Rating distribution guide

You are at a good level.

The problem is that there are still hundreds of thousands rated higher than you.

The average rapid rating is 894.

You can see the average in the leaderboard.

It means you’re better than 29% of players silly billy

You can push to 1300+ with ease, with no study or memberships and just grind play (which is +85% percentile; not great in the grand scheme of things and people in this rating range are not that "good" and we need a lot to improve on, I am stuck because I don't study chess; I was looking for an online chess notebook that is free to use). I exclusively use the Bird's eye opening as white (f4), I stopped playing Leningrad style Bird because it was too passive and weird to navigate for me. FEAR THE BIRD!

I really want a membership too, we normal folks only get 1 engine analysis a day and 1 puzzles/puzzle rush/puzzle battle & custom puzzle (I think with cuts on the amount if you used analysis for puzzles), and barely any drills etc. But I really want a Online Chess notebook to study with; without that I don't think I would improve that much. At least you get the free analysis, BTW you can set up games/positions custom in the analysis and upload them to finish vs computer to get moves analysis but it's 1 sided and it analyzes for that side that has the next turn when you upload it. AND I REALLY WANT THE EXPLORER FULLY

I got to 1500 without a membership it’s possible to get even further as well with work, example: YouTube playing higher rated people and using your resources that you have well.

@fools... 85% not great! If it were an exam I'd be elated. Chessable is worth a look. It uses spaced repetition to teach. You can have up to 5 free courses. The free courses are normally cut down versions of their main courses but they are still generous with the content.

Petre, roughly under 1400 is beginner, 1400 to 1800 is intermediate. When a player gets to 1400 they still have a lot to learn but have some basic skills and understand basic principles in all areas of the game.

Thank you for your useful responses guys.

I feel that the rating distributions here are disjointed from real life USCF or FIDE ratings. For example, to be within USCF's 95 percentile you need a 2000+ rating (link below). In FIDE a 95 percentile is equal to roughly 2300+ rating (link below). However, on chesscom you can achieve a 95 percentile with a 1700 rating which is absurd. 

Part of the problem is the player pool size, quality of players and calculation methodology. But that doesn't change the fact that ratings and distributions here are meaningless compared to real world ratings. 

http://www.uschess.org/archive/ratings/ratedist.php

https://en.chessbase.com/post/visual-presentation-of-world-chess-ratings

 

@theimmortalpatzer One thing to keep in mind is that these rating systems are relative. So, while the chances of a 1000 beating a 1200 will be roughly 25% in each system, a 1000 in one won’t be the same as a 1000 in another. I believe chess.com tries on an ongoing basis to make their ratings both statistically accurate and similar to equivalent ratings in FIDE and national federations, but with different player populations there’s not much that can be adjusted, especially if they want ratings to be accurately predictive of win chances (which is the rating system’s purpose.)

Also, though, it makes sense that FIDE’s 95th percentile would be a lot higher than chess.com’s. To be rated in FIDE at all you need to (1) play at least 5 games in a FIDE-rated tournament (with a FIDE-qualified arbiter), (2) At least earn 1/2 point by drawing one game, and (3) be rated 1000 or better. Just those last two push their distribution way up compared to chess.com.

Here, to be rated, you just need to play some chess games. Everyone gets a rating.

I'm aware that the rating systems are designed as relative ratings. I think the issue here is that chesscom has chosen a rating methodology that resembles national federations and FIDE rating systems too closely and people expect the ratings across organizations to be similar. I think that the blitz format has a horrible distribution on chesscom and should be addressed. Based on your comment it sounds like chesscom is aware of the issue and they adjust their rating system on occasion. 

Important to note that the Chess.com player pool includes the huge number of players who join, play a few games, decide they suck/don't enjoy Chess, and leave with a rating around 500. I'd guess that well over half then player pool has never put any kind of effort into learning the game. I'm currently rated just under 1400, which is IIRC 91.4 Percentile, even though I consider my skill level embarrassing compared to how I played 20 years ago, which I'm realizing in hindsight wasn't all that good either. Scoring over 75% means you're one of the players who's starting to understand the game propely. Hooray! Your journey to true knowledge begins here...

 

According to chess.com, the percentile is calculated from active players, which means having played a game in the previous 90 days.  So, it doesn't include those who straight-up quit.  (It does include a lot who like playing the game but don't care to work at improving, though.)

You're only partially correct. Chesscom active players include anyone who played a game within the last 90 days. Considering the explosion of new players in 2020 imagine how many players joined the site and abandoned it after a handful of games. Recently, chesscom awarded 150 points to all active bullet players. I'm think blitz ratings need to be adjusted similarly. 

The last 90 days include the very tail end of the Queen’s Gambit hype. Things should be near a steady state now.

You would need to have historical data and a mathematical model to make such a claim. But whatever. 

If it were my job to make such judgments then I might benefit from having the data. As is, I’m judging by how much people have been talking about it lately.

I would like to know if the performance here at 1536 in blitz chess can really mean that I am in the 95.3 percentile. That doesn't sound like much to me, because on lichess.org I had a top percentile of around 90. Are there any more extensive stats available for, say, whole classes (wood group, stone group, etc.)? Thanks a lot Jaroslav