@likeyacutg said in #18:
> @OctoPinky I agree, and I think the rating system may break down when the rating deficit is too high, as often the lower rated player plays worse than they normally would because they are nervous to play such a strong opponent, and makes farming easier because most don't play at their true strength when faced against someone 600 points higher.
In fact, in OTB chess it works the other way around, higher rated players tend to lose rating to lower rated opponents on the whole. Some underdogs might be demotivated, others especially motivated, some are simply unrated. Very high rated players rather tend to be slightly overrated. It would be surprising if it worked differently here. You can make an experiment, announce its start in advance and then play 30-50 games against opponents rated 400-600 points lower than you. (Or play Lichess Bundesliga three times for a team of some of your friends in a lower division, if Lichess Bundesliga takes place in acceptable time for you. Or join a bigger arena long after its start.) My guess is that one is more likely to lose rating than to gain it, at least in the long run. Moreover, it's more interesting to play against similarly rated players, so the only reason for such experiments is to demonstrate that gaining rating this way is very difficult and not worth it.
If a player A plays many games against lower rated opponents B, C, D and keeps winning almost all the time (which indeed happens on Lichess), it doesn't mean that it'll work against lower rated opponents in general.
Look at this article, based on reliable data: www.chess.com/news/view/fide-mathematician-proposes-changes-to-improve-rating-accuracy
There're other articles by Jeff Sonas on this topic. The fact that high rated players were losing a lot of rating to lower rated opponents was the main argument for raising the lowest ELO from 1000 to 1400.
> @OctoPinky I agree, and I think the rating system may break down when the rating deficit is too high, as often the lower rated player plays worse than they normally would because they are nervous to play such a strong opponent, and makes farming easier because most don't play at their true strength when faced against someone 600 points higher.
In fact, in OTB chess it works the other way around, higher rated players tend to lose rating to lower rated opponents on the whole. Some underdogs might be demotivated, others especially motivated, some are simply unrated. Very high rated players rather tend to be slightly overrated. It would be surprising if it worked differently here. You can make an experiment, announce its start in advance and then play 30-50 games against opponents rated 400-600 points lower than you. (Or play Lichess Bundesliga three times for a team of some of your friends in a lower division, if Lichess Bundesliga takes place in acceptable time for you. Or join a bigger arena long after its start.) My guess is that one is more likely to lose rating than to gain it, at least in the long run. Moreover, it's more interesting to play against similarly rated players, so the only reason for such experiments is to demonstrate that gaining rating this way is very difficult and not worth it.
If a player A plays many games against lower rated opponents B, C, D and keeps winning almost all the time (which indeed happens on Lichess), it doesn't mean that it'll work against lower rated opponents in general.
Look at this article, based on reliable data: www.chess.com/news/view/fide-mathematician-proposes-changes-to-improve-rating-accuracy
There're other articles by Jeff Sonas on this topic. The fact that high rated players were losing a lot of rating to lower rated opponents was the main argument for raising the lowest ELO from 1000 to 1400.
