I look back to where this started, and this started with Fritz Zwicky noticing that galaxies and galaxy clusters are moving faster than expected. This is the root observation.

From there, a good question is "How do we model it?". And sqrt(GM/R^2 * cH/(2pi)) works really well.

By the way, cH/(2pi) = c^2 / (2pi * 13.8 bly), which looks like centripetal acceleration.