400-1000 PLAY THE SAME SKILL WISE

After being here for 3 years or so and playing 2,300 games. Bouncing around between 900-500 up and down multiple times (currently 600). I came to the conclusion that rating differences don't mean anything under 1000. For example a 300 rating point diffrence giving a 85% win rate simply does not apply here. This is based off first hand experience and viewing my insights. My win rate has been close to the same in all these rating ranges. Basically to get out of this beginner range, one must already be stronger than a sub 1000 to maintain a net positive win rate. I think this has to do with people starting off as an advanced player per the starting rating while not being very good. The same goes for people starting at 500 when they are much better players. What this does is jumble up the player base skill. This is why one game you can stomp a 850, the next game a 575 will outplay you entirely, NOTE: I get a large variety of ratings in my match making because I play rapid 20 min games. So the matchmaking has a limited pool of players. I don't think starting ratings should exist. EVERYONE should have to start at 100 and grind their way up. That way ratings would be much more consistent vs skill in the beginner ranges. Just my opinion. Would like to here other peoples experience. 

Hi there!

Your whole explanations are great! And they make a lot of sense!
But that sadly only applies to limiting the scope to the other players, shifting the focus away from yourself.
Have you considered that you might have fluctuations in your skill level? That that could be the root of your rating fluctuation too? By no means is the algorithm to put someone in a rating bracket 100% correct, there are smurfs and underrated alts too.

But I think that with the number of games you’ve played, it should have stayed roughly accurate.

Greetings and have a nice day!

Salve there, you are right that the sub-1000 experience is inconsistent, and "polluted" by new accounts plummeting from higher starting ratings, but you are wrong about the "math validity" and the proposed solutions. You claim that a 300-point rating difference resulting in an 85% win rate "simply does not apply" here: mathematically, it's incorrect. The ELO/Glicko systems are purely "predictive". If a 600 plays a 300, and the system says the 600 should win 85% of the time, but the user feels it's a 50/50 coin flip, the system will adjust the ratings after the game. If the 600 loses to the 300 often enough, the 600's rating will drop and the 300's will rise until they meet in the middle. Your perception is "skewed" by "variance"; in a sample size of 10 games, a 300-point gap might look invisible. Over 100 games, the 600 will certainly dominate. You're mistaking "high volatility" for a broken formula. Second point: you state "Basically to get out of this beginner range, one must already be stronger than a sub-1000 to maintain a net positive win rate". That's a "circular argument", a "tautology" (a statement that's true by necessity). Of course, to have a rating higher than 1000, you must play chess worthy of a rating higher than 1000. The rating system is a measuring tape. You seem to believe the rating acts like a barrier rather than a "measurement", which is a misunderstanding of how the metric works. Now, let's move on to your proposals: "Everyone should have to start at 100 and grind their way up". This sounds "fair" in a sporting sense, but it would be catastrophic for the user experience and would actually make the problem much worse, not better. If a GM joins the site and has to start at 100, he will have to play 200 games to reach his true level. In those 200 games, he will destroy absolute beginners. If you force everyone to start at 100, the "floor" becomes incredibly crowded. The math spread required to separate a novice from a beginner would be impossible to calculate amidst the "noise" of strong players rushing through the ranks. Besides, you mention playing 2300 games over 3 years and fluctuating between 500 and 900: well, this is the strongest evidence the system is working correctly. If you were truly an 1100 player stuck at 600 due to "unfair ratings", you would eventually gain rating simply by math probability of facing opponents who make more blunders than you do. The frustration comes from the fact that improving at chess is incredibly difficult, and it's easier to blame the matchmaking system than to confront the "plateau" in your own play. The rating system is accurately reflecting your skill level over 2300 games. The "inconsistency" you feel is not a bug in the rating software; it's a feature of beginner chess, where games are decided by luck and blunders rather than skill. The only way to fix that is to study and improve, not to "reset the scoreboard". Ciao

I honestly think both perspectives make some sense here. The sub-1000 range absolutely feels inconsistent at times, especially in longer rapid queues where the matchmaking pool is smaller and players can fluctuate wildly from game to game. Beginner chess is full of variance: one match looks like a tactical masterclass, the next looks like pure chaos (admittedly, that is an exaggeration). That can make rating gaps feel far less meaningful in practice.

At the same time, I don’t think the rating system itself is fundamentally broken either. Elo/Glicko are more like long-term measuring tools than absolute statements about any single game or short streak. Over small samples, a 300-point gap may look invisible, but over hundreds of games the trends usually become clearer. One shouldn’t blame the matchmaking system or the player:

I also wouldn’t frame it as blaming either the player or the matchmaking system. Chess improvement is rarely linear, and lower-rated games are naturally volatile because consistency is still developing. That unpredictability is frustrating sometimes, but it is also part of what makes chess in general entertaining and sometimes surprisingly competitive. Beneath all the rating swings, most people are simply trying to improve or enjoy the game.

So, just do the same, without blaming either party! It’s not about the ratings, it’s about the game.

So let’s all just have a great time playing chess !

I can't agree with that. I'm 500 rapid and can hold my own against 400-550 elo opponents, but I had a few 1000+ opponents and they were fearsome, I had no chance against them.

