Bot Ratings
Sort:
I think that it seems very likely people will mess with the ratings
Chess.com could punish those cheaters too.
JHACKIL wrote:
On chess.com, the bots don't have accurate ratings. Should players be able to play rated games to regulate the bot's ratings?
The bot used to be in the play interface and ratings did change based on their games against other players. People were using bots to farm rating and I believe that's one of the reasons the were moved to their own section.
I understand that people could receive unrightful ratings based on their games against bots, but:
(1) the ratings of bots will promptly reach their true value (vast amounts of people play against bots each day)
(2) the inflated ratings of the players will also eventually reach their true value (assuming the user is active on chess.com)
JHACKIL wrote:
I understand that people could receive unrightful ratings based on their games against bots, but:
(1) the ratings of bots will promptly reach their true value (vast amounts of people play against bots each day)
(2) the inflated ratings of the players will also eventually reach their true value (assuming the user is active on chess.com)
If someone played enough bot games, as a percentage of overall games, they likely will end up with inflated ratings.
Yes, but that is only temporary. As I have previously stated, the bots on chess.com will quickly reach their true strength, as many people play against them each day. After this, the players of chess.com would not be able to inflate their ratings with bots (the bots will have accurate ratings).
Martin_Stahl wrote:
JHACKIL wrote:
On chess.com, the bots don't have accurate ratings. Should players be able to play rated games to regulate the bot's ratings?
The bot used to be in the play interface and ratings did change based on their games against other players. People were using bots to farm rating and I believe that's one of the reasons the were moved to their own section.
So you could choose the same bot again and again? I can see ppl playing the weakest bot continually until they had wrung all the rating points from it and then move onto the next weakest bot. Is that what you meaning by "farming"?
@chrislamuk yes
It is the same with real people. I could, say, play a 250 player to earn rating points, but I won't receive as much rating as if I played an 850 rated user. I see no reason not to allow the users of chess.com to play rated against bots. Both the ratings of the bots and of the users will reach their true value.
Martin_Stahl wrote:
@chrislamuk yes
If the player is randomly assigned a rated bot to play then that would stop the problem. The player would have to play that bot to a result or maybe a new bot can be assigned after say 3 days. You can also keep the unrated bots separate as you do now. I would really like to see rated bots come back 
chrislamuk wrote:
Martin_Stahl wrote:
@chrislamuk yes
If the player is randomly assigned a rated bot to play then that would stop the problem. The player would have to play that bot to a result or maybe a new bot can be assigned after say 3 days. You can also keep the unrated bots separate as you do now. I would really like to see rated bots come back
One of the main reasons for bots is choosing the personality/style. I suggested a separate bot rating pool when the system changed but it wasn't something that really got any traction/support.
Bots are all like 500 rating points higher than they are to make u feel better about winning
"IntrovertedCPawn"
This is exactly why I would like bots to play rated against humans. Their ratings will reach a true value.
Makes sense
All bots, including martin are overrated.
Hopefully chess.com will implement my solution.
sbstnsrfss , stop lying. Bots are rated accurately. Martin is not overrated. No bot is overrated. JHACKIL, define "True value"? Chess.com ratings? Ha-ha. The most inaccurate rating on Earth after lichess. So much randomness, so much distortion.
Learn about Elo and what Elo means. Then learn about history of how engines got their ratings. Engine rating corresponds to true FIDE rating. Any engine with any setting can be precisely rated using engine vs engine tournaments. This scales up and down.
Next, on komodochess website read about skill level settings. Komodo is used here on chess.com. Chess.com displays ratings that are slightly different from ratings that Komodochess outlines. But still that's very close to true FIDE rating compared to 5|5 Blitz human performance.
One huge caveat: chess.com bots run locally using resources of your PC or smartphone. Exact configuration is uknown. IF there is a time limit on thinking then due to some background processes on your PC or weakness of your smartphone bots may make stupid moves. So for consistent results play bots using Desktop software. If you like Komodo - you can download a free version and use in any chess Desktop UI program - free or paid.
The "true value" of one's chess rating is the capacities in which the player in question operates, on average, in relation to others on a given platform. Secondly, it is deep within the realm of possibility that the ratings of chess.com bots are distorted; for example, I have defeated Li-BOT (2000 ELO, See My Profile) as a user rated significantly below 2000. Although Komodo may be a chess engine with a rating considered fairly accurate in relation to human players, it is simply not pragmatic to duplicate the ratings of humans in an engine. I, therefore, refute your above claims.
On chess.com, the bots don't have accurate ratings. Should players be able to play rated games to regulate the bot's ratings?