This section is basically about how we write up characters in RPGs, and how designers create processes for doing that. Some parts of this chapter aren’t quite there yet, but overall I think it’s coming together pretty well.
Given that the word “role” is in the name, we can assume that an RPG calls for having defined characters. The question is how we go about providing a way for players to effectively define their characters. It’s in the nature of RPGs that character creation is generally a combination of dreaming up a fictional character and assembling an abstract mechanical construct. These two aspects of RPG characters necessarily interrelate and have decidedly blurry edges.
Statistics
Most RPGs give characters some kind of numerical values, variously called stats, attributes, ability scores, etc., to represent how effective they are in various areas. Some games make the actual numbers for characters’ stats substantially larger or smaller than others. In my mind this is another case where designers have been a little too willing to base things on tradition and the apparent feel of things rather than what function they serve at the game table. D&D traditionally had stats that ranged from 3 to 18 (coming from a 3d6 roll), but in earlier editions they didn’t distinguish characters mechanically all that much, and in later editions the ability score modifier mechanic meant that they were a step removed from the smaller numbers you actually used in play.[1]
I won’t say that larger numbers are never justified, but it’s entirely possible to create a range of 4 to 6 values that provide sufficient distinctions between characters of different power levels, especially when you couple them with other traits that make characters potent in ways that run orthogonal to the raw numbers. Superheroes have such a huge diversity of abilities that superhero RPGs often have a massive range of numerical values. TSR’s Marvel game had a chart of values ranging from Feeble (2) to Unearthly (100), and then special values going up as high as 5,000, and then the “Beyond” level with an infinity symbol next to it. Mayfair Games’ DC Heroes RPG went as far as to use a logarithmic scale for stats, so that each stat level is twice as powerful as the previous one. In stark contrast, the more recent Marvel Heroic Roleplaying from Margaret Weis Productions rates traits by die type, so that the only numerical values for traits are d4, d6, d8, d10, and d12. It’s counterintuitive to have such small numerical distinctions between the likes of the Incredible Hulk and Black Widow, but in terms of fostering action like in a comic book, the game is a rousing success. This is partly because of the range of other traits that exist in the game, and partly because in superhero comic action, characters are often able to adapt whatever superpowers or gadgets they have to the particular situation. It’s an approach that wouldn’t work in every genre, but for an RPG intended to simulate Marvel comic books it’s pretty much ideal.
While it’s certainly a plus for them to be intuitively readable, to me the more important thing with character stats is how they affect the actual rules. That in turn means that the range of numbers they have depends on the mechanics they’re meant to tie into. Apocalypse World treats stats as small adjustments to a tightly-defined 2d6 roll, so while having starting characters’ stats top out at +2 or +3 might seem stingy if you’re used to other RPGs, it actually confers a huge advantage in the game. It also deliberately steps away from stats as indicators for a “physics engine,” as stats like Hard and Cool don’t need to correlate to how many pounds a character can bench press.
It’s notable that from 3rd Edition on, D&D’s ability scores mainly serve as a basis for deriving a modifier, typically from +0 to +4, so that in practical terms, for player characters the game has roughly 5 possible values for each stat, which happens to be the same as Apocalypse World.
Soft Distinctions
“Soft distinctions” are what I’m calling things that you write on the character sheet that don’t involve game mechanics. Some are just plain obvious, like giving a character a name or describing what they look like, while others can help reinforce the game’s specific themes. For example, in Don’t Rest Your Head, part of character creation is answering a series of questions, including “What’s keeping you awake?” Since the game is about insomniacs with superpowers on the edge of sanity, this becomes a deeply significant question.
