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

Elo punishment for parties with large skill gaps #145

Open
GODrums opened this issue Jun 3, 2023 · 2 comments
Open

Elo punishment for parties with large skill gaps #145

GODrums opened this issue Jun 3, 2023 · 2 comments

Comments

@GODrums
Copy link

GODrums commented Jun 3, 2023

The new elo gain/loss feature has a big issue unfortunately, which is the elo punishment in parties with large skill gaps, as this makes the displayed numbers extremly inaccurate.
I would love to see a revised version of the feature including this punishment in the next version. If you accept contributions, I could also work on this request myself.

Details

In my experience, elo punishment p influences the gain/loss d in the following way:

  • if everyone in the party is lvl10, there is no punishment -> p = 0
  • if the maximum elo difference is less than 1300, there is no punishment -> p = 0
  • if the maximum elo difference is between 1300 and 1600 -> p = 7
  • if the maximum elo difference is between 1600 and 2000 -> p = 14
  • if the maximum elo difference is between 2000 and 3000 -> p = 21
  • if the maximum elo difference is larger than 3000, there will be no gain/loss -> p = d

So the correct formula to calculate the gain/loss after the punishment would be similar to:
d - (d / |d|) * p

@timche
Copy link
Collaborator

timche commented Jun 3, 2023

Great find! This is amazing stuff. I'd love to work with you on this. Please add me on Discord if you can! timche#2098

@GODrums
Copy link
Author

GODrums commented Jun 4, 2023

Thanks, I added you (Rums#9989).
I did a bit more research on that topic and collected some sample data, seems like the algorithm is a little more complex than initially thought, but we should be able to make accurate predictions.

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

No branches or pull requests

2 participants