Glicko2

The Glicko-2 system is a rating system designed to measure the skill level of players in games of skill. It was developed by Mark Glickman and is an update to his earlier Glicko rating system. Glicko-2 aims to provide a more accurate and robust estimation of player skill compared to its predecessor.

A key feature of Glicko-2 is its consideration of rating deviation (RD). RD represents the uncertainty in a player's rating. A low RD indicates high confidence in the rating, while a high RD suggests that the rating is less reliable, often due to infrequent play. The system also incorporates rating volatility, which measures how much a player's skill is expected to change over time.

The Glicko-2 algorithm updates a player's rating, rating deviation, and rating volatility after each game or set of games. When a player wins, their rating increases, and their RD generally decreases, indicating greater certainty in their improved skill. Conversely, a loss leads to a rating decrease and a potential RD increase. The volatility parameter adjusts based on the consistency of a player's performance.

Glicko-2 is widely used in online gaming platforms and competitive environments to provide dynamic and accurate skill assessments. Its ability to account for both rating uncertainty and skill fluctuation makes it a sophisticated tool for ranking players.