Soft distinctions are basically a type of aesthetic mechanics for character creation, and they can be an excellent tool to help players develop their characters. RPGs have traditionally included some optional soft distinctions, as well as edge cases like alignment.[2]
These often run into the issue of being blank spaces. They offer freedom, but they also lack guidance. Apocalypse World presents one solution to this, which is that each playbook has lists from which you can circle one option. You can still just make up something on your own, but if you don’t feel strongly about it or just don’t have anything in mind, a list of options can help you quickly come up with something flavorful. For example, in AW the Brainer playbook (which is a creepy psychic) has five different lines from which you circle one each for your character’s Look. The second of these is for how they dress, and it reads:
High formal wear, clinical wear, fetish-bondage wear, or environmental wear improper to the local environment.
You may or may not have a clear idea of how your character dresses. This set of four choices lets you come up with something very quickly that is in line with the Brainer’s flavor, yet allows for a great deal of interpretation. And if you think the list doesn’t work and you want your Brainer to run around in a kigurumi bear costume, you can go ahead and write that down instead.
An approach that I’ve developed in my own games is to offer an optional d66 random table of possibilities for each soft distinction included in character creation. That lets me include things like the question “What does being a girl mean to you?” while including a list of 36 examples and providing an out for anyone who has trouble thinking of an answer on their own. It also offers the Maid RPG-like fun of rolling up a wholly random character and seeing what you make of what the dice generate.
Mechanical Distinctions
In most any RPG that has much in the way of mechanics to speak of, there are things that mechanically distinguish different characters. In that sense statistics are a type of mechanical distinction, but here I want to talk about traits that change or expand how a character interacts with the rules. The biggest example of a mechanical distinction from D&D is the character classes. Whether you play a fighter or a wizard is a multifaceted choice with a potentially massive number of secondary mechanics involved, such that two players can experience the game in substantially different ways depending on their choices of classes. D&D often gives different classes entirely different subsystems to play with, so that warriors typically have a smaller set of feats and/or other abilities while spellcasters have an entire spell slot system to work with.
Character traits are seldom completely mechanical. You see such things mainly in games like HERO System, which presents a variety of customizable abilities with minimal flavor, so for example you might buy the Energy Blast power (with whatever Advantages and Limitations you think appropriate) and flavor it as your character’s heat vision superpower, fireball spell, trusty laser rifle, or any of a zillion other things. But most of the time big and small traits both have mechanical significance and express something about the character, even if it’s something mundane, like “she’s especially good at fighting with a rapier.”
Mechanical distinctions are one of the major places where RPG design veers closer to board game design. They tend to be a mixture of traits that create new options and traits that modify existing ones (including modifying ones created by other special traits). A wizard who gains a fireball spell is gaining something novel that’s not really covered elsewhere in the rules, while a fighter who gets a power attack feat is getting a modification to the basic attack rules, and a wizard who gets a metamagic feat is getting a modification to the special spellcasting rules for their class. Collectively, mechanical distinctions can hugely change the overall gameplay experience. This is why although D&D 4th Edition originally had every class follow the same structure, different classes could provide very different experiences, even for classes with the same role.
Game balance is something that RPG people can be a bit weird about, I suspect because the most popular RPG is one that has kind of a strange history with the concept. I don’t think that balance needs to or even can be perfect even in a game that isn’t so full of variables and general fuzziness as your typical RPG, but putting some thought into basic fairness and functionality is always a good thing.
Magic: The Gathering is now one of the most interesting games out there from a design standpoint, especially so since the staff regularly blogs about the thought that goes into it. They came to think of the three major types of M:tG players in terms of the archetypes of Timmy, Johnny, and Spike. Timmy likes to try big, flashy things and plays for fun. Johnny treats the game as a creative outlet, and likes to find clever ways to play the game. Spike meanwhile ruthlessly plays to win, and for a while the staff simply called him “the tournament player.” They would try to give each set some new cards to appeal to each of the three archetypes, with the aim of allowing players coming at the game with any of those approaches to have a good time. In a blog post about “Ivory Tower Game Design,” Monte Cook seemed to have misinterpreted the whole concept, and said that D&D 3rd Edition was based on the premise that some options should just be inherently better than others, and it should be up to players to master the system and figure out which choices were optimal and which were outright traps. The Toughness feat for example, which gives a character an extra 5 hit points, was something they pointed to as an obvious choice for Timmy, except that once you know more about the game, you find out that in the long term it’s a decidedly suboptimal trap option. He later came to regret that though, and not just because the blog post got him so much flak.
