I spend a good chunk of my time playing bots because it's quick and i can resign whenever and practice my openings over and over. I know bots arent that accurate to real people at their supposed ELO, but for what it's worth, I got up to a 1500 bot and that's where i've hit my wall.
when i play 400 elo players, sometimes it feels like, well, accurate to 400. Sometimes it feels like im playing a 1000 elo bot.
Same here, I can beat the 1300 ELO queen abuse bot, but stay stuck at elo 500 with some games rated with over 90% accuracy and over 1000 actual elo.
The reason is that the bots are stupid and over rated, and so many chess channels that think 'So a 400 elo player wouldn't notice that queen hanging or blunder' .... Oh no believe me, I always notice it, right after playing blunder blunder blunder in too many games.
Most players in elo 400-600 know how to play chess well, they just can't stop blundering thus also lose a lot of games and never climb up in ranking.
Also the blunders that bots do make to 'simulate' their elo score are the most offensive and unrealistic garbage imaginable. They also refuse to play something different after a take back - they will repeatedly make the same mistake, hence that 1300 Elo bot, you can make him always and repeatedly blunder his queen within the first 5 moves.
Most of your games are simply determined by blunders.
Stop learning openings. Stop learning about pawn structures. Stop worrying about anything except the board and whether your pieces can be captured on the next move and try to avoid it.
You can play like Magnus Carlsen for the rest of the game, but if you give up a piece every tenth move, your rating will still be below 500. Play against real people.
If you want to be really kind and considerate, you can call that tool "experimental". The truth is probably closer to "nonsensical".
I try to remember the bot names but I can't. 1300 Nelson is a pushover once you learn how to develop one step ahead of his next queen attack, he is actually a fantastic bot for learning to counter that queen push strategy that a lot of 400-500 elo players like to use, but that is what the bots actual elo should be.
1400 elo 'boxbox' bot under creators I can beat easily as well, he plays very passively and develops weakly, its easy to penetrate its development.
The 1500 and 1600 bots - Antonio and Isobel, I still can't manage to beat them. It says Antonio plays defensively which is a lie, it will wreck my usual setups. Isobel likewise but more powerful, it knows how to completely wreck carro kan defenses, but tyring to get her to open with E4 to practice that needs way to many restarts meh.
Remember folks - bots are 'its', not real people.
Don't play with bots.
The way bots play is usually a little offset to their actual rating, I believe the rating of the bot is just there to help you get a good scale for how difficult it is, as for the rating prediction it's still a fairly new feature, and won't be completely accurate.
I disagree to not play against bots, at least just one in particular - Nelson.
You are going to want to master queen rush defense before going into elo 400-500 range as most players in that range use it.
All the lower rated bots are mostly junk and don't teach you anything.
Also I wish there was a bot that always played the fantasy variation of carro kan as white, as its one thing I'm struggling to learn to play against.
I disagree to not play against bots, at least just one in particular - Nelson.
You are going to want to master queen rush defense before going into elo 400-500 range as most players in that range use it.
All the lower rated bots are mostly junk and don't teach you anything.
Also I wish there was a bot that always played the fantasy variation of carro kan as white, as its one thing I'm struggling to learn to play against.
I've overcome 4 rating plateaus as an adult and not once did I seriously play against bots in the process. And not a single coach will recommend you to waste even one minute of your time on openings if your rating is below 750. And the reason is that all openings lead to the same result: Someone gives away all his pieces. At least that is the case 90% of the time.
And why should you learn openings for the remaining 10%, if it's even that much, when they might only account for 2%? Besides, if you have the option to abandon the game without losing rating or to take back a move, it is not a real game. So your brain doesn't release comparable stress hormones and dopamine while you play. That means you also just perform worse without the same feeling of reward.
Actually yes you are right, I spent a long time playing against bots because I felt like I'd be too crap against actual players, but playing actual chess now is a lot more rewarding even if I lose a lot.
Same with me. I am 1100-1150 and when I lose I get 1100-1300 estimate and when I win 1500-1900 estimate and draw 1400-1700. It happens for the same reason I can beat 1600 and sometimes even 1800 bots. The ratings are really inflated with the bots meaning the estimated elo will also be off.
Same there, 400 in my profile but in matches near my name or at the ending it'S 1500 something. And when I lose/win a match my elo on my profile doesn't change!
I played this game and it rated me 2300 and mech 2400, all I did was play a solid opening and get my butt handed to me in the middle game, why is the rating so inflated? (1050 actual rating)
I played this game and it rated me 2300 and mech 2400, all I did was play a solid opening and get my butt handed to me in the middle game, why is the rating so inflated? (1050 actual rating)
The estimated rating is an experimental feature. Less kindly put: it's complete nonsense. Chess.com should be ashamed to put out a product like that. (But of course they aren't).
Nonsense.
It's meant to provide feedback, in a more recognizable form, for one game, nothing more. It was never meant as an alternative means to gauge your skill... as its developers made clear when it was introduced. Enjoy it for what it is, like other Game Review features. Just don't take it too seriously.
"The Glicko system is a rating system designed to measure and track overall player skill over time, while a game review "estimated rating" is a temporary assessment of performance in a single game. They serve different purposes and are not meant to be directly compared as competing rating systems. Think of it this way: A high Glicko rating suggests you're a skilled player, while a high Game Review rating means the game under review is judged to have been well-played. They are distinct concepts serving entirely different objectives."
- chess.com
Nonsense.
It's meant to provide feedback, in a more recognizable form, for one game, nothing more. It was never meant as an alternative means to gauge your skill rating... as its developers made clear when it was introduced. Enjoy it for what it is, like other Game Review features. Just don't take it too seriously.
So this is useful feedback in your opinion? To claim that a 1000-rated player is playing on a 2400 level? Really?
So I've been stuck in the 400s. Some matches i play are really easy, maybe I'll stalemate on accident, but SOMETIMES I come across a match that ends up being decently difficult, maybe only one or 2 blunders that snowball into checkmate. I'll check the calculated ELO for the match and in these more difficult matches, its rated REALLY HIGH.
is this normal??? why in the 400s do i end up playing these hard matchups where we both end up with a high calculated rating?
ive been studying my forks, pins, skewers, hanging pieces, "continuity", pawn structure, basic openings, checks, etc. Im not exactly great, I dont leave pieces hanging (often at least), Im just frustrated that Im stuck in 400, and i get these calculations and it makes me more frustrated.