non-distributed version control systems seem so much simpler
I find quite the opposite. The simplest case is one user, and a "distributed" VCS is clearly the easiest option in that case--no central repository needed, no environment variables to set, or separate paths to worry about. Just say "init", and you're off and running. (At least with Mercurial or Git, the two DVCSes I have experience with.)
With more than one user, it's slightly more complicated, but not enough to worry about. It all boils down to the distinction between "save this change" and "share my changes with my co-workers". Having those as separate commands really isn't that confusing, and once you're used to it (which should not take long), you'll have a hard time remembering how you lived without it! And that really is the entire difference, fundamentally, between distributed and non-distributed VCSes.
(Most of the things that are great about Git are unrelated to the distributed/non-distributed aspect, or at best tangential to it. For me, the big wins of either Git or Mercurial over, say, Subversion, are how much better/faster/easier/more powerful branching is, which doesn't really have to do with being distributed or not, and how much faster the whole thing is, overall, without all those network round-trips, which does.)
I started out somewhat skeptical, like you, but after my first pilot project, using Mercurial, I was a complete convert! YMMV but it Works For Me(tm)! :)