I made one of these things once ... - page 15

New comment
 
sanyooooook:
Found an interesting pattern (EA)
If no details are expected, it's easier to show the balance graph at once :)
 
sanyooooook:
Found an interesting pattern (EA)

What is it?
 

sanyooooook jealously protects the forum from spam and advertising, but is too lazy to post his "miracle" code. DDD

PR for obscure critters...

And not a word about number theory and its surroundings.

Odd.

;)

 

You're right, there's nothing interesting there.

If I wanted to post code openly I would have used CodeBase. I didn't see any spam or advertising in it, although you can see a speck in someone else's eye...

The thread doesn't say it's about number theory:

I made such a thing once ...

That's what I did once made such a thing.

For a new branch seemed a little material, and posted here. And the idea, if you understand this screenshot, I once posted it once


 
sanyooooook:

And the idea, if you get it, is a screenshot, I posted it once before

The level system is kind of tricky. It seems to divide a range in half all the time, but the definition of the range and its binding is obscure.
 

At the beginning we define 2 points 1 - local minimum, 2 - local maximum (for now I do it visually) the middle point between them is usually a support level, we will use it instead of point 1. We wait for the formation of a new local maximum, looking for the midpoint between the new maximum and the previous midpoint, etc. This sequence has been observed by me practically on all TFs.


 
sanyooooook:

while I'm doing it visually.

Visually, of course... But if you set the ranges according to the zigzag, I think, it will not always be so beautiful. At least, my experiments with fibs showed exactly that: sometimes the alignment is good, sometimes it is completely useless. It means I need some additional analysis to decide if the top is good enough for binding or not. The topic is interesting, but so far I haven't had any particular thoughts about such analysis.

It is also unusual that the next range is anchored to the middle of the previous one, maybe there is a risk of divergence or collapse on the machine.

 
Candid:

Visually, of course... But if the ranges are zig-zagged, I don't think they always turn out so nicely. At least, my experiments with fibs have shown exactly this: sometimes the binding is good, sometimes it is completely useless. It means I need some additional analysis to decide if the top is good enough for binding or not. The subject is interesting, but I haven't had any particular thoughts about such an analysis.

It's also unusual that the next range is pegged to the middle of the previous one, perhaps there is a risk of divergence or collapse on the machine.

I was looking for the range by mcd: points 1 and 2 should lie in different zones of the oscillator


 
sanyooooook:

I searched for a range by mcd: points 1 and 2 should lie in different zones of the oscillator

Well this is equivalent to a zigzag with MAKD switching, it is also not quite clear what to consider an extremum of the MAKD. Again, it depends on the parameters.

Although there is an additional detail in this picture, the *right" extrema are indeed quite correct in shape.

 
Candid:

Well this is equivalent to a MAKD switched zigzag, it is also not quite clear what to count as an MAKD extremum. Again, it depends on the parameters.

Although there is an additional detail in this picture, the *right* extrema are indeed quite correct in shape.

EM, I am not looking for MAKD extrema, I am looking for price extrema using it. I consider an extremum formed when the signal line has crossed the main line. The minimum is sought in the range where the MAKD is in the negative zone; the highs are sought in the range where the MAKD is higher than zero on the condition that the main line is higher than the signal line and the maximum detected is higher than the previous one. I agree that there may be inaccuracies in identifying extrema.
1...891011121314151617
New comment
Reason: