Swiftguard TTRPG Design Log #3

Abandoning Classes for Character Progression

Welcome to the Swiftguard TTRPG Design Log, a series of posts logging the development of my new d20 fantasy TTRPG, Swiftguard.

As of right now, I’m in the process of laying out Swiftguard’s Playtest Packet 1 content. It’s about 25% of the way there. As predicted, formatting tables is what takes the most amount of time. As soon as I moved to sections without tables, the layout process quickly accelerated. I’m hoping to have it released by early June, but as I’ve said in the past, I have no guarantees.

For now, let’s unpack the second major design departure from SRD 5.2, the rules that Swiftguard is largely built on. Why did I choose an “à la carte” leveling system rather than a traditional class system?

The Class / Subclass Template is Unnecessarily Restrictive and Distracting.

A few weeks ago, Wizards of the Coast released a new Unearthed Arcana playtest packet for D&D 2024 - Horror Subclasses - which presents six redesigned subclasses and two new ones: the Reanimator Artificer and the Hollow Warden Ranger. Curious to see the direction that D&D was going, I watched the reveal interview between Todd Kenreck and designer Makenzie De Armas before reading the content for myself. Of all the options they discussed, the one that grabbed my attention was the Hollow Warden.

De Armas explains, “I had a stint recently where I was really into old folk tales, cryptids, the very imagery of the forest shrouded in fog with the eyes just in the night…and so this subclass is built around sort of drawing on the cryptids and forest terrors, and these ancient forest guardians and really looking at the ranger’s relationship to nature at its most bloodthirsty and violent.”

Before moving on, I want to say that it was incredibly smart to pass Jeremy Crawford’s previous role of “designer who excitedly explains the playtest” to De Armas. Her passion is infectious, and I’ve noticed that the community is very supportive of her.

That being said, I’ve followed Unearthed Arcana for years. I know that often, something that sounds compelling in flavor often falls flat when it comes to its game mechanics, and unfortunately, the Hollow Warden is no exception.

For me, a subclass’s design really comes down to its 3rd level features, since those are the features that are most likely to see play. If a subclass isn’t able to fulfill its story promises in that level, even if the features are powerful, it’s already failed an aspect of its design, in my opinion.

This is where the Hollow Warden falls. Its features aren’t bad by any means, but mechanically, they’re just kind of dull. Again, let’s look at DeArmas’s story promise with the subclass.

“Ancient forest guardian.”

“Imagery of the forest shrouded in fog with the eyes just in the night.”

“Really looking at the ranger’s relationship to nature at its most bloodthirsty and violent.”

Just hearing those descriptors, what kind of mechanics would you expect? For me, it’s damage, fear, and stealth. So what does the Hollow Warden actually get?

At 3rd level, Wrath of the Wild grants two benefits: Ancient Armor, which increases the user’s Armor Class, and Unnerving Aura, which restricts an enemy’s action economy. Sure, both of these features have creepy descriptions, but by themselves, I’m not really sure they get the point across. Unnerving Aura is close - it’s kind of a fear effect - but it doesn’t convey “bloodthirsty and violent”. The other feature, Hollow Warden Spells, offers Wrathful Smite, a Paladin spell that does offer both bonus damage and fear. Sure, it checks those boxes, but unlike Wrath of the Wild, there’s no flavor text to convey it. And I guess that’s my point. D&D design has often had these examples, where there’s some sort of story promise, but the mechanics don’t really line up with the story that’s promised. Oh, and Wrath of the Wild? That only activates when you cast Hunter’s Mark. Which I guess is my overall point. The Hollow Warden’s concept sounds awesome, but it has to fit within the framework of one of the core classes, and as such is dependent on that core class for its mechanics to work.

Again, just hearing “Ancient forest guardian shrouded in fog with the eyes just in the night”, I could also imagine other classes fulfilling that same story. It may be a Druid, a shapeshifter that uses violent plant magic to crush trespassers (like the quest “In the Heart of the Woods” from The Witcher 3). It could be a Rogue, a non-magical forest guardian that efficiently assassinates enemies of the woods and has had superstitious folktales told about them. Maybe it’s a Bard, who stays hidden in the fog, using Mind Magic to disorient targets and mauling them with mind-controlled beasts.

My point is that there are a lot of ways to fulfill this concept, but as soon as there’s a class / subclass combination that prescribes it, players are often railroaded in their thinking to that one answer, not to mention that the actual design becomes constrained by the class’s template as well.

