Machine learning in trading: theory, models, practice and algo-trading - page 3430

New comment
 
fxsaber #:

Virtual is written in MQL5. It should be ported to C++ without any problems. I think it is one of the fastest tick backtesters in the world.

What about OHLC?

Do you have limits, stops?

and a link?

 
The backtester needs his own, it's the first test job for any quant :)
 
Maxim Dmitrievsky #:
You need your own backtester, this is the first test task for any quantum :)

Yes I have already written a lot of them simple, but with the limiters there is more than two lines to write and think )))))

It's hard for me, and I don't want to, and debug, and time ...


Although I've already spent the whole day searching for it(())

I'll have to write it, I guess.

 
mytarmailS #:

Yes, I have already written a bunch of them simple, but with the limiters there are more than two lines to write and think ))))

it's hard for me, and I don't want to, and I don't want to debug, and time ...

So there is no difference limitki or marquette, just a condition for opening.

You can't write it yourself - it takes even longer to understand someone else's :)
 
Maxim Dmitrievsky #:
So it doesn't make any difference whether it's limited or marquee, just a condition for opening

If you don't write it yourself, it takes you even longer to figure it out :)

It's clear that it's just a condition, but a backtester with market entry via vectorisation is written in one line of code....

And there are already limits stops stops takes, it's another much more complex and slower structure, and also the architecture of the code must be +- competently organised, that is a task far from an hour of work + debugging and other hemorrhoids.



And I see that after R-key I can't think in the concept of loops at all, only tabular tidydata structures.

 
At Saber all codes are scripted by Enigma machine. It is possible to understand them approximately never :) taking into account peculiarities of mql
 
mytarmailS #:

it's clear that it's just a condition, but a backtester with market input through vectorisation is written in one line of code...

And there are already limit stops and take-outs, it's another much more complex and slow structure, and the code architecture should be organised +- competently, it's a task not for an hour of work + debugging and so on.



And I see that after R-key I can't think in the concept of loops at all, only tabular tidydata structures.

Oh, it seems that way. That's a start.
 
mytarmailS #:

and OHLC?

No.

Are there any limiteds, stops?

Yes.

Is there a link?

Profile->Publications.

 
fxsaber #:

No.

Yes.

Profile->Publications.

thanks

 

The multicurrency is ready, so far this crooked one will do.

What's left is the best part.


1...342334243425342634273428342934303431343234333434343534363437...3699
New comment