Magical vs. Divine vs. Supernatural - Differentiating types of mystical abilities

[Greywander]

I'm working on a big overhaul to 5e that is focused on making big changes to the class system and character progression. As a consequence of that, I'll have to redo a lot of the spellcasting system, too. Basically, for reasons, I've found it necessary to make all classes "full casters", and wanted to offer a way to distinguish different types of powers so that not everyone is literally a mage.

One important thing is that there is a distinction between the source of power, and the method of executing that power. What I want is for any character to be able to choose any source, while different methods will be locked behind different classes. For example, taking the artificer class would allow you to use your powers by putting them into items (spell-storing items, basically), but you would also have to decide the nature of those powers. For example, maybe these are holy relics blessed by the saints, and thus use divine magic.

I'm primarily here to talk about sources, not methods. I think I've boiled it down to four broad types of sources:

Mundane
Anything that's just your character being really skilled, using the aid of a mundane item, or perhaps technology. This is for martial characters.

Some ideas I have are that mundane techniques generally aren't limited in the same way that other sources might be. They don't have the same restrictions or counters. It just works. At the same time, they can't overcome effects that require magic (unless you're using a magic weapon with that power). Mundane powers might have their cost increased to offset the lack of restrictions. You could combine this with a method geared toward low level power spam to get something closer to a true martial experience, where you're using primarily weaker at-will abilities.

Arcane
Magical knowledge gained by study, and carried out using specific rituals. So things like wizards, bards, and some warlocks.

For arcane spells, this is likely where you'll have to strictly adhere to things like spell component requirements. You might have to gesticulate wildly, or loudly intone an incantation. Arcane magic might allow you to cast spells as time-consuming rituals without spending resources, but at the cost of greatly increasing the resource cost for on-demand spells. Then again, this does sound like it could be a method, rather than a source. Or just a class feature giving access to rituals.

Theurgic
Theurgy is power channeled through you by another entity. This might be a god, a fiend, nature spirits, or something else. This is for clerics, druids, rangers, paladins, and some warlocks.

One thought is that it might interact with alignment in some way. Another idea is that you have to write up a list of tenets to follow (either one for each power, or one for all your powers), and breaking those tenets will cause you to lose access to that power or increase the cost of that power until you make amends.

Supernatural
These are innate supernatural abilities that are just part of who, or what, you are. This is mostly for sorcerers, but also psionics and some warlocks. Maybe monks, too, if you don't count ki as mundane. Things like dragon's breath or unicorn magic and many other monster abilities might also fall under this label.

One idea I like for this is to make you choose a weakness (possibly with a preset list of "bloodlines" to choose from), such as fey magic being blocked by iron. Your weakness can be used to ward someone against your supernatural abilities, or deny you the ability to use any of your supernatural abilities if you come into contact with it. Some supernatural abilities might come from different bloodlines, so while you couldn't use fey magic while in iron manacles, that wouldn't prevent you from using draconic abilities. Defining what can or can't be a weakness will be tricky, which makes preset bloodlines more appealing.

---

My plan is that you can select a source for each power (powers are "spells", by the way), making it easier to customize your character to fit your concept. For example, a paladin might be split between some mundane martial powers and some divine theurgic powers. I'm not sure what I want to do with Counterspell, but I was thinking of making three different versions of the anti-magic field, each with their own unique effect. For example, one might create a dimensionally locked area, preventing teleportation and summoning, and coincidentally also interfering with the channeling of theurgic magic.

One of the big challenges is the difference between sources and methods. Methods dictate how you use your powers, but sources dictate the nature of that power. Some of the things we could do for sources would be things better suited to methods. The source shouldn't care whether we're using spell points or spell slots or pact slots or putting spells into items or any of that. I think the artificer example is a good demonstration, because it's rather out of the ordinary and it's something that our sources would need to accommodate.

So if we think about an artificer that is storing their spells/powers in items (with charge that corresponds to spell points), then here's how each of these might like. Artificer items can be mundane inventions, like bombs or gadgets. They might also be arcane items that can be activated safely as a ritual (no resource cost), or activated instantly at the risk of damaging the item (costs limited resources to use). They might be holy relics that won't impart their blessing if you break the tenets, or they might have fickle spirits sealed within that must be appeased. They might be magical items crafted from monster parts, allowing you to use that monster's abilities for yourself. All four of these examples seem to be compatible with the artificer's unorthodox methods for using powers, so this looks pretty good.

Thoughts on this? Is there another way you might handle these different power sources? Are there any other power sources I might have missed?

[Maat Mons]

Personally, my preferred trifecta of power sources is Arcane, (Un)Holy, and Nature. I also like to match power source to ability score. Arcane uses Intelligence. (Un)Holy uses Charisma. And Nature uses Wisdom.

[Anonymouswizard]

Why do they have to be different mechanically? Is there a reason why a character casting Martial Meteor Swarm needs to have different mechanics to one using Homeopathic Meteor Swarm?

For Theurgy the one people always turn a blind eye to is sacrifice. It's everywhere in the sources, but nowhere in the games, mostly replaced with prayer and the idea of sin removing your powers. But nah, if the monkey wants to keep or recharge his magical mojo they'd better be burning some sheep kidneys on a regular basis.

[Greywander]

Personally, my preferred trifecta of power sources is Arcane, (Un)Holy, and Nature. I also like to match power source to ability score. Arcane uses Intelligence. (Un)Holy uses Charisma. And Nature uses Wisdom.

Since I'm working within the framework of D&D 5e, I don't want to get too crazy, but at some point I want to make a wholely original system with distinct magic systems. In such a system, non-divine theurgy would be more like calling up a spirit to use their supernatural abilities on your behalf. So the supernatural source in the OP sort of corresponds to "nature".

This trio can also be matched to ability scores, with arcane using INT, theurgy using WIS, and supernatural using CHA. But I don't know that I want to force players into a specific stat. It's something I need to give more consideration to.

Why do they have to be different mechanically? Is there a reason why a character casting Martial Meteor Swarm needs to have different mechanics to one using Homeopathic Meteor Swarm?

Because the fighter doesn't want to have to yell out an anime attack name while fondling dried goat turds just to do an extra fancy sword swing?

More specifically, I want to avoid making all characters feel like magic users. I guess 4e did something similar, and it made characters feel kind of same-y.

For Theurgy the one people always turn a blind eye to is sacrifice. It's everywhere in the sources, but nowhere in the games, mostly replaced with prayer and the idea of sin removing your powers. But nah, if the monkey wants to keep or recharge his magical mojo they'd better be burning some sheep kidneys on a regular basis.

One thing I wanted to do whenever I get to making the original system mentioned above is that non-divine theurgy would involve writing up contracts with various spirits. They'd be written more like a class feature instead of being all legalese. The two most important parts of the contract would be Services, which lays out what the spirit will do for you, and Payment, which lays out what you'll do for the spirit. This might be a cost paid regularly, or every time you call on that spirit. I think it could be a really fun magic system, and some contracts might offer considerable power, but at a steep cost.

But making an original system is a lot of work, so for now I'll stick to working with D&D. A lot of the aspects of this overhaul only make sense as a mod to D&D; in an original system I'd do something completely different.

[noob]

Why do they have to be different mechanically? Is there a reason why a character casting Martial Meteor Swarm needs to have different mechanics to one using Homeopathic Meteor Swarm?

For Theurgy the one people always turn a blind eye to is sacrifice. It's everywhere in the sources, but nowhere in the games, mostly replaced with prayer and the idea of sin removing your powers. But nah, if the monkey wants to keep or recharge his magical mojo they'd better be burning some sheep kidneys on a regular basis.

More than a few modern religions turned away from the practice of sacrifice preferring to insist on contrition, obedience to the divine precepts and prayer.(there is still a lot of modern religions with sacrifices, they just happen to usually not fit the vision of the "dnd cleric" people have(life cleric))
Those religions have an indirect impact on fantasy clerics.
What would make the most sense would be for the religious customs clerics follows to depend on the god.
Many fantasy gods would be justified in asking for specific sacrifices in exchange for keeping their clerics powerful according to their precepts.
That could be fun to have some specific god ask their clerics to be virgins all their lives so that they can work for them harder or seeing a god of wealth directly grant their clerics divine power in accordance to how much money they except their clerics to bring back.

[Anonymouswizard]

Because the fighter doesn't want to have to yell out an anime attack name while fondling dried goat turds just to do an extra fancy sword swing?

Wow. You seem to be watching the wrong anime.

I'm also pretty sure that the Wizard doesn't want to do that.

More specifically, I want to avoid making all characters feel like magic users. I guess 4e did something similar, and it made characters feel kind of same-y.

Then why did you make all the characters magic users?

More seriously you seem to be engaging with what people actually disliked about D&D 4e: putting (almost) everybody on the same resource system. That's not bad, many games do it for their magical characters (and IIRC LFG does it for it's nonmagical characters), but people tend to not look past the power structure to see the powers if literally everybody uses it.

I don't know. Maybe it's because the d20 hack I've begun designing doesn't care about the origin of your powers beyond 'it's supernatural okay'. But admittedly it's chasing different design goals (there's literally no mundane PCs, picking your classis picking your magic).

[LibraryOgre]

One thing Ars Magica does is that the various power sources (Divine, Magic, Faerie, and Infernal) have differing effects on each other. A Divine area really helps Divine powers, but tends to suppress others, especially infernal. Magic and Faerie help each other, a bit, and themselves a fair amount.

