Skip to content

Calculating the rating for clans!

User Avatar
#44
Auto-translated
AlexeyPank;66299
Here, I suggest the following rating calculation system for clans. Thus, the clan rating (Rklan) consists of two components:
1 - Non-linear component Rnel=∑R(n), where R(n) is the rating of each clan member according to the non-linear system.
2 - Linear component Rlin=∑M(n), where M(n) is the number of points received by each for their place in a tournament. M(n) is calculated via a geometric progression (this is done so that the gap created by stronger players who took higher places is significant. This allows small but strong clans to occupy high places in the rating). Thus, M(n)=M(n+1)*q, where q is the progression multiplier. M(n)min=1, q=2. The maximum number of points for first place is limited by the number of participants, but only those belonging to clans (i.e., if players not belonging to clans participate in the tournament, their place is not counted as existing). The progression multiplier is chosen as 2 as the minimum that yields integer results.
Rklan=(Rnel+Rlin)/2 This will allow strong players to easily outperform the "Chinese," since the rating of one strong player exceeds the sum of the ratings of all those below.
All this is a rough draft; I await crushing criticism.
:) I can't say anything bad yet. We need Kyle to recalculate the rating for all clans again, just to see in reality what the results will be; maybe this time they won't forget to include me in the preliminary calculation :)

PS. I read the thread "Order of the Phoenix challenges to a fight" and came across an interesting post by AmberSolera:

-----------------------------------------------------------------------------------------------------
As for "place or days," it's more customary to consider the place, of course. But it depends on other "outside" participants. Well, for example, I trailed you by only 2 days. But 30 people jumped in between us... So, compare our places? While at the same time you trail the leader by 10 days, but in terms of places, you are right behind him... An imbalance occurs.
--------------------------------------------------------------------------------------

Draw your own conclusions regarding the non-linear and linear parts of the formula for calculating the personal player rating :) The conclusion was made by AmberSolera himself, however, for some reason, he forgot it in the discussion about personal player ratings, even though I was talking about exactly this flaw that the linear part of the formula introduces into the non-linear one :)

And as a result, we see that for this type of Competitive event, they decided to determine the winner by the minimum sum of Walkthrough days without considering places. And in the clan rating, Kyle performed the recalculation once more, but still did not use the oscillator and the bonus coefficient. I personally consider these decisions correct. Further comments on this matter are apparently redundant.
AmberSoler;66326
At that moment, the conversation was not about player ratings. Therefore, there was no other method of evaluation at all. Players did not have a rating as an asset per se. And how else to compare six results with each other if the results of other players were not taken into account by definition?
So what? The conversation was about how to determine the results of several players and your speech was about how the places held create an imbalance, therefore they decided to count by days spent on the Walkthrough and that is sufficient :)
AmberSoler;66326

I hear a bell, but I don't know where it is... A good saying and very appropriate in this case... And our logic hasn't started improving yet... :)
And that's where I see that you hear the bell but don't know where it is. You admitted it yourself honestly :) And your logic hasn't started improving either :)
Kyle

Secondly, I have already explained to you here why the clan rating was calculated according to your system.
Further discussion in this thread will be considered off-topic and deleted
Dear Kyle, I do not intend to argue with you further and even less to off-topic. I only want to know from you why you calculate the PRELIMINARY rating for clans using a formula that is knowingly disapproved of by you, because the Best Of Best rating was posted on the 3rd. You said that the official version would be calculated from the beginning of this year and old tournaments would not count. What then is this preliminary rating for? How do you intend to calculate the official version? Please explain to everyone, because it is not clear. Will you calculate it the same as the preliminary one or differently? Will you preliminarily test the NEW coefficient proposed by AlexPank and the oscillator in the calculations? If YES, then why do you and we need preliminary calculations for clans based only on the non-linear formula?
That is what interests me, and probably other players as well.