Thoughts on Spell Lists

Spell lists should be designed around the needs of a campaign instead of a system.

A campaign where you fight evil creatures can have detect evil, but such a spell is actively harmful if you're going for a more realistic campaign with moral ambiguity; on the other hand, a campaign where you fight insane robots could have detect robot (or detect electric signature), something you'll never find in any D&D supplement (except maybe Spelljammer?). A campaign where entire dungeons will be underwater should have access to water breathing spells. A campaign where water is used mostly as a dungeon obstacle should not have water breathing spells. If the campaign takes place inland and there isn't likely to be much water at all, you also shouldn't have water breathing spells, especially as a player option that they can spend resources to take. There is only disappointment if you let the player actively choose an option you know will be useless.

Plane Shift is a classic example of a spell that can ruin campaigns. In a planar campaign it is super helpful, maybe even a virtual requirement. In another campaign, where the planes exist but most of the action is concentrated on earth, it's a "nice to have". On the other hand, on a campaign where the players are trapped in the Underdark, or supposed to be exploring the ruins of lost cities, Plane Shift can actively get in the way of the focus of the game. Or maybe the GM just doesn't want to design planar environments. In that case, the spell shouldn't be presented as an option. It's better GMing to get rid of the option entirely than come up with ways to artificially keep your players from using the option. The latter approach tends to feel like railroading.

On the other hand, maybe I, the GM, do like the idea of a our reality bordering an elemental plane of fire, and only that plane of fire. I could see a custom spell like "Travel to Plane of Fire", or perhaps a more fun name like "Hell Bound". So instead of dropping Plane Shift entirely, you can customize the nature of the spell. This ISN'T railroading, but world design.

So far this is nothing that can't be solved with some house rules. But if you're doing a lot of changes to the base D&D assumptions, house rules get over-complicated quickly. And really, that requires that you know what you'll need and what you won't early on. House rules presented before the game starts are often better received than house rules presented afterwards. But this isn't practical. I'm not arguing for you to figure out which spells are best suited for your campaign and which ones aren't before you've even started it! That way lies madness and wasted prep.

My favored ruleset, Knave, solves this problem quite elegantly.  Because spells are items, the GM has near total control over what spells exist in the world, and can add spells to the world as needed.  You can also take advantage of Guildhall systems- that is, requiring magic users to stop by at the local guild to learn spells-  to get the same effect with traditional Vancian magic- a guildhall may only be capable of teaching a limited set of spells, but different guildhalls teach different spells. This requires you not allow the player to get free spells per level, though.This is how B/X D&D actually works, but rarely how it's actually played. ... I think.

This incremental adding to the spells available in the world is a pretty stress free way to both limit spells that would cause problems as well as allow for thematic spell choices. Consider a guild made of descendants of a crashed spaceship- such a guild could teach unique science fantasy themed spells, like "Black Hole", "Invert Gravity", or the aforementioned "Detect Robot". Or if using Knave, you could place spell items with this theme inside a "Barrier Peaks" type science fantasy dungeon- if the players interact with the science fantasy elements of the world, their spell list will echo this, and if they don't, they'll never know those spells existed. Not all campaigns will need this particular example, in fact, most probably won't. But when using this style, you can build your spell list organically, to suit YOUR needs. As well as the needs of your players. Listen to your players, what they're interested in. You don't need to give them exactly what they want, but if you have a player trying to become a druid, it might be a good idea to start sketching up some nature spells.

Naturally, this also allows for third party spells to integrate into the game much better than having to add them to a game with a core spell list. When there's a core spell list, the extra spells picked up from blogs, modules, or even secondary canon material (such as Xanather's in 5E D&D) will always feel extra. If the chief way to get spells is by a player choosing them, the core spells will feel right in a way that the extra spells don't. But if the spells are all in a jumble being introduced via interaction with the world, then secondary source spells basically get to be on equal footing where the players are concerned.

- - -

To be clear, I'm not saying this is the only way to manage magic, or the best way. If the game you want to run happens to align with the D&D core assumptions (or the core assumptions of whatever game you're running), then you can use that spell list without problems. A lot of games are fine with this. Some people use magic word systems, which allow players to have a ton of power in creating spells- if you want that to be the feel of the campaign, where magic can truly do just about anything, that's fine, though it's not my preferred style.

There's also the GLOG, where you can pick the spell list of the world by picking the Wizard Schools available. It helps that GLOG spells are often very flavorful. Going with a flavorful spell list and seeing what that does to influence your worldbuilding is a perfectly valid approach, as long as you are willing to deal with the consequences.

And that speaks to a cost of the approach I have been suggesting here. A good spell list presented to the players is a marketing point. It gets players excited, which can help start your game on the right foot. In fact, when you're new to D&D, even the basic spells us veterans are tired of seem really cool and magical. But there's a cost to that as well. The excitement will fade. If you want players to be excited whenever they find a new spell, you need to hide the spell list. If you want them to be more excited at the beginning of the game, to give it more starting kinetic energy, you need to display your spell list from the start. Just don't be surprised if that enthusiasm lags. There is no right answer here.

So, to sum up:

• Pick a normal D&D spell list if you want a normal D&D game and don't need as much room to deviate later.
• If you want room to deviate from those norms and have include more unusual ideas, building a custom spell list is a good idea.
• The easiest way to do this is by NOT giving the players access to all the spells in the game at the start but building the list organically as the players earn their spells.
• Alternatively, coming up with a cool spell list and presenting it up front, like GLOG spells, can be a great marketing tool. 
• Or you can just let players collect magic words and create spells from scratch if you're ok with a more chaotic game. 

And I'm sure there's other methods that I'm not thinking about. For one thing, this assumes specific spells doing specific things. Open ended magic users like Skerple's Sorcerer, for example, are a different ball game entirely.

But I really do find the hidden spell list, organically built up over the course of the game, to be the most intriguing and useful of these options, at least for my own purposes. It's something I will explore further.