However, I think your general idea of power sources works, and I think "all classes are full casters" works, too... it's more or less what 4e did, with fighter "spells" being super-cool attacks, rather than actual wizardry.

One option to balance them could be somewhat thematic... Mundane requires an item, a lot of the time, but might be done without them (i.e. "I have the 'Cleave' ability, but no weapon... can I do it with an unarmed attack?"). Arcane usually requires a combination of words, gestures, and material components, meaning spells have "ammo". Theurgic requires rituals... prayer, non-consumable holy items, and time. Supernatural might have a very limited set of capabilities... almost a domain. Within the domain, they're cool, and require mostly concentration (not Concentration, just "thinking about the thing"), but outside their domain, they're a lot more limited.

So, a fighter might have abilities to trip people, shove them, or hit lots of people. Thieves might have crippling powers and parkour stuff. Mages (and artificers and alchemists, if you like that sort of thing) have their spellbooks and spell components. Priests have little they can do without time to pray, but aren't limited to a small spell list (in 5e terms, all their spells are rituals, but they have all the cleric spells). Warlocks can do like 5 things, but they do them well.

[Herbert_W]

I think I see what problem you're trying to solve here. I can see this working very well in a complete original system, but I'm not convinced that this is a good solution to that problem in a 5e adaptation.

I'll spell out the problem that this solves: most classes in DnD (and indeed pretty much any system that uses classes) are an answer to two different questions at once:

• Why does this person have special abilities? ('Source,' in your terminology.)
• What can this person do with their special abilities? ('Method,' in your terminology.)

This can lock players in to an answer to one of these questions that they don't want based on their answer to another question. For example, I can easily see a player wanting a character with the source of a sorcerer but the method of a warlock, or vice versa.

By separating source and method, you separate these questions and allow for the full range of combinations of answers.

The first "gotcha" that I think you're going to hit is that many classes are tied to a specific source as an integral part of their class identity. Fighters are definitely Mundane, to the point where being Mundane yet (theoretically) keeping up with supernaturally empowered guys is the foundation of the fantasy and the appeal for many players. A Theurgic wizard doesn't even make sense, and is at best just a confusing way to say 'cleric' or 'warlock'. Complicating matters, other classes are tied to a specific combination of sources - e.g. a 3e duskblade is a specific blend of mundane and arcane that's greater than the sum of its parts, and a 3e factotum opportunistically draws from pretty much all sources. In fact, most classes will end up drawing from a present list of sources, and usually a short one. Very few classes will benefit from making this choice available to players. It makes more sense to wall off this choice to those classes, so that players who don't have to think about this choice won't need to.

This could be avoided if you abandon the normal lineup of classes and make new classes such that every source/method combination does in fact make sense. One potential source of inspiration for what these classes should be is the terminology that players use when describing existing classes - e.g. "prepared caster" vs. "spontaneous caster", "tank," "DPS main," etc. You could then present the classic lineup as recommended build packages. That's a whole lotta system soup though - we need to cut down on the soup by being very ruthless with either the number of sources of the number of classes that we allow in the game.

This also means that we've strayed very far away from being a 5e adaptation.

Is there another way you might handle these different power sources? Are there any other power sources I might have missed?

I'd take this in the other direction - compacting, instead of adding.

Much of the awkwardness that comes from allowing any source to combine with any method comes from crossing the martial/caster divide - and, specifically, it occurs whenever you consider Mundane as a source for methods that are normally associated with Arcane/Theurgic/Supernatural (i.e. 'magical') sources. Interestingly enough, the other way around works well. A fighter who hasn't really learned their art but gets their skills through a pact or ancestral spirit etc. could be a very interesting character concept.

I think the fundamental issue here is that Arcane and Mundane are actually the same thing: learning. "Mundane" is just learning with the assumption built-in that it's specifically nonmagical, which is why it doesn't fit magical methods. Likewise, Arcane is just learning that's specifically magical.

Theurgic could be renamed "granted," which means basically the same thing but without the unintended and confusing connotation that the ability is granted specifically by a god.

All of the supernatural abilities that you've listed are innate, and we need a category of racial abilities such as e.g. poisonous bites, so we can rename this category 'innate' and include them.

• Learned: Anyone could do this, but they'd have to learn how first. This encompasses what you've called Mundane and Arcane.
• Granted: Somebody else does the cool thing for the character. You've called this Theurgic.
• Innate: This ability is a consequence of what a character is. This would include racial and what you've called Supernatural abilities.

Since we want each source to have some mechanical effect that's, let's come up with a strength and a weakness for each category:

• Learned abilities have fewer or no limits on how often they can be used. A learned spell, for example, may be cast without expending personal power given enough time and material components. However, learned abilities often require certain tools. A learned spell may require material components, and a learned martial technique is only usable when armed appropriately. Learned abilities may be lost when a character is affected by a memory-altering spell such as feeblemind.
• Granted abilities have use limits that can be negotiated and circumstantially raised. However, they require the continues good will of whoever grants them. They also require the continued existence of whoever grants them. Character with granted abilities are at risk of being depowered over the long term as they appease or replace their patron.
• Innate abilities may be overused at the expense of a character's health. They may be lost when a character shapeshifts, or mentally projects outside of their body. Innate abilities may be suppressed by specific weaknesses - the idea of cold iron preventing all fey innate abilities from working is a good one.

But I don't know that I want to force players into a specific stat. It's something I need to give more consideration to.

One compromise might be to give each ability/spell/etc. a key ability score, and each class' access to it a key ability score which may be different. Players can use whichever of the two is higher. For example, a wizard casting charm person could use either charisma (becasue that's the key ability for that spell) or memory* (because they're casting it with wizardry) to determine the save DC. A wizard who specializes in enchantment could get by with either a high charisma or memory score, with both builds having their own advantages. For example, charisma could be important for more than just determining save DC - an opposed charisma check is needed when trying to control a target charmed by multiple casters, and you can't substitute memory for charisma there - but a wizard with a high intelligence will be able to make good use of a wider range of spells due to being able to substitute memory for whatever the various key abilities may be.

* What I think 'intelligence' should be called, for reasons that (I think) I've covered in another of your threads.

[Greywander]

Thanks for everyone who's replied so far. I know I can sometimes seem like I'm arguing against whatever someone might be saying, but I really do appreciate it when someone takes the time to post. I think this sort of discussion helps to draw out the underlying assumptions, goals, and other points that we might not be consciously aware of at first. So if I'm arguing against what you said, I think it's more just me exposing another layer of my thoughts, opening them up to be analyzed critically. Eventually, we'll cut through the things we think are important and get to the things that are actually important, if that makes sense.

What would make the most sense would be for the religious customs clerics follows to depend on the god.

Indeed, this is why I think the idea of writing up a set of three or four tenets to follow would work well. Something like sacrificial offerings could be included as a tenet.

For theurgy-as-witchcraft, you could instead choose to write up a single tenet (or two) for each power/spell you have, which represents your contract with the spirit granting you that specific power. Breaking that tenet displeases that spirit and they will refuse to grant you that power until you make amends, or ask for a greater cost to use it, but different spirits granting you other powers won't care.

Wow. You seem to be watching the wrong anime.

I'm also pretty sure that the Wizard doesn't want to do that.

You... realize I was just giving a colorful description of verbal and material components, right? Whether they want to or not, wizards do, in fact, do that.

Then why did you make all the characters magic users?

I know this was said in jest, but it really was the best solution I could come up with to the problem I was facing. I think this was my last thread on the subject, which includes links to previous threads, but here's a quick summary of the context:

Spoiler

Show

This overhaul is one I've mentioned a few times on the forums, usually referring to it as "stacking classes". The basic idea is that each class is cut down to just four levels, and you'll stack several classes together as you level up. My plan was, and still is, to take base classes from the PHB and rework them into two or three stacking classes, and likewise to rework PHB subclasses into full stacking classes.

Initially, the plan was for each class to get a "core feature", such as spellcasting, and by stacking classes with the same core feature it would advance that feature, like raising your caster level. But while caster classes would have been pretty well unified around spellcasting, martial classes would have gotten... uh, Sneak Attack? Ki? Combat Maneuvers? Part of my worry is that caster classes would easily stack to advance spellcasting, while martials would be stuck with many different weak core features that wouldn't stack with each other.

I was also playing with an idea for an analogue to multiclassing, which would look more like turning your character into a gestalt; basically, you'd be able to spend extra XP to get more classes than you'd normally have, which could in turn cause you to have more core features than you're supposed to. Suddenly, you'd be able to have full caster progression and pick up some martial core features, or double up and get caster progression beyond that of a full caster.

One thing that may have really brought things to a head was that I simply couldn't figure out what to do with half caster classes like paladins. Is their core feature spellcasting, or is it a martial ability?
If it's not spellcasting, how do they fuel their smites?

In the end, it simply made more sense to take whatever core features I would have given non-caster classes and simply turn those features into spells. For example, all smites are now just spells. Then, every class is running on the same progression system, and you're not punished for combining a martial and caster class together. By adding methods and sources, I can then create some distinction between characters, so that one character can feel more like a mage while another character can feel more like a warrior, despite them both technically running off of the same system.

More seriously you seem to be engaging with what people actually disliked about D&D 4e: putting (almost) everybody on the same resource system. That's not bad, many games do it for their magical characters (and IIRC LFG does it for it's nonmagical characters), but people tend to not look past the power structure to see the powers if literally everybody uses it.

