Score Margins (#165)

* Introduce score margins or margin of victory

* Update documentation

* Add changelog fragment

Author: vivekjoshy

.github/workflows/main.yml

@@ -80,8 +80,10 @@ jobs:
             cache-dependency-path: "**/pyproject.toml"
         - name: Install Tox
           run: python -m pip install tox-gh-actions>=3.2.0
+          continue-on-error: ${{ matrix.experimental }}
         - name: Setup Test Suite
           run: tox -vv --notest
+          continue-on-error: ${{ matrix.experimental }}
         - name: Run Tests
           run: |
               python -m pip install --upgrade pip
@@ -92,6 +94,7 @@ jobs:
           with:
             token: ${{ secrets.CODECOV_TOKEN }}
             fail_ci_if_error: ${{ !matrix.experimental }}
+          continue-on-error: ${{ matrix.experimental }}
   pypy:
     if:
       github.event_name == 'push' || github.event.pull_request.head.repo.full_name !=

CODE_OF_CONDUCT.md

@@ -37,9 +37,9 @@ We will not tolerate any behavior that:
 
 ### Reporting Violations
 
-If you experience or witness a violation of this Code of Conduct, please report it **privately** using the following form:
+If you experience or witness a violation of this Code of Conduct, please report it **privately** to this email:
 
-https://openskill.youtrack.cloud/form/f2b97d40-c27d-4617-9bea-7f2152a1db1f
+[coc@openskill.me](mailto:coc@openskill.com)
 
 All reports will be taken seriously and investigated promptly. Your identity will be kept confidential unless you choose to disclose it.
 

README.md

@@ -67,7 +67,7 @@ graph TD
     GF2[Multifaction]
     GF3[Asymmetric Faction Size]
     GF4[Predict Win, Draw, Rank]
-    GF5[Per Player Weights]
+    GF5[Weights, Score Margins]
     end
 
     subgraph Technical["Technical"]

benchmark/README.md

@@ -1,11 +1,12 @@
 # Benchmark Instructions
 
-Simply run the respective Jupyter Notebook files with a compatible Python version and choose the options.
+Run the respective Jupyter Notebook files with a compatible Python version and choose the options.
 
 ## Available Benchmarks
 
 ### Local
 
+- ``benchmark.py``: Performs k-fold cross-validation to compare models.
 - ``win.ipynb``: Compares win performance against TrueSkill.
 - ``draw.ipynb``: Predicts draws on standard chess matches.
 - ``rank.ipynb``: Uses rank prediction on a large dataset.
@@ -14,4 +15,4 @@ Simply run the respective Jupyter Notebook files with a compatible Python versio
 
 - [OpenSkill - Weights Benchmark](https://www.kaggle.com/code/daegontaven/openskill-weights-benchmark): A benchmark on CS:GO matches using the weights feature.
 
-Note that the rank benchmark downloads a large file (around 1 GB) and requires slightly more RAM if running locally.
+Note: The rank benchmark downloads a large file (around 1 GB) and requires slightly more RAM (>12GB) if running locally.