Designing D&D Subclasses with Jeremy Crawford Part 2

The following is a video transcript

Todd Kenreck: Today is part two of my conversation with Jeremy Crawford about designing your own subclasses. We talk about how he does it, and also the themes, making sure abilities match the theme of what you're going for, and in the end, how to balance all of it.

Jeremy Crawford: You've got to make sure that in the wording, that you're as clear as possible, that you're leaving as many important questions answered as possible. I say important because there's kind of a line we design toward, and that is we want to make sure that kind of we imagine a reasonable person, the kinds of questions a reasonable person and a typical D&D situation, the kinds of questions that might arise, either from that person or those situations, and we want to make sure the rule answers those questions.

I say that because D&D, being this wonderful, open-ended storytelling platform, can spark an unlimited number of questions, an unlimited number of corner cases. We have decided, in our design approach to 5th Edition, we're not going to try to answer all those questions, because the truth is, we never could. No role playing game every could, unless it was the rules were thousands and thousands and thousands and thousands of pages long, or going in the other direction, the core system was so simple enough that there were just a few sort of standard mechanics that could answer every question, but the way D&D is currently structured, which is this mix of kind of an elegant core system with a lot of exceptions built on top of it, we have to be very careful we don't get carried away addressing every possible corner case, because when you do that, you suddenly make the standard thing hard.

What I mean by that is you want the standard thing that people are going to do at the game table, with their class features, you want that thing that's going to come up session after session after session, you want it to be as easy for them to do as possible. You don't want it to be a pain point. You want it to be fun. You want it to be engaging. You want it to be intuitive, intuitive for that character archetype, which means it can, at times, be complex. If the archetype is about scholarship or being a tactical mastermind, it's okay if the features also involve some tactical meat, some little extra crunchiness, but you want to make sure it's appropriate to the character archetype, and you want to make sure it's not going to bog down play.

In trying to hit those goals, it's why you have to resist trying to answer every single corner case, because then, suddenly the corner case answers will distract the reader from what's actually important. And what we have found, if you're a designer, people, their eyes are eventually going to glaze over if the text is too dense, the mechanic is too complicated. You can go for rich mechanics. You can go for complexity, but you need to do it with a steady hand, with a clear eye, and make sure you are always communicating the most important thing for your reader to take away. That means if you start answering 10 corner case questions, it becomes harder and harder for your reader to grok, "What's important for me to take away from this class feature?"

You're also going to want to make sure it's just written well. Is it clear? Is it written the way other mechanics like it in D&D are written? Because here's another funny thing. You could write a mechanic in a variety of ways, two different class features that do exactly the same thing, but write them so differently that it could actually confuse people. So whenever possible, we try to write similar mechanics in a similar way, so that again, it's easy for a person to kind of on-ramp into the feature and say, "Oh, I got this. This is similar to this feature I've seen in this other class, but with this one or two little bits of difference."

So you need to polish it, make sure it's clear, make sure it's complete, make sure that if you refer to a condition, for instance, that you actually remember what that condition does. Make sure if a class feature refers to something like the dash action, you actually know how the dash action works.

Here's a great example of this, and this is not an example from a subclass, but just yesterday, for a book we have not announced, I was evaluating a legendary creature, that had a legendary action that allowed it to take the dash action or to disengage. Here's the thing. The disengage action, when you take it, makes it so that your movement does not provoke opportunity attacks until the end of the current turn. Taking that as a legendary action doesn't do anything, because unless that legendary action also involves movement, it's like congratulations. You disengaged, but you're not moving, so that just did nothing at all. That's a great example of you've got to make sure, when you're making a reference to something else in the system, you've got to make sure you know what that thing actually does, because you could basically be doing nothing at all, or you could be doing something entirely different from what you expected. So that's also a really big part of finishing a subclass, is making sure everything actually works the way it's supposed to with the rules that people have in front of them.

You'll notice I still haven't gotten to what people think of as balance. One of the last things we do when we're finishing up a subclass is what people traditionally think of as balance, and that is, is the power right? The reason why we do this last is honestly, of all the things we've talked about, getting the story so it's compelling, making sure the game mechanics are fun, telling the kind of story we want, making sure they're complete, and clear, and understandable. Balance, of all those things, for us, is the easiest step, because it involves a few simple things.

First, make sure the subclass as a whole is hitting a similar output target to the other subclasses of that class. Here's what I mean by an output target. That can include damage. That can include healing. That can include defenses. Look at, overall, the bonuses to damage it might be providing, the increases to AC, the ability to control other people. See, is it roughly comparable, as a whole, to the other subclasses of that class? Keeping in mind that the power does not need to ramp up identically in every subclass. You might have a subclass that gets a lot of its power up front and very little later on. Now, you don't want to overdo that. You don't want to have all the power front-loaded, because A, you could end up with somebody who's way too powerful for that low level. It also becomes this temptation for multi-classers just to dip and snatch that juicy nugget of power that's sitting at a low level. So first off is just kind of eyeball it. Compare it to other subclasses in the game.