This is were methods will come in to play. There will be a default method, which will probably look like spell points or spell slots or something. All characters have access to the default method. Some classes (but not all) give you access to other methods, and you can allot what are basically "caster levels" to different methods (including doing splits). So the fighter, for example, might (not sure yet) give you a bunch of weaker spell slots that refresh on a 1 minute rest, making them limited use in combat, but more or less refreshing completely between combat. A rogue might (again, not sure) give you at-will casting for weak spells, with a few higher level slots that refresh on a long rest. Warlock would give you pact magic slots that are higher level than normal slots and refresh more quickly, but are more limited in number. Artificer puts spells into items, each with their own spell point charge, which can then be handed out to party members. And so on.

I won't know how well it will work in practice until I try it, but my hope is that methods will create enough distinction for characters to feel truly different from each other.

One thing Ars Magica does is that the various power sources (Divine, Magic, Faerie, and Infernal) have differing effects on each other. A Divine area really helps Divine powers, but tends to suppress others, especially infernal. Magic and Faerie help each other, a bit, and themselves a fair amount.

I like the idea, and I've even considered how one might do a magical rock-paper-scissors game, the issue is that I'm not sure how you'd get it to work. If, say, sorcery beats wizardry, what does that mean? Does simply being a wizard make you more susceptible to sorcery? Spellcasting is usually a one-sided interaction, and rock-paper-scissors only makes sense if there's a two-way opposing interaction, even if that interaction is simply defending yourself against something. For example, a pokemon's type comes into play when they take a hit. But again, simply being a wizard isn't enough. Wizardry is something you do, not something you passively are.

Another concern is that a PC caster might find themselves fully negated by an enemy caster of the "wrong" type, which wouldn't be too fun. I like the idea of soft counters, though, I'm just not sure how to best implement them. But I'll give it some thought.

One option to balance them could be somewhat thematic... Mundane requires an item, a lot of the time, but might be done without them (i.e. "I have the 'Cleave' ability, but no weapon... can I do it with an unarmed attack?"). Arcane usually requires a combination of words, gestures, and material components, meaning spells have "ammo". Theurgic requires rituals... prayer, non-consumable holy items, and time. Supernatural might have a very limited set of capabilities... almost a domain. Within the domain, they're cool, and require mostly concentration (not Concentration, just "thinking about the thing"), but outside their domain, they're a lot more limited.

I think this will largely come down to methods (e.g. the example of artificers putting spells into items), but with such a short list of sources, it should actually be feasible to write up separate lists of things like spell components for each source. Some methods might bypass some components (e.g. artificer items act as the material component).

I think I see what problem you're trying to solve here. I can see this working very well in a complete original system, but I'm not convinced that this is a good solution to that problem in a 5e adaptation.

If I made an original system, I would just make each different magic system completely separate. That's something on the backburner for now, though. The advantage of doing a mod of an existing system is that you can only tinker with the parts that interest you, and for everything else just say that it works like it does in the original.

The first "gotcha" that I think you're going to hit is that many classes are tied to a specific source as an integral part of their class identity.

This isn't a particular concern for me, as I take the view of a class as a package of features. Since I'm rewriting classes anyway (as that's the core of the overhaul), that will be the philosophy I use. As mentioned above, this is for the "stacking classes" overhaul I've talked about before, so mixing and matching thematically disparate classes is expected and encouraged, and will give a lot more freedom to players to build interesting characters.

As I've also mentioned, some classes will also unlock access to certain methods of using your powers/spells. My expectation is that methods, rather than sources, will play a greater role in creating class identity. For example, the artificer who puts spells into items will still feel like an artificer regardless of whether those items are mundane inventions, infused magic items, holy relics, or a homeopathic mishmash of monster parts.

A Theurgic wizard doesn't even make sense,

But a priest with a book of religious rituals does. The mix'n'match nature of stacking classes means it can make a lot of sense to take wizard as one of your classes just to get access to the spellbook, which can be refluffed to suit whatever character you're playing as. Maybe it's a book of rituals or prayers. Maybe it's a collection of blueprints for inventions.

I'm honestly kind of hyped to see what kind of cool combos people will be able to make once this is done. I'll have to be careful about how things can be used together.

I think the fundamental issue here is that Arcane and Mundane are actually the same thing: learning. "Mundane" is just learning with the assumption built-in that it's specifically nonmagical, which is why it doesn't fit magical methods. Likewise, Arcane is just learning that's specifically magical.

Maybe it's my own stubbornness, but I just have trouble seeing how a mundane ability could be shut down by an anti-magic field, or dispelled with Dispel Magic, or get counterspelled. At some point, it seems like it would be more logically consistent to simply say, "Yeah, it's all just magic." But there's definitely a niche for players who don't want to play a magical character. That doesn't mean they can't get cool abilities. Hercules wasn't exactly a wizard, though in all fairness he was a demigod. But at high levels, a mundane character will start to look more and more like a mythical hero such as Hercules, being able to achieve superhuman feats without the use of magic. I think a lot of people will want that, so it makes sense to keep mundane and arcane as separate categories.

• Learned abilities have fewer or no limits on how often they can be used. A learned spell, for example, may be cast without expending personal power given enough time and material components. However, learned abilities often require certain tools. A learned spell may require material components, and a learned martial technique is only usable when armed appropriately. Learned abilities may be lost when a character is affected by a memory-altering spell such as feeblemind.
• Granted abilities have use limits that can be negotiated and circumstantially raised. However, they require the continues good will of whoever grants them. They also require the continued existence of whoever grants them. Character with granted abilities are at risk of being depowered over the long term as they appease or replace their patron.
• Innate abilities may be overused at the expense of a character's health. They may be lost when a character shapeshifts, or mentally projects outside of their body. Innate abilities may be suppressed by specific weaknesses - the idea of cold iron preventing all fey innate abilities from working is a good one.

It seems you've basically circled back around to my original proposals, which I suppose means I might be on the right track. It's a tricky problem, so more thought will be required.

One compromise might be to give each ability/spell/etc. a key ability score, and each class' access to it a key ability score which may be different. Players can use whichever of the two is higher.

I think the problem here is that there's little incentive to use the key ability score for the spell, rather than your class one. If you're playing an INT class, then just use INT for everything. Now, this is something that might actually work out a little better for stacking classes, as you might want to stack a class with a key ability score that was a dump stat for you, but the spells you're interested in use a secondary stat as their key ability score. My worry is that this will just overcomplicate things, though. How I wanted to do it was that each class adds spells to your personal spell list, and then you just learn a certain number of spells from your list, so you're not actually getting spells from specific classes.

A while back, I did have the idea that different schools of magic might use different ability scores. So regardless of class, enchantments would use CHA, for example. I don't remember the exact list, but I think I had two schools each for CON, INT, WIS, and CHA (e.g. necromancy was CON-based).

Something I'm already planning to do is have INT determine how many powers/spells you get (currently thinking INT mod + proficiency, so INT is a big chunk). CHA will probably give you a bonus resource for extra casting. Not sure what WIS will do, but I might move concentration to using WIS instead of CON. This way, all three mental stats are important for power usage, even if you're using e.g. STR or CON as your casting stat. With that in mind, I think freely picking your casting stat would probably work alright.

[Herbert_W]

I know I can sometimes seem like I'm arguing against whatever someone might be saying, but[…] I think it's more just me exposing another layer of my thoughts, opening them up to be analyzed critically. Eventually, we'll cut through the things we think are important and get to the things that are actually important, if that makes sense.

No worries - I tend to do the same thing.

Maybe it's my own stubbornness, but I just have trouble seeing how a mundane ability could be shut down by an anti-magic field, or dispelled with Dispel Magic, or get counterspelled.

I’m not suggesting that they could. In fact, the fact that they can’t is kinda the point. There being no universal anti-Learning field (except for, perhaps, feebelemind effects) is one of the strengths of the Learning source.

But what about antimagic fields and learned magic? Well, antimagic fields could just target magic independent of the source behind it. An antimagic field doesn’t deprive a wizard of the knowledge of how to cast spells - it just prevents them from doing things with that knowledge, like a hacker whose laptop is dead.

The same could hold for counterspell, dispel magic, etc. Whether an effect can be countered with them is completely independent of whether it comes from Learning. A fireball is a fireball and can be counterspelled, regardless of whether it came from a wizard or a sorcerer or a cleric or a guy with a scroll.

There could also be, of course, other ways to shut down magic which do depend on the Source - but those don’t work on Learning.

I like the idea of each school of magic being a keyword and antimagic fields and the like affecting things based on those keywords. That makes not only possible but also mechanically simple to have, for example, an anti-illusion field or an antimagic-everything-but-illusion field.

I think the problem here is that there's little incentive to use the key ability score for the spell, rather than your class one.

Maybe I didn’t explain that idea very well . . .

Now, this is something that might actually work out a little better for stacking classes, as you might want to stack a class with a key ability score that was a dump stat for you, but the spells you're interested in use a secondary stat as their key ability score.

. . . but you’ve gone ahead and filled in the gaps. It’s also possible that your second class might have some spells whose spell-specific key ability score is a primary ability score for you.

