On Learning New Spells

In the past in my campaign, magic-users have not been a popular class choice. I’m OK with this actually, as I do like the idea that magic-users are a rare breed. On the other hand, I don’t want to give the magic-user such a raw deal that nobody ever plays one. Now that I’m watching some players play them I want to double check my rulings, especially as concerns learning new spells upon leveling. To do so, I will start by looking at what’s in the books, and then comparing that to what I’ve been doing.


In OD&D I see little mention of how magic-users obtain spells.  It almost seems to me that the expectation is that every magic-user has every spell on the list in his books, and interestingly it appears that clerics also must maintain spell books.  I see rules on how many books are required to contain the spells (one book per level), the cost of replacing lost books, and the cost of inventing entirely new spells.  Scrolls appear to be used only for casting stored spells, which actually makes the spell Read Magic make much more sense.  Read Magic is essentially the ‘scroll activation spell’, used on the fly when you want to pop some spells off a scroll (it specifically mentions that the spell lasts long enough just for “one or two readings”).


In B/X magic-users start with  a book containing one first level spell only, and gain new spells automatically:

When player characters gain a level of experience, they will return to their masters and be out of play for one “game-week” while they are learning their new spells. … Magic-users and elves are limited to the number of spells they may know, and their books will contain spells equal to the number and level of spells the caster can use in a single day.  – X11

Like in OD&D, scrolls appear to be only for single-use castings.  However, read-magic has been modified to act as an “unlocking” spell for scrolls, including the text:

once a scroll or runes are looked at with a read magic spell, the magic-user becomes able to understand and read that item later without the spell.

AD&D 1e

Magic-users in AD&D start with 4 spells in their book, read magic automatically, plus one each from three categories (offensive, defensive, and misc.) They gain a single additional spell automatically upon leveling.  Found spells must be rolled to see if they are understood (using a percentage based on Intelligence).  Spells copied from scrolls destroy the scroll, though the text would seem to imply that failing to understand the spell results in not being able to copy it at all, thus preserving the scroll.  I have not found a clear description (though it may exist) if the roll to know a spell is failed — I assume it means the magic-user will never be able to learn that spell.  (DMG 39 for all of the above.)

My Campaign

My rules are a hodge-podge of the above along with some custom house rules:

  1. To learn a spell, a magic user must roll equal to or less than his intelligence on a d20.  Failure indicates he cannot understand the spell, but he may try again after gaining a level of experience.  One check per level per spell is allowed, regardless of source of the spell.
  2. Starting out magic-users roll a check for every first level spell to determine what they start with in their spell book.  This does not count as a learning attempt for first level, it is assumed to have happened during their apprenticeship.  A mulligan is allowed if all spells are failed.
  3. Read magic is required to ‘unlock’ scrolls, however once unlocked the spell is understood and further castings of read magic are not required.
  4. Scrolls can be cast from directly, which requires only the read magic to unlock, and no roll to understand the spell.  Magic-users may even thus cast spells normally above their ability to cast (eg. a 1st level magic-user could cast a 3rd level spell directly off a scroll).  This destroys the scroll.
  5. Scrolls may be copied from the scroll to the magic-user’s spell book.  The magic-user must make his roll to understand the spell, and may only do this for spells that are within his normal ability to cast.  The scroll is destroyed either way, succeed or fail.

I guess actually my system is basically straight out of AD&D, with a few tweaks here and there.  Having written it all out, I think actually I’m OK with my system.  The only minor change I think I might consider is the destruction of a scroll when the magic-user fails to copy it into his book.  Perhaps the roll should symbolize ability to understand the spell, and then only once the spell is copied does the scroll destroy itself.  I will have to ponder on this.

Curious to hear how other folks run this.  I suspect this is an area where most DMs have their own house rules that vary slightly from any written source.

11 thoughts on “On Learning New Spells

  1. I’ve never liked (or used) the chance to know requirement before, but your tweak so that the magic user can re-try learning them each time he levels up makes me reconsider it. I’m still not completely sold, but I may test it out sometime to see how it works, or possibly only apply it to higher levels of spells.

    Concerning how I run magic user spells in my game, I give players the choice between having learned magic from an apprentice, being part of the guild or having gone to an academy. Apprentices are allowed to copy their master’s spells as they level, but have to follow their master’s orders (which makes plot hooks really easy) until they leave their master at 5th level. If they are on good terms with their master, they will usually receive a gift from their master when they “graduate,” and can probably call on their master for help in extreme situations, but they’re pretty much on their own. This is the only option that’s actually been chosen so far in my campaign.

    I have to get going, so I’ll cut this shorter than I was planning. If you’re interested, I’ve written more about the options in my game, as well as a few more, here: http://grognardling.blogspot.com/2011/07/twenty-questions-part-2-magic-user.html

  2. Experiencing the magic-user in a long campaign setting has been extremely fun so far. It definitely makes me turn off my design hat and contemplate some really difficult decisions (take the 12 int chance to learn a spell from a scroll or keep the scroll for a reactive situation).

    It pains my gut that I have to make that decision but it makes learning the spell much more meaningful. That being said it REALLY sucks when you fail a spell that is useful. Especially given the rarity of scrolls in our campaign.

    From a player perspective, it would be rather nice if there was a way to preserve the scroll rather than consume it(for example a 2nd intelligence roll, or some form of d6 roll that happens if I fail the roll). Taking more time is also another option that could be added as a means to be more careful with the scroll.

