I’m an RPG hacker, tinkerer, and designer. Most RPG systems have at least some appeal to me, and tons of systems have bits of them that are my favorite implementation of that bit. On top of that, even clearly flawed systems are often my right-now favorite while I’m playing them. So picking a favorite RPG system is quite possibly even harder than picking a favorite RPG. I’ll exclude anything I’ve designed, just on principle, and try to narrow it down at least.
A couple of my favorite games are such in part due to their systems. I’ve used the core of Ars Magica to build a general fantasy game, but I think most of what makes it appealing to me is specific to the setting—once you strip all that away, it’s a pretty simple system, which is mostly good, but also doesn’t make it very distinctive. Underground is a more novel core system—a human-scaled version of the system from the old DC Heroes game. It breaks DC Heroes’ lovely 3×3 stat symmetry (three each of mental, physical, and social/spiritual, each split into power, finesse, and resistance), but is otherwise an excellent refinement and expansion, turning it into a very elegant gamist/simulationist system that handles a wide range of numbers and facilitates easy conversion between units (i.e., from speed to distance or strength to mass, and the like).
Aria is another old system, quite crunchy and detailed, that I continue to love for the clevernesses in how it all comes together. A lot of people found the book “pretentious” due to its insistence on eschewing the traditional RPG jargon in favor of using standard dictionary words in accordance with their conventional definitions, but once you get past that to the underlying system, it is one of the better systems I’ve ever encountered. And it’s magic-building rules are probably the only thing to out-do both The Primal Order and Ars Magica.
Though these days, if I wanted a crunchy system, it would probably be Burning Wheel. It’s probably the only crunchy system I’ve actually enjoyed running. While it has a lot of detail, it all falls together very elegantly, all makes sense, and the numbers work out mathematically, do what they claim to do, and both make sense and are fun in play for the most part. No boring can’t-lose fights; no whiff-filled endless challenges. If I ever run it again, I’d like to try some sort of abstracted system for wises, like the Circles system, instead of tracking them individually, but that’s my only complaint. And the way all the bits interact just perfectly is amazing, and makes running and playing it fun. It’s probably the only game I’ve found that could handle a Cyrano-style simultaneous duel of wits and rapiers, and make both halves equally interesting and equally fraught with peril.
Generally, however, I go in for less-detailed systems. For a long time, Story Engine (originally used in Maelstrom Storytelling) was my favorite system. It uses an interesting structure, focused on scenes and narrative control rather than either simulation or difficulty, and pioneered a lot of the play style that was reinvented by the rise of Narrativist indie RPGs nearly a decade later. It’s also unusual in being structured such that cooperative rolls are the norm—typical difficulties are out of reach of a typical character acting by themselves.
In that vein, Primetime Adventures and The Shadow of Yesterday (TSoY) might be the best examples yet of two different developments of that Narrativist structure. I prefer Primetime Adventures, which is about as far as a system can go in focusing on authorial control while still having any character-differentiating mechanics at all. It’s the game system I choose when I have an interesting setting in mind and don’t want to create my own rules—it has replaced CORPS, Fudge, and Savage Worlds for me as my default generic system.
But I have to admire the amazing balancing act TSoY pulls off, sneaking authorial directives into its Keys under the guise of character motivations, and basing rolls on character capabilities but using the results of those rolls to determine narrative control rather than success. It’s the game system we just chose when we were playing something else and loved the game but the rules were not gelling for us, and it has been used as the basis for more than a few other games, most notably Lady Blackbird.
I also need to mention two systems that are amazing, but a bit tied to their settings. Zero has a very clever system that inherently trades off specialization and breadth: by using a fixed skill list of just the right length, the d6 x d6 roll perfectly scales such that the more skills you have the less good you are at each of them, without needing to fiddle with skill points or any other bookkeeping. While other parts are specific to its setting, that core is something I’ve reused elsewhere, because there’s really nothing else that does that so smoothly.
Meanwhile, Everway is simply in a world of its own. Every part of that system is awesome, from chargen based on inspirational images, through metaphorical primary traits, to the best way yet to quantify an infinite variety of special powers. [Is it frequently useful? versatile? powerful? Total the yeses up and you have your cost.] I’ve adapted it for space opera, and others have adapted it for supers, and I’m sure it could form the core of any number of other fantastical (as opposed to more realistic) settings, though with a little more work than for adapting a more traditional system.
But if I had to pick just one system to use from now on, I think it would be Over the Edge. For me, it’s the perfect balance of words (mostly) and numbers (just a few). It’s focused on character capability, but boils everything down to a small number of stats, with an easy default roll built in. The way that it uses dice is also genius, providing bonuses and penalties that don’t change the range of the roll and don’t require complex—or really any—math (beyond the basic addition you need for an unmodified roll). It’s much like TSoY in feel and goals, though without the excellent Keys mechanics, but simpler and more free-form, so easier to adapt.