The problem with Ivory Tower Game Design isn’t the idea of system mastery, but the way it punishes intuitive, innocent choices. Coming to better understand a game and mastering its complexities is one of the great pleasures of gaming, whereas arbitrarily punishing one of the three major player archetypes is just kind of stupid. When designing mechanical distinctions, you should strive to make them all fun and appealing, and to make them all useful and consequential.
Player Choices
One thing that’s important to understand is that when you give players a list of options to choose from, they will tend to treat all of the entries as potentially viable. Whatever the merits and flaws of D&D’s alignment system, you certainly get more Chaotic Neutral PCs than you would if you didn’t put those nine choices in front of people. If you wanted to specifically make a game primarily about virtuous heroes, it would be an outright mistake to oblige every player to make a choice that includes options for amoral or evil characters.
A more amusing example comes from the Slayers d20 game that Guardians of Order put out. There’s a Japanese writer who does a column about weird American RPGs, and he went as far as to rope the writer and artist of the original Slayers novels into playing Slayers d20. In the source material, the PC type characters are almost all human, but the book provides stats for basically every race that’s appeared in Slayers, including the animal people and fishmen that pop up as comic relief. Naturally their party consisted of a raccoon-man sorcerer, a disturbingly cute fish-boy, and a golem. They seemed to have had a lot of fun with the game, but it looked quite a bit different from the source material.
I don’t mean to say that you should unduly limit what choices you give players, but that you should be conscious of what effect your presentation of them will have. When I look through a Powered by the Apocalypse RPG, I like to see a list of playbooks that shows a coherent vision of what a typical group of PCs is going to be like. While there’s something to be said for the unlikely combinations that games like D&D can allow, if you do something like that in your game it should be on purpose.
Random Chargen
Although random character generation is not as much of a thing in RPGs as it used to be, it’s still a major current in the medium, owing to it being an important part of how D&D works. In D&D you start by rolling 3d6 for each of the six ability scores, which in turn determine your character’s basic talents and what classes they’re most suited to. The role of ability scores in the game mechanics has grown over successive editions, and it came with a proliferation of kinder rolling methods and even point-buy and array[3] systems.
The appeal of random chargen is in how it lets you discover your character rather than creating one whole cloth, and how it forces you to try out concepts you might not if left to your own devices. The guy who normally plays a fighter might roll a high INT and wind up trying out being a wizard for the first time. On the other hand, when you roll ability scores randomly, some PCs are going to be just plain better than others for no reason other than luck. While plenty of people simply roll with it, it’s entirely possible to (as Reign did) have random character generation rules that assign things randomly while giving each character the same overall power level.
While there are game groups that do just fine with PCs with disparate power levels, there isn’t a lot of positive justification for it. Power disparities naturally make some PCs better able to handle challenges, which gives other PCs less to do. Even if the players are exceptionally mature and don’t resent the players who got handed superior stats, they’re still arbitrarily going to have to work harder to fully participate in the game.
Personally, my objection to random chargen as it’s been presented in most RPGs isn’t that it exists, but simply that the random things it generates usually consist of nothing more than a few numbers. Those numbers are important and are things you can interpret to build up something of a picture of the character, but they’re still pretty limited in terms of developing a character. Maid RPG, with its copious tables of Special Qualities and such, uses random rolls to create weird, interesting characters, and while you might not have thought to make your fighter have a high Wisdom, you definitely didn’t think to have your maid be a cyborg mermaid.
