Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Stability weighting of highest strains in mania #2449

Closed
wants to merge 6 commits into from
Closed

Stability weighting of highest strains in mania #2449

wants to merge 6 commits into from

Conversation

Adrriii
Copy link

@Adrriii Adrriii commented Apr 24, 2018

Quote from the discord :

"So I have implemented a stability system that rewards constant difficulty and gives a small penalty to bursts around easier areas.
I am using a Queue where I store the previously calculated strains (unchanged), to which I apply a stability factor, calculated with the average strain of the previous sections.
The stability factor is decreased on overrated or underweighted sections, and increased when on an section that matches the current average difficulty.

The same implementation is possible using a global average instead of a moving average.

This adds a few constants to be able to tweak it :
memory (number of strain sections of 400ms (depending on section length constant) to keep in mind for the moving average)
stabilityStepDecrease (the step when encoutering either a positive or negative stability change)
stabilityStepIncrease
maxStability (aka max bonus for stability)
minStability
overweightFactor (the threshold representing when should we consider something unstable due to a high increase in strain)
underweightFactor"


The constants shown here can (and should) be played with to get and idea of the impact. (Currently, the star rating calculation will not be impacted on the maps library if you don't delete the maps, restart the game, and download them again).

The goal of this change is to normalize the difficulty, instead of rewarding isolated bursts in the map. In mania, combo is not important enough to consider bursts an important part of the play, compared to what a constant difficulty represents in terms of final score

With the current constants, I have been able to give, in my opinion, a better representation for my setlist of test maps, which are:
-Galaxy Collapse
-Aiae
-Blastix Riotz
-The Empress Scream off ver
-Triumph & Regret
-Eternal Drain
-Odin

This change should come along with a modification of strain values calculations, because at this day streams, stamina and reading (the hardest to calculate) aren't representated enough in comparison with jumptrills.

I do not intend to see this specific pull to be merged in the final version because it needs more work, but it can at least be an addition or merely a thought to the improvement of mania star rating calculation.

@peppy peppy added this to the Candidate Issues milestone Apr 25, 2018
@peppy peppy modified the milestone: Candidate Issues Feb 17, 2020
@smoogipoo
Copy link
Contributor

Going to close this as it's stale + general vibes are that this shouldn't go in as-is + WIP by others is changing everything here.

@smoogipoo smoogipoo closed this Dec 14, 2021
@Adrriii
Copy link
Author

Adrriii commented Dec 14, 2021

No probs, haven't been following the calc changes but I hope theres some kind of nerf for bursts :)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

4 participants