Machine learning in trading: theory, models, practice and algo-trading - page 3493

[Maxim Dmitrievsky]

Who wants a beauty? I'll give it to you for 16 :)

[lynxntech]

Maxim Dmitrievsky #:

Who wants a beauty? I'll give it to you for 16 :)

the car is glamorous, not interesting in general, if you bought it, fool as it was.

[Maxim Dmitrievsky]

lynxntech #:

the car is glamorous, not interesting in general, if you bought it, bollocks as it was.

I'm selling it for 16. 7k mileage, 5 litres, 8 pots, not beaten, not painted.

[mytarmailS]

Purchase via freelance, can be webmoney

[Maxim Dmitrievsky]

mytarmailS #:
Buying through freelance, you can webmani.

Then +20% and my commission of 1 million (I don't own the car) 😀

[mytarmailS]

What I once said about analysing the optimisation surface of model learning.

https://t.me/tslabprorugroup/155749/198482

[mytarmailS]

You have to look not only at the profit curve, but also at the way AO got to that curve.

[Maxim Dmitrievsky]

mytarmailS #:
You need to look not only at the profit curve, but also at the way AO got to this curve.

Well, you still need to look at the data first. It's just as a confirmation that the TS is designed normally.

[mytarmailS]

Maxim Dmitrievsky #:
Well, you still have to look at the data first. It's just as a confirmation that the TC is designed normally.

And how can it be seen except for the way that was in the video?

[mytarmailS]

СанСаныч Фоменко #:

For starters, before "finessed": approximately the same classification error on train, test, validaishin, and then a consistent run outside of these files, preferably in the MT5 tester. Everything else is nothing more than mental exercises.

I can fictionalise one sample so that it will pass the test on 5 more samples, but the real 6th sample will be a drain.

So this primitive does not work, and what the guy in the video says is similar to reality.