For a while now, I’ve felt that 5e’s Class / Subclass structure is a limiting design that often distracts players from the parts of their character that’s the most exciting. In general, I find that players love the stories promised by subclasses, but those specific stories are often lost in the broad mechanics of the archetypal base classes. If the Hollow Warden was to be printed as is, I’m sure it would be remembered as “the Ranger subclass with great defenses and Wrathful Smite”, not “the ancient forest guardian that’s bloodthirsty and violent.”

The Benefits to a Class System

Okay JOHN, we get it. You’ve done a lot of “poo-pooing” class systems by complaining about a fringe example in a playtest packet that hasn’t even been released with a finalized design. Obviously there’s got to be something to say for Class systems, right? After all, so many well known games use them!

Yes, like any design decision, class systems have benefits as well as drawbacks.

For me, the most compelling argument in favor of a class system is that it guides players through learning the game. With a set progression of features unlocked level-by-level, players can learn one tool at a time. As a Game Master who’s been described as having an “encyclopedic knowledge of the rules”, it also makes it easy for me to remind players of tools that are frequently forgotten (with some examples being “remember your Aura of Protection” or “remember you have Extra Attack”).

It also makes it easier to identify what each character’s role is probably going to be. If I told you that my next game has one player bringing a Barbarian, another a Rogue, and another a Bard, I’m guessing you’ll have a pretty easy time guessing how that team will work together.

Another benefit that’s often cited is that, from a designer’s perspective, classes (as opposed to gaining features “à la carte”) are easier to balance. Mike Mearls, the rules engineer for 5e 2014, said in one of his Happy Fun Hour streams that a small number of big choices are easier to balance than many little choices. For example, in earlier editions, feats were small bonus features gained frequently, usually every other level. This led to huge gaps in power between optimal “feat chains” and suboptimal combinations. That’s why in Fifth Edition, feats are available less often but as larger packages. According to Mearls, this made them easier to balance against one another. Notably, sentiment was echoed by Matt Colville in his video “What We Knew Before We Knew Anything” (regarding the design of Draw Steel!, which you can watch here).

Okay, I get the theory. I have no comment on Colville’s opinions or design, considering I haven’t seen the Draw Steel! playtest and the project hasn’t yet been officially released. I can completely believe that for that system, that design philosophy holds true.

However, I’ve seen my fair share of 5e homebrew, and from what I’ve seen, I think it’s a mistake to assume that 5e’s class system is easier to design for than something à la carte, especially for newer designers. Sure, if you’re the game’s lead designer, it may work better for you (even though D&D 2014, Mearls’s own game, was far from immaculately balanced), but I’ve found it’s actually harder on designers just trying to get started.

Homebrewing is Hard.

In some of my earliest D&D games as a player, I met a few GMs who were very enthusiastic about homebrew, including creating their own subclasses, equipment, and monsters. Of these elements, the monsters were the easiest to balance, especially on the fly. In terms of balance, they were followed by equipment, since the bonuses and features offered by equipment were relatively small to begin with. Subclasses were where things got out of hand, and it was because there was no real way to “start small”.

As I mentioned with the Hollow Warden, designing a 5e subclass means that one has to design the features to mesh with the base class. A great example of how difficult this can be can actually be found alongside the Hollow Warden in the Horror Subclasses playtest - the Hexblade Warlock. The playtest Hexblade has features that rely on the activation of the Hex spell, similar to the Hollow Warden’s reliance on Hunter’s Mark. The difference, however, is that Warlocks are dedicated spellcasters with much better choices for their Concentration. The way Hexblade plays is either you use Hex to get your subclass features or you use a better spell but have your subclass features locked away. The Hexblade subclass doesn’t mesh with the Warlock base class, and if the core D&D design team made this mistake, wouldn’t it be all too easy for an amateur designer to make a similar one?

In order to make a good D&D subclass, a designer has to account for meshing that subclass with the base class, which as the Hexblade shows, is already a tall order.

To make a great D&D subclass, a designer has to account for how the class / subclass combo meshes alongside the other classes. For example, one of my GMs designed a Cleric subclass that was so powerful, it essentially did everyone else’s job for them better than they could. At the time, I was playing a Fighter and the Cleric had a better AC, better damage options, and could heal all in the same turn. In a private conversation, I admitted to the GM that throughout one session, I couldn’t help but think “Why is my character even here?” Because the subclass was too powerful, it indirectly stole the joy from the other players at the table. We felt like lame sidekicks to this one character’s story.

