How do you practically assess the contribution of a "specific" input to the NS? - page 10
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
Well, there are probably hundreds of thousands of graduates in Russia who have studied the science of econometrics. You are the only one among them so far. And the rest do not go here, because they have long ago become millionaires, applying science.
And I even answered you once before. There are no other options anyway. Mathematical proof of the profitability of the system in the future is a matter of faith. Do not like the 2 figures, fine, I do not like it either, have 100 tests and 100 forwards, 200 (less than 200 - make 1000) numbers - and this is statistics, and statistics is already mathematics, and mathematics is little more than a belief.
Well, if you measure the rail between Moscow and St Petersburg with your feet, there is no question.
There are other methods, such as Monte Carlo and it's not the only one. Thousands of series of observations are fed into the model
Well, if you measure the Moscow-St. Petersburg rail with paws, there's no question.
There are other methods, for example, Monte Carlo, and it is not the only one. Thousands of series of observations are fed into the model input
Thousands are not enough. You need millions and billions.
Why be shy?
Thousands is not enough. You need millions of billions.
One more time about ZZ. I see I'm not understood.
We take a story and draw a ZZ. Nothing is redrawn on the story. Learn NS. We do a forward test. We do not know where the zones will be and in these conditions it is not a question of re-drawing. I hope that NS will indicate real reversals, because false reversals that we see during re-drawing were just not applied to the input of NS - NS does not know about such distress.
What does the NS have to learn? Some pattern before a true reversal that we don't see with our feeble minds.
That's the idea, not how to overcome over-drawing.
One more time about ZZ. I see I'm not understood.
We take a story and draw a ZZ. Nothing is redrawn on the story. Learn NS. We do a forward test. We do not know where the zones will be and in these conditions it is not a question of re-drawing. I hope that NS will indicate real reversals, because false reversals that we see during re-drawing were just not applied to the input of NS - NS does not know about such distress.
What does the NS have to learn? Some pattern before a true reversal that we don't see with our feeble minds.
That's the idea, not how to overcome over-drawing.
If you don't point out the pattern before the turn, she won't be able to learn it on her own. And any successful reversal will only be a coincidence.....
If you don't point to the pattern before the reversal, it will not be able to learn it by itself. And any successful reversal will only be a coincidence.....
It won't. The network will place orders retroactively, because it has learned a stationary ZZ, and in reality it draws.
Do not be lazy and try to trade with the drawing indicator yourself.
Personally for me this topic has been closed for a long time.
It is strange that anyone has any illusions.
If you don't point out the pattern before the turn, she won't be able to learn it on her own. And any successful reversal will only be a coincidence.....
I've been taught for years that it will formulate a pattern - have they fooled me?
If the input is informative, it will formulate.
You have garbage in - garbage out.
I was taught for years that it would formulate a pattern on its own - was it fooled?
To a certain extent no, NS can indeed form some patterns within itself, which it actually trades on. But this pattern will always be different, and not the one you want the net to see. unless you explicitly tell it to look for those exact patterns which you need.....