Tuesday, September 29, 2009

What do spell levels measure?

Today I'm once again venturing into the often ridiculous field of extrapolating gameworld reality from arbitrary mechanical conventions. For a long time I have operated under the assumption that spells level are roughly indicative of the power output the spell. Cure serious wounds comes after cure light wounds for a reason, after all. But it takes only a casual perusal of a few spell lists to call that definition into question.

Let's look at the first edition PHB to see what I'm talking about. How many players of fifth level magic-users would pick flame arrow over fireball? Certainly we can construct scenarios where flame arrow would be the more useful of the two, but that doesn't change the fact that fireball is the one to pick in general adventure situations. Similarly, newly minted fist level magic-users generally want sleep or magic missile or charm person.

Mending
is also a first level spell and though useful, it doesn't exactly resonate with power the way setting people on fire with burning hands does. I suppose you could argue that magically reassembling a shattered coffee cup actually requires more power than setting someone on fire, given that an act destruction is generally easier to accomplish than an act of creation. But that doesn't stop mending from being a lame choice for a 1st level M-U.

I suppose spell level could measure the energy input required to activate the spell. First level spells are first level spells because of the minimal mana needed to achieve those specific effects. Flame arrow is a third level spell simply because it is an inefficient transformer of mana into fire. But neither of these ideas gets me anywhere towards understanding why Phantasmal Force is a 1st level Illusionist spell but a 3rd level M-U spell.

Looking for a Vancian analysis, I think we could posit that spell levels as a measure of formulaic complexity. The properly trained M-U brain can only hold so many sonnets, whether they were written by sophomores or Shakespeare. Different classes learn different mnemonic exercises, allowing for different storage capacities of various types of spells.

If spell levels measure energy input or formulaic complexity then the ramifications for your campaign are enormous. Under these schemes there's nothing particularly fixed about the fact that fireball is a third level spell. The PCs might discover an ancient scroll with a previous version of the spell that's fourth or fifth level. They might then figure out that the fireball in the book merely represents the current state of the art in pyro-sphere technology. I don't know about your campaign's magic-users, but I'd personally blow a lot of time and gold pieces trying to crack the code for a second level version of fireball.

20 comments:

  1. The problem with formulaic complexity is that some some third level spells have only a verbal component, while others have a verbal and a somatic component, and yet others have a verbal, a somatic, and a material component. If all are meant to be of roughly equivalent complexity, why do some require so many more steps and mnemonics than others?

    I don't think this is a single solution problem where value X results in level Y. I think you have to plot spells on a graph with at least X, Y, and Z coordinates describing any single spell alpha.

    ReplyDelete
  2. You can rationalize it a hundred ways, but it's ultimately game balance.

    You might introduce a 4th level version of fireball for kicks, but I'd never do a 2nd level version. That's effectively an artifact. Of course an artifact is doable for a 1st level character, if you really want to base your campaign around it...

    ReplyDelete
  3. I assumed that spell levels were supposed to just space out different kinds of spells along the player's progression.

    Considering using the following rule in my next game--I thought you might like it (since the rule makes the game more insane):

    Imperfectly Memorized Spell

    This spell allows a low-level spellcaster to access *any* spell he or she wants, so long as it has a 50% of completely backfiring.

    It's like a last-ditch "cross-the-streams" gonzo effect. You want a fireball? You got it, but it might just fry your party instead. You need last minute healing but don't have access to that spell? Go ahead, but you might just heal all your enemies.

    A few mechanical details would have to be worked out for things like "charm monster", but I like the idea.

    ReplyDelete
  4. @ Cappadocius: What is more, if the only factor was formulaic complexity, then M-Us with a high enough Int could probably understand spells beyond the level they can cast.

    I would say that FC is an important factor, but I always pictured that Magic Users needed to have some kind of "mana attunement" in order to make certain esoteric concepts work properly. After all, one thing is to know something, and another entirely different thing is understanding it intimately, as I suspect Vance imagined his magic (although I have never touched any of those books. Spanish translations of old fantasy novels are quite scarce here).

    ReplyDelete
  5. Looking for a Vancian analysis, I think we could posit that spell levels as a measure of formulaic complexity. The properly trained M-U brain can only hold so many sonnets, whether they were written by sophomores or Shakespeare. Different classes learn different mnemonic exercises, allowing for different storage capacities of various types of spells.

