Machine learning in trading: theory, models, practice and algo-trading - page 3122

 
Rorschach #:

Shake my assemblage point completely. We can filter, approximate, extrapolate. Any other options?

Link to external signals.) It's complicated of course, but just the price series is too uninformative in my non-professional view))))

 

liked it from Simons, this is after 2010 of course

Section overview: Brown and Mercer realised the need to model realistic aspects of trading, such as market negotiation and trading costs.

Nyah)))))

 
Maxim Dmitrievsky #:

Preliminarily, the second version is not dramatically different from the first. Somewhere better, somewhere worse. We need to investigate in more detail.

example of the result of the second variant


The drawdown on the oos for how long? I have 1-1.5 years each.
 
Forester #:
Sagging on the oos for how long? I've had one to one and a half years.

There's a lot of variations, some yes, some no. I'm getting a lot of models spit out.

You only have one model?

 
Maxim Dmitrievsky #:

a lot of variations, some yes, some no. It spits out a bunch of models.

You only have one model?

Thousands. I can't get rid of these long drawdowns. And trading with them will be sad for 1.5 years. I'm afraid I won't have the patience for more than a month and will close my account.

 
Forester #:

Thousands. I can't get rid of such long drawdowns. And trading with them will be sad for 1.5 years. I'm afraid I won't have the patience for more than a month and will close my account.

It would be nice to understand how that piece differs from the others.

 
Maxim Dmitrievsky #:

it would be nice to know how that piece differs from the others.

I have 2 of them in a 5 year stretch of history. 1 and 1.5 years. Market volatility, I guess, or everything became completely randomised
 
Maxim Dmitrievsky #:

It would be nice to understand how that piece differs from the others.

Can the parameters of the row be generated?)
 
Valeriy Yastremskiy #:

Linking to external signals))))))) It is of course complicated, but just a number of prices are too uninformative in my non-professional opinion)))))

How tiresome to look at all this: research, puzzles and searches of the system....

The logic of market quotation could not have been complicated in 1970.

What does neural networks have to do with it, if the quote was launched at that time from the knees, from the paper written with a pencil and counted on the accounts?

So what if it's been 50 years.

The algorithm has not changed, I say as it is, 100%, checked!

Well, in 1970 a man couldn't invent something that a human couldn't understand, he couldn't!

 
Valeriy Yastremskiy #:
Can the parameters of the row be generated somehow?)
You can even output the difference and correct it through the MO.