comparing lichess vs chessc*m

[nicbentulan]

actually, a huge part of why i wanted to see the rating correlations in the 1st place is comparing the 2 sites. i think lichess has a bigger gap.

1 - in lichess, everyone starts at 1500 in chess960 regardless of their chess ratings.

1.1 - btw, this allows me a loophole that i call 'farmbitrage' ( look it up XD ). i just challenge people who are like 1000 and have never played chess960 and then get to 2000 easily. lol.

2 - in chessc*m, your starting rating depends on your other chess ratings. if your (bullet,blitz,rapid) = (850,920,1410) and you just start playing chess960 and your 1st game is blitz, then you'll be 920.

this makes me think that the gap (either absolute difference or percentage difference) in lichess will be larger than the gap in chesscm and thus there are more underrated problems in lichess than in chesscm eg if in lichess i'm 1720 blitz chess & 1540 blitz chess960 and then i play a 1510 blitz chess960, then they might be like 1910 blitz chess. - so for me who plays mainly chess960 and thus has a 'proper rating', i'm matched up against a lot of improperly rated players.

question:

apart from downloading the data sets and just getting computing these differences myself, is there a way to see this or something like this in your programme? eg idk maybe comparing the R squared or std error would tell us something?