Monday, 12 July 2010

Scaling Spells and the Power Curve

AD&D's approach to magic is a curious thing. Cleary, Gygax was getting a little concerned at the power levels of high-level magic-users compared to other classes; take the addition of magic resistance, for one. At the same time, there is so much more detail about spell acquisition and casting compared to the original game and, despite all the 1000-gp gem material components, not all of this is unfavorable to the wizard.

For example, there is a lot more scaling by caster level of spell duration, range and effect. Magic missile gets not just a smoother curve of additional missiles compared to Greyhawk, but extra range per level. Scaling makes sense, of course, but the problem with it in game terms is it adds to the already exponential increase in power that magic-users experience.

I won't bore you with the graphs and tables. Okay, maybe just one little one, based on the OSRIC rules. Suffice it to say that fighters can stay in the fight for only as long as their hit points hold out, so hit points are the main measure of their effectiveness (higher hit chances being balanced by tougher monster armor at higher levels, and vice versa). Even with extra attacks at level 7, a fighter's effectiveness therefore only goes up by a factor of 15 or so in 10 levels.

Meanwhile, we will assume that magic-users load up on quantifiable damage spells at each level they can, and the effect of spells scale to the spell's level, more or less. Starting with 1 die of damage spells (magic missile), a 5th level magic-user gets 3 castings of 3 magic missiles each plus, ideally, a fireball or lightning bolt which adds 5 more dice, plus whatever goodies are at 2nd level - traditionally, spells more oriented to helping friends or hindering foes. So, in 5 levels an AD&D magic-user has become more useful by a factor of 15 or even more, compared to a fighter who needs 10 levels to get there. 

In the chart above, we assume that the m-u knows the scaling spells magic missile and either fireball or lightning bolt, with effectiveness = spell's level normally, but considering any scaling damage factor of the spell. So, fireball at caster level 5 is about 5x as effective as magic missile at caster level 1, for example, not 3x.

A few ideas for the magic system I've been developing for a more OD&D-like system:

1. Tone down the scaling. This makes spells easier to resolve anyway. Ranges and effects can be improved by a means I'll show in the next post, so casters need not feel limited or predictable. I'm also considering having less intense multiplication of magic missiles, and keeping other damage spells at a fixed number of dice.

2. No duplicate spells memorized. This tones down the power curve and makes the magic user have to be more creative with the less obvious spells.

3. Contingent on 2, above, allow 1st level magic-users to have 1 extra 1st level spell memorized. This smooths the curve in the other direction, giving the magic-user one more encounter to be useful in.

A more radical solution (inspired by Ian's comment here), which also makes creating NPCs more manageable, is to not have level limits on spell slots; just allow 1-2 spells at start + 1 extra spell per character level, with new spell levels being learnable at each odd-numbered character level.

All this by way of prelude to the 2nd level spell lists, coming up next.


  1. I've used #2 for a while now. It works.

  2. I've never liked duplicate spell memorization because it conflicts so strongly with Jack Vance. Not just the "vancian magic" system in d&d, but literally with the characters in the short stories. Look, I like to memorize poetry, I've never once been able to memorize a single poem twice...

    Gygax literally says in the DMG that casting spells in combat is hard and spellcasters should use wands. Problem is Dungeon Masters the world over are niggardly with giving mu wands! All wizards are supposed to have a wand of magic missiles IMO.

    Spell slots are better off being used for a few combat spells, but mostly useful stuff like detect magic, and what not. My favorite example is Rhialto the Marvelous (I think) who--being the powerful magician that he is, can memorize, The excellent prismatic spray and two other spells.

    So Rhialto is 3rd level. Back to your point about scaling. Careful with your scaling. There are no magic items that make a magic-user's magic-missile more powerful, but picking up gauntlets of ogre power and a sword of sharpness increase the power curve of the fighter by quite a bit. I look forward to your follow-up post about your ideas!

  3. @jeff - Thanks, I thought it might.
    @UWS - And then again, magic-users get wands, so ... I do have some ways for spells to get better with levels, but I'm certainly not going to scale ranges, and still thinking about damage. I mean, when a wizard gets to "cone of cold" shouldn't he be happier than "Oh yeah, another 3.5 damage/level spell right next to my staple fireballs"?

  4. isn't your definition of effectiveness dependent on the assumption that the M-U doesn't get randomly-allocated spells? If you assume that an M-U's spells are randomly allocated, then the M-U's effectiveness needs to be adjusted for the probability of actually getting a combat apell (which is quite low).

    These models of effectiveness would be better done as survival studies. I have made a start on such work on my blog, comparing Warhammer 2nd ed with Pathfinder and a putative 3rd system (as an example). I aim to do more in the future but the upshot is, I really don't believe M-Us have the effectiveness you claim for them.

    For example, a first level magic user's magic missile is unlikely to knock a fighter down in one round (assuming they start combat not in melee), but a fighter has a high chance of wasting a M-U. You need to consider the expectations of their damage and hit points to see who is more effective, and its unlikely to scale in favour of the M-U.

  5. Thanks for your thoughtful comment, faustus.

    Random spells had crossed my mind. In the campaigns I played and gm'd there was always an understanding that m-u's would eventually get the "good" spells be it by choice, or planting scrolls as treasure ... What I am hoping to do with my spell list is achieve a little more of a balance so this isn't necessary.

    In any event, random spells does address the balance problem between m-u's and other classes. However, you could also argue it then makes the balance problem reappear, between m-u's who are lucky vs. unlucky.

    I guess a background assumption I neglected to spell out was that I am thinking of character power as contributing to a well-balanced party. I have no problem with the magic-user being a force multiplier who achieves full power only when defended by other character types. Now if you are running an arena game, a Monte Carlo toe-to-toe simulation may be appropriate, but in a cooperative party, power is going to center around who in the party contributes most to the action and can stick around when others flag.

  6. I agree entirely about mages as a force multiplier when properly defended (I prefer magic to be overpowered relative to other classes, because ... it's magic!), in fact I prefer them to be overpowered on their own too. But I'm not convinced that mages are, in D&D.

    I should get around to those simulations!