This has the added benefit of allowing players to play against type at the expense of versatility, regardless of whether you have stacking classes. Suppose that I want to play a socially awkward sorcerer. Normally, I couldn’t actually dump charisma and would have to merely roleplay the awkwardness with no mechanical representation - or else all of my spells would suck. Under this system I could choose some other ability score where there are spells available that use it (say for example, CON and various necromancy spells) and my socially awkward sorcerer will be fully competent in a limited and thematically coherent field.

Expanding on this: you could make any combination of class and primary ability score viable, by adding options for that class that have that as a key ability score. Say for example that you discover that there are no good combat options for sorcerers who focus on dexterity (and suppose that you see this as a problem; i.e. you want DEX sorcerers to be viable). You could fix this easily by adding some abilities with DEX as a key score to the options from which sorcerers select, such as spells that fire various arrow-like effects from a bow. This is a quick fix which doesn’t require overhauling the class.

There’s a sliding scale at work here. If you grant everyone the ability to use any ability score for any spell, the whole game becomes SAD, people choose which score to primary based on other factors, and you end up with some serious nonsense (“I use DEX to cast shout. Yes, I know there’s no somatic component. No, I don’t know how that makes sense. Look, the rules say I can, so I do, OK?”). On the other hand, locking characters into a single ability score per class (or set of ability scores with different features depending on each) makes for an inflexible system. There’s a happy medium in the middle and exactly where that medium is may depend on the design goals for your game.

I happen to think that having two options for each spell, one that makes sense for the spell itself and one based on the class used to access it, makes for a good balance which opens the gates wide enough to let in a lot of interesting character concepts and combinations while keeping the nonsense out.

How I wanted to do it was that each class adds spells to your personal spell list, and then you just learn a certain number of spells from your list, so you're not actually getting spells from specific classes.

That throws a spanner into the works for my idea of letting players choose between using a spell-specific or class-specific ability score.

Having spells not associated with a specific class is interesting. It’s a radical idea, and it would run into massive problems in the old paradigm where classes answer both the questions of “what can my character do?” and “why can they do it?” because in that paradigm it would leave the second question unanswered - but since that second question is answered by a Source rather than the class under your system, there’s no need to keep the close association between spells and classes around to answer it.

Since we’re entertaining the radical idea of decoupling spells from classes, why not also consider the radical idea of decoupling them from ability scores? In practice, ability scores only function as soft prerequisites anyways: if an e.g. spell uses charisma to determine save DC, that just means that the spell has “charisma is a primary or secondary ability score” as a prerequisite, with the spell sucking if the prerequisite is not met.

How about giving every spell a key ability score, but not for save DCs, such that players can choose a number of abilities with each score corresponding to their modifier in that score? That makes ability scores meaningful without forcing players into pump-and-dump archetypes.

If I made an original system, I would just make each different magic system completely separate. That's something on the backburner for now, though.

You could make a fairly straightforwards 5e adaptation where each school of magic is its own system with its own rules and idiosyncrasies. For example, I can easily imagine a system where enchantment spells always have verbal components which the target must be able to hear and which can be disguised as normal speech but not silenced (because it’s enchantment). I can also imagine a system where there’s coherent, sensible, unified rules that cover how all illusion spells work.

You don’t need to, but you totally could. Just sayin’.

There will be a default method, which will probably look like spell points or spell slots or something. All characters have access to the default method. Some classes (but not all) give you access to other methods, and you can allot what are basically "caster levels" to different methods (including doing splits). So the fighter, for example, might (not sure yet) give you a bunch of weaker spell slots that refresh on a 1 minute rest . . .

OK, so this is an important part of your system’s design that wasn’t at all clear from your initial post. If I understand correctly:

• When a character levels up, they gain an increase in “caster level” plus a level in a class which grants them class features.
• Those class features include Methods, which allow a character to trade away “caster level” (or trade away individual spell slots?) for other slots with different recharge requirements.
• Those features also include things that a character can do, which consume “spell slots”. These are somewhat confusingly also called Methods. Maybe that should change.
• Each feature has an associated Source, which explains where the feature comes from and determines which circumstances can cause a character to loose access to it and/or gain additional “free” (no spell slot required) uses.

Part of my worry is that caster classes would easily stack to advance spellcasting, while martials would be stuck with many different weak core features that wouldn't stack with each other.

I think we’re seeing convergent design here, because I have my own homebrew system that I’m slowly developing, and one of the fundamental design principles that I’m striving to follow is to avoid exactly this problem.

I have a somewhat different solution, though.

The main point of difference is that while you give all characters one central resource which can then be optionally traded away, I plan to give characters of different classes different types of resource to start. This makes all classes feel differentiated in terms of resource management right from the beginning. However, like you, I plan to have all class features depend on the expenditure (or investment) of resources.

To introduce some of my own terminology, a resource is considered “smooth” if it meets both of these criteria and “chunky” if it fails either of them:

• If you were to make a graph of a character’s level in classes that provide this resource vs its utility, it would be approximately a smooth line. There would be no large and sudden upticks, like the power boost that most classes see at 5th level in 5e.
• If you were to combine two classes which both grant this resource at the same rate, it won’t matter what the split between those classes is. For example, a multiclass sorcerer/wizard would have the same spells per day regardless of whether they’re a sor1/wizX or sorX/wiz1 or anything in between.

My plan for designing a system which remains balanced under copious multiclassing is to have all power increases come from smooth resources. That means that I need to design a good variety of resources that are all smooth and yet all interestingly different from each other.

Here’s what I have thus far:

• Caster level, i.e. spells par day, working pretty much as it does in 5e.
• Power points, which work a bit like 3e’s psionic power points and recharge on a short rest.
• Binder level, wherein a character can have binds (think “3e binder” or “that thing that homebrew shamans seem to always do”) whose total level equals their binder level.
• Extra attack threshold, wherein a character gets an extra attack if their unmodified D20 roll on their first attack is equal to or less than this threshold. This is not only a smooth version of extra attack, but it also evens out a character's damage per round as they’ll most often get a second attack when their first attack misses. If this threshold would be 20 or higher, it decreases by 20 and the character gains an automatic extra attack, potentially multiple times until the threshold is under 20.
• Maneuvers, which function in a manner analogous to cards in collectible card games. Characters have a “hand” of ready maneuvers, each of which activates and is dropped if a certain condition is met. Characters of a higher level have a larger maximum hand size. In order to keep things manageable for characters of high level, I plan to have many maneuvers that allow several copies to be used at once for greater effect. Characters can draw maneuvers from a “deck” in-combat, and I’m still working on the details of how that should work. In any case, having a larger hand means more opportunities for various conditions to be met and more effects when they are.
• Rogue’s luck, which is a dice pool of increasing size and number that has both offensive and defensive uses. If used defensively, the entire pool is rolled, any dice showing a 1 are dropped, and the highest single remaining die result is retroactively added to a save/check/AC/etc. When used offensively, a single dropped die is added back into the pool before rolling and the highest single result is added to an attack. This is essentially a combination of sneak attack and evasion, where a character needs to use sneak attack in order to stay capable of evasion.
• Rage dice, which are a set of dice of increasing number and size. Entering a rage means rolling a die, whose value then ticks down every round until it reaches zero and either the rage ends or another rage die is used. Various rage powers automatically expend a character’s current rage die, have an effect dependent on the number showing on that die when it is used, and allow another rage die to be rolled.
• Opportunism dice, which are a dice pool of increasing number and size. All of them are rolled at the beginning of an encounter. There’s a table of opportunism powers, all of which “cost” one opportunism die which must be showing exactly the number corresponding to that power. Better powers require higher numbers. One spent die may be rerolled as a mental action (which is a bit like a bonus action).
• Various types of damage buffer can absorb a limited amount of damage of specific types or from specific sources. For example, paladins may have a damage buffer that would protect them from damage caused by a demon but not from a hungry bear. Damage buffers clear on a short rest and can also be cleared via healing.

In order to make it easier to differentiate otherwise-similar classes, I plan to have most classes grant progression in multiple resources. Wizards and sorcerers both get caster level progression, obviously, but wizards also get binder level progression and sorcerers also get power point progression. Wizards use their binder level to prepare spells; a prepared spell counts as a bind (with a level equal to the number of metamagic techniques applied to it, minimum 1), can be cast once, and counts against spells-per-day when it is prepared. Wizards can also use spells-per-day to activate magic items instead of spending a charge. Sorcerers have a short and thematically coherent list of “bloodline spells” which they can cast using either power points or spells-per-day, in addition to the usual learned spells which must be cast from spells-per-day.

This means that multiclass characters can end up being very versatile in how they use their resources. For example, a multiclass wizard/occultist gets binder levels from both classes. That means that they could use their high binder level to bind many vestiges and conserve their spell slots for a long adventuring day, or prepare many spells at once for a short one.

That's probably more detail than is entirely necessary. Anyways, the point is that I'm sticking to a paradigm where all resources are smooth and all class features depend on resources. That's similar to what you're doing; we're just approaching the same problem in a different way.

[Goobahfish]

Hi,

Sounds like a bold endeavour. I have written my own RPG so I will give you my 2c.
It sounds broadly like you are making a defacto skill-based system (i.e. each class is thematically carved up into parallel strands which can be intermingled). That was my approach. It has worked thus far, so you are on the right track for allowing more interesting character creation options (IMO).

Now, to the meat (hopefully I don't go too abstract):

Part I:

The challenge.

