How Chess Elo Ratings Work: Expected Score, K-Factor and Updates
A FIDE rating estimates performance relative to a player pool; each result is compared with the score expected from the rating difference and the gap is multiplied by a development coefficient.
Timeline
- Before a rated game: The two ratings determine an expected fractional score under the governing rating table.
- After the result: Subtract expected score from 1 for a win, 0.5 for a draw or 0 for a loss, then apply the player’s K-factor.
- Rating list: FIDE aggregates rated results for the period, rounds the change and publishes an updated monthly list.
An Elo-style chess rating is a relative performance estimate, not a percentage, exam grade or permanent title. A player’s number is meaningful inside the rating pool and rules that produced it. FIDE maintains official over-the-board lists, while national federations and online platforms can use different player pools, formulas and anti-abuse rules, so equal numbers from two systems need not represent equal strength. [1][2][3]
The core idea compares result with expectation. A rating difference maps to an expected fractional score for each player. The actual score is 1 for a win, 0.5 for a draw and 0 for a loss. A result above expectation raises the rating; one below expectation lowers it. Beating a much stronger opponent therefore gains more than beating a much weaker one. [1][2]
FIDE calculates each game’s difference as actual score minus expected score, sums those differences for the rating period and multiplies the total by K, the development coefficient. The result is rounded under the regulations. This means the same win can produce different changes for players with different K values, and a tournament’s net change reflects all rated games rather than one headline upset. [1]
Under the cited standard-rating rules, K is generally 40 for a player new to the list until completing at least 30 games, 20 while a player remains below 2400, and 10 after the published rating has reached 2400. Players through the end of the year of their 18th birthday use K 40 while below 2300. A cap can reduce K in a very busy rating period. [1][3]
FIDE publishes its official list monthly. Only games in properly registered and reported events count, and unplayed games such as forfeits ordinarily do not. FIDE also maintains separate rapid and blitz rating regulations and lists. A player can therefore have different standard, rapid, blitz, national and online ratings without any contradiction. [1][4]
An initial FIDE rating uses rated opponents and the player’s score under the current formula, which includes two hypothetical opponents rated 1800 and drawn games for the calculation. Zero scores in a first event are disregarded under the stated rule, and the initial result is capped. These details are regulation-specific and should not be replaced by a generic online Elo calculator. [1][3]
Use ratings as estimates with uncertainty, especially for new, inactive or rapidly improving players. They predict average results across repeated competition better than one game. When checking a calculation, use the regulation effective for that event, the correct time-control pool, the player’s K-factor and the official pre-event ratings; FIDE rules have changed over time, including the treatment of large rating differences. [1][2][3][4]
Sources
- FIDE Handbook — Rating Regulations
- FIDE — Arpad Elo and the Rating System
- FIDE — New Rating and Title Regulations
- FIDE Handbook — Rapid and Blitz Rating Regulations