The market is a controlled dynamic system. - page 3
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
... ... but this picture is not very easy to handle later on, so we'll make it into a digestible form:
.
.
and we immediately notice in this picture an autonomous generator.
Then we build various indicators, then combine them into our trading system (it doesn't matter how sophisticated it is now) -- and hook it up to the generator.
The main thing is that the TS is the converter of its input vector Y into the output signal Z of order operation
.
.
We have an open system, and it is impossible to close it.
avtomat:
... and immediately notice the stand-alone generator in this figure...
I don't agree with this notion.
There are 2 streams in MT4/MT5.
The first one is the history and it is this that can be represented by as a sequence of vectors (OHLCV).
The second flow is ticks (asc and bid) when we work on the real account.
To my great regret these two streams have different characteristics. Completely different. At timeframes higher than H1 the differences can be neglected (sampling noise is less than 1%), but lower...you cannot. And it's not quite right to build a model which depends on timeframe (process sampling period).
Z.O. I probably shouldn't have persuaded you to open a thread here for discussion (research). I'll get a perpetual ban again for the commonplace truths and explanations to their developers.The history should also be ticked, otherwise we'll immediately step on a rake...
Any TC has inertia and introduces a lag
.
.
The quality and stability of the TS depends on the value of these parameters.
Exceeding the lag above the critical one leads to the loss of stability.
I disagree with this notion.
There are 2 streams in MT4/MT5
The first one is history, and it is what you can represent as a sequence of vectors (OHLCV).
The second flow is ticks (asc and bid) when we work on the real account.
To my great regret these two streams have different characteristics. Completely different. On timeframes higher than H1 the differences can be neglected (sampling noise is less than 1%), but lower...you cannot. And it's not quite right to build a model which depends on timeframe (process sampling period).
Z.O. I probably shouldn't have persuaded you to open a thread here for discussion (research). I'll get an eternal ban again for the commonplace truths and explanations to their developers.The story should be ticked too, or else you'll make a mistake...
It doesn't change the point. That is not the point. Not only that, we can represent vector Y as composite Y=[(O,H,L,C,V),(a,b)]
ps.
But you do not need a ban in any case. Let's work!
Now, with all this in mind, let's do a little hustle...
Without depriving the market of its main advantage of being a generator, let us introduce some structure into its description.
In general, let us imagine it in the following way:
.
.
Given a particular kind of transfer function W and a very approximate description of noise f, we have the problem of identifying the input process X.
(Note -- with some experience, it can be solved)
But let's not stop there and go further.
Let our process, described by vector Y, at different parts allow to be described by three (let us assume three for definiteness) essentially different functional structures, i.e. it has three structurally stable phase states:
.
.
Now we are faced with the problem of choosing the best structure by any criterion.
Well, we've done the first half of the journey - we've broken it down.
I hope I've made it all clear and understandable.
The task ahead is control synthesis.