We have two terms here, but we probably need four.
#1: Source: What is the in-world explanation for how this 'supernatural' thing comes about? (Eldritch parallel dimensions, the goat god?)
#2: Mechanism: How does the practitioner achieve the evocation of this 'supernatural' thing. (Read a book, sacrifice a goat?)
#3: Effect: What actually happens. (Fireball, illusory goat)
#4: Rules: How are the three above things represented by the game rules. (Resources, damage, states, saves)

My understanding is that you want a tight coupling between 2&4 and 'maybe' a coupling between 1&3 and NO coupling between 1&4? Please correct me if I am wrong.
This can work potentially, but runs into a metaphysical problem... and a rules problem.

The main complaints of 4th edition (as mentioned above) appear to include the homogeneity of different classes. If the goal is to avoid this homogeneity, you need explicit heterogeneity in one of these categories (specifically #4) which means you also need heterogeneity in one of the others (assuming some tight coupling somewhere).

The problem I see with this approach is that the thematics of 'source and effect' are usually tightly linked. Fey charm people's brains. If you are fey and aren't charming brains... you are 'failing' at 'feying'. This is a one-way interaction though, so if you are 'charming people's brains' you aren't necessarily 'feying'. So we can sort of say knowledge of #1 should give you a decent understanding of the content of #3 (but #3 does not imply #1). Dragon-power? You be breathing fire. Ancient Lich book? Life-drain and skeleton making etc.

Now we have another issue. The effect and the rules are quite tightly coupled too. If you are raising a skeleton... there should be one set of rules for how this happens in the game. Otherwise you are making a 'hard to understand' game. The only part of #4 you can play with is the rules for what you had to 'sacrifice' to get this happen in a game sense (i.e. the game representation of mechanic). Which sounds like resource management. I.e. Warlocks vs Wizards (rather than Wizards vs Sorcerers). If you separate out your resources, you are making 'creativity' harder (because you end up with two weak pools rather than one coherent pool) but usually improve heterogeneity (and 'feel'). This is fine to an extent depending on implementation. 5e does this reasonably well with spell slots. It does it awfully with things like Wild Shape, Rage, Bardic Inspiration etc.

--------

Part II:
Metaphysics

So... when I was building my system, I had a lot to think about and many different iterations to achieve the mighty 3 maxims (these were my design guidelines):
Does it allow creativity?
Is it simple?
Is it fair?

In terms of the first, we cover a lot of ground above. A bookish cleric researching healing spells in the cloistered library of the great temple is an archetype quite foreign to D&D but pretty cool (I have some work to do now...). A barboarian (a swine-themed shapeshifting rager) is another example. Or perhaps the illusionist swashbuckler who stabs while tricking foes into stumbling down stairs that looked like a solid floor. The possibilities are endless. D&D handles these badly.

Obviously simplicity and 'feel' are in tension with each other. If things work the same, they feel the same. Different abilities should 'feel' different to use. To that end I ended up divvying things up in a variety of different ways.

#1: Mechanistic magic (kind of a merger of your Mundane/Arcane). This is broadly magic which creates a physical effect for which the caster is the source of knowledge. "I make a pillar of stone to hold up the ceiling.", "I coat my weapons in shadows (by absorbing the light) to make it look cool and hard to judge." I need to have a vector to achieve this (which is usually mana, the force, the one power or some other magic source). The point here is that it is achieved by the caster using the caster's knowledge.
What are the limitations of Mechanistic magic? I couldn't create an autonomous creature. It would have to be a subset of my own consciousness or primitively programmed. But it couldn't really be an instinctual creature. I might open a gate and summon a creature, but then I wouldn't have control of it.
Could I charm someone?... not directly. If I had sufficient knowledge of how someone else's mind worked, at least sufficient to make them change their mind, why couldn't I just stop their heart beating instead? Fiddling with brains is an information-heavy task so if the caster is the source of knowledge, I am in big trouble. I could do some hypnotism-level stuff (see psychological magic below) but that is about the extent of it. Likewise my illusions will be heavily limited by my imagination and if I am bad at art, my illusions should look fake.
Spells like... comprehend languages are completely out here. Seriously... how can you 'magic' your way into understanding a language you don't understand... surely the spell to accomplish this would be inherently more complex than simply learning the languages...
There are many limitations on what mechanistic magic can do (mostly spells that don't make a whole lot of sense from a mechanical point of view - 'Find secret doors'... how does the spell know the door is a secret door?)

#2: Divine magic (broadly your Theurgy). So this is a broad category that can range from explicit gods to the spooky vibe in the forest. The metaphysical explanation is that gods are a kind of manifestation of the collective unconscious. Ergo, things that are part of the collective unconscious are also part of the world, so if an entire village thinks the forest is haunted by an evil spirit. It is (which becomes self-sustaining). Hence, animals have their own spirits that they don't explicitly pray to but you get the idea.

#3: Biological magic (broadly your Supernatural). This covers quite a bit. Some creatures could be biologically telepathic? So they can communicate in the same way that we speak without really knowing how we 'interpret' sounds (trust me speech recognition on computers demonstrates how non-trivial the reflexive parts of this kind of thing are). Shapeshifting falls under this category too. In my setting, druids are literally doing gene-activation to turn into animals (along with lots of stored energy). Firebreathing dragons, oozy oozes etc. All biological. It isn't magic so much as inherent properties of the substrate.

#4: Psychological magic (i.e. music...). So I wanted bards, but in the proper music/influence sense (rather than 9th level casters). Real armies have musicians etc so these are all 'Psychological magic'. The idea is that you are effectively hypnotising/influencing people through their senses (not directly hardwiring the brain). Most of these are very 'plausibly caused by music' effects so mostly psychological. You aren't casting fireballs by playing the guitar. You could be using a guitar to focus your mana to make a fireball, but guitars don't make fireballs.

#5: Feng Shui... so this kind of magic is about the natural 'flows' of the universe and to an extent supercedes the gods (read Daoism for an general vibe). The gods can interfere with these flows (as can mortals) as they will but that is what they are doing. Interfering. So monks are broadly in tune with the ebbs and flows of the universe and use that in their kung fu (I have sword kung fu, magic kung fu etc). It is a bit like the Force from Star Wars.

Now... all of these 5 have different mechanics. Mechanistic magic relies on an unreliable mana-resource which you get random amounts of each turn. They can also store some in their body during rest. So this kind of magic tends to be fairly unlimited in its down-time utility but not as reliable during 'peak times'.

Divine magic relies on a pool which regenerates slowly via prayer but also via 'good deeds' or 'evil sacrifices'. There are very few limitations on Divine magic because... like how could there be. Please god, cast this spell for me. Please god, find the secret door for me. I wanted a pool that ebbed and flowed on different days so roleplaying a 'crisis of faith' or 'the gods have abandoned me' or 'that was explicitly wrong' can be played out nicely. You can lose favour for doing the profane (we will come back later).

Biological Magic... is pretty limited... except for things like actual Fey where their magic is 'part of their biology'.

Psychological Magic. Songs are a resource. Once you start a song, you can change melodies half way at no cost. You can layer melodies together to make 'bigger songs'. Different melodies give you different once-per-turn abilities.

Feng Shui Magic... this isn't going to create fireballs. But it might help you with your fireball making. It might make you more in tune with the mana in the world. It might connect you to a local spirit. This one is currently kicking my arse. I was hoping to make a ying-yang kind of vibe but it became too mechanically messy to work with. I will probably have another crack at it (currently it is a generic Qi which you build up and expend by doing different things). The general idea is that you build up using simple moves to have a 'finisher'.

As you can see, my coupling goes the other way. The source, mechanic, effect and rule are pretty tightly coupled the whole way through. The way I got around this is that I have explicit ways of channelling one into the other which makes hybrid characters a lot easier to work with. Prayer can generate mana, songs can generate mana etc.

TL:DR
The biggest problem I can see is this point above (). The story around source is often very important. If you don't have some kind of tie-in between the source and the rules... you won't 'feel' authentic. It is partly why Warlocks and Clerics in 5e feel a bit BS at times.

Good clerics should lose spell slots when they 'take the easy path'. Warriors should lose honour when they flee. Wizards should occasion screw up a spell and blow up the house. Warlocks should be 'forced to do things they don't want to or lose their power' (it is a pact after all). If you do what you intend (liberating the source), you may end up shunting these things into roleplaying and DM-fiat which while it can work, can cause some serious problems for 'feel'. It's not an insurmountable problem but you are butting up against an inherent contradiction ('feel' vs 'flexibility'). It won't work out without some disappointing compromises I suspect.

--------

BTW: If you need inspiration: https://patera.fandom.com/wiki/Patera_Wiki

[Greywander]

Wow, two monstrous posts. I'll do my best to reply to them, but I'm going to condense as much as I can.

A fireball is a fireball and can be counterspelled, regardless of whether it came from a wizard or a sorcerer or a cleric or a guy with a scroll.

This is the position that D&D takes, and it's one I'm not particularly keen on myself. In D&D, a spell is a spell, and wizards, clerics, sorcerers, etc. just have different ways of accessing that spell. But at the end of the day, a Fireball is still a Fireball, and thus is subject to the same anti-magic fields and counterspells and such. This just doesn't make much sense to me, I'd think that e.g. divine power would have a fundamentally different nature compared to wizardry. A god can create an effect similar to a Fireball at the request of one of their clerics, but the means by which that effect is produced are fundamentally different to how the wizard produces that effect. It's like the difference between cooking on a gas or electric stove, or even using a solar cooker or microwave. The end result is, well, not the same (microwaving is very different from baking or boiling), but similar enough, but the means are very different. If there's a power outage, I can still cook on a gas stove.

