Forgive my ignorance. I don't understand the threat model being worried about here. If someone breaks into your CVS/SVN server and rewrites project history, h

Forgive my ignorance. I don't understand the threat model being worried about here. If someone breaks into your CVS/SVN server and rewrites project history, how is that less bad than someone breaking into your remote git repo and rewriting project history? Don't both attacks require a break-in at the server/remote? Or does CVS/SVN have a better method of detecting such a break-in?