It's a compatibilities / sensibilities approach. SysV-style init has been around for ages; it's what all of us old-school admins know; it is also a brittle, loathsome mess. However, systemd, for all of its perks, may not yet be ready for prime-time (I'm talking production environments - I have yet to implement a switch to it in such an environment, although I'm considering doing so at my upcoming job). So, they compromise. Hell, for a long time there was a compromise built into some distros and other UNIX-like systems, which would use either have a symlink named telinit that linked to init, with the capabilities of acting like telinit built in, or a standalone telinit that also had its utility built in to a standalone init. A lot of this double-duty crap is fairly common at the legacy systems level, because it is foundational and expected to work a certain way by a great number of stodgy admins with huge farms of established systems.
I would suggest you learn everything you can about systemd, since that is the future of stable production systems, and use it. Forget SysV - the reason it is so simple to set up and use is because it has to be. Exploiting the backwards compatibility of systemctl by calling it as telinit is only masking the complexity of what is really going on now.
/rant
http://0pointer.de/blog/projects/why.html