There’s a sliding scale at work here. If you grant everyone the ability to use any ability score for any spell, the whole game becomes SAD, people choose which score to primary based on other factors, and you end up with some serious nonsense (“I use DEX to cast shout. Yes, I know there’s no somatic component. No, I don’t know how that makes sense. Look, the rules say I can, so I do, OK?”). On the other hand, locking characters into a single ability score per class (or set of ability scores with different features depending on each) makes for an inflexible system. There’s a happy medium in the middle and exactly where that medium is may depend on the design goals for your game.

TBH, the idea of just picking a spellcasting ability score doesn't really sit well with me, either. If possible, I think I'd like to see if there's some scheme that allows you to cast a spell using any ability score, but the ability score you use has some kind of effect on the spell. In such cases, there would be times where you might want to cast a spell using a different ability score because of how it changes things. I'm just not sure what sort of effect the choice of spellcasting ability score would have on the spell. I'm not sure there's a good set of generic effects that would be applicable to every spell, and coming up with a set of unique effects for each spell seems really daunting.

Basically, what I'd like is for the choice to raise one ability score over another to sort of dictate what your specialization is when it comes to spellcasting. As a metaphor, we could imagine a game where you can choose to put points into moving faster so you can dodge easier, do more damage on a hit, attack faster, take less damage, and have more health, but you don't have enough points for everything so you have to specialize. You might, for example, build a character who is really slow but has maxed defenses and hits like a truck. That would make your character more effective against some enemies (e.g. those who can't deal enough damage to get around your defense) but leave you vulnerable to other enemies (e.g. a speedy glass cannon who can get through your defense and is to fast for you to hit).

Now, I already sort of have something like this, in that INT will determine how many spells you get, WIS will be used to maintain concentration, and CHA will give you bonus resources for spellcasting. STR, DEX, and CON have their own effects (e.g. AC and HP). But that still leaves the spellcasting ability free floating. There's not a compelling reason to choose a spellcasting ability that isn't your highest stat. Perhaps we could do something with the idea of casting a spell "forcefully" (using STR or CHA), "cunningly" (using DEX or INT), or "stoically" (using CON or WIS), and that having some kind of effect on the spell. Maybe there could be some kind of rock-paper-scissors effect where the target would gain a bonus or penalty to their defense, e.g. casting a spell "forcefully" would allow the target to add their CON or WIS modifier to their AC or saving throw. But that just sounds like saving throws with extra steps.

If we can find something that works for spells, perhaps the same system could also be applied to regular attacks, making STR and DEX martials fundamentally different (also, CON martials becoming a thing?).

Having spells not associated with a specific class is interesting. It’s a radical idea,

Not really? 5e already has racial spells, and spells from feats, and the like. Most of the time, a spell doesn't care about what class it comes from, only what spellcasting ability score it's using. A sorlock can use sorcerer and warlock spells more or less interchangeably. The class the spell came from generally only matters for certain ability scores (e.g. Potent Spellcasting) and when determining how many spells you get (since you prepare a list of spells separately for each class).

How about giving every spell a key ability score, but not for save DCs, such that players can choose a number of abilities with each score corresponding to their modifier in that score? That makes ability scores meaningful without forcing players into pump-and-dump archetypes.

Whoa... that would actually work scarily well with what I had in mind. Basically, spell levels would no longer exist. If a spell is on your list, you can just learn it. Each spell has a base effect with a cost in spell points, and then enhancements that cost additional spell points.

For example, let's say we have an Invisibility spell. The base effect only costs, say, 2 spell points. It makes the target invisible, but the invisibility breaks if the target does anything, which includes moving. For another, say, 2 points, you can move at half speed, and for 3 points you can move at full speed. You can target one additional creature for every 2 extra points you spend. By spending, say, 5 points, the invisibility no longer breaks on attacking or casting a spell. So you could, for example, cast the spell (2 points) allowing full speed (3 points) on yourself and two other targets (4 points) and allow attacking and spellcasting (5 points), for a total cost of 14 spell points, or a spell slot of equivalent value.

What I was planning to do instead of cantrip scaling was that at higher levels you'd get "free" spell points toward each casting, allowing you to purchase stronger effects while expending the same resource. Eventually, might be be able to cast the spell for free, even. For example, once you get 2 free points per casting, you could cast the above Invisibility spell with just its base effect at will. It shouldn't be too hard to have an ability score also contribute free spell points to each casting.

At the end of the day, though, we still need a way to calculate spell attack and spell save DC, so that problem hasn't gone away. One of the other things I'm wary about is that it might not always make sense to limit which ability score can be used for a spell. If you learned a Fireball spell by studying magic, then INT would make sense. If it's granted to you by your deity, the WIS would make sense. If it's a projection of sheer will, then CHA would make sense. If it's a dragon-like breath weapon, CON would make sense. If it's an improvised flamethrower using a lighter and compressed flammable gas (alternatively, a bomb/grenade), DEX could make sense. If you just carry around a baby fire dragon and you're wringing it out like a wet towel, STR might make sense. As you can see, there's a case to be made for using any ability score, so how would we narrow it down?

OK, so this is an important part of your system’s design that wasn’t at all clear from your initial post. If I understand correctly:

• When a character levels up, they gain an increase in “caster level” plus a level in a class which grants them class features.
• Those class features include Methods, which allow a character to trade away “caster level” (or trade away individual spell slots?) for other slots with different recharge requirements.
• Those features also include things that a character can do, which consume “spell slots”. These are somewhat confusingly also called Methods. Maybe that should change.
• Each feature has an associated Source, which explains where the feature comes from and determines which circumstances can cause a character to loose access to it and/or gain additional “free” (no spell slot required) uses.

Sorry, I didn't mention it because I didn't want to overload the OP with tangentially related information. I can see now that this is relevant to the discussion, and I should have included it.

To the first bullet point, more or less. I'm splitting it up by tiers instead of levels, with each tier being exactly 4 levels long. Each class then exactly fills one tier, and your tier determines your proficiency bonus. When you finish a class, you advance to the next tier and choose a new class. This gives you another caster tier. As you level up the class, you stay in the same tier, but gain the features of that class.

To the second bullet point, again, more or less. I'm not counting methods against the number of class features a class gives, since I don't think it makes you stronger, just different. But yeah, you could allocate your caster tiers to any methods you have unlocked, splitting them or stacking them in the same tier. If, for example, you wanted a character who was thematically an artificer, you'd only need to pick up one class with the put-spells-into-items method, and could then mix and match other classes as you pleased instead of needing to grab more classes with the artificer method. Maybe you want an artificer necromancer, so you take both those classes. The necromancer doesn't, by itself, have anything to do with putting spells into items, but nothing is stopping you from allocating all your caster tiers to the artificer method and putting spells like Raise Dead into items.

To the third bullet point, I'm not entirely sure what you're referring to. Are you talking about things like smites? Or do you mean the idea of artificers putting spells into items to cast them? This is based off of a 5e homebrew I made where all artificer spells are first put into items with limited charge, and then cast from the item. This does a much better job of making the artificer feel like an artificer, and not just another generic spellcaster. In any case, this is still a resource pool, it's just that the resource is in the item now.

To the fourth bullet point, I want to make it clear that the whole source/method stuff only applies to spells (or "powers", which is the generic name I'm using), not to all class features. But yeah, the source determines the nature of the power, how you can lose it, and how it can be countered.

I think we’re seeing convergent design here, because I have my own homebrew system that I’m slowly developing, and one of the fundamental design principles that I’m striving to follow is to avoid exactly this problem.

I have a somewhat different solution, though.

Sounds complicated! But it should allow you to do some things with your system that I couldn't in mine. I like your distinction between smooth and chunky resources; I haven't designed any methods yet, but it seems like it would be a good idea for me to make them all smooth so that doing splits doesn't hurt the overall power of the character. Though if individual powers are also assigned to a specific method, it might make sense to make split builds a bit stronger, since each method only has access to some of your powers, not all of them.

It sounds broadly like you are making a defacto skill-based system (i.e. each class is thematically carved up into parallel strands which can be intermingled). That was my approach. It has worked thus far, so you are on the right track for allowing more interesting character creation options (IMO).

I was actually thinking something similar. The stacking classes idea seems like a compromise between a class-based system like D&D and a skill-based system. When I eventually make an original system, I think I'd like to do it as a pure skill-based system, but I'll admit that I do find it fun coming up with builds in 5e. The restrictions in class-based games are what makes it fun to see what kind of characters you can build, whereas skill-based systems let you just get everything. So maybe even in a skill-based system I'd want to try to add some kind of class-like structure somewhere, e.g. to the magic system. Vampire: the Masquerade sort of does this with your vampire clan, and with mutually exclusive disciplines, e.g. you can't get both Celerity and Temporis (and only some clans have access to Temporis anyway).

We have two terms here, but we probably need four.
#1: Source: What is the in-world explanation for how this 'supernatural' thing comes about? (Eldritch parallel dimensions, the goat god?)
#2: Mechanism: How does the practitioner achieve the evocation of this 'supernatural' thing. (Read a book, sacrifice a goat?)
#3: Effect: What actually happens. (Fireball, illusory goat)
#4: Rules: How are the three above things represented by the game rules. (Resources, damage, states, saves)

My understanding is that you want a tight coupling between 2&4 and 'maybe' a coupling between 1&3 and NO coupling between 1&4? Please correct me if I am wrong.

