Apropos of a previous post, Josh W asks: How often do you find yourself budding systems off and turning them into other distinct games? If it happens often enough, how do you deduce the design through-line that subsystem fits?
For reasons implicit in your second question, this happens to me exactly never. I’ve certainly created sub-systems only to realize that they don’t fit the design through-line. Usually they interpret the design goals too literally, or are disappear down a rabbit-hole of unnecessary simulation. Because of this, they are also usually flawed in and of themselves and thus go to the great discard pile in the sky.
Never have I accidentally created a rules subset that would work fine, if only I were working on a different game. In my view, building a game and its core activity around a nifty rules idea is a recipe for disaster. Start with the activity you think will make for engaging play, then facilitate that with rules that enable it in as simple and elegant a manner as possible.
Sometimes I do encounter games that feel like they sparked first from a system idea, with the core activity yoked in to serve it. If it even has a clear core activity. When you ask a designer to describe his game and he starts with a game mechanic, that’s a warning sign of a game designed from an abstract rules concept outwards. (Or maybe he’s just lousy at pitching.) For a few categories, like abstract games, design from sub-system out may be desirable. Often, though, you'll get the feeling during play that you are working for the rules rather than having them work for you.