Discussion of article "Patterns with Examples (Part I): Multiple Top"

 

New article Patterns with Examples (Part I): Multiple Top has been published:

This is the first article in a series related to reversal patterns in the framework of algorithmic trading. We will begin with the most interesting pattern family, which originate from the Double Top and Double Bottom patterns.

Regarding the topic of the article, I will try to draw a diagram as a tree of patterns that starts from a double top. This will assist in understanding how broad the possibilities of this concept are:

Tree

I decided to combine the concept of several patterns with the assumption that they are based on approximately the same idea. This idea has a simple beginning - find a good movement in any direction and correctly determine the place where it is supposed to reverse. After visual contact with the proposed pattern, the trader should correctly draw some auxiliary lines, which should assist in evaluating whether the pattern meets certain criteria as well as in determining the market entry point along with the target and stop loss level. Take profit can be used here instead of the target.

Patterns can have some common construction principles, based on which the concept of these patterns can be combined. Such clear definition is what differs algorithmic traders from manual traders. Uncertainty and multiple interpretation of the same principles may lead to disappointing consequences.

The basic patters are as follows:

  1. Double Top
  2. Triple Top
  3. Head and Shoulders

Author: Evgeniy Ilin

 

Multiple tops? All patterns have been zigzagged long ago.

I looked at the beginning of the class header and got a convincing feeling of incompetence and meaninglessness of this code.

The root class should be interface = universal, but immediately a separate field for Head-Shoulder Excess, some kind of innovative approach? Like more articles, albeit full of hat, just so long as the views and traffic, like social networks?

 

Ban for a week for aggressive behaviour (post trimmed)

Be mutually polite, please.

 
Aleksey Mavrin:

Multiple tops? All patterns have been zigzagged for a long time.

I looked at the beginning of the class header and got a convincing feeling of incompetence and senselessness of this code.

The root class should be interface = universal, but immediately a separate field for Head-Shoulder Excess, some kind of innovative approach? Like more articles, albeit full of hat, just so long as the views and traffic, like social networks?

Well offer something of your own, what are you wasting your efforts? Write an article about zigzags and how they describe all the patterns, make a bunch of interfaces and we will all admire them and praise you in the comments.

 
Evgeniy Ilin:

Well, offer something of your own, why are you wasting your efforts? Write an article about zigzags and how they describe all patterns, make a bunch of interfaces and we will all admire them and praise you in the comments.

I also noticed that you didn't refer to any other works like this one. This apparently means that you have not studied the experience in this area.

That's why I understand where your position "I've done something, and you criticise" comes from.

I have studied the experience of others, and did not see much point in writing more articles (even if I wanted to write something), because almost everything was already there, just compiled and modified for specific tasks.

I didn't see any novelty in your work. It looks like "one of 1000 ways to cook potatoes". Works that structure and show all 1000 at once are valuable, I hope you will understand.

 
Aleksey Mavrin:

I also noticed that you did not refer to any other works like this one. This probably means that you have not studied the experience in this field.

So I understand where your position "I have done something, and you criticise" comes from.

I have studied the experience of others, and did not see much point in writing more articles (even if I wanted to write something), because almost everything was already there, just compiled and modified for specific tasks.

I didn't see any novelty in your work. It looks like "one of 1000 ways to cook potatoes". Works that structure and show all 1000 at once are valuable, I hope you will understand.

You may be right about something. Articles are written for different purposes. But just because you didn't see something new for yourself doesn't mean that someone else won't. I understand that you have a lot of experience in freelancing and in particular as a performer you are excellent. An article is still a little different. Not everyone is so savvy and I proceed from the fact that I want to show something as accessible as possible. I've written articles with a lot of maths, and experience shows that not everyone gets into it, or to be more precise, people just don't write because it's difficult. Sometimes you need something simpler. The article should not always be about "look, I made a grail". The material should be easily digestible. What good will it do if I make a huge mega article with super maths that 2-3 people will understand? There will be such an article, but that's not why articles are written in the first place. It's also a question of purpose. You should not write an article with the purpose of somehow boasting or showing off, an article is for the reader, the more people understand it the better. You should also understand this truth and take it into account. You'd better write in private and I would explain to you what and why. And so you should realise that it is not very nice to write in a public thread like this, here is a ban for a week for you to think about it