I don't think this is quite right. First, I'm not sure you can really split it up like this. I'd say you have the source right, and the mechanism is the method, and the spell/power is the effect. But all three tie into the rules in their own way. The spell's description details what happens when you cast it. The method dictates the resource cost, and how the spell is cast (e.g. from an item, channeled through a weapon, as a ritual, etc.). The source determines the nature of that spell, e.g. is it magic, or divine power, or an innate ability? The source can then interact with other things, such as a creature being resistant to magic but not to divine effects.

The problem I see with this approach is that the thematics of 'source and effect' are usually tightly linked. Fey charm people's brains. If you are fey and aren't charming brains... you are 'failing' at 'feying'. This is a one-way interaction though, so if you are 'charming people's brains' you aren't necessarily 'feying'. So we can sort of say knowledge of #1 should give you a decent understanding of the content of #3 (but #3 does not imply #1). Dragon-power? You be breathing fire. Ancient Lich book? Life-drain and skeleton making etc.

I think players who make characters using fey bloodline sorcery will likely be naturally inclined toward using enchantments and illusions. Trying to micromanage this and dictate which spells players can or can't use with different sources seems likely to cause more harm than good. I get what you're saying, but I think this is one place where it's just easier to trust the player than it is to force them.

Now we have another issue. The effect and the rules are quite tightly coupled too. If you are raising a skeleton... there should be one set of rules for how this happens in the game. Otherwise you are making a 'hard to understand' game. The only part of #4 you can play with is the rules for what you had to 'sacrifice' to get this happen in a game sense (i.e. the game representation of mechanic). Which sounds like resource management. I.e. Warlocks vs Wizards (rather than Wizards vs Sorcerers). If you separate out your resources, you are making 'creativity' harder (because you end up with two weak pools rather than one coherent pool) but usually improve heterogeneity (and 'feel'). This is fine to an extent depending on implementation. 5e does this reasonably well with spell slots. It does it awfully with things like Wild Shape, Rage, Bardic Inspiration etc.

Yup, since I'm going with stacking classes, then I'll be writing features like Wild Shape or Rage to either continue to scale automatically, to function in a way that scaling isn't required (e.g. like how non-damage cantrips don't need to scale), or turn them into spells that then scale as you get more resources.

As for split methods, that's something I'll need to consider. Will you get a lot more castings of weaker spells, but few/no castings of stronger spells? Or will there be some way to still cast higher level spells, but at the cost of burning through limited resources faster? A method that's just a bunch of spell points would allow the latter, since you can spend as many spell points as you need to to cast that high level spell. A method that uses spell slots, however, only let's you cast spells as strong as your strongest slot, which would then tend toward the former. Maybe that's a feature, not a bug? Some methods might work better in splits if you still want access to higher level spells, while other methods might work better if you just want more castings of weaker spells.

Obviously simplicity and 'feel' are in tension with each other. If things work the same, they feel the same. Different abilities should 'feel' different to use. To that end I ended up divvying things up in a variety of different ways.

This is quite interesting, and reminds me a lot of some of the ideas I have for magic systems for a completely original system. But D&D really wouldn't like some of these.

For example, Geomagy, geometric magic, would take the place of traditional wizardry. It's studied and learned, and it's 100% ritual-based. That means no busting out Fireballs mid-combat. Each spell is cast by drawing a magic circle with various symbols in a form of magical programming. The circle needs to be big enough to fit the target inside the circle, so often it's big enough (and complex enough) that you can't just draw one in a matter of seconds. What you can do is draw magic circles on things such as scrolls and carry them around with you. But again, the target has to fit inside, so we're talking big scrolls, as in multiple feet wide (so more of a rug). Once the circle is completed, the spell activates and stays on indefinitely, until part of the circle is erased, which breaks the spell logic (like deleting a line of code). Some spells consume material components, which must be placed in specific spots on the magic circle, and the spell will stop functioning if the components are removed or fully consumed. The strength of the spell increases with the size of the circle; you might even see cases where a city is built in such a way that its walls and streets form a magic circle.

Another magic system is witchcraft, where you seek out spirits and make contracts with them. Rather than writing up something in legalese, it should be written like a normal class feature. The contract has different sections to help guide the player in making sure they address the things they want to, but the most important parts are the Services, which describes what the spirit will do for you, and Payment, which describes what you do for the spirit. Other parts can include things like how the contract can be terminated, or if the spirit can be compelled to provide services against their will, or if you can be compelled to provide payment against your will, or what the punishment is for breaking the contract (e.g. failing to provide services or payment). You might have a spirit serve you as a familiar, and pay them in daily food and grooming. You might summon a monster to fight for you, giving it a payment each time you summon it. Or you might have a spirit possess you to grant you its power, and in return you might have to let it possess you fully for the same amount of time whenever it wants. The contract doesn't allow the spirit to do things it couldn't already do, so seeking out powerful spirits with strong abilities is important.

Sorcery is how you get "on-demand" spellcasting, but it's generally limited to just a few abilities. Dragons breath fire... and not much else, for example. Maybe you can teleport, or turn invisible, or shoot lightning. Sorceries are innate supernatural abilities, and generally act more like superpowers. With witchcraft above, you're typically contracting with a spirit to use its own sorceries on your behalf. Sorceries generally stem from a particular bloodline, depending on the supernatural source, and each one has a weakness. Again, the example of fey and iron.

I've played a bit with the concept of authority-based magic, where you command a spirit to do something, and it does it. I haven't given too much thought into how one obtains authority, or how to determine how much authority is required to do X, Y, or Z. But it's one I might develop more later on.

D&D isn't really built to handle systems like this, though, so I'll save them for an original system.

TL:DR
The biggest problem I can see is this point above (). The story around source is often very important. If you don't have some kind of tie-in between the source and the rules... you won't 'feel' authentic. It is partly why Warlocks and Clerics in 5e feel a bit BS at times.

Good clerics should lose spell slots when they 'take the easy path'. Warriors should lose honour when they flee. Wizards should occasion screw up a spell and blow up the house. Warlocks should be 'forced to do things they don't want to or lose their power' (it is a pact after all). If you do what you intend (liberating the source), you may end up shunting these things into roleplaying and DM-fiat which while it can work, can cause some serious problems for 'feel'. It's not an insurmountable problem but you are butting up against an inherent contradiction ('feel' vs 'flexibility'). It won't work out without some disappointing compromises I suspect.

I think having Theurgy require you to write up and follow a list of tenets could cover clerics and warlocks pretty well. It might be more roleplay-heavy, but so are the things like "taking the easy path" or being forced to do something for your patron. Wizards botching spells reminds me of the homebrew wild magic system I came up with... which will definitely find its way into this homebrew in some form, but probably as a method, not a source.

Anyway, I think I have a bit better idea of what to do for sources. Just keep it simple, but make them very different. There's still a lot I haven't figured out yet.

[Goobahfish]

I don't think this is quite right. First, I'm not sure you can really split it up like this.

Yeah, I was thinking about this afterwards and #4 can be sort of split up into 4.1, 4.2 and 4.3.

Example:
Fey Magic, Wizard, Charm
4.1: Has disadvantage when someone has 'brandished' cold iron within 30 ft.
4.2: You use scrolls to invoke this Fey magic which was... I dunno written in Fey blood.
4.3: Make a Will save. But I have resistance to Fey magic.

This example is perhaps a bit complex. Obviously 4.3 is where the 'work' is being done in terms of balance etc. 4.2 is just the resource management and 4.1... is quirks of the magic? This will be a pain to balance as it functionally becomes a huge table (sources x mechanics) which all need to glue together without contradiction. That said, such a table will usually make you think about combinations you hadn't thought about before.

I think players who make characters using fey bloodline sorcery will likely be naturally inclined toward using enchantments and illusions. Trying to micromanage this and dictate which spells players can or can't use with different sources seems likely to cause more harm than good. I get what you're saying, but I think this is one place where it's just easier to trust the player than it is to force them.

100% agree. I'm not a fan of straight-jacketing players. In my system I specifically have a talent (read feat) called 'idiosyncratic repertoire' which is designed to jump outside your skill choices just for that extra bit you can't stuff into your character.

Yup, since I'm going with stacking classes, then I'll be writing features like Wild Shape or Rage to either continue to scale automatically, to function in a way that scaling isn't required (e.g. like how non-damage cantrips don't need to scale), or turn them into spells that then scale as you get more resources.

So, there are a few ways to get around this issue. One thing I dislike about the way 5th edition handled these was:
#1: Druids get 2 wildshapes... then infinite wild shapes. Seriously WTF...
#2: Barbarians get 'proficiency' wild shapes... except not.
#3: Now everyone gets ability X which is proficiency times/day

The problem with #3 is the goal it to just try and grab the right 3-6 levels from each class (Hexblade being the worst offender in 5e I think) because you really just want ability X and you don't even have to invest in it properly any more.

I think 5e should have just had a 'Heroic Actions' pool which was your proficiency modifier. Then certain class abilities consume Heroic Actions (like Wild Shape, Rage, Heroic Surge etc). That way you could potentially create a character with a lot of options but they would cut across each other. If that is the build you were going for though, huzzahs. The question then is what do you sacrifice to get this flexibility.

