Glicko rating
A Glicko rating is a rating that comes with an uncertainty value, the rating deviation, so that a player with few or old results can move faster than one with a long recent record.
Also called: Glicko
Open a Lichess profile and the number you see is a Glicko-2 rating. New accounts start at 1500 with a wide uncertainty, and that uncertainty is what the system adds to a plain rating. It is called the rating deviation, or RD. A rating plus or minus twice the RD gives a rough 95 percent interval for true strength, so 1600 with an RD of 80 points to a strength somewhere between about 1440 and 1760.
Two consequences follow. A large RD means every result moves the rating a lot, while a small RD means the system already trusts the number and changes it gently. And RD is not fixed: it shrinks as you play and creeps back up during long absences. Lichess marks a rating provisional with a question mark when its RD is above 110.
Mark Glickman designed the original Glicko system, and Glicko-2 adds a third quantity, volatility, which tracks how erratic a player's results have been.
FIDE uses neither. It applies a K-factor set by status and carries no uncertainty value, and US Chess publishes its own separate method, also with a K that falls as games accumulate. A Glicko rating and a FIDE or US Chess rating therefore rest on different assumptions and different pools of players, so never compare them point for point. See the entries on K-factor, provisional rating and online rating.
A neighboring idea is Elo rating. An Elo rating is a number that estimates playing strength from results: the gap between two ratings predicts the expected score, and wins and losses move each player's rating up or down.
How "Glicko rating" applies to your own results depends on where you stand now, and the improvement plans by rating band show what players near your rating work on.
Related terms
The K-factor is the multiplier in an Elo-style rating update that sets how many points a rating can move after a result; a larger K makes the rating react faster.
A provisional rating is an early estimate of strength built from only a few games, which is allowed to move quickly until enough results exist to make it reliable.
An online rating is the number a website or app computes from games played on that platform; it belongs to that site's player pool and is separate from FIDE and national over-the-board ratings.
An Elo rating is a number that estimates playing strength from results: the gap between two ratings predicts the expected score, and wins and losses move each player's rating up or down.
A rating system is a method that turns game results into a number for each player, rising after wins against strong opponents and falling after losses, so that strength can be compared.
Guides that go further
What a rating measures, how FIDE, Lichess, US Chess and chess.com compute theirs, and why the numbers differ.
Average chess rating and a good chess rating
What share of active Lichess players are rated above 800, 1200, 1500, 1800 or 2000, by time control.
Work out how a FIDE rating changes after a run of games.
Sources
- Lichess FAQ, FAQ answer on ratings, read Oct 21, 2026.
- Mark E. Glickman, Example of the Glicko-2 system, 22 March 2022, Introduction, read Oct 21, 2026.
- FIDE Handbook B.02 FIDE Rating Regulations, effective 1 March 2024, Clause 8.3.3, read Oct 21, 2026.
- US Chess, The US Chess Rating System (Glickman and Doan), dated 6 April 2026, Section 4.2, read Oct 21, 2026.
Want to spot this in your own games?
A first lesson starts with your games: George reviews them and builds a study plan around what he finds.