Player rating for tournaments across all installments of Heroes.

What if the oscillator doesn't start from zero (as already suggested)? Then the underdogs might actually benefit. I suggested starting the linear part from 5 points; then, there wouldn't be such a disadvantage for the underdogs (especially with a large number of participants). So, the oscillator isn't unfair; we just need to figure out what value to start from (I believe it shouldn't be 0, because the non-linear component isn't zero).
So, there's no need to blame the oscillator; it just needs to be refined.

Why not start from 1000? AmberSoler claims that all numbers should be strictly justified. There's no need to refine anything :) The oscillator isn't unfair :) And the non-linear rating doesn't unfairly penalize anyone – everyone gets what they deserve. What you earn is what you get. I agree that the non-linear part is suitable for the Arena. The more people there are, the harder it is to win, meaning you have to personally defeat a certain number of players to take a certain place. But in offline tournaments, considering the number of players in the tournament is pointless because they don't interfere with each other's map completion. The number of days is enough to understand how well each player performed. Examples have already been given; what's the point of a player outperforming many players and taking second place if all the players performed poorly? Therefore, they can only realistically judge their result in relation to the leader's completion time in days. The leader is given acceptable limits for speedrunning. So, the meaning of your refinements is unclear to me. We can clearly see how the results of the underdogs are distorted; you want to start counting from a non-zero value. But so what? The distortion of the results in the middle of the table isn't as noticeable, but it exists (see examples where a weak player takes second place but spends an absurd number of days completing the map!).
Why refine and straighten out these results, especially since, in the described situation, straightening them out actually distorts them? The player in second place will still get a bonus, even though they performed poorly, which is evident from the leader's result. The non-linear system doesn't suffer from such glitches. Even the VDV system is better than the oscillator because it gives at least a fixed number of points for each placement without distorting that result with bonuses, although the non-linear system is still better because the issue with absurd results remains.