Chess Ratings Explained: Elo, Glicko and Why Your Numbers Differ Between Sites

How Elo and Glicko ratings work, what a rating gap means in expected score, and why 1500 on one site is not 1500 on another.

A chess rating is a prediction in disguise: it says how likely you are to beat someone of a different rating. Once you read it that way, many confusing things become clear — why your rating jumps after a single game, why it differs between websites, and what a 100-point gap actually means.

The Elo idea

The system was devised by the Hungarian-American physicist Arpad Elo and adopted by FIDE in 1970. Each player has a number. The difference between two numbers predicts the expected score, where a win is 1, a draw is 0.5 and a loss is 0. After a game, your rating changes by how much your result differs from that prediction: beat a stronger player and you gain a lot, lose to a weaker one and you drop a lot.

What a rating gap means

The expected score depends only on the difference. Some reference points:

• 0 points:50% — an even match.
• 100 points: about 64% for the stronger player.
• 200 points: about 76%.
• 400 points: about 91%.

These are scores, not win rates: a draw counts as half. So a 200-point favourite does not win three games in four — they score three points out of four, with some draws mixed in. You can try your own numbers in the Elo calculator.

The K-factor: how fast your rating moves

How big a swing a result produces is controlled by the K-factor. A high K lets new or improving players find their level quickly; a low K makes the ratings of established players stable. FIDE uses different K values for new players, for most rated players and for the very strongest, which is why a titled player loses only a few points for a bad result while a beginner can lose many.

Glicko: ratings with uncertainty

Online sites do not use plain Elo. Lichess uses Glicko-2, and Chess.com uses a Glicko-style system. Glicko was created by Mark Glickman and adds a second number, the rating deviation: how confident the system is about your rating. A new account has a high deviation, so its rating moves quickly; a player with many recent games has a low deviation and moves slowly. If you take a break, the deviation grows again, which is why returning players see big swings.

Why 1500 is not 1500 everywhere

A rating only makes sense inside the pool it was calculated from. The same player may be 1600 in Chess.com blitz, 1750 on Lichess blitz and 1450 in FIDE classical. Reasons include:

• different player pools and starting ratings,
• different rating systems and settings,
• different time controls (bullet, blitz, rapid and classical are separate ratings),
• and rating inflation or deflation in each pool over time.

So compare yourself to others within the same site and time control, not between sites.

Using a rating to improve

A rating is a result, not a goal. If it is stuck, the useful question is what you lose games to: tactics, time pressure, endgames or openings. That answer comes from your games, not from the number. Importing your recent games into Chesslume and looking at the accuracy scores and the blunder counts shows you which of those it is, so your study time goes where it matters.