    This is how I've always viewed it; and while it's a little before my time some older versions of D&D actually required longer periods of memorization for higher-level spells. (Which was a useful form of game balance, since taking an hour or more to re-memorize a single 6th or 7th level spell made you wary of casting them without good reason.)

    @cappadocious: The various components - verbal, somatic, material, etc. - aren't actually part of the memorization process; they're part of the final arcane act that completes the spell. The metamagic feats like silent spell in 3rd edition essentially said "I know a different, simpler way to complete the spell."

    As for "blowing a lot of time and resources trying to create a 2nd-level version of a spell", this is actually a useful way to deal with spells you believe are underpowered for their level. It's quite possible that arcane universities have proved that the common fireball spell is formulated in the most efficient way possible, but that there's still hidden inefficiencies in, say, the dream spell. (It's common in modern mathematics to be able to prove information about a mathematical problem without actually coming up with a solution.)

    ReplyDelete
  6. The problem with formulaic complexity is that some some third level spells have only a verbal component, while others have a verbal and a somatic component, and yet others have a verbal, a somatic, and a material component. If all are meant to be of roughly equivalent complexity, why do some require so many more steps and mnemonics than others?

    That's easily explained - the spells with more types of components have simpler formulae than spells with fewer components. A third level spell with only verbal components has a really complex verbal component, while one with both verbal and somatic components has a simpler verbal component at the expense of having to include a hand-dance to supplement it.

    But I'm not sure that I necessarily think that "level" should be due to a single attribute - there's no reason that level can't be a combination of energy input, formulaic complexity, and energy output after all and the theoretical thaumaturges just lump it all into something they call a "spell level".

    And of course Clerical spell levels are probably on a completely different scale. They certainly look like it when you look at the spell lists.

    ReplyDelete
  7. The problem with formulaic complexity is that some some third level spells have only a verbal component, while others have a verbal and a somatic component, and yet others have a verbal, a somatic, and a material component. If all are meant to be of roughly equivalent complexity, why do some require so many more steps and mnemonics than others?

    You are confusing complexity with number of elements.

    It is entirely possible to envision a high-level spell with a very complex verbal formula, consisting of various larynx-bending syllables that must be intoned in precisely the correct fashion in order to be efficacious, that does not happen to require material or somatic components.

    It is equally possible to envision a spell that consists of nothing more than a quick flick of the hand, a word, and a pinch of powder.

    I see nothing incongruous in assigning to the first a greater level of difficulty than the second.

    ReplyDelete
  8. Speaking to the larger question of the original post, it must be remembered that in a Vancian magic system, we are talking about mystical energies that are somehow imprinted upon the brain, and released through the casting of the spell.

    In such a context "level" means different things in different contexts.

    In the context of the knowledge needed to cast a spell, it refers to a general level of arcane knowledge. "You won't be able to cast the fireball spell until you really understand the relationship between the language of the spirits of the Elemental Plane of Fire and the psychic sigils that you create with your fingertips. It's more than just mumbling the right words and waving your hand, you know!"

    In the context of Vancian magic, the act of studying ones spell books impresses the mystical energies on ones mind. The mind can only absorb so much mystical energy at one time (hence the limit on spell memorization). However, with practice and study, it is possible to train the mind to be able to absorb more and more mystical energy. Hence we see the ability to memorize (and therefore cast) more and more spells, as the level of the caster increases.

    In other words, in a Vancian conception, don't think of it as a limit on what can be *cast*. Think of it as a limit on what can be *memorized*. Where, in this case, "memorized" means "the mystical energies are internalized after a period of study, and will be released after the spell is cast."

    Almost like a "pre-casting" of the spell.

    ReplyDelete
  9. This comment has been removed by the author.

    ReplyDelete
  10. The rationale I use is that D&D magic is essentially as goofy as Rust Monsters, level limits for species that live 5-10 times longer then a humans and all the other silly things that make it work as a game as long as you don't analyze the whys and hows too closely.

    I played one RPG, of Japanese origin I believe, that basically had every spell available at any level. A level 1 fireball is not as cool as a level 3 fireball which goes further and makes a bigger boom. If you detect magic at level 1 something magical ten feet in front of you might glow. At 8th level you can scan all objects in an 80ft radius and so forth.

