There are three topics that I want to discuss here, all of varying relevance to my next D&D campaign, which I expect to remain a very low-level game. (1) Limiting D&D's high-level game-breaking magic in a way that doesn't short-change the high-level mage. (2) Limiting a cleric's ability to turn undead, so that you can have undead attack a party with a cleric and not just have them blasted out of existence right away. (3) Making wands and staffs into something other than charged spell-batteries or "super-scrolls".
1. Nerfing the Archmage
This rule (meant for classic D&D) was partially inspired by the way spells are memorized in AD&D. In AD&D, it's a pain: each spell requires one full hour of study or meditation per spell level for any caster to memorize. By contrast, in classic D&D (and the d20 System editions), a caster's entire repertoire takes only one hour of study or meditation to refresh, following eight hours' rest.
Should I ever again run a high-level D&D game, I want to both scale this idea better, and really compound the problem for higher-level spells, to really make the use of high-level magic something a mage needs to think twice about every time.
But here are a couple of things that I don't want to do: I don't want to implement fiddly or excessively harsh penalties for high-level spells, like aging or Constitution loss, or casting costs in GP or XP. AD&D did that sometimes, and it's stupid. I don't want to make high-level spells damaging or corrupting, or difficult to cast, or likely to fail. That just screws over mages that might really need to bust out some magic in an emergency.
I rather like the feel of making high-level spells into rituals (kind of like d20 Modern, where spells up to 5th level are normal, but 6th level and higher spell effects can only be cast as lengthy rituals or incantations). But all that really does is make high-level combat magic useless, and blasting spells are the least game-breaking of all the spells. Spells like delayed blast fire ball and meteor swarm are the ones that I don't at all mind archmages slinging around with impunity! It's the polymorphs, teleports, and wishes that need to be clamped down on.
So here's the idea: spells are divided into three distinct tiers: low-level (1st to 3rd); mid-level (4th-6th); and high-level (7th to 9th). To fit this tier system a little better, the magic-user spell list receives a slight adjustment: haste moves up to 4th level, ice storm moves down to 3rd, and wall of ice remains at 4th level, as the reverse of wall of fire.
Now, low-level spells continue to work as normal: a caster can memorize his entire daily allotment of low-level spells with only an hour's study, and once cast, the mage can refresh his spells after a single night's rest. Everything from sleep up to fire ball works this way.
For mid-level spells, the rules are slightly different. Each individual mid-level spell requires a small ritual, an hour for each spell, to memorize. Thus, for example, a 12th level wizard or cleric (spells 4/4/4/3/2/1) needs about seven hours of study to memorize his full allotment of daily spells, one hour for the twelve low-level spells and six hours for the six mid-level spells. Furthermore, once cast, a mid-level spell slot takes a full three nights' rest to refresh. The caster can still exert himself or adventure as those three days pass, but the expended mid-level slots won't be available for re-memorizing spells into until the mage can rest for eight hours on three separate nights.
High-level spells are the most difficult to prepare. Each of these takes one hour per spell level to memorize, just as in AD&D (effectively a whole day, seven to nine hours). Wish is an exception, taking three full days of uninterrupted ritual just to memorize. Once a high-level spell has been cast, a full week must pass (in other words, the mage must get seven nights' rest) before the slots are refreshed and spells can be memorized into them again. Wish is, once again, an exception here: if wish is ever used to alter reality in some significant way (as opposed to just using it to mimic a lower-level spell, or temporarily conjure up a magic item, or one of the many limited uses for this spell), it takes a full lunar month before that spell slot becomes available again. (That should make using wishes to raise ability scores into a real dilemma!)
Note that when a mage does have all of his spells memorized and ready to go, he can still cast them just as easily as in the regular rules. But now, the use of high-level spells becomes analogous to the use of a paladin's cure disease per week, or a monk's quivering palm. It's not something you ever do lightly; because the opportunity cost for loosing the high-level spell must now be balanced against the thought that you may need that spell slot available tomorrow. I think this will give high-level mages the feel (and balance) that I want for D&D.
2. Limiting the Turning of Undead
In classic D&D, a cleric can run around turning undead all day long. It's a limitless, at-will ability. And even a 1st level cleric has a better than 50% chance of turning skeletons. From there, it just gets crazier. Mid-level clerics are automatically turning away or even destroying skeletons and zombies, making some of the coolest low-level monsters in the game of no consequence whatsoever to the typical mid-level adventuring party. Throw a goblin-horde at a 7th level party? Cool. Throw a zombie swarm? Pointless.
