>"proportional to mv²" also means "proportional to 1/2 mv²" and "proportional to 123.45 mv²".

Yes. This is my point. The error has never been found experimentally because everything eventually gets resolved to Newton's E = Fd. Dividing every result by 2 makes no difference because the proportions do not change.

I still don't see why gravitational mass is a problem but I'll keep looking into it.