    Basically the spells were based more on your character's power level then an arbitrary 'spell level'. It just felt like it made more sense to me.

    ReplyDelete
  11. I've always understood some of the "filler" spells to imply that NPC magic users do things besides kill goblins in dungeons. A first level apprentice magic user in a town quite likely memorizes "mending" every day... I've always assumed (perhaps naively) that some of the silly spell lists NPCs in modules had memorized were supposed to be mimicking that likelihood. The PCs are "loaded for bear", as it were, while the town-type NPCs were preparing for another day at the tower/marketplace/library/etc.

    ReplyDelete
  12. Maybe spell levels are like electron states, higher tiers require more energy?

    I big issue for me is whether to let PCs memorise lower level spells in higher slots. 3e allows this, earlier versions didn't. It makes a big difference - a 1e M-U could probably never cast more than a small number of fireballs per day, unless he had scrolls/wand of fireballs etc. A 3e Wizard can fill his upper tier spell slots with fireballs routinely, often giving much more destructive power.

    Most justifications for spell levels don't give any reason not to allow the 3e approach. But in 1e it clearly required work-arounds to the nature of magic, eg you needed Rary's Mnemonic Enhancer just to put Fireball in a 4th level slot.

    ReplyDelete
  13. "Maybe spell levels are like electron states, higher tiers require more energy?"

    And you have to fill your lower tiers before being able to access higher tiers? So you can't only have 9th level spells, you need to fill in your lower tiers too. Likewise you can't only have lower tier spells, because there are limited slots available - and this may require putting 'worse' spells in higher tiers.

    Hmm, I think something like this does start to provide a kind of rationale, anyway.

    ReplyDelete
  14. My response is here:

    Add More Flavor & Style to Your Wizards!

    http://gamedevonline.net/

    ReplyDelete
  15. Hey ... its magic. It's not supposed to make sense! ;-)

    ReplyDelete
  16. I like Formulaic Complexity.

    My usual explanation is that there's a preparation phase for each spell, rather than just memorization. In order to prepare a spell, you have to produce a material focus in addition to memorizing the arcanities necessary to release the spell.

    Spell levels are the technical skill required to accomplish the preparations.

    A fireball, for instance, might be a simple matter of mixing some volcanic ash with alcohol and saying a few words over it. Flame arrow, on the other hand, requires chipping a crude arrowhead from a piece of obsidian.

    ReplyDelete
  17. Further support of Jeff's concluding hypothesis of "best/common version currently known" is that in AD&D 1e there was a large chance a PC would not have access to each particular spell, at least when first gaining access to that level of spell.

    This was the biggest difference from clerics, whose deities normally granted them access to the whole salad bar, so to speak.

    So "game balance" was never intended. A magic-user who happened to (as the die declared) live or study where magic missile and fireball were known would have much more combat ability than his or her cousin who instead had only mending and flame arrow.

    To paraphrase, a significant part of "1st level spell" meant having many more chances to learn a specific, desired spell -- and less cost to do the research required for another die roll attempt to check.

    ReplyDelete
  18. For those who want more fireball in thier games, take a look at Hackmaster 4e. It effectively has some sort of fireball spell at each level, including some interesting variations.

    ReplyDelete
  19. I'm with those who've said, essentially, "why must there be a single determining factor?"

    ReplyDelete
  20. Anonymous5:43 PM

    They just represent the original designer's idea of how magic should work. The very notion of a mending spell is silly - in a system where the max number of 1st level spells you can have is 7, why would you ever have mending in your arsenal? It's just there to add colour and flavour.

    I also dispute that a 2nd level fireball spell would be that relevant. At 3rd level, a wizard is taking on creatures of less than 4 HD and is much better off using sleep, which is essentially a first level instant death spell. Sleep is way overpowered! Also charm person, which takes one target out of combat for the duration.

    I was always a bit disappointed by the available 4th and 5th level spells - some are useful out of combat, but none compared to fireball. It wasn't until D&D 3.5 (?maybe?) and horrid wilting that even 6th level spells compared. Basically every mage stocks up on stoneskin at 4th level and never looks at the other spells...

    ReplyDelete