Machine learning in trading: theory, models, practice and algo-trading - page 1214
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
If you do not know what to do with the market, then you will not know what to do with the market.
But even 95% accuracy of prediction on the OOS will not save from the sinking of the real, the market is the market, it often changes.
This is theoretical considerations or practice? If it's the practice - show me the signal for "normal model is at least 70% assurance". For you this is the norm - show at least one and stimulate the community to move towards the same goals.
My practice after giving up ZZ brought me back to 50% randomness. Well 55% sometimes - from which 5% will eat the spread.Is this theory or practice? If it is practice - then show me the signal for "a normal model is at least 70% ass". This is the norm for you - show at least one and stimulate the community to move towards the same goals.
My practice after giving up ZZ brought me back to 50% randomness. Well 55% sometimes - from which 5% will eat the spread.My practice. I do not trade with signals and I do not use them, I have no motivation, trader to trader is a wolf.
You get 50-55%, I get 70-95%, one gets a Zhiguli, the other a Bentley, people are different :)
Besides, I'll let you in on a little secret: with competent risk management you can make a profit on Random, on 50% of predictions, rolling strategies do not require predicting the direction, you only need the volatility, or rather the spread (ATR), ideally you also need a trend indicator, without direction, the two states "trend or flat", the volatility is projected quite well, the trend / flat state is worse, but also above 70%.
My practice. I do not trade with signals and I do not use them, I have no motivation, trader to trader is a wolf.
You have 50-55%, I have 70-95%, one has a Zhiguli, the other a Bentley, people are different :)
In this market the trader is a fellow trader, at least he can share his experience, no one is asking for ready-made solutions. What's the purpose of such meetings?
The above demotivators do not understand what goal they pursue, and when you dig deeper, you realize that they are not so smart.people are different :)
No, not different)))
For example all of you jabberwocky are the same.
Well, well, well, what does this have to do with the trade model and the model for finding good models? Or have you decided that I somehow miraculously put ZZ there - I can't even imagine how it could be done...
Or maybe I don't understand something) I just don't like zigzags
I may be missing something) I just don't like zigzagging
Judging by your examples and articles, your objectives are quite different - you take the type of training (forest, logit, etc.) of interest and make the EA learn and trade it,
And if you use ZigZag, then you should first make data mining and then do MO - all tops and bottoms of ZZ should not carry the same information, because depending on ZZ setup, the scale changes in bars and alternation of ZZ with big leverage and with small leverage form patterns
imho,@Maxim Dmitrievsky simplified the work with IR: you train and you get the result, or if there is no result, you may change the type of training and if there are no results, the data (predictors) hold no information for IR
Judging by your examples and articles, your tasks are quite different - you take the type of training (forest, logit, etc.) of interest and make the EA learn and trade it,
And if you use ZigZag, then you should first make data mining and then do MO - all tops and bottoms of ZZ should not carry the same information, because depending on ZZ setup, the scale changes in bars and alternation of ZZ with big leverage and with small leverage form patterns
imho,@Maxim Dmitrievsky simplified the work with IR, you train and get a result, or if there is no result, you may change the training type and if there are no results, the data (predictors) has no information for IR
In the end it all comes down to the banal trying of variants - with or without CA.
If the classical implementation on trading with the 1st instrumentHere trader is a fellow trader, at least you can share your experience, no one is asking for ready-made solutions. Otherwise, what's the purpose of communication?
I agree, comrade, but it is important to distinguish between trader and researcher, we are both researchers and traders, when there are unproven ideas, vague insights, it is logical to "digest" them in a public forum, because 95% of the ideas are either banal or boring, but when money has flowed in ... I'm not talking about yards or at least mio $, I'm talking about miserable kilobucks a month, which you can just live on, without the need to sell in a humiliating hiring, then by itself disappears motivation to yap at right and left, at least about the details of their dough mower.
Personally, I "found" and lost such a mower has several times, so I know what I'm talking about, although now I'm beginning to suspect that, and maybe there was no mower and everything just seemed, on the rollback strategies can make a profit long enough, purely by chance, probably in the future will think of this theoretical justification.
In the end it all comes down to the banal search of variants - with or without LA
If the classical implementation of trading with the 1st instrumentWhen MQL was beginning to develop there were some examples from Reshetov, but they were primitive, though )))
They may not close it, but open a hunt, like for a valuable fur-bearing animal, for java is a programming tool, a multiplatform defeat, threatening even giants with their own stuff in the field, who want to graze their own sheep :)
Yes, I just thought why create something myself, I am interested in the cause-effect relationship of the two variables my program is already able using Apache Lucene, JSOUP, JSON, Apache POI and so on technologies to recognize text anywhere in anything in the pictures to documents and so on (this is accompanied by information matrices (stored in a distributed database) according to which is indexed information recognized in graphic objects) if something can not - looking for a site to convert data into an acceptable format for recognition or itself if can transfer.
The thing is that I do not want to reinvent the wheel ... I just need to find a neural network capable of rapidly learning the input of two variables - the equity data, and the trend indicator.
(experience in Java EE for about 5 years, there are many projects already implemented).
I'm not even trying to attach neuronics to market trading. It is unnecessary and most likely impossible at the moment, since there was not at least one implementation of a stable earning neural network.
My Equity graph is not random and quite informative (i need to check it), i have learned to distinguish trends from flattens.
I have learned to distinguish flat trends. The trading is going on, but i need to improve efficiency.