Designing Roleplaying Games: Setting First, or Mechanics First?

I have spent a few months absorbing media on the topic of designing roleplaying games, and reading a wide range of published roleplaying games. I have been trying to figure out whether I should focus on:

1. Designing a full setting
2. Designing a complete game system worth of mechanics
3. Designing both setting and mechanics.

A discussion thread over at rpg.net has been helpful here. Its something of a chicken and egg question, as design is an iterative process that can go back and forth. But one of the answers suggested focusing on what you think your strongest idea is. At the moment I lean towards thinking I have better setting ideas, which suggests building off an existing rule set would be more likely to lead to success than trying to build a new game system. If I can express my ideas within an existing toolkit, is there a good reason to reinvent the wheel?

Mechanics

My goals for mechanics include:

1. Similar levels of complexity for social and exploration challenges, not just combat challenges. This reflects my GM preferences.
2. Some form of temptation/corruption system that acts as a conflict gauge (and my idea here is that a tension between three points in the game system is potentially more interesting than tension between two points as happens with Call of Cthulhu’s Sanity/Mythos Lore trade-off).
3. Highlighting the group as being more than just the sum of its parts, and links back to temptation to use group resources with a “ask for forgiveness, not for permission” rule.
4. Fantasy genre emulation – magic as something that escalates during a scene, not something that opens with a nuke. I have been thinking a lot about how the escalation die is used in 13th Age.
5. Fantasy genre emulation – splitting the rewards of success (gain in power) from the benefits of experience (improvement in skills). Rewards tend to strongly influence player behaviour in campaign games.

My observation from reading through a number of Indie designs, is that while they often have an innovative mechanic that works well for the published scenario, its often quite limited in what it can do outside of that scenario. Good for one-off convention play, not good for on-going campaign play. Every now and then, however, some new design manages to scratch a particular gaming itch, and ends up being hacked and ported into a wide number of settings. Apocalypse World seems to be flavour of the month in that regard.

Rolling Dice

I did some mucking around, trying to come up with a new way of rolling dice. I think that most of the easily discoverable ways of rolling and interpreting dice have been found and expressed in an existing roleplaying game system. The best articles I found on dice mechanics were at Darkshire, an rpg wiki, and RPG Stack Exchange.

One thing I found, was that I would often scribble a way of rolling dice that had occurred to me, and then a short time later I would read a description of that dice system in a published roleplaying game (the ways a d6 and a d12 are used in The One Ring was a prime example of this).

In terms of trying to analyse whether a particular way of rolling dice is any good, my take home lessons form this research were:

1. How easy is it to determine success/failure? This can also affect how transparent the system is.
2. What is the balance between random numbers and character skill?
3. What is the granuality of the system? Fine grain = 1d100, coarse grain = 1d4.
4. Availability: d6s beat a set of polyhedral dice, which in turn beat custom dice that only work with one game.
5. For ease and speed of use: Comparison > Addition > Subtraction > Multiplication > Division.
6. Adding more dice, more operations, and/or bigger numbers all slow play down.

Because I am thinking about mechanics involving temptation and shared resources, I thought delving a little more deeply into dice pool mechanics would be a good idea. This is because a bonus from a temptation can be easily represented by adding an extra die to a pool, which may be easier in play than changing difficulty target numbers or choosing to tweak the dice results after they have been rolled.

Dice Pools

The Main forms of dice pool seem to be:
1. Roll and add numbers
2. Roll and count successes
3. Roll and match patterns.

Rolling a bunch of dice can be fun, but there are some problems with dice pools:

1. Expertise does not scale well, the marginal utility of each extra die is small, and large numbers of dice may be needed to accurately model an increase in skill.
2. Too many dice can be a problem, five dice in a hand are okay, thirty-five dice not so much.
3. If using a step die scale (d4, d6, d8, d10, d12) you are locked into a coarse gain five tier system, unless you use weird dice (d14, d16, etc) or odd combinations (d8, d10, d12, d10+d4).
4. To adjust task difficulty you should either change the number of dice, or the target number, not both. Doing both makes the system too complicated for mortal minds to handle.

Of the different Dice Pool systems I looked at, the Cortex Plus approach (roll 3+ dice, keep best two) seemed a little easier than Savage Worlds (which has exploding dice). Cortex Plus felt a lot like the FATE game system, only with the full range of polyhedral dice rather than the +/- Fudge dice. Some other nice variants on dice pools that I read about were:

1. Roll and Pluck: take dice out of your pool and allocate them to the tasks at hand (such as initiative, defence, attack, and movement).
2. Momentum: success can carry over to future rounds (in the Firefly version of Cortex Plus, a big success allowed you to bank a “Big Damn Hero” die for future use).
3. Stake dice: in Houses of the Blooded you can improve the outcome of a success, by removing dice from the pool before they are rolled.
4. Floating dice: in Mythender successful dice (rolls of 4-6 in a d6 system) convert into a different tier of dice that allow different powers to be used, or to be exchanged for other game resources.

I think that is enough for this post. I’ll try posting some thoughts on trying to find an existing ruleset that matches my setting ideas later this week.