If I am reading your goals correctly, you'd largely skip over this in many ways because there wouldn't really be 'multiclassing' but every class would be a defacto multiclass. Your issue then is what gives you 'full progression' vs 'half progression' vs etc etc.

In my magic system I basically went: Add your Ranks in magic skills together. This determines your mana and mana/turn. So you can imagine stacking multiple magic schools together makes you a much more reliable spell-caster (more mana) in a somewhat non-linear way. If you get more mana/turn, you burn through your innate mana more slowly. That said, the first school of magic is probably the most valuable from a synergy point of view with each additional school adding less and less value (as they start to overlap/crowd).

In 5e... spell slots are kind of dumb (but 'easy to use'). If you had a 3 parts wizard-1 part fighter, how many spells slots could you get? What spell levels could you cast? It gets tricky because ultimately you don't want 3 parts wizard-1 part fighter to be better than 4 parts wizard. So 4 parts wizard needs to get something for their investment. Higher DCs is a bad idea. More spell slots is probably a decent idea. I guess you could just have the spells/level cap out at 3 for the 3/4mage and 4 for the 4/4 mage...

some of the ideas I have for magic systems for a completely original system

These are super cool. I thought about the Witchery mechanic and decided to leave it to roleplaying (rather than class) mostly because I figure a fighter can make a 'pact' for magic weaponry in the same way a wizard could make a 'pact' for spells. A bit like... magic items?

[Maat Mons]

Your description of every spell having a base cost, plus the option to spend extra spell points for enhanced effect sounds very similar to 3.5 psionics. Every power had a power-point cost, a baseline cost for a baseline benefit. And most had "augments," which were additional benefits that could be gained by spending additional power points.

Your description of abilities scores dictating area of specialization in magic sounds very similar to 3.0 psionics. Every "discipline" had a key ability score, and powers were grouped into disciplines by theme, the same way spells are grouped into schools. Whichever class you picked, you'd probably be locked out of large swaths of your class power list by virtue of having ley ability scores too low to make those disciplines viable. So between two different members of the same class, how they distributed their ability scores could make them play very differently. People hated it.

Honestly, I'm not sure how many people are interested in fiddling about with distributing points across a number of areas of expertise. That's basically how the 3e skill system worked, and 5e was generally praised for eliminating it. Usually, any system based on distributing points between multiple things devolves into picking which things will keep up with increasing enemy numbers, and which will keep falling further behind.

[Herbert_W]

But at the end of the day, a Fireball is still a Fireball, and thus is subject to the same anti-magic fields and counterspells and such. This just doesn't make much sense to me . . .

Maybe we mean a different thing when we say “fireball”. To me, a fireball is a bead of concentrated magical fire that follows a trajectory and then explodes. If it’s not magical fire, then it’s not a fireball.

I see a clear distinction between shutting down a character’s ability to produce an effect and removing the effect after it is produced. The former would be analogous to cutting the power to an electric stove - and the important bit of this analogy is that it won’t work on a gas stove. The latter would be like moving the food away from the stove and putting it into a fridge - and the important bit of this analogy is that this will work just as well against any stove.

I’m suggesting that feebelemind falls into the first of these categories: it shuts down the ability to cast spells that depend on Learning. I’m also suggesting that antimagic field falls into the second: it instantly removes beads of concentrated magical fire as soon as they begin a trajectory, and at that point the way in which they were created no longer matters.

If [fireball is] a dragon-like breath weapon, CON would make sense. If it's an improvised flamethrower using a lighter and compressed flammable gas (alternatively, a bomb/grenade), DEX could make sense. If you just carry around a baby fire dragon and you're wringing it out like a wet towel, STR might make sense.

Welp, that confirms it. We mean different things when we say “fireball”. That word covers way more conceptual space when you say it. Your various balls of fire aren’t the same effect with different Sources. They’re fundamentally different effects.

It’s worth giving some thought to just how much conceptual space each spell should cover. At what point do you say “This isn’t a fireball any more, so let’s give it another name.”? I have a suggestion for an answer, and I’ll get to that later.

It sounds like you’re allowing players to make some pretty significant adjustments to how each spell works. The degree of flexibility that you’re aiming for remains me of the way that GURPS allows players to modify advantages and powers. That flexibility comes at a cost. In GURPS, that cost is complexity. You may instead choose to pay in vagueness.

Taking inspiration from GURPS, I’ll suggest allowing players to give each spell perks and drawbacks. These would apply on a per-spell basis; allowing a player to pay for loading one spell with perks by loading others with drawbacks would be ripe for abuse. Sources would then apply some number of automatic perks and drawbacks.

Since we’re reframing Sources in terms of perks and drawbacks, there’s no reason why a spell can’t have several sources to represent dependency on multiple contributing factors. Say, for example, a cleric who can call up fire from their volcano god (Granted) might also needs complex arcane motions to properly direct that fire (Learned).

As you can see, there's a case to be made for using any ability score, so how would we narrow it down?

Earlier, I posed the question of at what point you say “This isn’t a fireball any more, so let’s give it another name.” Here’s a suggested answer: if altering a spell in a given way would change it’s key ability score, then the altered version should be listed as a separate spell. There’s probably going to be other differences as well that make a separate listing worthwhile. For example, a traditional arcane fireball (INT) is a ball and subject to modification via metamagic, whereas a dragon-like breath weapon (CON) is a cone or line, and a mental image of fire that’s forced into reality (CHA) could affect an area of any shape.

This also ties in to . . .

Whoa... that [giving players one spell known per point of each ability score, selected among spells keyed to that ability score] would actually work scarily well with what I had in mind.

. . . because if players can change the key ability score of a spell under this system, they could just redefine all of the most useful spells to key from whatever abilities they have. So, lets make each spell’s key ability fixed.

In order to keep different ability scores meaningfully distinct, I recommend establishing design guidelines that apply to all spells keyed to each ability score. Since players can select any spell keyed to their positive ability modifiers, what an ability modifier’s keyed spells can’t do will have a greater impact than what they can.

I’ll make another amendment to my suggestion, and that’s that the granted spells that I just described are what I’d suggest giving a character at first level. Characters are expected to grow in versatility and complexity as they level up, and gaining more spells on their spell lists through classes won’t do them any good if they can’t learn more spells. The rate at which characters learn more spells at they level up will probably end up outstripping ability score increases. We might (or might not) also want ability scores to be relevant in spell selection at high levels.

I think the simplest solution would be to just give each character one spell from a class’ spell list every time they gain a level in that class, regardless of key ability score. That makes ability scores less and less relevant at higher levels and class selection more relevant, but that might not be a problem.

Not really? 5e already has racial spells, and spells from feats, and the like.

Perhaps I wasn’t very clear. Spells that don’t come from classes in 5e still all come from some specific source that answers the question of “why can my character do this?” Having that question answered in a way that’s forwardly acknowledged as separate from the balance-enforcing mechanics that gate access to the spell is radical.

TBH, the idea of just picking a spellcasting ability score doesn't really sit well with me, either. . . . At the end of the day, though, we still need a way to calculate spell attack and spell save DC, so that problem hasn't gone away.

Why does an ability score need to be part of that calculation? It’s tradition, sure, and there’s a lot of inertia that we’re fighting against here, but what purpose does this serve other than fulfilling expectations set by that tradition and often-misguided ideas about realism? (That’s a rhetorical question, and the implied answer is “none at all.”)

It might be a bit offensive to our sensibilities to have every charm spell have the same save DC, but . . . well, we’re already considering attaching perks and drawbacks to spells. Having a higher save DC could be one of those perks. Heck, having a variable save DC that’s boosted based on circumstance or spending an extra spell slot could be a perk.

You almost never see characters in DnD who are OK but not great in every ability score - and when you do, they’re punishingly underpowered. Removing the need to pump and dump ability scores opens up a much wider range of character concepts.

Honestly, I'm not sure how many people are interested in fiddling about with distributing points across a number of areas of expertise. That's basically how the 3e skill system worked, and 5e was generally praised for eliminating it. Usually, any system based on distributing points between multiple things devolves into picking which things will keep up with increasing enemy numbers, and which will keep falling further behind.

This is entirely true and a good point if those invested “points” give a character bigger numbers in ways that cannot be readily combined. This problem can be avoided if those “points” grant versatility while big numbers come from elsewhere, or if they grant bigger numbers that can be combined.

For an example of the first case, let's suppose that there’s some homebrew Vancian system (i.e. each prepared spell may be cast once per preparation) where each school of magic is a skill and the number of spells that a wizard can have prepared at once depends on their skill in each school, while their maximum spell level depends on the spell slots that they have available. A low level wizard with a single point in evocation might prepare burning hands or magic missile. A higher level wizard who still only has one point in evocation could prepare meteor swarm instead. Both are limited in the same way; they cannot sling fire all day. Yet, both can sling enough fire to be worth the action investment at their level.

For an example of the second case, let's suppose that there's some homebrew system where characters have several pools of power points:

• A very small pool which refreshes every round,
• and larger pools which refresh less frequently, e.g. on a short or long rest.

We run into the problem that you described if all of a character's can only draw from one of these pools each round - because if they draw from their larger pools, they waste the power points that they could have taken from the per-round pool. We don't run into the problem if we don't do that. We could allow characters to mix-and-match power points between pools, or we could allow a character to cast an extra spell per round from the per-round pool.

The problem that you point out is real but it's also pretty easy to avoid if you know what to look out for. It's certainly not inherent to all systems that involve distributing points across a number of areas of expertise.