A purely theoretical question for mathematicians. With the possibility of moving to the practical plane. - page 2

 
Aleksey Nikolayev:

The Dunning-Kruger effect.

I don't know what you mean? But probably has nothing to do with this topic.

=============

Okay. In the simplest approximation, my "hypothesis" looks like this:

Price movement (direction) alternates, trending up, trending down. We will not consider breaking it into smaller trends inside as well as correction with flat, at least at the initial stage.

This movement shows a zig-zag beautifully:

But it cannot be used in real time that easily, because the next "top" may not turn out to be such.

as in this case the "top" did not turn out to be the top.

I was thinking, what if we find some dependence (coefficient...) between already stable tops? And using it, predict the "stability" of the last vertex?

Is this problem solvable?

 
Сергей Таболин:

I don't know, what do you mean?

=============

I was thinking, what if we find some relation (coefficient...) between already stable tops? And using it, predict "stability" of the last vertex?

Is this task solvable?

Ignore the haters, stand your ground. Really, there's another problem that could arise there.
----
Well, take the "tops" to 0-1 and see... I guess the haters were right)))

 
Сергей Таболин:

I don't know what you mean? But it probably has nothing to do with this topic.

=============

So here goes. In the simplest approximation, my "hypothesis" looks like this:

Price movement (direction) alternates, trending up, trending down. We will not consider break-up into smaller trends inside, as well as correction with flat, at least at the initial stage.

This movement shows a zig-zag beautifully:

But it cannot be used in real time that easily, because the next "top" may not turn out to be such.

as in this case the "top" did not turn out to be the top.

I was thinking, what if we find some dependence (coefficient...) between already stable tops? And using it, predict the "stability" of the last vertex?

Is this problem solvable?

You have a well-deserved place on Yusuf's board of honour. Are you by any chance brothers?

 
Nikolai Semko:
You're justifiably hanging on Yusuf's board of honour. You're not brothers by any chance, are you?

Have you read anywhere in this thread that I am saying this idea is 100% correct? Do you have any idea what I'm asking? Or have you "deservedly" decided to flub here as well?

 
Сергей Таболин:

Have you read anywhere in this thread that I am saying this idea is 100% correct? Do you have any idea what I'm asking? Or have you "deservedly" decided to flub here as well?

No you have not. Simply, like Yusuf, you are looking for something that cannot be found where it does not exist.

And in doing so, you bring that nonsense out into the open, creating a flood in this intellectual forum.

Yusuf's thread is an intellectual flubber, as is yours.

 
Nikolai Semko:

No, you didn't write that. It's just that, like Yusuf, you're looking for something you can't find where it doesn't exist.

.........

Whether something is there or not can only be proven. Personally, I think it is possible to identify mathematical patterns, but I don't know how. That is why I am asking.

You say you can't. Thank you for your opinion. I hope you won't repeat yourself.

 
Сергей Таболин:

Personally, I think it is possible to identify mathematical patterns, but I don't know how. That's why I'm asking.

I think you've already been told.


I've read over a hundred books in a few years, including books on mathematics, markets, and psychology, and it's become a hobby to read and check something interesting.)

You don't want to read, try to check your hypothesis with the help of MQL, I think you've already made friends with MQL.

If you are interested in a fast implementation in MQL, then the EA calls the PP indicator and writes in the file a counter which increases when you redraw PP and resets the counter when the opposite top of PP appears (up / dn)

Run the EA in the tester and get the result - how many times was a stable top not stable ))

This problem can be solved within one hour

 
Сергей Таболин:

I don't know what you mean? But it probably has nothing to do with this topic.

=============

So here goes. In the simplest approximation, my "hypothesis" looks like this:

Price movement (direction) alternates, trending up, trending down. We will not consider break-up into smaller trends inside, as well as correction with flat, at least at the initial stage.

This movement shows a zig-zag beautifully:

But it cannot be used in real time that easily, because the next "top" may not turn out to be such.

as in this case the "top" did not turn out to be the top.

I was thinking, what if we find some dependence (coefficient...) between already stable tops? And using it, predict the "stability" of the last vertex?

Is this task solvable?

The obvious variant for the null statistical hypothesis here would be the assumption that prices represent a SB (Wiener process). In this case the sequence of knee values (or more precisely z/z0-1 where z is a knee and z0 is the minimum knee value) of the zig-zag is a realization of a sequence of (a) independent, (b) equally distributed by (c) an exponential law with (d) a single parameter, random variables.

It is not very clear how to construct an alternative to the null statistical hypothesis here, since (b) - (non-stationarity of prices) is exactly violated and therefore a single available realization is insufficient to test (a).

 
Aleksey Nikolayev:

The obvious variant for the null statistical hypothesis here would be to assume that prices are SB (a Wiener process). In this case, the sequence of knee values (more precisely z/z0-1, where z is a knee and z0 is the minimum knee value) of the zig-zag is a realization of a sequence of (a) independent, (b) equally distributed by (c) an exponential law with (d) a single parameter, random variables.

It is not very clear how to construct an alternative to the null statistical hypothesis here, since item (b) - (non-stationarity of prices) is exactly violated and therefore the only available realization is insufficient to test item (a).

Unfortunately, as early as in 8th class I spat on a gold medal and completely abandoned my studies. Thank you for your explanations, of course they are meaningful for knowledgeable people, but for me they say nothing.

Therefore I repeat my initial question:how to calculate regularities in a number of values? For example +165, -240, +18, -378, +681, -115.... ?

Can anyone, without being too snooty, suggest a formula for solving this problem?

 
Сергей Таболин:

Unfortunately, back in 8th grade I gave up on the gold medal and completely abandoned my studies. Thank you for your explanations, they are certainly meaningful to knowledgeable people, but for me, they say nothing.

Therefore I repeat my initial question:how to calculate regularities in a number of values? For example +165, -240, +18, -378, +681, -115.... ?

Can anyone, without being too snooty, suggest a formula for solving this problem?

Sergey, OK, I'll answer without any rhetoric.

There are an infinite number of formulas to generate the sequence you cited.

Which one of them will satisfactorily predict the next terms of this sequence - XZ.

To create an exact formula for a pattern that takes into account its future terms as well, you can't do without a time machine.

Your question can be rephrased as follows:

Can anyone, without being too snooty, suggest a working grail formula?

So the problem is in the last 4 points (....) of the sequence you cite.