From theory to practice - page 1104

 
Alexander_K:

Y-axis is S=+-sqrt((Tn-Tn-1)^2+(PRICEn-PRICEn-1)^2)


This is a clear sign that the graph is not a "random coin flip" :-)
and one may tear oneself to pieces over analytical proof and physical substantiation. Maybe they will give a Nobel, or maybe someone already checked in a public and diligence is secondary.

this is a thankless place - here everyone expects a practically tangible result :-)

 

Bloody hell... Now I'm shaking with anticipation of cash in my purse...

Let's look again, so there's no mistake:

Top chart - AUDCAD price not this week (sometime Tuesday-Thursday, did a whole chunk of data cut without fail...)

Middle chart - just the sum of price increments with start = 0 at 00.00 Tuesday. We see an ordinary SB with no return to the average, with no obvious patterns. We got out of the 0 point and went to who knows where...

The lower chart is the sum of distances (hypotenuse, with + or - sign) of tick data, i.e. when our exponential time intervals between ticks were taken into account in the increments. We have a radically different view of the graph...

I'll go and prepare my pockets... Finally!

 

I'm tired of messing around in Yexel. In WCSim it looks like this:

At the top is the price. At the bottom is the sum of the hypotenuse.

Who the hell knows... The bottom graph, of course, is different from the price. But, what type of processes it belongs to and whether it is worth mixing with it - I do not know yet.

I will try to look for the Grail in Minkowski space today.

 
Yuriy Asaulenko:
It's the process that counts! You know, I looked at your $20 signals.
Don't offend the artists, it's easy to offend them.

What are the signals - there are contest accounts, look at them crookedly, go deeper and look more substantially in that case.

They are not artists, but creators!

 
Yuriy Asaulenko:
It's the process that counts! You know, looked at your $20 signals.
Don't offend the artists, it's easy to offend them.

process for process's sake - here too, I'm prepared to inform you without going too deep that it's masturbation.

Writing for the sake of the writer.

"Artists" - don't steamroll the left, who and when to offend and which artists and what degree of offence...

 

Yuriy Asaulenko:
1. Важен процесс!

2. You know it,

3. Looked at your $20 signals.

4. Don't offend the artists, it's easy to offend them.

1. Process is important. But don't make it a cult!

2. don't know,

3. crooked $20 look.

4. I respect artists.

 
Renat Akhtyamov:

Oh, no.

Alexander is the kindest of men.

Yeah, that's not the point.

The fact is?

 
Roman Shiredchenko:


Roma is in his repertoire. I like this dude))

 

We need a sine wave!


 
Renat Akhtyamov:

You need a sinusoid!


The sinusoid sits in the dispersion of the process.

That is, the half-width of the wave packet(probability density) varies ("shakes") over the course of a day strictly periodically.

Some uncle in one of the threads said that if you can predict the variance relative to the mean, then predicting the price is nothing.

From the pages of this forum I once again say to this uncle - if we consider the variance of the process not from STO price vs. average, but from STO distribution of increments in the sliding window = 8 hours, then the variance(the probable deviation of price from average) varies sinusoidally. So?