Chess Ratings Explained

What a chess rating is

A chess rating is a number that estimates how strong a player is, based on results against other rated players. A higher number generally means a stronger player. The number does not measure talent or intelligence directly. It is a statistical estimate built from wins, draws and losses, and it only has meaning relative to the pool of players who share the same rating system.

The practical use is prediction. If two players have different ratings, the system can estimate how many points the stronger player should score over a number of games. That is also how the system updates: you gain points when you do better than expected and lose points when you do worse.

How the Elo system works

The best-known system is named after Arpad Elo, a physicist and chess player. The US Chess Federation adopted an Elo-based system around 1960, and FIDE, the international federation, adopted it in 1970. Many variations exist today.

The core idea is the rating gap:

Equal ratings: each player is expected to score about 50 percent.
A gap of about 100 points: the higher-rated player is expected to score roughly 64 percent.
A gap of about 200 points: roughly 76 percent.
A gap of about 400 points: roughly 90 percent.

These are expectations over many games, not guarantees. A 1600 player will still lose or draw some games against a 1500 player, because anyone can play a bad game or make one decisive mistake.

After a game, your rating changes by an amount that depends on the result and on how surprising it was. Beating a much stronger player earns many points. Beating a much weaker player earns very few. A setting called the K-factor controls how large the swings are. Newer or less-established players typically have a higher K-factor, so their ratings move faster, while established players’ ratings move more slowly. The exact values depend on the organization.

Glicko and rating deviation

Chess.com and Lichess use systems from the Glicko family, developed by Mark Glickman as an extension of Elo. Lichess uses Glicko-2. The main addition is a measure of uncertainty, often called rating deviation. A new account, or a player who has not played in a while, has a high deviation, which means the system is not confident about the number yet and lets it move quickly. As you play more games, the deviation shrinks and your rating settles.

This is why your first few online games can cause large jumps, and why a provisional rating should be taken lightly.

FIDE, national federations and online ratings

There is no single “chess rating”. Several separate rating pools exist:

FIDE ratings are used for international over-the-board tournaments, with separate lists for classical, rapid and blitz.
National federations, such as US Chess or the English Chess Federation, keep their own lists, with their own formulas.
Online platforms, such as Chess.com and Lichess, keep separate ratings for each time control: bullet, blitz, rapid and often classical.

Because each pool has different players, a different formula and a different starting point, the same person can hold quite different numbers in each. Online ratings are commonly said to differ from over-the-board ratings, and the two big sites differ from each other too. See our Lichess to Chess.com rating converter guide for rough comparisons.

Rating levels in rough terms

Different sources draw the lines in slightly different places, so treat this as a loose guide for online and club play:

Under about 800: beginner, still learning the rules, basic tactics and how to avoid blunders.
About 800 to 1200: novice to improving player; knows the rules and basic tactics, but still hangs pieces fairly often.
About 1200 to 1600: club-level or intermediate; spots most simple tactics and has some opening and endgame knowledge.
About 1600 to 2000: strong club or advanced tournament player.
2000 and above: expert territory and beyond. Titles such as master and grandmaster have their own requirements, which are more than a rating threshold, and grandmasters are typically rated around 2500 or higher.

For scale, Magnus Carlsen’s peak FIDE rating of 2882 is the highest ever recorded for a human. Top chess engines are rated far above any human, which is why humans do not beat them in normal games.

What do 300, 800, 1000 and 1200 mean?

Readers often search for a specific number. Here is a cautious reading of each, remembering that the meaning depends on the platform.

300: Very low on most online sites. It usually points to a brand-new player who is still learning how the pieces move, or a site whose scale runs low for beginners. It is not a judgment of ability, only a starting point.
800: A beginner who knows the rules and plays real games but often loses pieces to simple threats.
1000: An improving beginner. On many platforms this is somewhere around or below the middle of the active player pool, but exact percentiles vary by site and time control and change over time, so check the site’s own rating distribution if you want a real figure.
1200: A solid casual player who has moved beyond the basics. Whether it counts as “good” depends on where you play. On some sites it is near average, on others somewhat below.

The more useful question is what you can do at your level. Moving from 800 to 1200 is usually about not hanging pieces, learning common tactics and finishing basic endgames.

Why ratings differ between sites

Several factors explain why a 1500 on one site is not a 1500 on another:

Different player pools, with different average strength.
Different formulas and different default starting ratings.
Different time controls, since a blitz rating and a classical rating measure different skills.
Different rules for how fast ratings move.

Compare ratings only within one pool and one time control when possible.

What ratings cannot tell you

Arpad Elo himself noted that ratings are a comparison of performances and nothing more. A rating does not measure intelligence, and it can be affected by the number of games played, the time control and how rested you are. People often link chess strength with IQ, but the relationship is more modest than popular claims suggest. See our article on the IQ of chess players for what the evidence supports.

How ratings go up

Ratings rise when you consistently score better than the system expects. In practice that tends to mean:

Playing regularly, with enough thinking time to avoid simple blunders.
Reviewing your losses to find the recurring mistake.
Learning common tactical patterns and basic endgames.
Playing opponents close to your own level, since results against much weaker or much stronger players teach less and move the number less.

Do not worry about short-term swings. A rating is an estimate that becomes more reliable over many games.