Page MenuHomeFeedback Tracker

ArmA Reforger Ranking System
Feedback, NormalPublic

Description

Game Version number:
1.2.1.189
Modded?: (Yes/No)
No

Issue Description:
Intelligently ranking players according to skill level would have many knock-on benefits for ArmA Reforger. Currently, there is no skill-based system available.

Microsoft released a research paper recently describing their TrueSkill system. This might be a good place to commence

https://www.microsoft.com/en-us/research/project/trueskill-ranking-system/

There is an open source version with a paper here:
https://arxiv.org/html/2401.05451v1

There is also Glicko2 which incorporates volatility:
http://glicko.net/glicko.html

Ranking could help:

  • encourage more competitive play
  • facilitate team balancing
  • motivate players to improve
  • form a basis for tournaments
  • stimulate esports coverage

Details

Severity
Feature
Resolution
Open
Reproducibility
N/A
Operating System
Windows 10 x64
Category
General
Steps To Reproduce

Try and find which is the strongest side when starting a game or try and see who are the most skilled players.
No information is available

Additional Information

Elo Rating System: https://en.wikipedia.org/wiki/Elo_rating_system

https://github.com/topics/trueskill

TrueSkill (Python):
Original Implementation: This is the official Python package implementing the TrueSkill rating system developed by Microsoft Research.
Flexibility: Supports various match rules, including 1 vs. 1, team games, and free-for-all scenarios.
Bayesian Inference: Utilizes Bayesian inference to quantify players' skill points, starting from a Gaussian distribution where
\mu
represents the average skill and
\sigma
represents the confidence in the rating.
Documentation: Comes with detailed documentation and examples for usage, including shortcuts for 1 vs. 1 matches.
License: Open-sourced under BSD, but the TrueSkill™ brand has usage restrictions for commercial projects.

ts-trueskill (TypeScript):
Portability: A TypeScript port of the Python TrueSkill package, making it accessible for JavaScript/TypeScript environments.
Ease of Use: Offers easy-to-use functions for 1 vs. 1 matches and supports team ratings.
Typings: Comes with TypeScript typings for better type checking and IDE support.
License: Licensed under MIT, but references the original BSD license of the Python package.

TrueSkill Through Time (Python):
Temporal Modeling: Corrects biases in traditional skill estimators by modeling the entire history of activities using a single Bayesian network, allowing for better initial estimates and comparability over time.
Scalability: Efficient algorithm that scales to millions of observations quickly.
Multi-Dimensional Skills: Supports estimating skills in different contexts or environments, like different types of tennis surfaces.
Documentation: Offers scientific documentation and is part of a broader initiative including Julia and R packages.

TrueSkill-Calculator (JavaScript):
User Interface: Provides an online calculator for TrueSkill ratings, making it accessible for non-programmers to see how TrueSkill ratings work.
Ease of Use: Allows adding or removing teams, assigning ranks, and weights to simulate different game scenarios.
Historical: Created after the official Microsoft calculator became unavailable, serving as a community alternative.

node-trueskill (JavaScript):
Integration: A JavaScript implementation for Node.js, useful for integrating TrueSkill into web applications or server-side logic.
Compatibility: Aims to mimic the output of the original Microsoft TrueSkill calculator for validation.
Simple API: Provides an easy-to-use API for adjusting player skills based on game outcomes.

Event Timeline

tartan created this task.Fri, Jan 31, 1:50 PM
tartan edited Additional Information. (Show Details)Fri, Jan 31, 2:01 PM
tartan updated the task description. (Show Details)Fri, Jan 31, 2:09 PM
Geez changed the task status from New to Feedback.Fri, Jan 31, 3:06 PM