I've defeated (3200) rating chess.com engine by accident.

Sort:
andrewfried

My guess from playing bots a bunch of times is that chess.com has to program a bot to make a certain amount of errors over time. And those errors happen randomly, at times when an equivalent human-rated player would not do that. And so this game is just an example of this bot's needing to catch up with the number of mistakes it was supposed to be making in other games. You'll note the Review Tab says the Maximus bot played like a 2350 rated player in this particular game. Any particular game it is probably impossible to play like some particular rating "3200". The programmers just have to try get as close to this as possible and sometimes the results are way off.

This happens at all levels of bots. For me, I've played a bunch of beginner and intermediate bots, with some stretches where I thought they played well (to my inexperienced mind) and then other times they would just give me free pieces for no reason at all. It's kind of like on any single move or game it's impossible to behave exactly like whatever rating it's supposed to be, but over time, it just has to try to correct itself to get closer to that. Some games, like this one, happened to have to have it veer off course.

Divesh_Chauhan
andrewfried wrote:

My guess from playing bots a bunch of times is that chess.com has to program a bot to make a certain amount of errors over time. And those errors happen randomly, at times when an equivalent human-rated player would not do that. And so this game is just an example of this bot's needing to catch up with the number of mistakes it was supposed to be making in other games. You'll note the Review Tab says the Maximus bot played like a 2350 rated player in this particular game. Any particular game it is probably impossible to play like some particular rating "3200". The programmers just have to try get as close to this as possible and sometimes the results are way off.

This happens at all levels of bots. For me, I've played a bunch of beginner and intermediate bots, with some stretches where I thought they played well (to my inexperienced mind) and then other times they would just give me free pieces for no reason at all. It's kind of like on any single move or game it's impossible to behave exactly like whatever rating it's supposed to be, but over time, it just has to try to correct itself to get closer to that. Some games, like this one, happened to have to have it veer off course.

Finally a smart person, who don't blame other's for cheating without any proof.

Divesh_Chauhan

But I didn't do multiple rematches with the AI. The match I posted above is the first match between me and the 3200 ELO AI.

king2f8
Divesh_99 wrote:
andrewfried wrote:

My guess from playing bots a bunch of times is that chess.com has to program a bot to make a certain amount of errors over time. And those errors happen randomly, at times when an equivalent human-rated player would not do that. And so this game is just an example of this bot's needing to catch up with the number of mistakes it was supposed to be making in other games. You'll note the Review Tab says the Maximus bot played like a 2350 rated player in this particular game. Any particular game it is probably impossible to play like some particular rating "3200". The programmers just have to try get as close to this as possible and sometimes the results are way off.

This happens at all levels of bots. For me, I've played a bunch of beginner and intermediate bots, with some stretches where I thought they played well (to my inexperienced mind) and then other times they would just give me free pieces for no reason at all. It's kind of like on any single move or game it's impossible to behave exactly like whatever rating it's supposed to be, but over time, it just has to try to correct itself to get closer to that. Some games, like this one, happened to have to have it veer off course.

Finally a smart person, who don't blame other's for cheating without any proof.

bro actually photoshopped martin ☠️☠️☠️☠️☠️

Devin0922
I think you actually just adjust the bot rating to 250 and then it played a dumb move then you fork and you adjusted the bot rating to 3200 so that’s technically cheating
AntzSK
I know why
AntzSK
The engine was playing as white, and he was playing as black, he simply just flipped the board
AntzSK
Oh wait never mind
albusbluderdore
Bruh he’s what 1600 in rapid? How tf are you beating a 3200 CHESS BOT WITH 1600 IN RAPID????
albusbluderdore
Are you sure your finger didn’t slip and change it like 2500? But even then, you’re only 1600?
albusbluderdore
How did u do that?
Divesh_Chauhan
vinaygore wrote:
How did u do that?

IDK

Divesh_Chauhan
Devin0922 wrote:
I think you actually just adjust the bot rating to 250 and then it played a dumb move then you fork and you adjusted the bot rating to 3200 so that’s technically cheating

It is not possible on chess.com to adjust the rating of a bot mid-game. We can only adjust the bot rating before starting the match.

Divesh_Chauhan
vinaygore wrote:
Are you sure your finger didn’t slip and change it like 2500? But even then, you’re only 1600?

First of all, rating isn't everything. There are professional FIDE players who don't have a chess.com account. So technically if that FIDE player will make a new account on chess.com then the default rating will be 250 but that doesn't mean that FIDE player don't play good just because the account have 250 rating. I personally think we should not judge someone chess skills just by looking at their chess.com or lichess account ELO (we should only consider the ELO given by the FIDE) we have to play thousands of games to improve our ELO. I have not played that many of the chess games that's why my ELO is only 1600, but in reality I'm really good at chess.