Machine learning in trading: theory, models, practice and algo-trading - page 2570

 
Aleksey Vyazmikin #:

Yes, I understand that it's not easy to understand the method right off the bat. I plan to describe it in more detail. And the data are processed before feeding them to CatBoost, but I want to make my own algorithm, which will take into account the changing dynamics of "decay", as you called it, when splitting.

Well yes, without at least a rough description of the whole pipeline it's hard to understand anything.

Aleksey Vyazmikin #:

Maybe there isn't anything, but how do you check it?

Probably using forward or inclusion of time in the predictors.

 
vladavd #:

Well, what's in the picture? The glass and the sanction.

I don't get it, where did you look? What site?

 
mytarmailS #:

I don't get it, where did you look? What site?

The same as in the screenshot, it was called sharkfx a long time ago, and even earlier it was a desktop program. In the original one too, but it's more convenient and there are more derivatives.

 
elibrarius #:
What secrets. Write directly - Oanda.

What secrets? Everybody knows about it, it's just that brokers are not supposed to be named here.

 
vladavd #:

Karoch, I'll tell you this, I have a system (or vision) that works pretty well and I look at the same "useless chart" that you said ...

There are things that you do not see, although you look at them closely.

I practically do not look at the order book anymore and of course I look at the real volumes with the CME... everything

 

I think it is not very correct to focus on the volumes/orders of one brokerage company, which is unlikely to pass through itself 0.01% or 0.001% of all forex trading.

If many hamsters are guided by this data, the sharks can easily "draw" any picture there, for them it is a trifle. And then take the cost of "drawing" there, from the rest of the market.

It is better to take information from CME, the volume is bigger. But also a little and can "draw".

 
elibrarius #:

I think it is not very correct to focus on the volumes/orders of one brokerage company, which is unlikely to pass through itself 0.01% or 0.001% of all forex trading.

If many hamsters are guided by this data, the sharks can easily "draw" any picture there, for them it is a trifle. And then take the cost of "drawing" there, from the rest of the market.

It is better to take information from CME, the volume is bigger. But it is also small and they can "draw".

You're wrong.

 
mytarmailS #:

you are mistaken

What exactly are you wrong about?
 
elibrarius #:
What exactly?
elibrarius #:

I think it is not very correct to orient on the volumes/orders of one brokerage company, which is unlikely to pass through itself 0,01% or 0,001% of all forex trading.

I'm watching 10 brokerage companies at the same time, I think that's enough.

Few people understand that you do not need to watch all traders of the world to understand the market segment. The law of normal distribution says that it is enough to see, say, 10k traders to understand that other 100k traders think the same way, and it does not matter if they trade on СМЕ or Forex, the psychology is the same, same books, same patterns, same rationale.

When they test a new vaccine, nobody vaccinates the whole planet for the test, they make a sample of 1-10 thousand different people and vaccinate them, and that is enough to draw a conclusion...

The same principle works here.

elibrarius #:

If a lot of hamsters are guided by this data, the sharks can easily "draw" any picture, for them it is a trifle. And then take the cost of "drawing" there, from the rest of the market.

Well, this is just a fantasy, a hallucination.

elibrarius #:

Better to take information from CME, the volume is bigger. But it is also not enough and they can "draw".

Again, hallucinations...

You're comparing a soft thing with a warm one, which is a violation of the laws of logic.


Or don't you understand what we're talking about?

 
mytarmailS #:

Contrast real volumes with market sentiment, you're comparing the soft to the warm, a violation of the laws of logic.

I'm not comparing them, I'm saying that both can be "drawn". And whether they do or don't, neither you nor I know 100%. I think that sometimes they draw and sometimes they don't. And in the moments when they don't draw, we can earn something.

mytarmailS #:

I am watching 10 DTs at the same time, I think that is enough.

Except for Onda and CME I don't know. The others have only price chart and tick volumes.