Not the Grail, just a regular one - Bablokos!!! - page 259

 
b2v2:
What is 3 on 1.
Who needs:
1. take 4 instruments and do a linear regression of 3 on 1 of them. For example gbpusd, audusd, nzdusd on eurusd. There are exactly 4 options, as you can easily guess. You can choose 4 out of 7 majors 35. Total of 140 variants.
2. Regression can be done on a straight line y=ax+b.
3. Regression can be done on a sine wave or on +1,-1,+1,-1.
4. Dick solves another problem - making a spread from instruments with minimal variance.

Regression can be done on ANY function.

To quote Joker again:

There are several approaches, the main ones are:

1. Normalizing all instruments relative to any one instrument

2. Normalizing all instruments in relation to some independent function

All methods are ultimately based on regression analysis and the link to the matlib for calculating this I have already given

I use rationing relative to an independent function...

From which it follows that 3 by 1 is not our method. And no 140 options.

 
it seems to me a transcend indicator would be suitable for this task, what do you think ?
b2v2:
What is 3 on 1.
Who needs it:
1. take 4 instruments and do a linear regression of 3 on 1 of them. For example gbpusd, audusd, nzdusd on eurusd. There are exactly 4 options, as you can easily guess. You can choose 4 out of 7 majors 35. Total of 140 variants.
2. Regression can be done on a straight line y=ax+b.
3. Regression can be done on a sine wave or on +1,-1,+1,-1.
4. Dick solves another problem - making a spread from instruments with minimal variance.

Regression can be done on ANY function.
 
seedormatrasch:
What is the method of normalisation?

I normalised by a constant and also by a straight line (with a positive coefficient - i.e. pointing from the bottom upwards - well, we need profit...)

In other words, when I say "normalized", I do the following:

A1*SPRED1+A2*SPRED2+A3*SPRED3+A4*SPRED4 = TIME, where TIME - increasing numbers from 1 to whatever

Generally speaking, the HORIZONTAL channel, as far as I understand it, implies normalization to a constant.

 
On a directtranscender indicator, it can too. It does almost everything except the optimum Dick spread. It even did an oscillator.
In general, an independent function could be anything. Alas, just think and search...
 
b2v2:
On a directtranscender indicator, it can too. It does almost everything except the optimum Dick spread. It even did an oscillator.
In general, an independent function could be anything. Alas, just think and search...

Let us have a transcendrimer indicator, but how to build all synthetics in one window to select the right ones later ?
 
Stells:

Let's assume we have a transcend indicator, but how can we plot all the synthetics in one window and then select the necessary ones?

In Excel, the LINEAR function draws a linear regression. You can draw as many as 10000 on one graph.

In mql5, I don't see much sense in programming something until you fully understand the idea.

 
Stells:

we have a transcendrimer indicator, but how do we build all the synthetics in one window so that we can select the right ones later?
Why a transcendrimer? What is it ?
We need to get a SUSTAINABLE channel. One that won't fall apart after the cointegration section.
I.e., just off the top of my head - regression on a constant (so that the channel is horizontal).
Or straight line.
If we initially do regression on some "crazy" f-function (i.e. the channel will not even be straight but will be floating, initially) - what will we get AFTER the cointegration section? And how can we work with it?
Or am I missing something?
 

transcender in the track

if i understand correctly only 8 lines in 1 indicator can be created in MT4

i.e. to either switch to mt5 or overlay multiple copies of the indicator in 1 window

 

IronBird:
Почему трансцендримера? Что это?

it's me, my restless spirit has been hovering and I've decided to show up

We need to get a SUSTAINABLE channel. One that won't fall apart after the cointegration plot.

That's not how it works, it's bound to fall apart.

I.e. first of all, regression to a constant (so that the channel is horizontal).

Or you can do it on a straight line.
If we initially do regression on some "crazy" f-f-function (i.e. the channel will not even be straight but will initially be floating) - what will we get AFTER the cointegration section? And how can we work with it?
Or am I missing something?

You can regress to a sine wave, for example.

 
b2v2:

In Excel, the LINEAR function draws a linear regression. You can draw as many as 10000 on one graph.

I don't see much sense in programming something in mql5 until you fully understand the idea.

There is also an addon in Excel called Regression, but you have to enable it first.

the thing is microsoft knows about the regression grail and they shut it down for the uninitiated

the knowledgeable know to go to file - options - addons - analysis package

Reason: