Author Topic: Implications for Spell List Mechanics  (Read 4237 times)

Offline Nifft

  • Sr. Member
  • ***
  • Posts: 324
  • Bad At Lurking
    • View Profile
Implications for Spell List Mechanics
« on: July 26, 2017, 11:32:25 PM »
I'm thinking about a couple of different ways to slice up spell access.

Shared List, Explicit Access (like D&D 3.x & 5e) - aka "Arbitrary Access" - Spells are on spell-lists, but there's no mechanical rhyme or reason to where or why. The class lists can't be simplified into a few rules -- the list itself is the rule by which you know if a character has access to a spell. New spells must explicitly know which lists they're on; new classes must either share a spell list with a pre-existing class, or define a new access list, which may or may not be supported by future supplements.

The problems with this access mechanism ought to be familiar to anyone who has been playing D&D for a while.


Disjoint Access - Each class has a distinct spell list. Some basic functionality may be duplicated, but overall a class has a distinct feel based on a totally exclusive spell list. Lists are never shared -- a new class must have its own list. This is how 4e did things, and it worked okay for 4e.

The problem of uneven supplement support persists, and there's a new problem: new classes (and homebrew classes) require a significant effort to create, since the author to draft new spell (or power) lists comparable to those of each base class. You can't easily bridge a concept like "swords + magic" by giving access to some pre-existing sword powers and some pre-written magic spells. You need to write a whole new set of balanced powers to represent that synergy. This can be difficult.


Sublist Access (like Tome of Battle) - Each "spell" list is distinct, with no overlap. Your class has access to one or more lists. Some lists are more common than others, and some lists have spells which are similar to spells on other lists, but the important thing is that each spell appears on exactly one list. Supplements are structured as new lists, and balance with old classes can be maintained by keeping rough parity in the number of lists to which each character has access -- in other words, you can write a new ToB discipline and allow characters to trade out one of their core disciplines for the new one, which creates novel play styles without guaranteed power creep. (Power creep can still happen, of course, if the new discipline is stronger than the discarded one...)

New classes are pretty easy: give some class features, and define a distinct sub-set of "spell" lists for your class. It's also much easier to introduce one new distinct, thematic list and augment that with a few pre-written, pre-balanced lists.

ToB mechanics make expansion & customization relatively easy.


Tiered Access (like Monte Cook's Arcana Evolved) - There are three types of spells:
- Simple (everyone has access to this small list)
- Complex (each class gets a subset of these, often based on descriptors)
- Exotic (nobody gets these by default; you have to expend character resources to pay for one)

One could look at 3.5e Psionics as this type of system, as well, at least from the Psion's perspective. There is a (large) group of basic Psion powers, then a sublist which each discipline gets, and then you can use a feat to buy an "off-list" power (from e.g. PsyWar or Ardent or Lurk), which you might not have been able to get any other way.

The neat thing about this is that supplements are much easier to balance. You can introduce new Exotic spells without much impact -- those are supposed to be rare, and must be paid for by each individual character. (In Arcana Evolved, for example, each Exotic spell cost a feat, similar to the 3.5e XPH Expanded Knowledge [Psionic] feat.)

Complex spells can be added as long as you ensure that each package of new Complex spells is balanced across all descriptors. So, if you know the game has descriptors like Fire / Storm / Ice / Metal / Wood, then you can ensure each supplement introduces new content for each descriptor in roughly proportionate measure.

The most dangerous area for supplements is new Simple spells, because everyone automatically has access to those. This list should be augmented very carefully, and very seldom.

Finally, the issue of new classes, with access to a different set of descriptors -- these interact very well with new spells, if the supplements are always careful to balance new spells among all the various descriptors.

Adding a new descriptor is problematic, of course, so you should take care in covering as much conceptual space as the game is expected to support across its entire history.


You can also do the Descriptor / Keyword thing without any Tiers, too. I'll call this Descriptor Access. Under that sort of system, you could have descriptors which are relatively abstract, and by using keywords to grant access, you can pretty easily create classes that are thematic, yet have very compact mechanics, and furthermore their mechanics will define access that includes new supplementary content automatically.

Here's an example of a few effects translated from a few 3.5e, level 2 spells:
(click to show/hide)




IMHO the two most interesting & viable mechanics are the Sublist Access (like ToB), Tiered Access (like Arcana Evolved), and Descriptor Access.

So, which is better?

One distinction between the two is what happens when you try to augment a subset of effects with a customization option. For example, let's say you want to represent a character whose fire effects feel holy, so you add a class feature / talent / feat like:

Generic Cleansing Sunfire - Your fire effects are more damaging to undead & fiends, and when you use an effect that targets yourself, you can remove some detrimental conditions.

In the case of Sublist Access, you've generally got access to a sublist which favors that descriptor, or you don't. For example, in the ToB itself, there's exactly one discipline which features [Fire] effects, so you either have that discipline on your class list, or you don't.


Under Sublist Access, you'd probably write an option that targets a specific sublist:

Sublist Cleansing Sunfire - When you damage a fiend or undead creature with a Desert Wind strike, you deal +X damage. When you use a Desert Wind boost, you may remove one of the following conditions from yourself: poisoned, shaken, hexed, ugly.

The upside of this sort of enhancement is that it's "self-contained", by being limited to one sublist, so you're unlikely to see unexpected interactions with a future supplement.

The downside is that the self-contained nature of the sublist means that you're unlikely to see any unexpected interactions in general.


Under Tiered Access or Descriptor Access, there are a lot of descriptors / keywords, so you can write a more general feature:

Descriptor Cleansing Sunfire - When you damage an undead creature or a fiend with a [Light] or [Fire] effect, you deal +X damage. When you use a [Fire] or [Light] effect on yourself, you may dispel one detrimental condition with any of these descriptors: poison, fear, necromancy.

The immediate benefit is that your Desert Wind feat(ure) can improve your non-maneuver abilities, such as fire & light spells (for a Jade Phoenix Mage or a Ruby Knight Vindicator), or class features that are light effects (for a Shadow Sun Ninja).

Also, of course, a future discipline focused around Light or Fire would be able to benefit from the feat(ure) without modification.




So, which is better?

IMHO it's hands-down Descriptor Access.

The flexibility of access, flexibility of augmentation, and flexibility of addition means that there's a huge design space available.

If there's interest in this sort of thing, I can show some of my ideas for what makes good descriptor design & helpful keyword choices.

Offline Stratovarius

  • Forum Host
  • Administrator
  • *****
  • Posts: 7678
  • Arhosan Emperor
    • View Profile
Re: Implications for Spell List Mechanics
« Reply #1 on: August 02, 2017, 03:52:22 PM »
I'm curious to see what the list access would look like if you coded all of the PhB spells via descriptor now.

Offline Raineh Daze

  • Epic Member
  • ****
  • Posts: 10577
  • hi
    • View Profile
Re: Implications for Spell List Mechanics
« Reply #2 on: August 02, 2017, 04:13:29 PM »
I think it's worth noting that the Descriptor is more like an expansion on the sublist thing--just substituting descriptors for the appropriate lists. It would obviously be a terrible idea to give any one class free pickings from everything except in a very constrained system else you end up back with the "wizards are best" problem anyway.

But I'm interested.

Offline Nifft

  • Sr. Member
  • ***
  • Posts: 324
  • Bad At Lurking
    • View Profile
Re: Implications for Spell List Mechanics
« Reply #3 on: August 02, 2017, 06:37:33 PM »
I'm curious to see what the list access would look like if you coded all of the PhB spells via descriptor now.
I could probably do a few levels.

One thing to consider, though, is that the set of descriptors which I choose is going to deeply influence the allocation, because part of the descriptor access system's goal is to present fair choices.

For example, let's say I have a bunch of spells that in 3.5e would be [Fire] & [Cold], and I decide to create these descriptors:
- Fire
- Sun
- Summer
- Water
- Night
- Winter

Those 3.5e [Cold] spells are probably going to go under Winter, not Water, since there are legitimate 3.5e [Water] spells to populate the Water descriptor.

On the other hand, if I only have Fire and Water, then [Cold] spells would probably get stuffed into Water.


For D&D in specific, there are spells that can't reasonably be severed from a class. For a Wizard, this includes stuff like detect magic, read magic, Rary's mnemonic enhancer, and permanency. Those should probably go in a special category -- Common spells for the first two, and a Wizard-specific category ("Arcana") for the latter two.

IMHO, the D&D list is a poor fit, but that's probably what makes this an interesting exercise.

I think it's worth noting that the Descriptor is more like an expansion on the sublist thing--just substituting descriptors for the appropriate lists. It would obviously be a terrible idea to give any one class free pickings from everything except in a very constrained system else you end up back with the "wizards are best" problem anyway.

But I'm interested.
Yeah, you don't want to give access to everything.


Offline Raineh Daze

  • Epic Member
  • ****
  • Posts: 10577
  • hi
    • View Profile
Re: Implications for Spell List Mechanics
« Reply #4 on: August 02, 2017, 06:50:01 PM »
Just be careful to not completely bone spontaneous casters in the process. :P

Offline Eldritch_Lord

  • Full Member
  • **
  • Posts: 173
  • Master of Magic
    • View Profile
Re: Implications for Spell List Mechanics
« Reply #5 on: August 02, 2017, 10:23:04 PM »
For D&D in specific, there are spells that can't reasonably be severed from a class. For a Wizard, this includes stuff like detect magic, read magic, Rary's mnemonic enhancer, and permanency. Those should probably go in a special category -- Common spells for the first two, and a Wizard-specific category ("Arcana") for the latter two.

IMHO, the D&D list is a poor fit, but that's probably what makes this an interesting exercise.

I'd say the D&D list fits the idea just fine, if you look at things right.  Those spells aren't necessarily either associated with the wizard or associated with each other (permanency doesn't scream "wizard" so much as "self-buffing full caster"); I could easily see splitting "spells that deal with other spells" into at least three groups: Symbols for spells dealing with written magic (read magic, transcribe symbol, glyph of warding, and so forth), Metaspells for spells dealing with prepared spells and spell slots directly (Rary's mnemonic enhancer, Mordenkainen's lucubration, imbue with spell ability, and so forth), and Enhancement for spells dealing with enhancing other spells (permanency, finding the center, spell enhancer, and so forth).

The wizard would obviously want all three of those, but Metaspells and Enhancement fit the sorcerer nicely with its whole "I control magic intuitively" schtick and Symbols would fit the archivist and wu jen given their use of written magic.  Detect magic is something every caster would want, but could just as easily fit in a "spells that detect things" category alongside detect evil, detect poison, and discern lies rather than a "spells that deal with magic" category.

That's not to say that there shouldn't be individual class lists--psions having disciplines, clerics having domains, and the like is a good way to differentiate members of those classes--but there are enough spells out there that I'd say every spell can can grouped with a pretty large number of other spells without needing to go in a class-specific list, depending on how finely you want to divide things.

Just as a rule of thumb, 2e divided all the priest spells into roughly 27 spheres and the current arcane spell schools can be divided into 3 or 4 distinct-feeling subschools on average, so between 24 and 30 descriptors totals is probably where you want to aim for so they feel neither too vague or too restrictive.

Offline Archon

  • Sr. Member
  • ***
  • Posts: 334
    • View Profile
Re: Implications for Spell List Mechanics
« Reply #6 on: August 03, 2017, 03:33:14 AM »
Descriptors probably work better from a perspective of supporting existing classes through expansions, but maintaining balance can still be hard - not all books can add to all descriptors evenly, if there are thematic restrictions (or just antsy writers). Sub-Lists are really easy to write, in many ways (Think of the number of TOB home-brew disciplines knocking around), and that means that they will be common, even if any one is not so well supported. I like it as a elegant system for avoiding the issues around spell lists, without the headache of constantly balancing expanding descriptor lists.

Also, and issue that might come up with descriptor lists is the confusion between the lists form and function - does the [water] descriptor have healing spells? damage spells? utility? Whichever you pick, any water-themed class will have all of those spells and abilities - conversely, does a Ice witch gain access to the [debuff] descriptor (even though there are fire or life themed debuff with that tag). So you will have big complicated descriptors, or lots of them, and will be spending the rest of eternity debating the relative merit and balance of [Dragon] over [Terrain Modification], both as a writer and a player. Of course, this might be a good thing - talk and thought is a good thing for a game to have.

TL;DR: Sublists are easy, elegant, and promote writing, Descriptors are complicated, and promote discussion, both are good.

Offline PlzBreakMyCampaign

  • DnD Handbook Writer
  • ****
  • Posts: 1962
  • Immune to Critical Hits as a Fairness Elemental
    • View Profile
Re: Implications for Spell List Mechanics
« Reply #7 on: August 03, 2017, 06:39:05 PM »
*Pokes head in*
I'll add a pet peeve of mine: ToB is a nightmare if you are lucky enough to be looking up a specific maneuver. "Uh, what discipline was that in again? Grrr time to go look through many pages rather one spot like any other PHB spell..."

Design space is good, but playing a game works like this:
Quote
1) Read over abilities in whatever order they are provided. HTMLis the wonderful for this.
2) "Oh I like that spell, how do I get that?" aka its a wizard spell. If you're working off descriptors, how the heck is a player going to know what classes have access to what descriptors once they've found a nice thing that they want?
3) Planning begins on getting nice things. This stage "kills" a lot of character ideas for systems not flexible enough to accommodate the player.
4) Play begins so players can work up to having fun with those nice things.

D&D is great at doing the above. Well, the SRD is. Otherwise you need to use my sorted spells spreadsheet. In fact, I was doing this process just today and forgot about it until writing this sentence. So, uh, can someone elaborate on "the problems with this access mechanism"? Is it just the fact that the designers didn't know their own game (aka all the base classes that had casting when adding spells or all the spells when adding base classes)? Is it because players enjoy making "all the things" builds where they have lots of options?

We have to be careful that we aren't blaming the structure of 3.5's casting system, rather than its actual spell text. Clearly there are a lot of broken spells.

Offline Nifft

  • Sr. Member
  • ***
  • Posts: 324
  • Bad At Lurking
    • View Profile
Re: Implications for Spell List Mechanics
« Reply #8 on: August 04, 2017, 01:41:32 PM »
*Pokes head in*
I'll add a pet peeve of mine: ToB is a nightmare if you are lucky enough to be looking up a specific maneuver. "Uh, what discipline was that in again? Grrr time to go look through many pages rather one spot like any other PHB spell..."

I feel like that's more a factor of media than system.

If you use sites like dndtools (which does give you online search for Maneuvers) then you'd have the same convenience -- albeit illegally.

The 3.5e PHB lists spells together, but alphabetically -- it's not like 1e PHB which listed spells by class:level, so you really did have all of your choices in one place. So you'd be turning pages in the 3.5e PHB to find two different spells anyway.


IMHO, the best systems would be "open-source" enough that custom search tools could be legally built.

I'd like to be able to build a filter for classes / options / descriptors / characters which would allow you to see exactly what your choices mean, and that includes knowing what your choices are.

Just be careful to not completely bone spontaneous casters in the process. :P
If we're restricted to D&D classes, I'd be more in favor of 5e-style Cleric / Druid / Wizard casting, where you prep a list of spells each day and then spontaneously cast from that list all day. It combines the better parts of Sorcerer (limited number of tactical choices) with the better parts of Wizard (easily swap choices around on a strategic time scale, away from the table).

But for me, this is more about designing new games which aren't limited to re-skinning D&D -- of course you could also use this to make a new D&D, and lots of people like D&D including me so that's cool too. It's just not what I'm aiming for.

Offline Bronzebeard

  • Full Member
  • **
  • Posts: 145
  • HELP! I'm lost!
    • View Profile
Re: Implications for Spell List Mechanics
« Reply #9 on: December 09, 2017, 06:39:12 PM »
...
You touched on the problem of forwards compatibility; where old classes have to be revised when new material comes out in order to use it.
For that reason alone I prefer a list or a sublist connection.

Also, as Maelphaxerazz eloquently put it, spells problematic nature comes from what they can do. I'm trying to fix it by deconstructing spells and limit what each character can do so each player will bring something to the table (not trying to steal your thunder or anything, but if you have a minute check it out. I also am using keywords).

I'd say the D&D list fits the idea just fine, if you look at things right.
I second Lord's notion. Not only you should prefer the underlying mechanic over the overlaying fluffy part for the sake of betterment of the game, you should also not stick to habits just because these are the ways things supposed to be. Heck, take 20 years back and you play a game where elves can't play wizards. It's blasphemy!

Just as a rule of thumb, 2e divided all the priest spells into roughly 27 spheres and the current arcane spell schools can be divided into 3 or 4 distinct-feeling subschools on average, so between 24 and 30 descriptors totals is probably where you want to aim for so they feel neither too vague or too restrictive.
I really liked the 2e sphere thingy. I was gravely saddened when 3rd came around inrespect to cleric.


I'd love to hear where you got to.