From theory to practice - page 398

New comment
 
Alexander_K2:

Here, I will not be lazy to do the distribution of time intervals between ticks for the AUDCHF pair for the past week.

Here, so far, literally after processing the 10 min data.

It's an Erlang flow! Can't you see it?

But, its parameters vary depending on the time of day. And there has to be a strict, direct correlation between astronomical and this "Forex time". You cannot use only one of them. IMHO.

Wow !

397 pages just to see the obvious... And what other flow were you expecting to see ????? independent order flows, a buffer in the form of a cup and netting ... there by the physics of price formation from orders and generation of ticks - at the output erlang ...

but the maximum benefit from this is that in theory it is possible to identify a dishonest dealer who cheats unnecessarily. that's all

 
Alexander_K2:

Yes, gentlemen.

Finita la comedy.

I have analysed again and again the available data on tick increments, on price increments within high order Erlang flows, on price deviations from moving averages.

There is little doubt left.

In the market, in one form or another, we are witnessing the so-called Laplace motion. There is a Laplace distribution in some average form everywhere and everywhere.

Now we will start to shake the Forex market like a pear.

See you there.

Ptz, 250 years ago already said that...

Application

The distribution is applied to modelling of signal processing, in modelling of biological processes (when other distributions cannot be fitted at all), economics and finance. Distributions can be applied to:

credit risks

insurance claims

in work with Kalman filter

 
Maxim Kuznetsov:

That's amazing!

397 pages just to see the obvious... And what other flow were you expecting to see ?????, independent order flows, a buffer in the form of a cup and netting...there by the physics of price formation from orders and tick generation - the output is erlang...

But the maximum benefit from it is that, in theory, you can identify the dishonest dealer who is cheating unnecessarily. everything

And I wish to see and work in the simplest flow!!! In the progenitor of this Erlang flow. In the stream "without memory". Then all the mathematics and physics of Markovian diffusion processes will work. Otherwise, it's all to hell.

 
Alexander_K2:

And I wish to see and work in the simplest stream!!! In the progenitor of this Erlang flow. In a stream "without memory". Then all mathematics and physics of Markovian diffusion processes will work. Otherwise, it's all to hell.

To work in a noise stream it is necessary to separate and cut off a useful signal (trend). Otherwise, you get nonsense and disconnected from reality.

The frequency of ticks does not play any role at all, give it a rest, don't be stubborn about this nonsense.

 
Alexander_K2:

In a 'no memory' flow. Then all the mathematics and physics of Markovian diffusion processes will work.

Two for you, sit down)
Mathematics says that you cannot make money in a stream with no memory.
 
Alexander_K2:

And I wish to see and work in the simplest stream!!! In the progenitor of this Erlang flow. In a stream "without memory". Then all mathematics and physics of Markovian diffusion processes will work. Otherwise, it's all to hell.

Who told you that the process is without memory?

Open an order, then look at the profit after an hour, and, depending on the state of the order, make a decision to close or to continue holding the position.

Don't you think that this simple trading algorithm already contains memory?

Now multiply this simple model by millions of positions and simulate the process of market pricing in general.

And how is it that from a cloud of micro-models with memory, you get a general model without memory?

 
Nikolay Demko:

Who told you that the process has no memory?

Open an order, then after one hour look at the profit, depending on the state of the order decide whether to close or to continue holding the position.

Don't you think that this simple trading algorithm already contains memory?

Now multiply this simple model by millions of positions and simulate the whole process of market pricing.

And how is it that from a cloud of micro-models with memory, you get a general model with no memory?

The flow of events "without memory" is what I need. As a model of chaotic collision of molecules (read - chaotic actions of market participants). But the memory of the content of these events, expressed in the presence of "heavy" tails, remains and we use it unrestrained. Well, like if a heavy particle experiences chaotic collisions but stubbornly moves forward until it changes direction.

 
In fact, it is only now that I am beginning to understand Automat, who suggested some tractor instead of a model, and Yusuf with his hyperbolic megabucks and superbears. Expecting the support of the masses, getting slapped in the face all the time. The oil painting...
 
Alexander_K2:

I need the flow of events "without memory". As a model of chaotic collision of molecules (read - chaotic actions of market participants). But the memory of the contents of these events, expressed in the presence of "heavy" tails, remains and we make unrestrained use of it. Well, as if a heavy particle experiences chaotic collisions but stubbornly moves forward until it changes direction.

Really?

Do you think traders randomly make decisions?

And I'm going to shit all over the depo... I'm going to roll the dice now... ))))

That's hilarious.

 
Alexander_K2:

Well, like if a heavy particle experiences chaotic collisions but stubbornly moves forward until it changes direction.

Well, then count the direction of the particle (trend) and trade on it - nothing more chaotic is needed.

1...391392393394395396397398399400401402403404405...1981
New comment