From theory to practice - page 1557

[Грааль]

Alexander_K:

I don't need 5-10-20% a month. I earn as much as many people would never dream of. Why would I fuss over a spear?! It's all or nothing- there's no other way.

boogaga! boogaga! eh... young blood! i envy you in some ways))

[Vizard_]

Renat Akhtyamov:

Give Sanka the formula already)))

[Alexander_K]

Vizard_:

There is no formula, Maestro... Not from him, not from anyone else. I somehow realized it yesterday. I'd better go to the factory - at least they feed me there.

[[Deleted]]

Олег avtomat:

Ask the weathermen what they know and what you don't know.

They know that maths is not everything

[Aleksey Nikolayev]

A few more words about the challenge of the breakdown. The suggestion to read Shiryaev was mostly a joke. What we can use is the well-known CUSUM algorithm or some variation of it. Within the framework of price modelling by regression this algorithm can be used to determine the moments when it is time to recalculate the regression coefficients. Obviously (as always in a matstat) errors of the first and second kind are possible in this case.

The main thing is that the discontinuity is not predicted in the future, but is looked for in the recent past.

[Evgeniy Chumakov]

Alexander_K:

There is no formula, Maestro... Not from him, not from anyone else. I somehow realized it yesterday. I'd better go to the factory - at least they feed me there.

Did you check the type of price movement before the signal? Like Che something with an exponent?

[Evgeniy Chumakov]

Alexander_K:

There is no formula, Maestro... Not from him, not from anyone else. I somehow realized it yesterday. I'd better go to the factory - at least they feed me there.

I doubt that anyone who has it will put it out at once.

[Maxim Dmitrievsky]

Aleksey Nikolayev:

A few more words about the challenge of the breakdown. The suggestion to read Shiryaev was mostly a joke. What we can use is the well-known CUSUM algorithm or some variation of it. Within the framework of price modelling by regression this algorithm can be used to determine the moments when it is time to recalculate the regression coefficients. Obviously (as always in a matstat) errors of the first and second kind are possible in this case.

The main thing is that the discontinuity is not predicted in the future, but is looked for in the recent past.

http://www.thealgoengineer.com/2014/online_linear_regression_kalman_filter/

there are also rolling regression and other modifications

[Alexander_K]

Aleksey Nikolayev:

A few more words about the challenge of the breakdown. The suggestion to read Shiryaev was mostly a joke. What we can use is the well-known CUSUM algorithm or some variation of it. Within the framework of price modelling by regression this algorithm can be used to determine the moments when it is time to recalculate the regression coefficients. Obviously (as always in a matstat) errors of the first and second kind are possible in this case.

The main thing is that the discontinuity is not predicted in the future, but is looked for in the recent past.

Specific research is needed, Alexei. CUSUM, Schuchart maps, etc., if you are interested and close to it.

I alone stupidly don't have time to do everything. And there is less and less hope for the forum members. One - quotes Vysotsky, the other - philosophizes and follows some signals, as if this will bring them closer to the goal. Some kind of theatre of the absurd.

[Anatolii Zainchkovskii]

Alexander_K:

As far as I remember, there was a desire to find a fairly stationary channel in which to trade. You and many others tried to build a TF with gaps, but as I understand the gaps were of the same size.My suggestion is to build dynamic timeframe, i.e. if we look at it visually, each candle has its own number of minutes (seconds, hours, weeks). How to do it?I mean from the break to the break it is one bar.