The idea with the transition matrix no longer worthwhile - MQL4 and MetaTrader 4

Neutron 2007.11.16 10:38 #111

Who are the contra-motors?

Sceptic Philozoff 2007.11.16 10:42 #112

It's the parrot from Strugatsky's Monday Starts on Saturday, with the arrow of time pointing backwards.

Neutron 2007.11.16 10:49 #113

Ahhhh, last time I read it was 10-15 years ago... I can't remember the classics, and my ears are full of hair now.

Sceptic Philozoff 2007.11.16 11:13 #114

Generally speaking, the idea with the transition matrix no longer seems worthwhile to me, because in order for matrix F in this equation to have no contra-literate properties, it should have an almost trivial structure reflecting only the autoregression of the last returns (predicted) relative to some number of previous returns. That is, the nonzero non-trivial coefficients will be only in its first row, and below it will be something like a diagonal submatrix with ones on the diagonal and the rest zeros.

Any questions from newcomers Strategy tester give the magic number

rsi 2007.11.16 11:28 #115

The course of the discussion shows that Prival put an "u" on the first two points of his Plan early on :-). However, it seems to me that Markovian processes to formalize the flow of quotes would make sense to check the adequacy of the model. But then of course the vector L(k) is not a set of quotes "out of the window", but the last value of the price (or returns), i.e. - a scalar. It may become a vector, if, for example, we take the last quotes from several instruments (say, looped ones). But then the question of portfolio testing becomes very topical - the problem again becomes immense :-(.

Forget random quotes Any questions from newcomers Problems checking for open

Prival 2007.11.16 15:53 #116

Gentlemen a little wrong, I thought I had posted this material.

We are investigating a flow that has characteristics of velocity and acceleration

So in the simplest case our state L(k) is a matrix

where V(t) is velocity and a(t) is acceleration.

For a discrete process L(k)=[V(k),a(k)]

But most likely it is like this

I'm attaching my articles on this subject for you to look at (at those times they tried to simplify them or to reduce them on 386-machines because of great computational expenses, maybe it was the needs of the military that made multiprocessor machines appear). There's detail on how the F matrix turns out, as it's for the aircraft just throw away the range.

If anything I'm in touch.

P.S. And grandfather Tikhonov Vasily Ivanovich died 1.5 years ago, I did not know. Remember with a kind word was a great scientist. (There are references to him in the articles).

Files:

perehodna_matrica.zip 83 kb

MT5&MQL5 in Linux The truth about councillors! The market is a

Prival 2007.11.16 16:25 #117

Does the fog of the matrix formulas dissipate ? (Close[i]-Close[i+1])/(time[i]-time[i+1])=Скорость

Komposter joined and said he will try to handle AKF. I wish the "Great Stakhanovite" would join us. I have a task worthy of a master's hand, to write a Kalman filter in MQL.

Files:

kalman.zip 13 kb

Discussion of article "Experiments How to write the Author's dialogue. Alexander Smirnov.

Slobodov Gleb 2007.11.16 16:46 #118

I've read this branch of the forum, but there's not much to understand...

If someone could write down the system of differential equations that are to be investigated.

Prival 2007.11.16 17:07 #119

shobvas:
I've read this branch of the forum, but there's not much to understand...

If someone could write down the system of differential equations that are to be investigated.

Download the Wordov file just above. Perehodna_matrica.zip

Slobodov Gleb 2007.11.16 17:38 #120

I doubt we can talk about speed, let alone acceleration... at least in the same vein as the speed and acceleration of aircraft.

Random Flow Theory and FOREX - page 12