Here’s a blast from the past, something I wrote ten years ago on what I was looking for in a generic RPG system. I still agree with a good bit of it, though some of it I’m less certain about, and about one particular issue I think I was just wrong.
Subject: Desirable Generic RPG Qualities
Based on some of the recent discussion, here are some of my thoughts on qualities that I would like in generic RPG rules, broken down into the categories:
- Character Generation
- Character Advancement
- Task Resolution
Desirable Qualities by Category
It should be possible to go from a description of what the character is capable of to a codification of the character in game terms, without the system requiring modifications to the character to fit certain genres, power levels or preconceptions of the game designer as to what combinations/levels of ability/backgrounds are permissible. It should be possible to describe the character as it is now, without having to reconstruct the development or career path of the character up to this point (if you want to that’s a different story entirely).
Should have few, if any, subtle emergent properties. The obvious way to build a character should be just as useful/efficient as a more complex way. Character building expertise, rather than character description, shouldn’t be rewarded.
In a point-build system, prices should be based on relative utility of a power/level of skill/attribute, not based on rarity. Thus total points should represent how effective the character will be in the setting, not how unusual (although it’s reasonable to increase the price if rarity itself increases the utility, e.g. possession of psychic powers in a setting where nobody else knows they exist).
Levels of ability should have specific measures, so that it is possible to work backwards from real-world descriptions to ability levels. E.g. if you know that you want the character to be as strong as a weightlifter, and that a weightlifter can lift 1000 lbs, then it should be possible to work out in game terms what STR is required to lift 1000 lbs.
The system should be capable of making fine distinctions between similar skills/attributes/powers, without requiring them where unnecessary. E.g. it should be possible to build a character who is particularly good at endurance sports, without being particularly resistant to disease, without requiring every character to separately determine how good they are at endurance tasks and disease resistance.
The system should be able to handle a wide range of power levels and genres without breaking, even when the power levels are mixed in a single setting, and without rendering characters’ abilities at one end of the scale indistinguishable from each other or irrelevant.
(possibly w/optional random generation, but if so random generation should only come up with characters that are legal under deterministic generation)
The process of building a basic character should be short enough that you don’t have to cut corners to create an average (or even skilled) NPC, and require little math or extensive consultations of the rules. A spread-sheet or character generation program should be sheer overkill.
There should be a way to improve characters over the course of play
Preservative of niches
The system should preserve the relative rank order of specific abilities among characters, presuming equal initial talent and equal attention to advancement. I.e. if one character starts out more stealthy than another, or a better shot, it shouldn’t be possible for the less skilled character to overtake the more skilled one by accumulating equal experience, unless the more skilled one neglects to advance that skill, or was deliberately bought as less naturally talented at it.
Insensitive to timing
The system shouldn’t distinguish between character that have advanced through experience and characters that are simply created as being more experienced. Order that abilities are acquired/improved shouldn’t make a difference to outcome (possible exeption: abilities that improve the learning of new abilities).
Equivalent to training
Although for some fields, experience attainable through the school of hard knocks ought to translate to experience from adventuring, for many abilities non-adventuring time spent training or on the job ought to be treated equivalently, and the system should provide for it. E.g. it should be perfectly possible to design a bright NPC high-school student, calculate how much experience she would get from attending college, entering graduate school, completing her PhD, and spending twenty years as a professor, apply it to the character, and arrive at an expert in the field. (It should also be possible to simply buy an NPC as that in the first place, but that’s an issue for character generation.)
Adjustable level of detail
Ideally it should be possible to fill out interpretations of rules results to as much detail as is desirable, while not requiring that you generate more detail than you want at the moment. For instance, when determining hit location the rules should allow for anywhere from straight success/failure down to “you hit his left index finger” depending upon circumstances.
Gives results that can be interpreted in quantitative game-world terms. E.g. an attempt to throw an object as far as you can should return results that can be interpreted as a specific distance (whether it’s 1 meter, 1 kilometer, or 1 light-year), not “that was really far, but just short of extremely far”.
gives reasonable results at all power levels and combinations of power levels handles unlikely cases as well as likely ones.
Easy to extrapolate
It should be easy to intuit the probabilities of any simple course of action, given familiarity with the game system. (I.e. the player shouldn’t have to be an expert mathematician, or perform an elaborate calculation, in order to get a good sense of the chances of success that a character ought to be able to tell at a glance, such as whether a particular ditch can be easily jumped.) The system should have few, if any, strongly counter-intuitive properties (such as novices being just as good at defense as experts), and any such should be clearly labeled and justified.
Shouldn’t involve more math than the players can easily do in their heads, shouldn’t involve looking up rules except for the occasional truly obscure case (which ought to be easily interpolated from known cases anyway), shouldn’t take a long time even when doing simple math (e.g. adding 20d6 is, to my taste, too much)
To such an extent as is possible. Since different types of tasks sometimes require different levels of detail (even if the requirement is merely the desire of the players to have more detail), there may well be a limit to just how unified the mechanics can be and still satisfy.