Custom Distinctions
There are a number of games that include character traits that straddle the line between the other categories of distinctions. In Fate, characters have “aspects” that both define them thematically and tie into the Fate Point economy. In Risus, the only actual character traits are “clichés,” which are a combination of a narrative description and a numerical value. In both cases a player might devise something like “Master of the Deadly Tiger Fist Style,” and by writing it on the character sheet, they’ve given their character a mechanically significant trait that the game designer never dreamed of.
These kinds of traits have the advantage of being extremely flexible. They’re a big part of why Risus can handle practically any kind of subject matter. There are a couple of drawbacks, and while they’re not a reason to totally abandon the concept, they’re something important to be aware of.
- The first is simply that they present a blank space that players will have to come up with something for. This isn’t in itself a deal-breaker, but there are going to be cases where it makes things grind to a halt.
- The second is that since which trait you use for a given action is so much a matter of interpretation that it becomes unduly tempting to push to use your best trait(s). This is especially true when it comes to traits for things like magic that have the potential to be applicable to a huge range of situations.
Relationships
Relationship mechanics are a more recent innovation in RPGs, but a pretty interesting one. There are any number of genres and styles where the connections between characters are of vital importance, so it’s only natural to want to find ways to make them matter to the rules. Probably the most flagrant example is the Smallville RPG, where for a typical roll you’ll use the die for one of your character’s Values and the die for a relevant relationship to another character. Smallville is the story of the young Superman through the lens of a CW drama, so it’s eminently appropriate to have relationships inform virtually every die roll in the game.
Overall, I’ve found that relationship mechanics are useful but tricky. They’re the sort of idea that has a certain intuitive appeal, but there’s the question of how to have your game handle them without too much overhead and use them well. Not every game lends itself to having relationships be a key part of a typical die roll, so if you want to use relationship mechanics you’ll have to figure out other ways to make them useful.
In Media Res Characters
I’ve only really seen it done as an optional rule in a handful of games, but it’s entirely possible to have character creation be something that players do over the course of the game rather than all at once.
Without Character Creation
We generally operate on the assumption that playing an RPG involves each player creating an original character. There are a lot of good reasons for this, but of course there are cases where it’s not the best option. Creating your own character creates a sense of ownership that you don’t really have with a character the game designer handed you, much less a character that’s been part of a popular work for several decades. While some people really like “fanfic” role-playing, it’s hard to feel you “own” Iron Man, least of all next to the likes of Stan Lee and Robert Downey Jr. There’s also the fact that where RPGs have provided pre-made characters, they’re routinely made in a sub-optimal way.[4]
It’s hard to fault gamers who prefer to make their own characters, but on the other hand there are advantages to having (competently made) pregen characters with certain games. They make it possible to jump into playing the game a lot faster, and potentially lets the designer craft a more cohesive scenario for the game. This works especially well for games designed for short-term play in the first place.
[1] In this respect GURPS has some interesting design choices. Attributes default to a value of 10, and you can spend points to make them higher or set them lower in exchange for extra points to spend elsewhere. The game’s action resolution mechanics have you roll 3d6, and you need to get at or below your attribute value to succeed. That makes the game’s wider range of values (hypothetically from 1 to 20, though in practice characters stick a bit closer to 10) immediately functional in a way the 3-18 ability scores aren’t in D&D. On the other hand, using the actual D&D ability scores as the basis of a check was a common house rule and is the core mechanic of The Black Hack.
[2] Alignment is an extremely rare mechanic outside of D&D, and in D&D it has gradually lost the limited mechanical significance that it once had. I think alignment is a concept that could be interesting if a game gave it more cosmic significance like in the source material that originally inspired the concept.
[3] “Arrays” are where you have a set of specific scores that you distribute among the stats. For example, in D&D4e the standard array was 16, 14, 13, 12, 11, and 10. It has the advantages of being a bit faster than some other non-random methods, and cutting down on fiddly min-maxing.
[4] Aside from them being so ridiculously time-consuming, demo games of D&D4e were always frustrating because it would’ve been trivially easy for me to build more effective characters than the ones Wizards of the Coast provided.
Yaruki Zero Games 