From theory to practice - page 472

[Uladzimir Izerski]

Igor Makanu:

I see, from a stochastic process to make its own stochastic process, well, there is a rationale here, in uncertainty to trade uncertainty

i don't know, i tend to think that if i transform price charts then there should be a lossless inverse transformation or the trading signals should be identical on both the main and transformed charts. i have thrown a couple of screenshots of my charts, i will not hide it is a conversion into decibels relative to the beginning of the day, it is hard to tell, but somehow price has levels relative to the day open - level repeatability is close to 100%, another issue is that levels alternate during the week

here are such thoughts

This is the question we can discuss. There is a ready answer.

[Igor Makanu]

Maxim Dmitrievsky:
So, I'm making a non-alternative AI that searches for the best solution

I do not know, I have long been inclined to believe that in order to teach a machine to do something, you need to be able to do it yourself, i.e. if you teach a machine to make market entries, you need to give it signals to enter, which are already formalised - i.e. I think it won't - that's what binary logic is - it's either yes or no, and it doesn't matter how this binary logic looks in a neural network or in a random forest, but it is anyway rigidly formalized rules, it's not a man that can change his mind.

Uladzimir Izerski:That's a question we can discuss. There is a ready answer.

Yes, I'd like to discuss it.

[Maxim Dmitrievsky]

Igor Makanu:

I don't know, I have long been inclined to believe that in order to teach a machine to do something, you need to be able to do it yourself, i.e. if you teach a machine to make market entries, you need to give it signals to enter, which are already formalised - i.e. I think it won't - that's what binary logic is - either yes or no, and it doesn't matter how this binary logic looks like in a neural network or in a random forest, but it's still a rigidly formalized rules, it is not a person who can change his mind

yes, I'd like to discuss

is already sorting itself out, it's a question of improvement

[Aleksey Nikolayev]

By the way, I keep forgetting to write here that for approximation of any objects depending on the increments (ACF, distribution function, moments, etc.) by their sample analogues on one realization the stationarity of these increments is necessary, but not sufficient. Their ergodicity is also required.

[Evgeniy Chumakov]

Alexander_K2:

Bye, that's all for now - gentlemen. I'll be away from the forum for the next few days.

Nah, you should be on the forum more often. Otherwise you might miss something interesting.

Yesterday I couldn't sleep and lay awake... ...thinking how to calculate the speed of the process.

Today I have calculated, and to my surprise (though why and without surprise, all logical) speed is a uniformly decaying (I won't cross it out or they will confuse it with something rotten and smelly) decelerating process with instant acceleration and then again deceleration evenly.

It is not said for nothing that most of the time it is a flat.

There was a lot of debate about what is a trend and what is a flat. It was even said that there is no flat, there is only a trend.

Or maybe a flat is a slowdown, while a trend is an acceleration of speed.

Or a trend is a price movement?

Or maybe a trend is when you buy more or sell more.....

You are looking for memory in ACF or something else. I think there is no memory there, I am sure of it.

You asked what invisible force was pulling the pair down, maybe it was the memory that was pulling them down or maybe it was the level of demand as I calculated it (there were more sellers willing to sell than buyers willing to buy).

Conclusion: Process memory is really memory, not ACF or something else. Well it's like you can't pour more into a 1 litre jug, it pours out and leaks.

Now, I'm waiting for criticism from local gurus!

[Evgeniy Chumakov]

Igor Makanu:

I now want to decompose the ZigZag by this principle, by the ZigZag angle increments

Here are the screenshots I posted above of the increments ranging from -45 to 45.

What's next and the big question is how do you calculate the angles? I didn't use price in the angle calculation.

[Renat Akhtyamov]

Evgeniy Chumakov:

Here are the screenshots I posted above of the increments ranging from -45 to 45.

What's next and the big question is how do you calculate the angles? I didn't use price in the angle calculation.

What did you use?

[Evgeniy Chumakov]

Renat Akhtyamov:
What did you use?

And what is on the chart apart from the price.

[Igor Makanu]

Evgeniy Chumakov:

Here are the screenshots I posted above of the increments ranging from -45 to 45.

What's next and the big question is how do you calculate the angles? I didn't use price in the angle calculation.

i think it does not matter whether it is angles or grapes, here is my old indicator, the first one is drawing the ZigZag from MT4 delivery, but it is drawing on points by calculating the grapes using the formula y=kx+b

and the second one plots this grading, i.e. dy, on the basis of this indicator it is possible to try to analyse the difference of angles (let it be dy), it is better than a standard one from ZigZag delivery and I should take it and twist

[Evgeniy Chumakov]

A couple of pictures of two different pairs from yesterday.

The red and green parallel lines are the dispersion channel. Only I have it straight as a pipe, I mean fixed.