I've started working on a new game, and one of my goals is to make it as light as possible to play without players really noticing that there's very little going on under the hood.
One result of this is a hybrid attribute system, which is my attempt to take a design that I'd normally build across three or four design layers and instead build into a two-layer design.
This will not be the lightest game ever, but it is going to be light for how it feels (if I succeed).
I base my dice mechanics around the idea that you want to have almost any conceivable event handled by a single roll of the dice. I don't like hybrid systems, and while I'm fine with pool-based systems where players roll varying amounts of dice I want to go with something that's super fast at the table.
I ruled out pools for this game because there's really two common ways to handle pools:
I've done a lot of the first type of pooled dice systems, and I'm frankly just a little bored with them. My last game I did was for the CODECO game design competition and it used a pool based system (and took the Best Overall Game award!) so I'm ready for something new.
Plus, the probability curve with thresholded success and added dice can be a little screwy in some edge-cases.
The second method, dice pool with the added sum results, isn't something I object to on principle (the WEG D6 lines and Open Legend both use this method, albeit with radically different dice), but I feel like there's a room to go lighter than that without losing the full feel of a game.
Also, you never get bogged down with the "which dice do I roll?" problem. I'm tolerant of a little run-time addition, since most players can handle it easily, but for a lite game that goes above its weight class I don't think it adds more than the added weight of handling multiple dice during play as a practical measure.
I want people to argue over whether they're playing a delightfully simple game or a refreshingly deep game. Counting up dice gives the latter a clear win.
Also, I'm using a system that already involves math, since players add up multiple attributes, a potential special ability bonus, and a potential difficulty modifier to even attempt a roll.
This isn't so different from most systems, but it's not a fluid part of the game and the whole point here is to create a game where characters feel distinctive and there's room to grow and develop them without forcing players to pick up a lot of rules or bear a lot of bookkeeping burden at the table.
Because even though increments of 5% are not terribly granular, they're good enough for storytelling work.
Also, because I've seen wonderful games use a single die to great effect. Symbaroum sticks out as a game that bears a lot of weight on its d20, but they don't stick to it with as much fervor as I am.
However, one way in which I'm going for a light approach is with our resources. Because these are in "check-box" size ranges (e.g. 3-6), there's never going to be a time in which I would have players, say, roll for d6 damage to see if they lose a variable amount of HP. Each unit of a resource is too valuable to be left to chance.
I'm not a fan of such things in heavier systems, preferring flat numbers with margins of success to change results, but there's something to random results in certain weight classes (or if randomness is desirable; I always prefer speed to randomness except for basic success/failure ambiguity).
I like d100 (my game Hammercalled is built on it), but it's something that I find more useful in a heavier system where you've got mechanics to interact with the granularity.
Theoretically there could be a cognitive study on how adding up the d00 and d10 work, but I think that's probably a moot point in the grand scheme of things.
D&D is probably the best known d20-based game and uses a comparison between characters' rolls (modified by their ability) and a flat difficulty number.
There are a few things about this that have some value, the most obvious being that high results are "good" in most players' minds as part of a psychological phenomenon that I'm pretty sure exists but I'm too lazy to investigate.
However, this is working with two numbers when one would suffice.
For this system I've got known inputs, namely attributes. I'm not capping attributes arbitrarily (except in character creation, because characters should suffer at the start of their journey), but I can manipulate the expected range of values.
Players will always add two attributes. On a d20, you want to give plenty of room for movement to let player characters feel powerful while still giving the GM the ability to define threats as dangerous or easy.
Characters get attributes that range from 3 to 7 during character creation (special attributes start at 0, but they're special so I'm not paying their potential deficiency too much heed).
This means that the worst-case combination is 6 (30% success with no modifiers), and the best-case combination is 14 (70% success with no modifiers). This seems acceptable to me, especially since players might conceivably get up to a +3 bonus from special abilities during character creation. Difficulty modifiers are unlimited.
I want to make sure that the characters have perceived room to grow from this point without being too weak.
I'm going to use a roll-under system for my central mechanic. The primary advantage here is crunching everything down to a single number (because you add up the stuff for your threshold and then roll under that). It's free of any magic numbers and other weird things that require special explanation and add undue weight to the game.
I'm fond of using blackjack results, with the higher numbers that don't go over the threshold giving special bonuses, but that will not be a major part of this game because it's not built to function at that level of granularity.
The other neat thing here is that I can have an over-threshold value, in this case five points above the threshold, for special effects.
That's a failure by the central mechanic logic, but it's a potential success with consequences.
This system is actually part of the core system design, and ties into the character resources. If the GM can't think of an easy consequence, the players can spend one of their resources to succeed.
Going with a single d20 roll-under dice mechanic gives me a skeleton that's fast and fluid that I can layer individual elements on.
When you're going for a system that you want to explain to a small child in a minute but lure in hardcore mechanic wonks with, having this central element of the game be streamlined and dynamic is a great idea.
I'm a little leery of the over-threshold effect, but that's something that I might spin out of the core mechanics section and instead make a special rule (basically "you can spend resources to recover from a failure so long as you're not more than X over the number") and which might actually serve as a function for game mechanics to build upon.
