A standalone application for managing a set of players given a history of games.
The Glicko rating system and Glicko-2 rating system are methods for assessing a player's strength in games of skill, such as chess and go.
The Glicko-2 rating system is intended to estimate the skill of player performance in head-to-head competition in a variety of games.
Description of Glicko-2 system
For more details, please, look at the Mark Glickman site and the example of the Glicko-2 system
Team implementation is based off of information from rhetoricstudios.com paper
Java must be installed on the system. Scala is required for building.
sbt assembly will build the app into ./target...
build.sh will compile the app and move it to the glicko2team directory
- Add matches to match-history.csv
- Configure using glicko2team.properties
## pass in the properties file
sh ./buildAndRun.sh glicko2team.properties
## OR the script will look for it in the same directory as the script.
sh ./buildAndRun.shThe app is configured using a properties file. By default it is glicko2team.properties.
logback.path
the logback configuration
matchSource.path
Path to the match history csv file
matchSource.csv.dateFormat
Date format of date/time for match times in the matchSource csv (uses java formatting syntax)
See SimpleDateformat
glicko2.team.ratingPeriod.games
Rating period can be either in days or number of games the number of games within a rating period.
player settings
player.default.rating=1500
player.default.deviation=350
player.default.volatility=0.06
glicko2.team.algorithm
the glicko2 algorithm for team evaluation.
COMPOSITE (default)
The Composite Opponent Update Method considers each outcome as a match against asingle player possessing the average rating and deviation of the opposing team players. Ineffect, the update method creates a composite player out of the opposing team and uses thisplayer’s resulting rating and deviation when updating a player.8
For example, if Alice competes against a team consisting of Bob, Carol, Dan, and Eve, the game will update Alice’s rating by considering the outcome as a 1versus1 match againsta composite player having the average rating and deviation of Bob, Carol, Dan, and Eve.
INDIVIDUAL
The Individual Update Method takes the naive approach to updating players after a game. This method treats each match outcome as a set of wins or losses, one for each member of the opposing team. This method takes advantage of rating systems like Elo or Glicko by abstracting a team match as a set of 1versus1 matches.
For example, consider a game where Alice participates on a team versus an opposing team consisting of Bob, Carol, Dan, and Eve. If Alice’s team wins, the game updates Alice’srating by considering this game as four wins: one against Bob, one against Carol, one againstDan, and one against Eve.
This software builds upon the work of Andriy Kuba who developed the initial scala-Glicko2 implementation that is used within within this application.