From theory to practice - page 231

 
Alexander_K2:

Formula (14) is the probability density function of price increments. Confirmed experimentally.

There (14) are two parameters C and lambda (the characteristic value of the jump) that are always positive. How can the trend direction be encoded in this even function?
 
Aleksey Ivanov:
There (14) are two parameters C and the lambda (the characteristic value of the jump), which is always positive. How can the trend direction be encoded in this function?

It is better to read it this way. The lambda is the sum of the increments of the number of ticks received N over the time interval T. T/N is the rate of trading.

 
Alexander_K2:

It is better to read it this way. The lambda is the sum of the increments of the number of ticks received N over the time interval T. T/N is the rate of trading.

i.e. the Lambda can also be negative?
 
Alexander_K2:

It is better to read it this way. Lambda is the sum of the increments of the number of ticks received N over the time interval T. T/N is the rate of trading.

I looked it up too, for the first time by the way...

Isn't it easier: take the sum of iVolume() ticks and approximate to 3*10^8, that's the window width

//as a version of course
 
Aleksey Ivanov:
So the lambda can also be negative?

Yes

 
Alexander_K2:

Yes

it's sadomasochism over understanding the Schroedinger equation //phew-phew, don't let it roll over

well, how can the free path of a quantum particle from a barrier breakdown to the next barrier be negative?

 
Renat Akhtyamov:

it's sadomasochism over understanding the Schrödinger equation //phew-phew, don't let it roll over

well how can the free path length of a quantum particle from a barrier breakdown to the next barrier be negative?

It's a model, of course. If anyone thinks it's a market formula, they're wrong, of course.

 
Alexander_K2:

It is a model, naturally. If anyone thinks it's a market formula, they're wrong, of course.

Yeah, that's what I'm saying.

the more so Schroedinger himself described the breakdown model very cool and further perversions of his equation are absolutely unnecessary and the model will be cooler

three steps, three components...

you can try that, it's a fact.

all microelectronics nowadays are based on this...

 
And I consider my results as proof.
 
Maybe someone will program this function in a thorough and complete way - what if it produces incredibly accurate predictions? I don't know... I have only used a small part of Shelepin's theory...