Player rating for tournaments across all installments of Heroes.

to IronAxe: You mentioned something here earlier.
And what do we get in the end: four people (AmberSoler, AlexeyPank, Sage and I) expressed the opinion that the tournament rating should be calculated using an oscillator, they all argued their point of view, but you simply don't want to hear them and take it into account... because your system is supposedly better a priori! It's not right to act like this...

The thing is - I gave examples with numbers where the oscillator specifically distorts the player's strength rating, artificially increasing it for some players (leaders) and decreasing it for others (outsiders). With a non-linear system, everyone gets what they earned in terms of days and how close they got to the leader. Then I provided results of calculating two abstract tournaments, where this linear distortion shows strange results, when a player from the second tournament takes much longer in terms of days, but by taking second place, he surpasses a player from the first tournament who plays better, as can be seen from his results. Conclusion: I showed you a situation where the linear system works incorrectly, and the non-linear one works correctly. No one has shown me another situation where it would be the other way around, either in numbers or otherwise. In the discussion, one obvious flaw of the non-linear system was revealed, but it also exists in the linear system. I introduced acceptable limits and slightly smoothed out this situation, but this drawback remained in the linear system and even in the oscillator. I am not forcing you to calculate the ranking in Kanighem's Tournaments using my system – calculate it however you want, and you can use the non-linear part of the formula that I proposed as you wish. In VDV's 3rd Heroes Tournaments, I will keep my own rating. This is agreed with VDV and Prophet. I believe that there are no such flaws in this calculation as with the oscillator, and this rating more realistically reflects the playing strength of players in Heroes. Since I did not see any corresponding examples in the figures comparing the non-linear system with the oscillator in favor of the oscillator.

Further about aging and current rating: I have already expressed myself on this issue, and AmberSoler too. But I repeat: combining a "player result" and his "activity" into one number is not the right solution! Or rather, it’s completely wrong. Why? I answered this question here. Therefore, we need to abandon the current rating in favor of, for example, an "accumulative annual rating" or something else, but it must be limited in time!

Aging reflects not only activity. If a player played worse compared to his previous game, and his tournament rating cannot cover the aging, then his rating may also decrease slightly - which is also correct. The rating has the property that it does not always increase, but can also decrease depending on the shown game and results. Missing tournaments should also not be encouraged. But I went halfway and agreed to age the rating by 10% instead of 33% in regular tournaments. In super tournaments, I think it should be aged by 1/3.
You can take on keeping an accumulative rating or any other you like. I undertook to calculate the current rating because, in my opinion, it most realistically reflects the current playing strength of a player on this resource. I repeat once again that the cumulative system will work on Kukle and Grails; why count the rating there as well? But I'm not forbidding anyone who wants to do so. Personally, I don't need the same balls, just from different sides. I hope I have clearly expressed my position. Understand my idea, and I will understand yours too.

Added after 31 minutes

to IronAxe: You mentioned something here earlier.
And what do we get in the end: four people (AmberSoler, AlexeyPank, Sage and I) expressed the opinion that the tournament rating should be calculated using an oscillator, they all argued their point of view, but you simply don't want to hear them and take it into account... because your system is supposedly better a priori! It's not right to act like this...

Regarding the spontaneously formed vote here, I can only answer one thing. Even if all four of you say that the oscillator is better a priori, but cannot refute the flawed examples I provided, then this means nothing. Moreover, Sage has not appeared for a long time and did not express his final opinion, and AlexPank seems to doubt it - since he plans to calculate Kanighem's Tournament using both systems. Only AmberSoler and you clearly spoke out in favor of the oscillator, without refuting the incorrect operation of the oscillator in the examples I provided.

I have come to a final conclusion for myself, which sounds approximately like this: The specifics of playing Heroes at speed involve non-linear results, expressed in days passed, which more realistically shows the playing strength of each player. Places are always linear and introduce strong distortions. A player in second place, having taken 1000 days, will receive a rating of at least 50 :eek: No, second place shouldn't be worth so much for a player who plays like that!