Once upon a time, when my only RPG experiences were fairly crunchy games and my own campaigns were heavy Sims, I believed that there was a spectrum of design when it came to mechanics, and that spectrum ran from "playability" to "specificity."
I was wrong.
Let me clarify; The spectrum described above definitely exists, but it does not exist across all games. It exists across each game, considered as an act of design, all by itself. It makes sense to compare your first draft with your second draft, in terms of this spectrum. But it does not make sense to compare two different games along this spectrum (unless they use the exact same rules, which would make the exercise pointless anyway).
There is another spectrum, however, that I want to talk about. This one actually does exist across all games; it represents a heterodoxy of design approaches, and it's much more interesting.
I don't have a good word for it. Maybe something having to do with Empirical Agency. But basically, this spectrum positions a game mechanic in terms of where the empirical* detail comes from: i.e. the system on one end, or the humans (either GMs or Players) on the other. Note that this is not a binary, and it might have different values for different subsystem: most "crunchy" games handle different types of activity or challenges at different places on this spectrum.
* Fictioneers will of course understand that when I say "empirical" in this sense, I'm talking about the game world, not the one where my character sheet has a coffee stain on it.
Over the years, philosophically, I began thinking of interactive fiction in so many ways, working with and emulating so many stories, genres, and media forms, I came to a place where - let's face it - a Difficulty is a Difficulty regardless of its sphere, and a small handful of basic dynamic structures can be seen to lie beneath all interactions (just as the "Hegelian Dynamic" can be seen in all manner of interactions between creatures, objects, forces, and reality). I began looking at things like Task Resolution Systems in a much more abstract way, like these intangible dynamic structures. Deleuze would call these "abstract machines," but we call them "game mechanical subsystems." These machines can be super-complex, handing up large amounts of highly specific detail with different subsystems for different types of situations, or very simple, maybe even "generic," governing every type of situation with a single machine and handing up only a "Yes" or "No" ... or anywhere in between (most are in between). But in an abstract sense they are all alike: these little structures run their algorithms (using dice or whatever) and then hand up a result including X amount of empirical detail, which then needs to be interpreted in terms of the current fictional situation.
Anyway, my position today is that the output from these machines only needs enough diversity and specificity to permit (and suggest) logical narrative interpretations with a minimum of ambiguity. And further: if your playstyle is flexible enough, ambiguity can be seen as not a problem but an opportunity for injecting narrative detail that the system could not have possibly foreseen; content that perfectly suits the world, the situation, the narrative, and the die roll.
On this view, the complexity of your mechanics should be just enough to prompt the creation of details that aren't so big they strain your creative powers, so vague they can be easily misinterpreted, or so complex that they take an undue amount of time.
All of those values, however (comfortable creative grainsize, narrative authority, and time) are totally subjective, and they provide different sorts of experiences. They may even differ within the same individual under different circumstances, when playing different games, or at different times in their life.
So. Yeah. There's something to think about.