 
Evgeniy Ilin:

You may be right about something. Articles are written for different purposes, but just because you didn't see something new for yourself doesn't mean that someone else won't. I understand that you have a lot of experience in freelancing and in particular as a performer you are excellent. An article is still a little different. Not everyone is so savvy and I proceed from the fact that I want to show something as accessible as possible. I've written articles with a lot of maths, and experience shows that not everyone gets into it, or to be more precise, people just don't write because it's difficult. Sometimes you need something simpler. The article should not always be about "look, I made a grail". The material should be easily digestible. What good will it do if I make a huge mega article with super maths that 2-3 people will understand? There will be such an article, but that's not why articles are written in the first place. It's also a question of purpose. You should not write an article with the purpose of somehow boasting or showing off, an article is for the reader, the more people understand it the better. You should also understand this truth and take it into account. You'd better write in private and I would explain to you what and why. And so you should realise that it is not very nice to write in a public thread like this, here is a ban for a week for you to think about it

You may be right in something too, at least probably so - most people need something to work, and thoughtfulness and structured code and common approach are tenth or unfamiliar words at all.

But I remain in my opinion, or simply - to do either well or not at all. I can't help noting that there are articles of higher quality and that are quite clearly written.

Nothing personal as you understand, I wish you success and write more, just try harder ;) A link to such works is definitely necessary.

Sorry for the emotionality, your article was the second one in a row. And I do not think that this is something that should be written in a private, not in a general discussion.

 

I consider this article useful for beginners - it is colourful, sequential, and generally interesting.

Of course, it is possible to search for these TA figures in a simpler way - search for fractals in a loop with a counter of their number (the algorithm, for example, is implemented in "Chuvashov's Fork").

But this, as they say, is a matter of taste.

The main value of such figures is that they form a strong resistance belt for the trend, which is an important preliminary signal.

 
Aleksandr Masterskikh:

I find the article useful for beginners - designed colourfully, consistently, generally interesting.

Of course, these TA figures can be searched in a simpler way - search for fractals in a loop with a counter of their number (the algorithm, for example, is implemented in "Chuvashov's Fork").

But this, as they say, is a matter of taste.

The main value of such figures is that they form a strong resistance belt for the trend, which is an important preliminary signal.

Thank you, it's for beginners. Yes, for these reasons I chose this family of patterns. About fractals, yes, in fact fractals are almost the same as mine, but there is a slightly different logic of finding the tops in my opinion, these walls falling and growing should be left and right adjacent to the central bar, which actually should be the top, and I am allowed some noise between the walls. Well, here you can experiment with fractals, but it is necessary to use built-in functions, and I do not even remember if there is such a function, or iCustom and drag the indicator behind it, and if you want to write the code on one terminal and the second terminal does not have such a function, and you want to minimally deal with it, it is better to make your own logic, in this case everything will work the same. It is still necessary to connect levels and refine the code to find the whole family of these patterns, and then you can check it in the tester. I think I will do it in the next article, I am working on a very large article on probability theory. I will probably continue after it.

 

Hello,

Can using fuzzy logic for these patterns improve the detection accuracy of go patterns and simplify the code?

 
Aleksey Mavrin:

Multiple tops? All patterns have been zigzagged for a long time.

I looked at the beginning of the class header and got a convincing feeling of incompetence and senselessness of this code.

The root class should be interface = universal, but immediately a separate field for Head-Shoulder Excess, some kind of innovative approach? Like more articles, albeit full of hat, just so long as the views and traffic, like social networks?

I agree!

But!

Many different good articles are published here on the forum, but there is no evaluation from the financial point of view, i.e. economics.
Let's say, it's great that there is auto-detection of figures: double top, triple top, head-shoulders.
But where is the evaluation of their application on 5 minutes, 30 minutes or on older timeframes, after all, this is a traders' forum.

This website uses cookies. Learn more about our Cookies Policy.