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

RFC: increase initial Glicko deviation #7309

Merged
merged 3 commits into from
Sep 21, 2020
Merged

RFC: increase initial Glicko deviation #7309

merged 3 commits into from
Sep 21, 2020

Conversation

ornicar
Copy link
Collaborator

@ornicar ornicar commented Sep 13, 2020

As to figure out the rating of a new player quicker.

The intended result is for noobs to lose fewer games
before they start encountering relevant opponents.

for #7273

Another thing to consider is hiding ratings
until a given number of games has been played,
so people don't get emotionally attached to
their meaningless provisional rating.

As to figure out the rating of a new player quicker.

The intended result is for noobs to lose fewer games
before they start encountering relevant opponents.

for #7273

Another thing to consider is hiding ratings
until a given number of games has been played,
so people don't get emotionally attached to
their meaningless provisional rating.
@niklasf
Copy link
Member

niklasf commented Sep 17, 2020

Should also change maxDeviation accordingly. In fact the initial deviation should probably always be the same as maxDeviation.

I tried to find justifications for the default value:

If the player is unrated, set the rating to 1500 and the RD to 350. These are
default reasonable choices, but the RD of 350 in particular can be determined
through optimizing predictability of game outcomes (not described here).
-- http://www.glicko.net/glicko/glicko.pdf

Doing the optimization on 5 million Lichess game results (using the technique from #6818 (comment)) yields:

image

This is without touching volatility.

@niklasf
Copy link
Member

niklasf commented Sep 17, 2020

Looks like the volatility change goes also in the right direction. Just a few samples:

image

* master:
  scalafmt 2.7.1
  dgt play style tweaks
  rewrite dgt play help zone
  Ready to start testing on Lichess.dev
  Finish porting to chessops and removed all other dependencies
  rm outdated nginx docs
  rm outdated iptables doc (no longer using extra ports)
  rm outdated manifest doc that now has its own route
  update some old links
  Change timeout draw separator
  New Crowdin updates (#7307)
  Migration from axios to fetch almost done
@ornicar ornicar merged commit a229913 into master Sep 21, 2020
@niklasf niklasf deleted the initial-glicko branch October 29, 2024 19:39
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

Successfully merging this pull request may close these issues.

2 participants