The d20 system introduced a limit on clerical turning, 3 per day + (I forget whether it was Wisdom or Charisma bonus), plus an extra 3 attempts with the Extra Turning feat. For my part, I think it's enough to give the cleric three turning attempts per day at 1st level, raising it up to four per day at 5th level and eventually five per day at 9th level.
So what should turning do, if it's now so limited? An ability that can't just be used at will shouldn't have the chance to fail as often as regular turning does, but neither do I want it to be so automatic where lower-level undead are concerned. So... what will turning do, and how will it work?
How it works is the easiest question to answer. D&D already has a mechanic for determining whether a magical effect works on a creature (the saving throw!), and yet another mechanic for determining whether monsters run away from battle (the morale check!). These mechanics ought to be more than sufficient as building blocks for a revised turning undead.
As to what it will do, I'm going to redefine the concept of turning damage. In the regular rules, each round that a cleric successfully turns undead, 2d6 hit dice worth of undead (the so-called "turning damage") are made to flee or be destroyed. This becomes 3d6 and then 4d6 at very high levels, for the weakest of undead.
Instead, I want turning to actually deal damage. Let's say 2d4 at 1st level, 2d6 at 3rd, 2d8 at 5th, 2d10 at 7th, and 2d12 at 9th (note that the cleric's turning damage is improving on levels when the cleric does not get a new spell level). At very high levels, this can become 3d12 (15th level) and 4d12 (26th level) -- after all, epic level clerics are epic. In any case, the damage dealt also takes the cleric's Wisdom bonus, if any.
Having turning actually damage the undead is fine, if turning is limited in uses per day. A cleric that uses up all his turning in one encounter may still face undead in the next, after all...
So, turning will cause this damage to all undead in a 60' long, 60' wide cone in front of the cleric. Each creature affected may save for half damage. Any undead that saves can remain in the battle and continue to fight. But those that fail the save will also flee from the battle, as if they had failed a morale check. (Many undead have morale 12 and thus can't fail morale checks, but that doesn't matter---failing a save against turning makes them flee as if they had.)
Any high-level undead that demonstrate a resistance to turning (things like phantoms, ghosts, revenants, nightshades, and liches) automatically take half damage from turning even on a failed save, and only one-quarter damage if they save. A high-level, intelligent undead like this that fails its saving throw against turning is not automatically compelled to flee; instead, it must make an immediate morale check at a penalty equal to the cleric's Wisdom bonus, and it flees only if this check is failed. (High-level intelligent undead usually have a morale score in the 9-11 range, so this mechanic ought to work just fine.)
3. Wand, Staff, Rod... Scroll, Tome, Libram
Nowhere in fantasy literature are wands and staffs ever used as storage batteries for multiple castings of the same spell. The concept of a wand of fire balls is strictly a D&D construct. And I hate it.
In fantasy, a wand or staff is usually an implement that a wizard needs in order to cast spells at all, just like a cleric's holy symbol. It's a focus. I quite like how 5th edition D&D does this: if a cleric has a holy symbol, or a wizard has a wand, they can cast spells without needing any inexpensive material components (e.g. no need for bat guano to cast a fire ball if you happen to have a wand handy). That's a very simple and elegant rule that also serves to simplify the game under most circumstances, while leaving open the somewhat fraught possibility that a wizard might lose his wand and have to start scrounging around for material components for a while. That's cool.
But what about magic wands? Well, wands as storage batteries are out. We already have spell-scrolls for that sort of thing. If you need 30 charges of fire ball, why not a magic book that has 30 fire ball scrolls penned into the pages? Call it a tome of fire ball and you're good. Likewise magic "librams" to replace the more powerful charge-battery magic staffs. Imagine finding a libram of the magi... that would be both awesome and flavorful.
Instead, I'd rather have magic wands and staffs be weapons. A staff is already a weapon, a rod is about as damaging as a mace, and a wand can function as a club in melee. But a magical staff or wand is at least +1 to hit and damage, and also carries within it an elemental missile attack, a bolt of force or fire or cold or holy power or whatever. A wand causes 1d6 damage (plus magic bonus, plus the wielder's Charisma modifier) and a staff causes 1d8 damage, with both requiring a missile attack roll in order to hit (the wand's magical bonus applies to this attack roll), the wand having a 30'/60'/90' range and the staff having 40'/80'/120'. The "ammunition" in a wand or staff is near limitless, with a wand having 30 charges when full (it regains 1d3 per hour that it goes unused) and a staff having 40 (regain 1d4 per hour), with one shot consuming one charge. Any character capable of casting spells (including thieves with the ability to read scrolls) can invoke the blasting power of a wand or staff.