How do chess ratings work?
Chess uses the Elo rating system, devised by the physicist Arpad Elo. A rating is a number that estimates playing strength, and it moves after every game based on the result and the strength of the opponent.
The core idea is that the rating difference between two players predicts the expected result. A 200-point gap implies the stronger player should score about 76% across many games. A 400-point gap implies about 91%.
After a game, each player's rating changes by the difference between what they actually scored and what they were expected to score, multiplied by a factor called K. Beat someone rated far above you and you gain a lot, because the result was unexpected. Beat someone far below you and you gain almost nothing, because you were supposed to.
K is larger for new and junior players, so their ratings move quickly toward their real strength, and smaller for established players, so a single bad day does not swing a settled rating.
The scale has rough conventional bands: below 1200 beginner, 1200–1600 casual club level, 1600–2000 strong club, 2000–2200 expert, 2200+ candidate master, 2500+ grandmaster level. Titles such as Grandmaster are awarded by FIDE and require both rating and specific tournament performances, so a title is permanent while a rating fluctuates.
Ratings from different pools are not comparable. Online ratings on chess sites generally run higher than FIDE ratings, and each site's pool is separate.
Elo is not chess-specific. Variants of it are used in tennis, football rankings, competitive video games and many other rated systems.