The Construction of Character Classes

Premise: a character class provides players with an interface to the game world and a place for their characters within it. (See yesterday’s post on classes.)

There are a couple of major schema for constructing character classes. Most games use more than one.

Class by Combat Role

You separate characters by how they fight. This is the simplest distinction between Dungeons & Dragons‘ fighting men and magic users, and it goes back to Chainmail. The fighter/wizard divide (or lack thereof) is arguably the definitive problem in character class design.

Basically, in Chainmail, there were guys who fought other guys with swords and polearms. Then Gygax added wizards, who functioned as artillery. So, going into Dave Arneson’s original game, you had these two basic types of units.

In 1974 D&D, they were represented by two overlapping but distinct systems: combat and magic.

In tabletop games, D&D 4e is probably the most rigorous example of this approach. Not only is every class and monster defined by a combat role, but those roles are transparently explained to the players right there in the rulebook.

Virtually every Everquest-descended MMO is built like this. (The extent to which Everquest itself follows this model is a longer discussion.) The most transparent about it is probably City of Heroes, which marries Everquest combat roles to character concepts while staying fairly abstract about both.

This is a somewhat circular methodology, though. Since D&D, it’s become imperative to have certain classes if you’re going to do a class-based fantasy game. And if you’re going to have combat, there’s a strong push to give each class a unique place in combat.

(That last bit’s especially true of MMOs, where problem solving comes down to money or combat, no matter what style the game adopts otherwise. )

Classes by Concept

You have a fictional persona in mind, so you make a class for it. By volume, this is probably the biggest driver in class design.

D&D’s original classes (fighting man, magic user, cleric) were driven by combat role and need. The first supplementary class,* the thief, appears to be driven more by concept.

The thief appears to be loosely modeled on characters like Fafhrd, the Mouser, and maybe Bilbo Baggins. (Gygax didn’t like Tolkien all that much, but that didn’t stop him from using Middle Earth stuff where he liked it or thought it would be commercial.)

He has the skills one would guess a thief might need, as well as one which seems rather particularly borrowed from the Mouser: the ability to use magic scrolls.** In the process, he introduces a third system to D&D. Mechanics supporting fiction, we are go.

Classes to Fill Holes

Your system’s missing something, so you make a class to fix it. D&D’s cleric is the classic example of a class created to fill a hole in a game or campaign. Dave Arneson had one player playing a vampire. As Mike Mornard recounts:

Well, after a time, nobody could touch Sir Fang. Yes, that was his name.

To fix the threatened end of the game they came up with a character that was, at first, a ‘vampire hunter’. Peter Cushing in the same films.

As the rough specs were drawn up, comments about the need for healing and for curing disease came up.

This is pretty much always a supplement for another design scheme, and many games that start strictly adhering to another approach evolve in this direction over time.

Classes as a Template Layer

Your system operates on a point-buy or other system, but classes ease new users into creating a character.

D&D 3.5 almost works this way. One of my favorite video games of all time, Hero’s Quest/Quest for Glory I, works this way. Ultima Online‘s current character generation process works this way.

Vampire: The Masquerade‘s clans were intended to operate this way, but quickly became vital fictional concepts in and of themselves. By the end of the Revised Masquerade line, fictional concept drove class design completely. How that played into the approach taken to Vampire: The Requiem, and how the vision changed over the course of that line, is the subject for another article.

Classes on a Spectrum

You have certain “pure” mechanical concepts (melee, magic, stealth) with classes that represent them (fighter, mage, thief) and then a series of classes which exist between those.

This is the fighter-cleric-mage spectrum of the original D&D. The Mass Effect series is also built this way, and quite elegantly. Tunnels & Trolls primarily follows this method, though it’s also evolved quite a bit over the years.

Classes by Convergence

Your game has no formal classes, but character optimization favors a small number of builds. You probably lie awake at night worrying about tank mages. Later designs building on the same system end up centering around these builds more than the original open, underlying system (usually based on skills).

Ultima Online is the poster child for this. UO wasn’t designed for classes, but they perpetually emerged from each revision of the skill system.

GURPS is often played this way, though I’d be a little afraid to say so on a message board.

