From theory to practice - page 1710

 

Probably the best book on proper time for dynamic systems:

Files:
 
Alexander_K:

Probably the best book on proper time of dynamic systems:

You clearly have a confusion with understanding the influence of time - in probabilistic models time and time sequence of events does not play a role, but only the final result of the event and general statistics - that is why the drawdown can be any, before the desired event happens. It seems obvious.

To take time into account, you have to apply a completely different mathematical apparatus, from the field of digital signal processing - where time is already the main factor, e.g. the arrival time of the signal or the phase.

I can suppose that physicists usually are quite illiterate in DSP and besides Brownian motion they do not know anything else about the world around them, hence their humble trust in Russian luck and other illuminating forces.

 
Alexander_K:

Probably the best book on proper time for dynamical systems:

Does it make sense to interpolate the curves as in the linked article? Will it work in a sliding window?

this seems to be the way to get a row close to the SB, e.g.
 
Maxim Dmitrievsky:

Does it make sense to interpolate the curves as in the article I downloaded? Will it work in the sliding window?

This seems to be the way to get a series close to SB, e.g.

In the article you posted? I haven't read it, sorry Max. No time. Got carried away with my own market timing, in which the Grail hides...

A cursory glance at the drawings sees nothing more than the vectors of a random process demolition. Demolition is a small part of the Grail and time is the Grail itself.

 
Alexander_K:

In the article you posted? I didn't read it, sorry, Max. No time. Got carried away with my own market timing, in which the Grail hides...

A cursory glance at the drawings sees nothing more than the vectors of a random process demolition, nothing more. Demolition is a small part of the Grail, and time is the Grail itself.

Yeah. Well, like less volatility, less outliers

It's still not clear how to recalculate correctly in the sk. window.

 
Cowen has a good attempt at linking price and time
 
Maxim Dmitrievsky:


I would now be interested in the following experiment.

1. take a BP on OPEN M1 or CLOSE M1 of some currency pair agreed upon between us. We agree on which time period of this data we will look at the results.

2. We look at the test results of your best neural network model.

3. For the same pair, for the same time period, look at the results on tick data.

4. Same for BP, which I'll give you.

5. Compare, draw conclusions.

If you're interested, let me know.

 
Alexander_K:

I would now be interested in the following experiment.

1. take a BP on OPEN M1 or CLOSE M1 of some currency pair agreed upon between us. We agree on which time period of this data we will look at the results.

2. We look at the test results of your best neural network model.

3. For the same pair, for the same time period, look at the results on tick data.

4. Same for BP, which I'll give you.

5. Compare, draw conclusions.

If you are interested, let me know.

I have a framework for any BP ready on python (I showed it in my videos), I can also use your data

I can even record a video.)
 
Maxim Dmitrievsky:

I have a framework for any BP ready in python (showed in the videos), I can also use your data

OK. On Saturday-Sunday I will send you my GBPUSD series for November 5-15 of this year.

If there is a real improvement in the result, I will tell you how it was obtained.

 
Alexander_K:

OK. On Saturday-Sunday I'll post my GBPUSD series for November 5-15 this year, or do you need a larger sample?

If there is a real improvement in the result, I will tell you how it was obtained.

It would be better to have a bigger sample. What TF is this series associated with?

Reason: