* If tomatometer & audience score are within 5% of each other, you can trust the ratings to give you a decent indiciation of movie quality.
* If tomatometer is more than 15%+ higher than audience score, it means it's an artsy fartsy movie that critics like and movies don't.
* If audience score is 15%+ higher than tomatometer, it's a fun movie even if it's not oscar worthy. (https://www.rottentomatoes.com/m/old_school is a perfect example)
Lost Highway (1997) - 68% Tomatometer - 87% Audience Score
Fight Club (1999) - 79% Tomatometer - 96% Audience Score
American Psycho (2000) - 68% Tomatometer - 85% Audience Score
Requiem for a Dream (2000) - 78% Tomatometer - 93% Audience Score
Dancer in the Dark (2000) - 69% Tomatometer - 91% Audience Score
Oldboy (2003) - 82% Tomatometer - 94% Audience Score
The Prestige (2006) - 77% Tomatometer - 92% Audience Score
Joker (2019) - 69% Tomatometer - 88% Audience Score
A lot of the old movies you picked are famous and popular in movie pop culture. Audience scoring this in RT probably went out of their way to watch these films, they are not as organic as recent scores as you have a larger number audience scores created by movie lovers.
If you find examples post 2015 when RT became a mainstream scoring system that would be great.
Only movie that's current in your list is "The Joker" which among critics is considered to be a copycat of other critically acclaimed films (taxi driver, the comedian). This is a film that tried hard to look artsy fartsy but was not.