Machine learning in trading: theory, models, practice and algo-trading - page 3423

 
mytarmailS #:
And this drawdown you have, is it within the framework of the trade or what went wrong?

That's how it closed on a manual trade.

 
mytarmailS #:
This is easy to check if you take one single pattern and calculate statistics. And it will be seen that there are patterns and they are repeated, but the statistics on them is not repeated.

The market rather consists of an aditive structure those patterns from different components are superimposed on each other, and their sum is the price. If you look at the sum, almost nothing ever repeats if you look at the individual components, it is already better. Here will help rsa decomposition, fourier decomposition and so on.

Usually they are similar patterns, not exactly the same.

 
Aleksey Vyazmikin #:

Will you try to prove it? ;)

That's the way it is in essence - there are patterns with low response in leaves.

Prove the rules of the Russian language? )

 
Maxim Dmitrievsky #:

Usually it's similar patterns, not exact-to-similar patterns.

To make it exactly right, you need to solve the invariance/mashtable problem in addition to the decomposition, roughly speaking, so that the model can see the same thing both on the minute and on the day.

There are many different approaches here, again the same Fourier can help, interpolations of all kinds or in general primitive candles from different TFs or zigzag + normalisation by y, of course.

So the right answer is decomposition + invariance.
 
Aleksey Vyazmikin #:

That's how it closed when trading manually.

To be honest, this is even worse, trading without a stop, with an incomprehensible risk and a tiny take. Very very dangerous trading, I would not sign up for this.

There is no demand from the robot, he is stupid, but to trade with your hands like that.... better a robot.
 
Maxim Dmitrievsky #:

Prove the rules of the Russian language? )

Yep - I doubt there is a logical basis and continuity in the spelling of endings. It may certainly have been invented that way in the 20th century, but that doesn't mean it's correct.

 
mytarmailS #:
To be honest this is even worse, trading without a stop with incomprehensible risk and a tiny take profit. Very very dangerous trading, I would not sign up for this.
.

No demand from the robot, he's stupid, but to trade with your hands like that.... better a robot.

Oh, how you like to discuss my "failures" in the signal!

When trading without stops, you exit a position when certain events occur.

Entries are based on the expectation of a trend change, so stops are simply unreasonable - a game against probability.

 
Aleksey Vyazmikin #:

Yep - I doubt there's a logical rationale and continuity in the spelling of endings. It may certainly have been invented that way in the 20th century, but that doesn't mean it's correct.

Oh, how many wonderful moments we've had. )

2nd declension, masculine, neuter case. Endings I, Y.
 
mytarmailS #:
In addition to decomposition, we need to solve the problem of invariance/mashtable, roughly speaking, so that the model can see the same thing both on the minute and on the day.

There are many different approaches here, again the same Fourier can help, interpolations of all kinds or in general primitive candles from different TFs or zigzag+normalisation by y naturally

So the right answer is decomposition + invariance.

Well, it's still a search in the symbol history, where they may be under-represented in the right amount.

 
Aleksey Vyazmikin #:

Oh how you like to discuss my "failures" in signalling!

I don't like it at all. You have no idea how painful it was for me to look at that drawdown and how worried I was.