I have an answer to your question.
After searching the Internet for an answer to this question, I still did not find exactly what I was looking for. So, my colleagues and I decided to experiment using the Youtube comment system.
First of all, we sorted what we considered popular videos into one section, the average video in another, and less popular in the latter. There were 200 videos in each section, and after several days of study, we began to notice a pattern. We found that you were right about the three required things, but we also went deeper and found an additional variable.
Youtube's comment system depends on four things:
1) The time when it was published,
2) Like / dislike ratio in a comment,
3) Number of responses
4) And, believe it or not, WHO published it.
The average like / dislike ratio of every public comment you've ever posted is embedded in it because (as we expected) they believe that people with a low like / dislike ratio will post comments that many do not like or simply disagree.
There is an algorithm in this, and it is much simpler than you think. Essentially, there are such things that we called "modular points", and you get a certain one based on these four factors. Firstly, here is what you need to know about converting points of a module with TWO factors:
For a like / dislike ratio in a comment, multiply that number by ten.
There are two modular points on the number of answers (NOT from the original author) that are in the commentary.
These are the two main factors that determine the number of modular points in a comment.
For example, if a comment had 27 likes and 8 dislikes, then the ratio would be 3.375. Multiplying by 10, you get 33.75 modular points. Using the following factor, the number of answers, say, this comment has 4 direct answers to it. Multiplying 2 by 4, we get 8. This is the part where you add 8 to cumulative modular points, which gives you a total of 41.75 modular points.
But we are not done yet; this is where it gets complicated.
Using the average like / dislike ratio of the total number of comments they ever posted, we found that the formula added to the module cumulative points is this:
C = MP(R/3) + (MP/10)
where C = Comment Position Variable; MP = Module Points; R = Person total like/dislike ratio
Believe me, we spend DAYS only on this part, which was probably the most frustrating. Even if 3 and 10 in this equation seem random and unnecessary, so far all the comments on which we tested this equation passed the test, but did not pass the test when these two variables were removed. Once this equation is completed, it gives you the number we called the Position Variable .
However, we have not even finished yet, we have not talked about time yet.
In fact, I was very surprised that this part did not take as much time as I expected, but, of course, it was painful to make this equation every time for every comment we tested. At first, during testing, we decided that it was time to overcome the barrier if 2 comments had equal variable positions.
In fact, I almost called it the end of the experiment when it happened, but upon further verification we found that there was still much to be done. We found that some comments were superior to each other with the same variable position, but time seemed random! After several days of verification, here is where the final result comes in:
There is ANOTHER equation that we must find before applying the 4th variable. Using another separate equation, this is what our algebraic conclusions came to:
X = 1/3(S/10 + A) x [absolute value of](A - 3S)
where X = Timing Variable; S = How long ago the video was posted in minutes; A = How long ago the comment was posted in minutes
I wish I came up with this, but, unfortunately, the system is so complicated. There are other mathematical reasons behind the other variables, but they are too complicated to explain, it will probably take at least three paragraphs. We tested this equation with over 150 comments, all of which were validated.
Once you find X , which we called a temporary variable , all you have to do is apply it to the following equation:
N = X(C/4 + 1)
where X = Timing Variable; C = Positioning Variable
N is the answer to all your problems.
This is the final equation, the final answer. A simple conclusion: the higher N, the higher the comment.
Note: Special thanks to my colleagues: David Mattison, Josh Williams, Diego Mendiet, Steven Orsett and Kyle Shropshire. I would never know about it without them and the work that they put into it.