Machine learning in trading: theory, models, practice and algo-trading - page 3155

[Maxim Dmitrievsky]

So what's the bottom line? ) It goes on to offer

Non-Scientific Double/Debiased ML#

which is almost a complete copy of my approach, but without the important, from my point of view, additions

And it works better.

Bingo! I've reached the top of the kozul myself. Couldn't be happier.

And just as expresses delight with this picture:

now at least I know what to call my approach :)

I'll have to pay him for such nice moments.

[mytarmailS]

doesn't make any sense

[Maxim Dmitrievsky]

mytarmailS #:
It doesn't make any sense.

but very interesting.)

[mytarmailS]

Maxim Dmitrievsky #:

but very interesting )

as much interest as understanding)))

[Evgeni Gavrilovi]

Maxim Dmitrievsky #:
So what's the bottom line? ) It goes on to offer

Non-Scientific Double/Debiased ML#

Test it on other currency pairs, at least 10. If it works on all of them, congratulations

[Maxim Dmitrievsky]

mytarmailS #:

interest as much as understanding))

well this quest I've passed, left to find another or wait for patches

[Maxim Dmitrievsky]

Evgeni Gavrilovi #:

Check on other currency pairs, at least 10. If it works on all of them, congratulations

You can check it yourself, I've developed a little bit of kozul theme, so you won't get bored.

[Ivan Butko]

Maxim Dmitrievsky #:

You can check it out for yourself, I've developed a bit of a kozul theme so you don't get bored.

Maxim, can you give me a thesis:

MO works on quotations?

[Maxim Dmitrievsky]

Ivan Butko #:

Maxim, can you give me a thesis:

MO works on quotations?

Self no, because it is a tool, and another level of thinking is needed, with ML categories.

People have only recently realised the need to move from ML to ML. And different new directions have started to emerge where stability of results in complex systems is required. With variable success.

[СанСаныч Фоменко]

Evgeni Gavrilovi #:

Check on other currency pairs, at least 10. If it works on all of them, congratulations

Actually, Maxim wrote above that it does NOT work on OOS anywhere. Naked overfitting. Just a little more and we will start to rejoice at the extraordinary results of optimisation in the tester.