There is a common trope that says that a the difference between a 2700 player and a 2300 player is much bigger than the difference between a 1400 player and a 1000 player. I personally think it's horseshit and I want to prove things either way.
I, and the people I have in mind who perpetuate the trope, are not talking about the amount of work or genius it takes to go from one rating to another, thy are genuinely saying the 1000 player will win 1 time out of ten against a 1400, but the 2300 will simply never ever win against a 2700. I on the other hand, put my thrust in the Elo rating system, and a 400 point difference is a 400 point difference, the weak player is expected to win 8 out of 100 times either way, or they wouldn't be at the rating they are. Mind you, I (and the rating system) counts draws as half wins, so the 2300 player only really need to draw 16 games out of 100.
Now there is a simple way to test who is wrong. Look very strong player and see whether, on average, they win points against much weaker opponent. Any rating system isn't perfect, and tend to lose in precision as the difference between player grow, so it might very well be the case that 2700 seldom ever lose to 2300. But if that is the case, since they'd win more than the system expected, they should win rating when playing against much weaker player. If they do, I was wrong, if they don't, I am right.
Which is why I am looking for profiles with public insight available, or even just a screenshot of the preset "do I win more rating points against weaker or stronger opponent". High ranking preferred, but any welcome. If for example everybody gains a bit more Elo against weaker player, it's also valuable information, it'd mean it's not just a high Elo thing. High Elo or high Glicko, the interpretations should be the same.
Alternatively, if someone know of an app which can generate this kind of insight, I am also interested, because for some reason, while the default setting for insight is private, games played are public and anybody can mine the games of anyone. Fortunately, I a competent enough to bake my own script and generate that insight myself, but since it's already been invented, I'd rather borrow someone's wheel rather than make my own.
I, and the people I have in mind who perpetuate the trope, are not talking about the amount of work or genius it takes to go from one rating to another, thy are genuinely saying the 1000 player will win 1 time out of ten against a 1400, but the 2300 will simply never ever win against a 2700. I on the other hand, put my thrust in the Elo rating system, and a 400 point difference is a 400 point difference, the weak player is expected to win 8 out of 100 times either way, or they wouldn't be at the rating they are. Mind you, I (and the rating system) counts draws as half wins, so the 2300 player only really need to draw 16 games out of 100.
Now there is a simple way to test who is wrong. Look very strong player and see whether, on average, they win points against much weaker opponent. Any rating system isn't perfect, and tend to lose in precision as the difference between player grow, so it might very well be the case that 2700 seldom ever lose to 2300. But if that is the case, since they'd win more than the system expected, they should win rating when playing against much weaker player. If they do, I was wrong, if they don't, I am right.
Which is why I am looking for profiles with public insight available, or even just a screenshot of the preset "do I win more rating points against weaker or stronger opponent". High ranking preferred, but any welcome. If for example everybody gains a bit more Elo against weaker player, it's also valuable information, it'd mean it's not just a high Elo thing. High Elo or high Glicko, the interpretations should be the same.
Alternatively, if someone know of an app which can generate this kind of insight, I am also interested, because for some reason, while the default setting for insight is private, games played are public and anybody can mine the games of anyone. Fortunately, I a competent enough to bake my own script and generate that insight myself, but since it's already been invented, I'd rather borrow someone's wheel rather than make my own.