Skip to content

Latest commit

 

History

History
288 lines (161 loc) · 7.98 KB

CHANGELOG.rst

File metadata and controls

288 lines (161 loc) · 7.98 KB
Raw

All notable changes to this openskill.py will be documented in this file.

This file is updated every release since v1.0.0 with the use of towncrier from the fragments found under changes/

Openskill 6.0.2 (2024-12-25)

Bugfixes

  • Update Dependencies + Add GitHub as Trusted Publisher (#152)

Openskill 6.0.1 (2024-11-02)

Bugfixes

  • Fix rate mutating rating objects when limit_sigma enabled. (#141)
  • Use correct constants in Thurstone-Mosteller models. (#148)
  • Scale beta**2 by 2 instead of by total player count. (#148)

Openskill 6.0.0 (2024-07-15)

Breaking Changes

  • Prediction methods better account for uncertainty. (#138)

Features

  • Support Per Player Weights and Balance Flag. (#138)
  • Prediction methods have faster runtime. (#138)

Bugfixes

  • Fix issue causing limit_sigma not being set implicitly. (#132)

Openskill 5.1.1 (2024-02-03)

Bugfixes

  • Resolves incorrect predictions for matches of size greater than 2. (#124)

Openskill 5.1.0 (2023-08-24)

Features

  • Optimize rating objects for deepcopy (#108)

Bugfixes

  • model.rating() methods no longer revert to default model parameters when arguments are 0 (#109)

Openskill 5.0.2 (2023-08-13)

Documentation Improvements

  • Improve SEO of Documentation (#107)

Openskill 5.0.1 (2023-08-12)

Bugfixes

Documentation Improvements

  • Integrates Algolia Docsearch to Sphinx Docs (#105)

Openskill 5.0.0 (2023-07-21)

Breaking Changes

  • All top level functions are now methods that must be called from model once it's been initialized. (#101)

Features

  • Add PEP-517 and PEP-518 compliance. (#82)
  • Everything is strictly type hinted to let you use your IDE's autocomplete features. (#101)

Documentation Improvements

  • All functions, methods and classes have docstrings now. There are also LaTeX equations in docstrings when necessary. (#101)

Openskill 4.0.0 (2022-12-11)

Breaking Changes

  • Drop support for Python 3.7 (#80)

Features

  • Add support for PyPy 3.8 and 3.9 (#80)
  • Drop scipy as dependency (#80)

Documentation Improvements

  • Improve SEO of Documentation (#80)

Openskill 3.3.0 (2022-12-06)

Documentation Improvements

  • Improve SEO of documentation.

Openskill 3.2.0 (2022-12-06)

Features

  • Predict ranks and their odds of entire match outcome. (#74)

Openskill 3.1.0 (2022-12-03)

Documentation Improvements

  • Add more details to the documentation, fixes broken links and point a custom domain name to the docs. (#73)

Openskill 3.0.0 (2022-11-30)

Breaking Changes

  • Modify default tau value to 25/300 #61 (#61)

Openskill 2.5.1 (2022-11-11)

Bugfixes

  • Allow setting mu and sigma to 0 for Rating objects. (#60)

Openskill 2.5.0 (2022-10-26)

Features

  • Support Python 3.11.0 Officially (#56)

Bugfixes

  • Fixes issue where equal ranks below zero don't draw (#54)

Openskill 2.4.0 (2022-06-08)

Features

  • Add more comparison magic methods to the Rating object.

Documentation Improvements

  • Add documentation about advanced usage.
  • Add documentation about future update to the default value of tau.

Openskill 2.3.0 (2022-05-14)

Features

  • Add support for python 3.7+ (#52)

Openskill 2.2.0 (2022-03-18)

Features

  • tau (defaults to 0): Additive dynamics factor, which keeps a player's rating from getting stuck at a level. Normally, a player's sigma will only decrease as we gain more information about their performance. This option will put some pressure on this back up. This default will change to be sigma/100 with v3, to be more congruent with TrueSkill, but higher may make your rating system more exciting. (#50)
  • prevent_sigma_increase (defaults to False): for a tau > 0, it is possible that a player could play someone with a low enough rating that even if they win, their ordinal rating will still go down slightly. If your players have no agency in matchmaking, it is not desirable to have a situation where a player goes down on the leaderboard even though they win. (#50)

Openskill 2.1.0 (2022-03-10)

Features

  • Faster runtime of predict_win and predict_draw. (#48)

Openskill 2.0.0 (2022-02-20)

Breaking Changes

  • The rate function now returns Rating objects. (#46)
  • Changes ordinal to accept both Rating objects and lists or tuples of 2 floats. (#46)

Features

  • Add a function to predict draws. (#45)
  • create_rating now checks if the argument is the correct type. (#46)

Openskill 1.0.2 (2022-02-09)

Features

  • Updates scipy to 1.8.0 (#37)

Openskill 1.0.1 (2022-02-04)

Features

  • Update development status to "Stable" (#34)

Openskill 1.0.0 (2022-02-04)

Features

  • Capability to predict winners of match given a set of teams. (#27)