What is the average chess rating? Why there isn't one number, and where to find the real distributions

People ask for one average because a single number feels like a grade. The pools are not one group. A Chess.com blitz rating, a Lichess rapid rating, and a FIDE standard rating describe different players under different rules. Last checked: September 25, 2026.

If you want the comparison that matches you, use what is a good chess rating after you have your own percentile. This page is about why the average itself is a slippery word.

The average (the mean) adds every rating and divides by the number of players. A handful of very high or very low accounts can pull it. The median is the middle player: half the pool is above, half below. A percentile says what share of the chosen group is at or below you.

Chess.com's own definition, from the help article updated February 9, 2026, is the percentage of Chess.com members with a rating equal to or lower than yours, per game type. That is more useful than a site-wide mean, and it still depends on who is included. Their article requires an account at least 7 days old, 20 games of that type with at least one in the past 90 days, and an activated account. Closed accounts are left out.

A worked comparison without fake numbers

Imagine three players in one time control. Their ratings are 800, 1100, and 2000. The mean is 1300. The middle player is 1100. Nobody in that toy pool is "average" in a way that describes a real site, and this page is not claiming those are real site ratings. The point is only that mean and middle answer different questions. On a real site, read the distribution instead of repeating a mean from a blog.

Open your Stats page and use the game-type menu. The percentile there is the official figure for your account. This guide does not print a Chess.com average, median, or distribution number. The public help page explains the definition. It does not list an average rating, and a site-wide stats page could not be quoted from a stable official URL on September 25, 2026.

1. Log in and open Stats.
2. Choose rapid, blitz, or bullet. Do not mix them.
3. Read the percentile for that type and write down the date.
4. Ignore any "average rating" screenshot that has no date and no link to Chess.com.

Lichess publishes a weekly chart for each time control: bullet, blitz, rapid, and classical. Open the one you play and read it there. On September 25, 2026 the logged-out rapid page did not give a mean or a median we could cite, so none is printed here.

New Lichess ratings start at 1500 ± 1000. The FAQ says the system is 95% sure a brand-new rating is between 500 and 2500, so the first results move a lot. That start point also means a raw average of all Lichess accounts is a poor description of "a typical player," because uncertain new ratings sit on 1500 by design.

FIDE and national federations rate over-the-board events under their own rules and their own player pools. The official FIDE list is ratings.fide.com. This page does not average that list and does not convert an online rating into a FIDE rating. Online and over-the-board chess are different games of time, conditions, and who shows up. See Chess.com vs Lichess for why online numbers diverge from each other before you even reach FIDE.

Searchers often want an average by age, or an average after one year of play. No official, current source for those figures was found for this page. Age-based averages you see online are unreliable unless they name the dataset, the pool, and the date. A year of play also means different things: one time control or four, reviewed games or hundreds of unreviewed bullet games.

Band notes such as 1000 and 1200 describe habits, not an average age or a guaranteed year of progress.

Once you know your percentile in one pool, the next question is why the rating sits there. How to get better at chess starts from the leaks in your own games rather than from a global mean.

A drill

1. Write four lines: site, time control, rating, percentile or "not shown yet."
2. If the percentile is missing, check the Chess.com requirements (account age, 20 games, one game in 90 days) or open the matching Lichess chart.
3. Repeat in 30 days. Keep the sheet. Throw away any undated average you copied from a forum.

Common mistakes

• Quoting one average for "chess" with no site and no time control.
• Comparing your mean-seeking number with someone else's percentile.
• Treating a new Lichess 1500 as an average-player result. It is the starting point, with a very wide interval.
• Repeating an undated distribution number from a search snippet as if Chess.com had published it.