To make an excellent D&D subclass, a designer also has to account for multiclassing. Now depending on your table, your GM may not allow multiclassing, but in my experience, this is an exception and not the rule. Depending on what features are gained in early levels, some subclasses may offer wildly disproportionate power to certain builds. For example, in 5e 2014, the Hexblade subclass from Xanathar’s Guide to Everything offered excellent offensive and defensive benefits to Charisma casters, making it a favorite among optimizers. Whereas the Cleric example was a homebrew subclass that stole joy by being too powerful, the Hexblade stole fun by being so powerful and so widely available it felt “necessary” to take on many builds, reducing the variety one was likely to see at certain tables. Again, this was an official option released by the core team, so if they’re having trouble, how much more trouble would an inexperienced designer have?

My success with Homebrewing.

At the same time as I was seeing my GM struggle with balancing their homebrew content with the rest of the game’s ecosystem, I was trying a different approach. In a campaign I was GMing, I would occasionally give a player a bonus homebrew feature based on their character’s narrative achievements.

For example, one player created a Dragonborn Monk that had been raised by the Cult of Tiamat to be a pit fighter. After escaping, he discovered a newfound purpose by worshipping the goddess Sarenrae. In the middle of the campaign, the character was distraught after losing a major duel against one of the Cult’s champions, and he sought out a temple of Sarenrae to pray. After a cathartic roleplaying scene, I granted the character a Bonus Action to activate translucent angel wings, similar to an aasimar. It ended up being a small but incredibly rewarding mechanical feature that allowed the player to celebrate his character’s narrative milestone, and consequently it encouraged other players to similarly explore their characters knowing that there was a small bonus on the other end.

The operative word here is that these features were small. They weren’t entire subclasses or spells or even magic items. They were small little powers that boosted an individual character’s mechanical performance while fitting to that character’s theme. The few times one of these powers broke the game, I would have a frank conversation with said player admitting my mistake and together we’d agree on a more balanced version. It let me get a feel for how one mechanical addition would pull on the other strings of the system.

For example, with this Dragonborn character, something that began to stand out was that Monks were very dependent on their Bonus Action. Even though flight is very powerful in 5e’s rules, activating the wings would come at the cost of that turn’s damage output. Like Matt Mercer’s “Vestiges of Divergence”, starting small and restrained also let me “level up” these custom abilities over time. The Dragonborn’s wings started out as a “once per long rest” feature, but as the character leveled up, I increased their frequency to “once per short rest” and eventually on demand.

From this approach, 5e’s class system eventually starts to look less like a string of thematically coherent mechanics and more like prepackaged proficiencies and features. Bard level 1 is a package of Spellcasting and Bardic Inspiration. Fighter level 1 is a package of a Fighting Style, Weapon Masteries, Second Wind, and some strong equipment proficiencies.

As the characters I wanted to roleplay became more specific and unique, I started to ask myself “What would happen if these packages were opened up and their contents switched around?” According to Mearls and Colville, this would create a system that was difficult to balance. Again, no comment on Colville, but with Mearls, was the alternative (D&D 2014) really a strong countercase in this regard?

In my experience with almost a decade playing and GMing the system (much of which was professional), my answer is “No”. Multiclassing and feat design was incredibly unbalanced! Some subclasses, like Hexblade and Twilight Domain, were so powerful they were either “required” or banned at certain tables, with others like Battlerager and Purple Dragon Knight being so unbelievably weak that I never saw one brought to the table.

I think the strongest case of a system that showcases how class systems can be balanced is actually D&D 2024, which many community members are criticizing for being “bland” (myself included). Then, when it comes to homebrew, the options are also wildly out of whack. Designing a subclass is an all or nothing approach. There’s no easy way to test something small unless you disregard the class system entirely, which is why Swiftguard’s character progression is à la carte - at each level up, features are “bought” with Build Points that are gained at each level.

Let’s go back to the Hollow Warden to use as an example. Rather than assign it as a “Ranger Subclass”, in Swiftguard, Hollow Warden might be a feature that any character can gain. I might design the feature to have a transformation that says something like: “As an action, you summon the spirit of an ancient forest guardian, which transforms your body such as turning your skin to bark or growing bloody antlers. To maintain the transformation, you must Concentrate as if Concentrating on a spell. When you transform, a 10-foot-sphere fog appears centered on you. The fog moves with you, and the area in it is Heavily Obscured to any creature except you. You have Advantage on attacks against any creature that has less than its original Hit Point maximum. Bloodied creatures in the fog are Frightened of you.”