Oh and 800-900 can be hard for me too

@gawonnie

Yes, you are fundamentally right. You correctly identified the "friction" between "math expectation" (that is, how ratings should work) and "human experience" (how games feel). The rating system is not broken; it's simply being asked to measure something that's inherently unstable (human learning in its early stages). The best approach, as you suggested, is to accept the "variance". In the sub-1000 range, you're not just playing chess; you are conducting "experiments". And the rating is just the "lab notebook"; it should not be confused with the experiment itself. Said that, I'd like to add a few "nuances" to the discussion, regarding the psychological aspect and the definition of "meaningful" games. We often treat the rating number as a "rank" rather than a "confidence interval". We should view sub-1000 ratings less as a "hierarchy" and more as a "diagnostic category". Being rated 800 does not mean you are the 800th best player on Chess dotcom; it means you are currently in the phase of learning where you understand basic development but struggle with tactical awareness. Consider the specific number as a "noise", and the "band" as a signal. If you focus on the rating, you are focusing on an outcome you can't fully control, while if you focus on the game, you focus on the only thing you can control: your next move. The frustration comes from expecting the rating to validate self-worth. If the sub-1000 range is "inconsistent", then it's the worst possible place to look for validation. It's like trying to measure your height during a growth spurt using a ruler that's bouncing up and down. Ciao

I would assume that many players here are underrated and overrated.

That explains quite a few games I had here.

#5 I agree

(1) there's a maths problem with the idea of everyone starting at 100 and grinding their way up. Since the winner gains the same number of points as the loser loses, everyone who joins at 100 and heads upwards is removing points from the pool, so if Chess.com implemented forcing everyone to start at 100, the more people joined, the more everyone's ratings would drop, which would upset people.

(2) it's an oversimplification to assume that a chess player's chance of winning can be summed up in a single number that's fairly constant. It might be true of seasoned competitors, I don't know. But most of us have bad days when we're tired, when we're pessimistic. I will lose a string of games because I'm in a self-destructive mood, and then play better for a bit. My chess ability is not actually what limits my rating. There must be others like me.

(3) But basically I agree. In my experience, as a drastic example of a rating-bouncer, it's as tough to win against a 200 as a 700. I suspect part of it is that the true depths down in the 200s are inhabited by others like me, who are down there because they don't care about winning, even pride themselves on being failures, who have self-destructive streaks, and some enjoy being anarchists who can prove that a 200 can win against a 900 - so if you show signs of playing good chess down there, you can rouse the inner competitor and suddenly find you're playing someone who can kick out 85% accuracy or higher, and who is no easy opponent! For that reason, it's easier to plunge into the depths than re-emerge. Also, for all of us under 900, the main reason for losing is usually some stupid blunder, and they're a fairly random matter for a lot of us.

The "800-900 is hard" is an excuse, that's the problem. If you are truly a 500, an 800 should beat you roughly 85-90% of the time. A 300-400 point rating gap is statistically massive. You are losing because you are relying on your opponents to make mistakes to win. And when you face an 800 who makes fewer mistakes than a 400, you have no resources to win. Calling 1000+ opponents "fearsome" is a "defeatist mindset". A 1000-rated player is still a beginner. They miss basic mates, hang pieces regularly, and have no understanding of pawn structure or positional play. If you find a 1000 player "fearsome", is because you have no foundation. You're likely moving pieces without a plan and hoping for luck. When you face someone who has a plan (even a bad one), you crumble because you are reacting rather than acting. You are losing to "structure", not to skill. Besides, you claim to "hold your own" against 400-550 opponents: there is no such a thing as "holding your own" at 500 level. You're simply engaging in a race to the bottom to see who commits suicide first. If you play a 400 and win, it does not mean you played well; it just means you played less terribly. Taking pride in "holding your own" against 400s is setting the bar so low it's underground. You seem surprised a 1000 player is unbeatable for you: a 1000 player sees the board completely different than a 500player. To a 1000, a 500 is just a random piece mover. If you reviewed your losses against 800s or 1000s, you would see you lost because of elementary tactics (forks, pins, hanging pieces, etc.). You survive against 400s because chaos favors the unprepared, but you perish against anyone who plays slightly coherent chess. The 900-1000 bracket should not be "hard", it should be a "wake-up call" that simply waiting for the opponents to blunder is a failing strategy. The reason 800-900 players are "hard" and 1000 players are "fearsome" is that you are playing "Russian Roulette" with your pieces, while they are actually playing chess. You don't need complex strategy theories, you need to stop throwing away points. Spend 50% of your chess time doing basic tactics puzzles (forks, pins, skewers, mate in 1, mate in 2, etc.): when you see these patterns in puzzles, you will start seeing them in your games. This is how you beat 800-900 players. And always review your losses (I know this is the "painful" part, but it's necessary): 1000 opponents are not fearsome and you have a chance. Stop fearing the 1000s. Respect the board: if you stop leaving your pieces "en prise", the fearsome 1000 players will suddenly start looking very beatable, because at that level, they're still waiting for you to mess up. Don't give 'em that satisfaction. Ciao

Obviously you don't know how I'm thinking.

It's the other way around, I'm feeling I would have to play special moves to surprise them because they're close to being intermediate. I never expect them to make easy mistakes when I play them. It's rather that I respect them TOO MUCH for their elo.

Yeah I know 300-400 elo higher opponents should beat me almost always. I have read about that even if I didn't know the exact %.

Sorry that this is the only part I could reply to, as the wall of text is too much for me as I'm autistic. I don't want to get sensory overload again. So I hope you don't take it personal. Maybe you could edit your post and make it easier readable. Sorry I don't know the exact english word, make more gaps or so, so that it has smaller parts.