Freedom vs. Structure in Class Design

The 5th Edition Q&A for last week got me thinking. In an increasingly rare case, it wasn’t about how I’m upset with how the game is shaping up, but rather that Wizards is taking an approach to the design that merits discussion:

…the important thing to know about a subclass is that it’s not about mechanics, it’s about the archetype. That’s why we want to look at things like “Knight” for a subclass, not “Defender;” the word “Knight” puts a face on the subclass and describes its place in the world.

This is in keeping with what D&D has tended to be over the years, a late-medieval western-Europe version of high fantasy with all of its associated tropes. It’s important to Wizards that each subclass, like the classes, paragon paths, prestige classes, etc. before it, be some role within the world. These options then give the players the chance to play this role in their unique (or, if the character optimization boards are any indication, ridiculously specific) way. In this sense, it makes a lot of sense to have subclasses named “Knight” or “Gladiator” or what have you.

But something about this felt a little off, and I couldn’t quite put my finger on it until somebody in the comments said it for me:

I think saying “Knight” when you mean “Defender” is a bad idea. Putting the face on our characters is our job, not yours. I’d rather have the mechanics straight forward and then I can say “This guy is a knight, because he defends the weak.” Remember, D&D is simply a mechanical guide to loop a groups imagination together…keep mechanics free from “loose labels” and “options”, just paint it out loud without hiding behind other terms, and we’ll handle what’s a Knight and what isn’t.

This also makes sense. A “fighter” is a rough model* of a character’s mechanical abilities, one that can be modified with feats, skills, items, selections within the class, and other choices. The way a player runs his or her character also affects the model, in the same way that a character wielding a trident feels different from a character wielding a longsword even though every single other option may be the same. But players expect that “fighter” is just a starting point and the character is defined by the other choices they make, the roles they fill, and they way they’re played. If that’s the goal, why is it that subclasses are explicitly intended to define the character rather than just give a slightly modified foundation on which a player can build?

It’s a fairly common debate, freedom versus structure. On the one hand, a class named “defender” gives players the option to put any spin on it that they want within the limits of the class. They could use that class to build a noble knight or a holy paladin or an armored warlord. The system gives the loose rules so that everybody agrees on what they’re doing, but it’s up to the player to take those rules and use them to envision a character. On the other hand, a class named “knight” brings forth an image immediately, and it lets others know exactly what sort of character this is down to their party role and something of their personality. It doesn’t makes sense that the knight, the paladin, and the warlord all share the same set of abilities, and defining them differently lets each shine on their own without all feeling like the same thing.

There’s not one good answer here. There are players who like the freedom of taking “defender” as a skeleton and building their own creature. This is why there are classless systems, a lot of them. There are also players who want to understand something about a character without hearing their entire backstory. If I say that my party contains a vampire, an ex-slave gladiator, a ballroom dancer, and Tony Stark, there’s no way to know anything about the characters’ roles, their system, their personalities, even whether they’re a group at all (hint: they are). To some players, that’s great. But D&D is specifically and totally about a balanced party and it’s designed for players who expect that.

There’s also a learning curve here regarding shared language. One of the stated goals of 5th Edition is to let the system be simple enough for any age or experience level; it’s why skills and feats are fully-optional mechanics. A new player might not know what a “cleric” is, but a “knight” means something and gives some direction regarding what they do and how they act. I saw this a lot during Delve Night, both in the characters brought to the table and the ones we provided. Handing a player (or describing your character to other players as) “Donald of Kent, a ghost possessing a suit of armor” leaves a lot of unanswered questions, but handing the player “a shardmind defender and ranged striker” at least gives them something they can understand even if the specifics are slightly different. The freedom to design something without a class is great, but it runs the risk of confusing and alienating people new to the system.

Then again, if you’re looking at it from a “get new players” perspective, you also have to take into account that people may want options you haven’t presented. A “defender” is vague and provides little direction, but it works as a dwarven fighter, a feudal-era samurai, a modern soldier, a space-age warrior, or any number of characters. A “knight” is really just one thing, a knight. If the system provide so much structure that it’s restrictive, players won’t be able to play the characters they want unless there are more books with more options. Ask anybody who’s played 4E and wanted to build a Dex-based character that wasn’t a striker. That concept wasn’t something the designers thought of and released (until Essentials, where we got a ranger who could be a controller as long as he loves archery), so it doesn’t exist no matter how much somebody might want it.

I like freedom, so it appeals to me that there’s a “defender”, a generic idea that lets me tweak it to fit whatever character I want. But given how easy it is to reskin things now that I’m used to it, I can do the same thing with a “knight”. The difference is that a “defender” tends to expect that I’m mixing-and-matching traits until I have what I want. This means more options for me, but it greatly increases the chance that there are combinations of abilities that break the game wide open; the solution to this from a design standpoint is to either playtest everything to the hilt (prohibitive; I learned this in World of Warcraft) or provide only the limited options you’re certain are balanced (boring; I learned this by reading Unearthed Arcana). A “knight” expects that I have everything from the following list, whether or not it fits my character concept. In a recent comment I described how frustrating it is that I love the cavalier except for the non-negotiable mount, and I’m not the only one who feels this way.

It’s kind of a moot point for D&D, a system that has always had classes and likely always will. They’re not going to stick to a generic fighter / rogue / magic-user set of classes, they’re going to present any number of things that fit with the lore and seem different enough from each other. I don’t much care which way they go as long as it’s interesting. Building interesting classes, however, is a lot harder than it looks, and that’s something I’ll discuss in a future post.

* — I really want to use the word “archetype” here, but that has a specific meaning within Pathfinder. In fact, it’s the equivalent of “subclass”, and this post didn’t need to get more confusing.