> When a game is over it might say -8 to the human, but if you look at the rating itself, it remains unchanged.
Well, I have lost some bullet rating points to Leela for real. Maybe there are some other conditions where this is not true.
@Wolfram_EP thats was 4 days ago.
Try it again. Today my rating don't change after playing against a BOT.
Seems like they changed this then, cause 4 days ago, I played 1 Rated UltraBullet game against BOT UltraIngot and lost 17 points (1741 to 1724)... every other game I played against BOTs were casual cause I didn't wanna lose any more rating against them...
If our ratings won't change anymore after playing Rated games against BOTs, I definitely support this!
Yeah, playing Ultrabullet against bots is a bad idea. Unless you're Penguin of course.
If ratings don't change from rated games, people won't try to win.
Ok, great. So developers have already noticed the danger and addressed it.
The problem of changing the human ratings is that once one human learns how to beat some high rated non-learning computer, he may then repeat it consistently using same opening and strategy, pushing himself to another level. Other humans may then eventually reach that same level by playing against him. I believe this is how a few GMs got such a high rating in ICC. Reaching 1600 blitz there is much harder than here, yet top players are always above 3600.
I agree there is potential for abusing the system, but making all games against bots casual for humans will make it much less likely for high rated players to challenge bots, thus making the whole lichess bot API less important, - bots can already play other bots in GUIs and bot tournaments, perhaps allowing rating change for select neural network would be a good idea since they play more human like, but since it appears this was already once changed to make it forced casual then I guess the issue won't be addressed any more
for example I just learned that rating doesn't change for me against bots and it made me like four times less interested in playing against them - there is little that I gain in drawing against them, I can just play them offline if I wish, there are things that I think are skewing ratings far more than bots ever could, like 0.5 and 0+4 both being bullets, berserk mode in tournaments (so many have lost all their rating this way) etc.
I disagree that there's any more potential for abuse against bots than there is against willing human opponents; and safeguards are already in place to both report and detect boosting or sandbagging.
whether there is potential or is not this is sad for the whole bot integration, I am sure players especially high rated like 2300+ who have actually taken effort to get high rating will quickly lose interest and what we will see is mostly 1500ish players with almost 100% predictable results, people won't even watch bot games anymore when that happens - having Leela AI chess play against humans was one of best things lichess ever had, sad that it will probably come to a halt now just because some were worried in-advance about abuse potential, I think we should have waited to see if there is any real cases of rampant abuse before taking such steps, but again there will probably no longer be any discussion on the dev level since the decision has been made
You're suggesting that players will blatantly violate Lichess ToS & not be playbanned, and that bots won't be fixed (or punished), both of which seem implausible.
