I’ve heard of something similar that is able to predict an effect of dark matter (the rotation curves), but AFAIK it couldn’t match other observations (bullet clusters, etc.) correctly.

Do you have a link for the model you’re talking about. I’m curious.