That’s a lot of features! Thankfully, with the Build Point system, it can be assigned more or less points based on how much it gives you. Not every feature is weighed equally. Also remember that with the three-action economy, a one action cost is equivalent to a Bonus Action in 5e. And, it can coincide with a variety of archetypes, from a character similar to a Ranger, Druid, Rogue, etc.

Of course, this is a major departure from 5e’s well known class system. When I was considering whether abandoning the class system was worth it, I found myself asking three big questions:

What would players get?

What would players lose?

Can players live with what they would lose?

What Players Get.

With an à la carte system more similar to Skyrim or Fallout 4, players get freedom. Their character performs the way they want without having to take features they don’t.

When I first played 5e, I wanted to create a character that was similar to Geralt of Rivia from The Witcher. Thematically, Ranger’s description matched my vision best. Mechanically, it ended up being “a fighter with some Druid spells”, and I learned that a Rogue / Fighter multiclass actually ended up closer. I didn’t want Druid spells, like goodberry or cure wounds for my character. I wanted an expert swordsman with high bonuses to lore skills to create informed strategies of how to deal with various monsters. While the 2014 Rogue entry discusses them being burglars, cutpurses, and con artists, their Expertise feature just means they roll high on those checks. Ironically, this meant the Rogue had an easier time getting high Nature and Survival checks than the Ranger, whose story promised them.

With Swiftguard’s à la carte system, the abilities do what they say. Want to be a master archer? Pick up Archer’s Mark. Do you want to cast fire spells? Take Fire Magic. This clarity in abilities that are built for clear function trusts players to fill in the blanks with stories they want, not stories that are prescribed and unfulfilled.

Of course, freedom comes at a cost.

What Players Lose.

To reiterate, in my opinion, the most compelling argument in favor of a class system is how it guides players through the class’s progression. Having a bunch of disconnected features spread out on a blank canvas can be overwhelming when players just want to make a dwarf that swings a hammer.

In response to this early criticism, Playtest Packet 1 will include nine “archetypes”, suggested builds that hold a new player’s hand through learning the ropes. If a player wants to play something similar to a 5e Barbarian, there’s the Champion archetype. If they want an experience of a fairy-tale knight that leaps in front of frailer friends, there’s the Guardian. If they want the know-it-all Wizard, there’s the Scholar.

Of course, these builds are just suggestions, and are welcoming players to make individual tweaks to get a personalized experience. If a player essentially wants a fighter with a little bit of magic, they don’t need to commit to the whole Eldritch Knight subclass or multiclass into a spellcaster - they can just spend some build points on Magic Training and pick the ability that gives them the spells they want. After all, a warrior with a little fire magic plays differently than a warrior that can teleport.

Players also aren’t being promised balance. While I think this system is easier to homebrew, it’s harder to promise balance. I’m sure even in the limited selections being included in Playtest Packet 1 that there are certain broken combos I haven’t considered. That’s one of the joys (and I mean that sincerely) of a designer releasing their creation to a community. That being said, I’m not convinced that “balance” is even what players want.

The number of comments I read lamenting the loss of the old 2014 Hexblade far outweighs the few of us celebrating the “healthier balance” of 2024. In my opinion, D&D 2024 is a much more balanced offering, and yet there are many detractors that think it’s unexciting. I think if players had to choose between balance and salient flavor, they’ll lean toward flavor.

Can Players Live with what they’ll Lose?

I don’t know. That’s the point of a playtest. Just as I may have come off skeptical of the idea that class systems were easier to balance while citing D&D 2014 as a counter-example, my ideas are just as theoretical. This process could prove me wrong, that my preferences are mine alone.

As we get closer to Playtest Packet 1’s inevitable release, I’m curious what you think.

Does an à la carte system sound intriguing, or would you prefer games that stick to the tried-and-true class system?

If you’re a D&D fan, what did you think of the Horror Subclasses UA?

How important are a feature’s mechanics versus its flavor text, and do you think matching them is important?

Let me know in the comments or in the Swiftguard Discord at: https://discord.gg/mtS6Yb4N

And hey, if you didn’t know, I also have a YouTube channel - youtube.com/@dragonmindttrpgs - where I’ll be posting additional updates once Playtest Packet 1 releases.

Thank you for your support, and remember, Your Story matters.

-John