Elo rating system

Definition The Elo rating system is a method for calculating the relative skill levels of players in competitor-versus-competitor games such as chess. It is a zero-sum system, meaning points gained by one player are lost by another, aiming to assign a numerical rating that reflects a player's performance relative to other players within a given pool.

Overview The Elo system assigns a numerical rating to each player, which is then adjusted after every game based on the game's outcome (win, loss, or draw) and the difference in ratings between the two competitors. If a high-rated player defeats a low-rated player, the rating change will be small. However, if the lower-rated player wins, their rating will increase significantly, and the higher-rated player's rating will decrease substantially. The system's primary goal is to predict the outcome of future games and to provide a dynamic measure of a player's current skill level. While most famously used in chess, variations of the Elo system have been adopted across a wide range of competitive activities, including esports, various sports, and online matchmaking systems.

Etymology/Origin The system was invented by Arpad Elo (1903–1992), a Hungarian-American physics professor and master chess player. Elo developed the system in the late 1950s and early 1960s as a statistical improvement over the existing Harkness system used by the United States Chess Federation (USCF). He published his definitive work, "The Rating of Chessplayers, Past and Present," in 1978. The USCF adopted the Elo system in 1960, and it was subsequently adopted by the Fédération Internationale des Échecs (FIDE), the world governing body of chess, in 1970, replacing their earlier rating systems.

Characteristics

• Mathematical Model: The Elo system uses a logistic distribution model to calculate the probability of a player winning against another. The expected score for each player in a game is derived from the difference in their ratings, represented by the formula Ea = 1 / (1 + 10^((Rb - Ra)/400)) and Eb = 1 / (1 + 10^((Ra - Rb)/400)), where Ra and Rb are the respective ratings of players A and B.
• Rating Adjustment Formula: After a game, a player's new rating (R') is calculated using the formula R' = R + K * (S - E), where:
  • R is the player's current rating.
  • K is the K-factor, a constant that determines the maximum possible rating change for a single game. Higher K-factors lead to more volatile rating changes, typically used for newer players or those with lower ratings, while lower K-factors are used for established or highly rated players (e.g., FIDE uses K=40 for new players, K=20 for players below 2400, and K=10 for players 2400 and above).
  • S is the actual score for the game (1 for a win, 0.5 for a draw, 0 for a loss).
  • E is the expected score for the game, as calculated by the logistic model.
• Zero-Sum Principle: In a two-player game, the points gained by one player are exactly equal to the points lost by the other, ensuring that the total sum of ratings within a closed system remains relatively stable (excluding points entering or leaving the system due to new players or player retirement).
• Predictive Power: The system's strength lies in its ability to estimate the probability of one player defeating another, allowing for fair matchmaking and competitive balance.
• Rating Inflation/Deflation: Over time, the average rating within a system can drift due to factors like the introduction of new players at a base rating, players retiring, or changes in K-factors, leading to observed rating inflation or deflation in some competitive environments.

Related Topics

• Glicko Rating System: A more sophisticated rating system developed by Mark Glickman that builds upon Elo by incorporating a "rating deviation" (RD) and "rating volatility" to provide a more accurate measure of a player's uncertainty and consistency.
• TrueSkill: A Bayesian skill rating system developed by Microsoft Research for competitive multiplayer video games, also based on the Elo model but designed for multi-player games and incorporating skill uncertainty.
• Chess: The original and most prominent application of the Elo rating system.
• Esports: Widely adopted in professional gaming for player rankings and matchmaking in titles like League of Legends, Dota 2, Overwatch, and Counter-Strike: Global Offensive.
• Matchmaking Systems: Elo is a fundamental component of many online matchmaking algorithms, aiming to create balanced and enjoyable matches by pitting players of similar skill levels against each other.