Exalted looks like it’s a class-by-concept scheme, with all of the D&D standards scaled up, but many people play it as a convergence scheme, and that’s the current approach to its mechanical design.

____

* Class was a slippery concept in 1975, something I’ll treat in the future. But by the definitions later games would extrapolate from D&D, Greyhawk D&D has four classes.

** This seems to me to be directly from “The Lords of Quarmall“.

4 thoughts on “The Construction of Character Classes”

  1. Yeah, class-by-convergence would have been my argument on the last post, but I read this post half-way through writing that response…

    Good typology here. When you mentioned lying awake at night worrying about tank mages, I could relate. Whatever you do to make something open, these ‘builds’ or patterns will arise. Then, you can come out with a second edition that plays off those builds or one that undermines them. Really, the fact that an open system allows for a handful of ‘builds’ that get players excited is great. If a game had 6 organic ‘builds’ to it, I would consider that a success. If the designers only thought of 4 of them ahead of time… well, then, lucky for them.

    Good eyes on Exalted 1e and 2e; I would say the same thing.

  2. Re: GURPS & Convergence

    Oh you won’t get any complaint on that from me, and I *like* GURPS way more than you do. Not to excuse lazy design (*cough* Blizzard with your clearly incorrect Talent options *cough*), but any system can be optimized by a sufficiently motivated player community. While Ultima Online is a less punitive game than say EverQuest; players were nonetheless very motivated to not accidentally handicap their characters (because deliberately unstructured competitive gameplay creates it’s own nuclear arms race). So, as these things go, broad paths of character types were gradually optimized into ‘correct’ builds. In GURPS, because it is possible to build characters which are objectively less powerful than other characters, the power-gaming quotient within a play group will average up to the most wanktastic member. In D&D 4e, it’s remarkably possible to simply play around your resident power gamer. In GURPS it absolutely isn’t, not because you built your character wrong (unlike in D&D 3.5), but because a convergence-built GURPS character is a monstrous amalgamation of force-multiplied-until-they-cried broken mechanics and the pristine blankness of a character without persona.

    (To that point, GURPS & Exalted are actually also excellent examples for how you can take a ‘all of these are actually good choices’ system and ruin it by devoting no effort to balance your mechanics at both high and low ‘level’, or with respect to play-styles.)

  3. I agree with most of this. If a game is classless and supports many VIABLE builds, then I would put a crown on it and parade it through town. If a classless game turns into an “arms race” where the “resident power gamer” forces EVERYONE to get two or three fighting arts/skills/powers just to keep up, then you have a game that wasn’t balanced well. The same goes for games in which players can handicap themselves unknowingly.

    I’m reminded of Exalted E1 where the rules tell you, “Hey, you better get DODGE or you’ll suck.” If the players NEED a particular skill/art/charm to even exist in your world, just give it to them already! Damn.

    I still think a classless system is worth the risk. And, maybe it is good to put it out and let people break it… E2 will be that much better for it.

  4. @Calvino
    “Really, the fact that an open system allows for a handful of ‘builds’ that get players excited is great.”

    Yeah, that’s often my feeling. On the other hand, you get this:

    @TerraObscura
    “In GURPS it absolutely isn’t, not because you built your character wrong (unlike in D&D 3.5), but because a convergence-built GURPS character is a monstrous amalgamation of force-multiplied-until-they-cried broken mechanics and the pristine blankness of a character without persona.”

    Unexpected convergence (or convergence that you simply ask players politely to avoid) can often throw off group dynamics. It’s not necessarily bad, but it’s a complicating factor in multiplayer games.

    Also, people will call your game broken on the Internet. But I don’t know of any cases where that’s actually hurt a game’s playing audience.

    “While Ultima Online is a less punitive game than say EverQuest; players were nonetheless very motivated to not accidentally handicap their characters (because deliberately unstructured competitive gameplay creates it’s own nuclear arms race). So, as these things go, broad paths of character types were gradually optimized into ‘correct’ builds.”

    Before Ultima Online introduced its PvP switch (and simultaneously redid all combat mechanics), there was pretty much one correct combat build, and one correct set of combat tactics. Variations were small, and more about player error and martyrdom than strategic choice.

Leave a Reply

Your email address will not be published. Required fields are marked *