Now, people will often wonder, "How do you determine the power of something that doesn't deal damage, that doesn't increase accuracy, that doesn't increase your armor class?" One of the simplest ways to do it is to think of anything in the game as being equivalent to a spell. Our spells are one of the places in the game that we have most carefully set up a sort of set of tiers of power, where damage goes up at certain rates, healing goes up at certain rates, but also, certain non-damaging and non-healing effects go up at certain rates.

Here's what I mean by this. If you want to see like what's roughly the damage equivalent of an ability that paralyzes people? Here's what you can do. Look for the lowest-level spell in the Player's Handbook that paralyzes somebody, and does it reliably. What I mean is don't look for a spell that maybe, as some side effect, it might impose some condition on somebody. No, look for like this thing does this thing, and that's it's shtick. So with paralyzing people, our main, low-level example of that is Hold Person.

Here's how you can translate, then, that paralysis into what we refer to as virtual damage. Look for, then, a juicy damage spell of the same level of the spell you just found, Hold Person, second level spell. So look for a juicy second level damage spell. The one we go to is Scorching Ray, because Scorching Ray is juicy and a lot of people take it. So in a lot of our calculations, if we just want to eyeball things, and again, we can get more precise if we have to, but if we just want to do a quick sniff test, we essentially translate an effect that might ... like a class feature that might paralyze people. We translate that into virtual damage equivalent to the damage dealt by Scorching Ray.

So when you're looking for a damage spell that's of the same level as the non-damaging effect that you're creating, do make sure you pick a damaging spell that primarily deals damage, not something that it might ... you know, it conjures up a tree and does a little bit of damage. No, going back to the example I gave before, pick as your damage spell something like Scorching Ray, where damage is the name of its game, and that will help you translate effects into virtual damage.

Now, there is another little bit to balance, specific to subclasses. Once you've done all this work, or close to the end ... The precise order is not actually that important. What's important is just that you do all of these steps, and you do it carefully, always with your players' fun as your guiding principle. Some point toward the end, you've got to do a multi-classing reality check. Multi-classing is an optional rule, so we don't design for it. What we do is once we've designed, we then do a multi-classing check to see, "All right, if I come in to this class as a multi-class character, what's the experience going to be like? How quickly am I going to be able to snag these various features? What could that potentially do, both positively and negatively, to the play experience of different classes pairing up with this class and this new subclass?"

What that almost always results in us doing is realizing we front-loaded the class a little bit. We dangled something maybe a tad too good, and we might take something that we were giving away at a low level and push it up a little bit. Now, this has a benefit for the subclass itself, not just for making sure the multi-classing experience plays out the way we want, because we find, over and over again, that players like a sense of real advancement. It's exciting when you get to a higher level and you get more powerful, so you want to make sure you don't give all the best stuff right away, because a great part of the D&D experience is, "I get some cool stuff now, but there's that neat stuff later on that I also want." You want that sense of aspiration.

But, you also don't want to go too far, where there's nothing juicy at the beginning, and it's all later on, because also, part of the reality is a lot of group tend to play D&D between the levels of about first level and 10th level, so they might get to that higher-level stuff, but often, that higher-level stuff is more there as inspiration for what this person can grow into than it is something that will ever actually see play.

Now, what often I would say to someone, though, that's no excuse for not getting that high-level thing right, because when the group does get there, you want them to have an awesome play experience. You want that feature to feel as polished and as ready for play as the lower-level ones. But again, I bring this up because you don't want to put the most fun, most exciting aspect of the subclass at like 17th level, because the fact of the matter is, many groups are not going to get there, except for the rare group that plays all the way to 20th level, or what's often more likely, is actually people will do one-shots or mini-campaigns, where they just start at high level, like, "Hey, you know, we're going to have a four-session mini-campaign where we start at level 15, and we're going to gain a level every session," you know?

One of the fun things about D&D is you can structure a campaign, or a mini-campaign, or an evening of adventure in many different ways. Groups certainly don't have to start at first level, and in fact, we expect a typical group, once they're familiar with the game, to actually start mini-campaigns around third level, because third level is sort of when all of your baseline abilities click into place, which is also an important thing to do a reality check on, is make sure, in that early part, that if there's something foundational for the subclass, for its identity in the world, and for the play experience, if it's foundational for it, do make sure that you're getting it early on.

Now, if that foundational thing is maybe too powerful for low level, what you can do is give a less powerful version of it and then have that thing grow in power over time, and that's something we've actually done in a number of our subclasses, where you get like mini version of something, and then as you level up, that thing gets more and more powerful. So, there are lots of tricks you can use to deliver the concept that you're going for with a particular design.

Todd Kenreck: Thank you, Jeremy Crawford, for being on D&D Beyond. I'm Todd Kenreck, your host. Thank you so much for watching.



  • To post a comment, please or register a new account.
Posts Quoted:
Clear All Quotes