Chesscom rating vs bots vs problems

Hello,

I did around 1200 games in 2 years, mostly blitz/bullet but now Im more into rapid 10min games. 4000 problems (57% wins).
But only 523 in bullet, 424 blitz, 771 in daily, ~900 in rapid, level 2000/2150 in problems.

In rapid i got 54% win / 4% draw / 42% lost. But Im progressing so slowly, I dont understand why.

When I play against bots, I almost always win against 1400, I win more than 50% of the time against 1500, maybe 33% ot the time against 1600 and still not against 1700 but its getting close.

Why such a difference ?

Regards

.

??

!

Because those are different things, requiring different skills.

Moreover, puzzle score is completely different and the developers didn't even make an effort to make the numbers similar. The same is true for "bot ratings", that are not ratings at all, just a number written there.

Thank you Magipi.

I understand the problems. Although often the level correlates, but let's admit it. What bothers me the most is for the bots. I follow the chess streamers, they fight against bots around their Chesscom rating. For example, Vitualis (Adventures of a Chess Noob) is ~1200 rapid rating, fighting against a bot at 1200. 1200 bot is a child's play for me. Even 1400. I don't understand...

Forget bot ratings. They are irrelevant. Most of the time bots are overrated, but there is no guarantee for that either. One 1200-"rated" bot may be super weak, while another 1200-"rated" bot may be extremely strong. Unless you are playing the exact same bot, that story means nothing.

Took a look at the way you play.

In the opening you can play E2-E4 and then E4-E5 sacrificing development. You move your bishops to bad squares in the opening, often having them blocked by your pawns or blocking your pawns. For example moving your white bishop to E2 or D3.

Also saw quite a lot of moments when you hanged your pieces.

You should work on your chess habits, developing your pieces to good squares and play at a time control where you have time to check instead of doing one move blunders.

OP, these weaker bots, they are much weaker than human counterparts of the same rating. The difference can easily be 1 000 points in some cases.

Bot ratings and live ratings earned by playing against people are completely different things. You should completely disregard bot ratings and do not view them seriously.
-------------------------------
But Im progressing so slowly, I dont understand why.

---------------------------------

Bullet games are useless for improvement, and blitz games are just marginally better for most people. Some people will improve playing blitz, most will not have significant improvement. You need to think about your moves, playing the first thing that came to your head will not get you better.

Longer games, where you use your time and do not blitz out moves (15|10 at the very least), plus a lot of effort outside of playing is what is usually needed to improve.

Thank you. What do you mean by a lot of effort outside of playing ? Course lessons ? Book ?

Do you think I should learn my openings ? or just play longer game (f.e 15/10) more focusing on basics ?

Routine advice ? Cause I play everyday, but I have no practice routine.

Thank you.

Yeah I know I often end by an avoidable blunder. I'm trying to play each game by only developing to castle fast, with no unprotected piece. Looking for an adverse unprotected piece to push on this weak square when its possible. But indeed I often end short in time and making hawful blunder.

Do you think I should learn my openings ? or just play longer game (f.e 15/10 like trewphiobsfgd suggested) more focusing on basics ?

Routine advice ? Cause I play everyday, but I have no practice routine.

Thank you.

In short: You play, then you go through your games in order to learn from your mistakes. But at first going through your games will help only slightly because you don't know much about chess.

So what then? Then you do some passive learning. You have many options for this, and some are free, some are not. Your options are books (you need to find something in accordance with your rating), coaches or video lessons (you have some free material on YouTube or paid ones from chessable for instance). Along with that, you should always drill tactics daily preferably, at least a little bit.

It works like a synergy. Then you play more, are able to understand more from going through your games. Then you learn more advanced concepts from passive learning, and on and on it goes.

Notice that this can take hundreds, then thousands of hours, and you will still just scratch the surface. It takes a really long time to improve to some high level.

At first, I would concentrate at playing longer games and using the time I have. In games aim for a quicker development (learn opening principles and apply them, unless there is a specific reason not to), and aim not to hang pieces. Learn to blunder check before you make a move. That is the most important thing at the start. These are the building blocks you need.

As for specific openings, you are free to check how you played openings after the game - there are many opening databases around. But, I wouldn't spend any significant time for memorization specific lines, people on this level do not know them, and will play some random moves as a response. So, it is ok to review openings but do not spend too much time on them.

I would choose a variation I play against various opponent's responses and just build upon them gradually.

I usually just link this on questions on how to improve:

https://www.chess.com/blog/nklristic/the-beginners-tale-first-steps-to-chess-improvement

ignore magipi. the site just has a rigged "player" pool

It is common knowledge that on chess.com the vast majority of your opponents are either adaptive bots or Rule-breaking League players (they think it's cool to be in the League of Legends, but in reality it doesn't give anything, and they greatly underestimate their rating to gain points). So you get an opponent with 1150, but he cooled his rating in the League by 400-500 points, and you will lose to such a player (according to the Wikipedia article on Elo rating in 99% of cases)

Опишу по-русски это непрекращающееся блядство на chess.com!
1. Для высокой игровой доступности, платформа автоматически при отсутствии подходящего соперника генерирует адаптивного бота (одним из свойств которого является способность “поддать гари” если вы выиграете несколько фигур, показывая чудеса вроде перфоманса 1950! Вы систематически проигрываете, топчетесь на своём рейтинге и вынуждены купить платную подписку.
2. Опять же для популярности придумали эти Лиги тщеславия. В итоге, все хотят пройти в Лигу Легенд (которая вообще ничего не даёт) и начинают механически играть и количеством партий набирают нужные очки, но тем самым сильно охлаждая свой рейтинг (на 400-500 пунктов, что является нарушением Правил). В итоге вы получаете соперника 1150 с реальной силой игры 1600-1700, и вы проиграете в 99% случаев (откройте статью Википедии про рейтинг Эло, там описаны шансы при игре соперников с разницей 100/200/300/400/500 и .).

Get a coach

magipi is one of the most informed forum participants i have ever read. we dont always agree, but his logic and wit are undeniable.

Thank you. It was not common knowledge for me. I play and I didnt look into stuff like this tbh. 
That being said, I dont get the point of "farming" fast games to get better league... is there a rating from where this kind of mentality disapear and we get more "real" rating ?

What that's mean for me ? That my "true" rating is more about 1200/1300 actualy ?

Bots are not overrated. Chess.com ratings are severely deflated.