It’s hard to get too far down the road of designing an RPG without discussing the action resolution mechanic. Having played almost exclusively D20 games, it’s hard to imagine using something else, but I’ve been spending a lot of time thinking about this lately.
I’m vaguely aware that debating the merits of linear vs non-linear dice resolution systems in certain corners of the internet is tantamount to feeding trolls. Conceptually, I like the idea of a non-linear resolution mechanic. Intuitively, it feels like the further a result is from the mean, the less likely it should be to occur. Also, when the resolution mechanic tends toward the mean, more emphasis gets placed on character ability. Linear resolutions though are more “cinematic” in that great success and failure is more common. Linear resolutions generally require just 1 die to roll, and most of the math associated with generating a bell-curve can be moved to the target numbers if needed.
d20: One issue I have with the D20 design is that the randomization element tends to overwhelm character ability at early levels. Take a look at Zak’s post on setting the dials which illustrates the point well. At first level, D&D 3.5 and Pathfinder look something like the following for a first level fighter:
So why stick with rolling a d20? It’s iconic, for one. It’s easy to roll and read, it’s just one die, and it makes the math pretty easy. Since it’s linear, it has the advantage of being a little more cinematic – great and poor results are just as likely as average results. If you are making a simple pass/fail test, then the bell curve isn’t as relevant anyway since extraordinary results aren’t any different from results close to the mean. It’s also a little more forgiving of encounter balance, since the randomization element plays such a large role.
d100: If you’re going to go with a linear resolution method, the d20 seems pretty good. The other option would be a percentile system. This has the advantage of being a little more straightforward about the math. Rolling it though requires two dice (admittedly a minor complaint) or waiting for this ridiculous thing to settle on a number. I don’t know that I need a system with that level of granularity in the core mechanic (am I ever going to use percentages more granular than 5%).
2d10: is either “the best of both worlds” or “an unfortunate compromise,” and to be honest I’m not quite sure which. The distribution isn’t quite a bell curve, but it is non-linear. It is somewhere in the middle of the realism of a bell curve and the cinematic even distribution. In addition, it has the nice benefit of keeping nearly the same range as the classic d20.
3d6: If you’re going to go for a bell curve, dungeons and dragons has been using a simple one since the beginning. 3d6 gives you a similar range to the d20, but provides a truer bell curve than 2d6. It has the advantage of using the most accessible dice. The downside is that you’ve got to roll three dice for every resolution and you’ve got to do some (admittedly simple) mental calculation every time.
FUDGE Dice Pool: Another nice bell curve is the system used in FUDGE and FATE and the derivative systems. This dice pool system is pretty simple, doesn’t involve much math, but does ask you to roll four! dice to determine the resolution of actions. You can use specially made FUDGE dice or standard d6s.
Fudge characters can also have Gifts and Faults, which are positive and negative traits that do not fit into the adjective scale.
Fudge uses customized “Fudge dice” which have an equal number of plus, minus and blank sides. A number of these dice are rolled, usually four at a time (“4dF” in Fudge dice notation), and for every plus side that comes up the result of using the Trait is considered one step higher (e.g. from Fair to Good) and for every minus side that comes up the result is considered one step lower. The goal is to match or surpass the difficulty level, also on the adjective scale, of the test. Thus, a Good attribute is considered to be Great if you were to roll two plus sides, one minus side, and one blank—the minus side cancels out one of the plus sides and the remaining plus side raises the result by one step. The same Good attribute would be considered Poor if you were to roll three minus sides and one blank. The same dice roll can be achieved with six-sided dice, treating a 1-2 as [-], a 3-4 as [ ] and a 5-6 as [+].
Dice Steps (d4, d6, d8, d10, d12, d20): I’m not quite sure what to call this method. Basically, as you increase in skill, rather than increase your modifier, you change the die you roll. Novices would roll a d4, and experts could roll a d12 or even d20. This mechanic reduces the math since skill changes the die rather than acting as another number to add. It also lets you use all of your dice, and nicely ties skill to the dice visually. The downsides: the more skill you have the wider range of variance in your results (not very intuitive). There’s also not much difference between one step to the next (until you get to a d20 that is).
Custom d20s: An idea I’ve been fiddling with is to use something of a combination of the stepped dice and a curve or pyramid distribution. My thought is to have a “novice” d20 that has a distribution from 1 to 10, with numbers in the center of the distribution appearing more frequently than numbers at either end. The next step up might include numbers from 1 to 15 with mid numbers appearing more frequently. The next step might be a standard d20. Moving up you might have an expert and master d20 that couldn’t roll lower than say a 5 for expert or 10 for master. The master d20 might even go to numbers above 20 (maybe in the 24-25 range).
For this to work, you’d almost have to have all “novice” dice one color, all “master” dice another color, etc. You could have odd faces painted one color and even faces painted another, and use the color result as a nice even/odd trigger (you could do this pretty easily anyway, but if you’re already having to create custom dice, you might as well right?). The downside is obvious, you have to spend money to create custom dice. Also, determining the exact distributions of numbers would have to be pretty much built into the math of the entire system, since it’s not easy to just swap them out later. The benefits are that you get to roll one die, the icon d20 at that, you get a visual representation of skill, and you can have a non-linear distribution.
Fate Deck (Cards): I’ve seen a few games now using a deck of cards as the core mechanic randomization element. The Runewars board game uses this mechanic, as does the skirmish game Malifaux. Can I really live with a core mechanic that is diceless?
Ok, so I would need some pretty good reasons to go this route. I think there are a few though. First, you have very fine control over the distribution, and since you’re drawing cards, you can have a deck of 50-80 cards before it starts getting too big. You could even model it after a standard deck of playing cards (although you’d be going linear in that case) like Malifaux does. Additionally, you can use the cards to generate other effects as well. You can randomize any number of elements with the exact same cards. In Runewars, for example, cards have numbers (approx 1-30), they have success/neutral/fail (non-linear results), and they have combat resolution damage for 4 different types of units (non-linear as well!). That’s some real efficiency that’s hard to argue with.
The down side? No more dice and a lot of shuffling unless you want card counters to start gaming the core mechanic. I suppose wear and tear could “mark” some of the cards, although you’d probably end up putting these cards in sleeves with the usage they’d get.
Resource Bidding: Give every player some “game currency.” Rather than randomizing for action resolution, blind bid the currency between the participants. I’ve never seen an RPG do this, but I’ve always been fascinated by the idea. You could have mechanics for restoring your game currency and monkeying with the bid process.
The first downside I can think of for this method is that players are going to take forever to decide what they want to bid, and with so many bids, I can see the game slowing to a crawl (maybe a shot clock here?). Secondly, this is getting pretty close to a story-telling game mechanic rather than a role-playing game mechanic. I really like the idea of resource bidding as a resolution mechanic, but I may use it for something other than an RPG.
In summary, for the core mechanic of Lost Worlds I’m considering the following:
Non-Linear (Bell Curve-ish)
- FUDGE/FATE Dice Pool
- Dice Steps: d4, d6, d8, d10, d12, d20 based on mastery
- Custom D20s
- Cards (Fate Deck)
- Resource Bidding