I have made a shitload of money in trading, and you are just a loser.... Preferably with a screenshot of the broker's report and a chart under + 1000%.
But so far I haven't seen any of them...
I'd believe it if I saw one, most likely :)
Believe no one but you and me))))))))))))))))))))))))))))))))))))))
No of course, there are explainable and understandable connections before, and after, and in between a black box with a random outcome. Children under 3 to 7 years old and after 15 to 21 years old. Everything seems to be the same, but the cuts are different... fates, I mean...
Not gone, but something she realised but couldn't bring it to the nobility))))
So the level of medicine and pedagogy now is such that children 3-7 years old are all the same "sort of"?
Well, then I understand why you like Chernigovskaya.
in my opinion, any adequate person will say that all children and people are different, that's why their fates are different.So the level of medicine and pedagogy now is such that children 3-7 years old are all the same "sort of"?
Well, then I understand why you like Chernigovskaya.
I think any adequate person would say that all children and people are different, that's why their fates are different.It's a strange denial of the obvious. Science today can't explain quite a lot of things. It's kind of like history. Or do you have the opinion that everything is explainable and provable today?
I would very much like to meet a person here who will say - thanks to MO I have made a shitload of money in trading, and you are just a loser.... Preferably with a screenshot of the broker's report and a chart under + 1000%.
But so far I haven't seen any of them...
I'd believe it if I saw one, most likely :)
MO is a tool, you need a model...
MO is multivariate optimisation,
i.e. search for unobservable parameters,
i.e. fitting...
if you search for parameters for a non-existent vacuum, the result is one, if you adjust the model, the result is different....
You need a model, already working, MO improves the model by 10-30%, if there is no model, the result will always be minus.
I need a model that's already working
I.e., if there is already a strategy that works, you can try to improve it with MO and other data. Something like a smart filter.
But there will be an issue of stretching the algorithm on the data.
It's a good option, though. I've never thought about it that way.
But there the issue of stretching the algorithm over the data will immediately come to a head.
In this variant, the working one improves (not butter).
And in the generally accepted approach, the algorithm is stretched on the market and settings on the algorithm, that is a complete fitting of everything with everything, and the MO can easily do it, so many function minima will find that the test will be good and will pass the test and valid... but only the test will not pass by reality....
I am also interested in creating "complex rules" through the fitness function, it is a global search, but there are many nuances....
The problem with MO (in trading) is that the final EA is likely to have no MO left.
I cannot agree. Basic strategy + MO model + MM - this is approximately how an EA model looks like in my understanding, and even individual tree leaves can be included in MO.
My point is that it can be better:
1. Select the really important data from the data. Right here, which exactly and precisely affect.
2. Understand why they are important (have an impact on price movement). If there are no real reasons why this data influences the price - most likely it is just a coincidence.
3. Based on point 2, write a TS that uses the data from point 1. Debug it for a long time in the tester, watching each trade. Then in real life, seeing real trades and glitches, which in theory and in the tester it was simply not realistic to take into account.
In the 2nd and 3rd step MO will only hinder.
1. Yes, we can identify such rules. At the same time, we will mostly skip those that have a complex structure, for example, even such rules as "Sell on Thursday at 16 o'clock, if the daily bar is growing".
2. What "real reasons" mean - it is not clear here.
3. Any strategy has unfavourable periods, which can be prolonged, so you need to have a zoo of such strategies, and to create it manually - or be a genius, or live a very long time. I just got involved in MO after developing such a genius strategy, which took about two years, and every sneeze of it I could logically justify, and when I put it to work, I just got an unfavourable period for it, which was smoothed out on the adjusted data in the tester.
Returning to the second point - I am looking to find similarities in the behaviour of binary predictors that will continue to be effective in the foreseeable future. To do this, I want to identify specific predictors. I concede that we need to take into account the cyclicality of the relationships between predictors, which any of the automata models I know of do not do.