Calculating the rating for clans!

If everything is compared only to the leader, the meaning of the other results is lost. If there is one player who will outperform everyone anyway, what is the point of comparing only with them? I already know that I am outclassed by them. But if I outperformed someone based on the tournament results, or they outperformed me, it is much more interesting for me to compare myself to them than to an unreachable leader.

Just look at the actual calculation results for the non-linear part alone. What, can you really not compare your result with other players? I'm surprised :eek: The point is this: each player receives their rating relative to the leader's result. But nothing stops you from then visually comparing the players' ratings AMONG themselves. Moreover, these will be real results without any straightening or distortion :) If the leader finished in 10, you in 900, and someone else in 1000, will you really not see from the rating who was worth what and be unable to compare your result with the player who finished in 1000 days? Furthermore, the non-linear part will show that your rating is significantly closer to the player who took 1000 days and lags far behind the leader — everything is strictly proportional; whoever played for what, received that. And the oscillator is precisely what tries to smooth out these moments. So, what you wrote applies more to the oscillator than to the non-linear system :) Why are you confusing people?

And this is where the non-linear component will say only one thing: we both play worse than the leader. And so?

"And so" what? Explaining is probably useless, BUT I WILL! Yes, the non-linear result will say that you both play worse, because that's how it is. Just please do not claim that in my system you and the player who took 1000 days will have an equal rating. So don't go on about how you can only compare your result with the leader's result :) As confirmation of my words, look at the actual rating calculation results carried out using the non-linear system. Now, the linear component will start to straighten the result obtained by the non-linear system, to say what? That you play ONLY SLIGHTLY :) worse than the leader because you took 2nd and 3rd place (ha-ha three times)? Even though this contradicts the facts expressed actually in Walkthrough days? And what now?

PS. And if we now move to the next tournament, where players finish close to each other, and neither of you is there. Then a player who takes the last place in this tournament, say 10th, but trails the leader (the leader finished in 1 day) by ONLY 10 days, will receive a rating in the final table lower than or equal to YOURS! And who turns out to be the free-loader in this example? This is what you achieve by "improving" my system. This is the kind of distortion we can get from your smoothing of my sharp corners :) To hell with such improvements :)

For the first tournament, the Player who took 1000 days will get 1 point from the non-linear and 50 from the linear. Total -> oscil. = 25.5
In the second tournament, the Player in 10th (10 days vs the leader's 1 day) will get 10 from the non-linear and 0 from the linear — oh, whatever, let's say they get 5. The oscil gives 7.5. This situation clearly shows the flaw related to accounting for placement and the number of players. And you cannot remove it with any bonus! AlexPank - is this how you want players to compare their ratings among themselves? :) As for myself, I've already said where I've seen such rating comparisons.
How will you explain to the player with a rating of 7.5 why a weak player's rating of 25.5 is THREE times higher than theirs, even though that player clearly plays worse? I don't even want to hear nonsense about different maps — we rely on the leaders' results and the players' personal results, even if they are on different maps.

Specifically for Kyle: The example uses two similar tournaments for the three-map set — 1st and 2nd.

Indeed, the non-linear system is simple. But it is the kind of simplicity that is worse than theft, IMHO. That is why suggestions arise to introduce additional parameters: player placement, number of players in the tournament, player's rating in similar tournaments, average tournament rating, etc. This is not a complication of the formula, but its supplementation and development.

Regarding theft, look at the example — how cunningly you snatch part of the rating :) from normal players, as well as for the very last place :)

And the non-linear part itself is, undoubtedly, key, but not the sole determinant.

I absolutely agree — the non-linear part cannot determine that a weak player has a higher rating than a strong player in the final rating table. Your refinements can do this with ease. And if these blunders are not clearly visible, it doesn't mean they aren't there in the real calculations :)

... the non-linear component is the basic foundation of the rating.

If you have no constructive suggestions other than primitively "dividing one by the other",

This most primitive "dividing one by the other" was called by you above: by AlexPank as , and by you as :) Remove this part from the formula and you will be left with one big BLUNDER :) Although you've already "perfected" it so much that in your system it is neither key nor basic :)
Therefore, my constructive proposal is that Carthage be destroyed :)