  3. It’s an interesting challenge to read OD&D without bringing the baggage that we’re familiar with from later editions. I agree with your reading that everyone in OD&D LBBs has spellbooks with all spells available. It hadn’t occurred to me that Read Magic might be necessary at any time that a scroll is used.

    You might also want to add a section on “Greyhawk OD&D”, where they first added the “chance to know” table (a very significant change!), pretty clearly a one-time only chance. It doesn’t specifically use the “spellbook” terminology, but it does introduce the distinction of clerics knowing everything (p. 7-8). It’s interesting to consider how this was copied into the AD&D PHB, but then the DMG later layered on its own spellbook rules in a way that wasn’t totally compatible.

    My opinion is that some mechanic is needed to deal with expanding spell rosters over time (other than “everyone knows everything”). This could be (a) the “chance to know mechanic” (and thus automatically in your book), or (b) a “spell acquisition” mechanic, but I feel that the two mix together poorly.

    What I currently have in my house rules is that every 1st level wizard starts with all 8 1st-level OD&D 1st-level spells in their book (this saves generation time, gives the starting wizard more options to play with, and simulates a “here’s all the basic knowledge” milieu). Later stuff has to get picked up from adventuring; roll a Target 20 check (d20 + level + Int bonus >= 20) to add a found spell to your book, but this is mostly just a time-consideration (takes n days [n = spell level]; try again immediately if you like).

  4. My opinion is that some mechanic is needed to deal with expanding spell rosters over time (other than “everyone knows everything”).

    I think this is exactly why we see the shifting rules from OD&D to Basic to AD&D. It’s especially interesting to note that OD&D includes rules for inventing new spells, which is gated by a percentage roll and includes costs in both in-game time and money. It seems to me like in the early days the spell list was expanded by player invention rather than supplemental publications. The fact that Greyhawk adds both lots of spells and the chance to know roll is quite telling.

  5. Totally agreed. And I can totally understand not initially anticipating the issue (like could you guess having 30 1st level spells in the PHB a few years later?)

  6. Delta, I like your formula/ruleset. At lower levels the ‘one chance per level’ adds interesting flavor and personality to different Magic-users. I don’t know if it scales well in a game where levels/scrolls are a challenge to acquire.

    At a certain level failing to learn a spell has drastic effects on an MU’s progression. The opportunity to fail the same spell multiple times in a row (in theory, a chance to fail every spell) is potentially crippling. (Considering an average INT score for an MU 10.5)

    I guess it’s all dependent on how the campaign is run (does the magic-user run across enough spells to beat the odds of failing?). That being said I thought about all of the ‘costs’ associated with adding new(non-researched)spells to a MU’s repertoire and was surprised by the list I came up with.

    Reaching required level
    Acquiring copy of the spell (if it’s in scroll form, determine to save it)
    Spending time to research (including casting Read Magic if necessary)
    Intelligence roll (Can’t recheck until next level if failed)

    I really like the epic-ness of learning a new level of spells and the ‘Oh yes!’ when I’m successful at scribing one. In hindsight though my only high level character has been successful in his rolls and hasn’t run into any significant road bumps (even with missing Charm Person for 3 levels).

    Just some observations.

  7. OK, let me give some insight on why I like what’s in the game now. Keep in mind that as player and DM we have slightly different goals: you have an eye on what makes your character more enjoyable to play, while my focus is on what best represents the game world I’m trying to portray.

    1. INT rolls. I prefer the idea of magic-users as the intellectual elite of the world. By requiring INT rolls to learn spells, it encourages players to only play magic-users with a high INT. As written without this, there is no penalty to playing an INT 3 magic-user, which I find very odd. In fact, I’d say without this requirement low INT magic-users are even more common, as magic-user would make an excellent choice for a character with all low scores (generally stays out of combat, so low STR, DEX, and CON don’t come into play much, but all the same spell power as an INT 18 magic-user).

    You mention that 10.5 is the average INT for a magic-user, but I think it’s actually a bit higher than that. Sure, 10.5 is the average roll of 3d6, but players both get 1 stat swap by my rules, plus they can always choose to play a different class. I think players are unlikely to choose magic-user unless they can get a reasonably high score in INT, and thus all magic-users tend to have higher than average INT, which was exactly my goal in the first place.

    2. Finding scrolls: Ultimately I look at this as sort of the release valve for the rate at which new magic is introduced into the game. As DM I can vary how frequently such shows up in treasure if I think the rate needs adjusting. As a player, if you feel the rate is too slow, you can focus your efforts on searching out such treasure (though I suppose you’ve got to convince the rest of the party first). Point is, this is easily adjustable as we play to reach the best level of equilibrium.

    3. Read Magic: Ultimately I think the point of this spell is simply to encourage expansion of the spellbook only during down time, as who would waste a memorized spell on Read Magic while exploring a dangerous location? This leads to what I think is a nice flow: the magic-users have a certain set of spells for an adventure, then go back to town and expand that spell list, then go on the next adventure.

    4. New Spell Levels: The only thing that gives me slight pause is that I would be disappointed, both as DM and player, should a magic-user attain a new spell level and end up with a useless spell slot. I’m sure it would be super annoying to hit 5th level and fail to learn fireball, especially were that the only 3rd level scroll you were toting around. That said, I fully endorse the “rulings not rules” mantra of the old school. Any time a player decides something is important enough to invest in-game resources (time and money) I am happy to take that into account. In this case, should you find yourself in the situation where you gain a new spell level and have only a single spell to try and learn, you might decide that a trip to the big city is in order such that you can pay to be tutored by a higher level magic-user, or perhaps gain access to the Magician’s Guild’s library and labs to aid your research. I have no problem giving some bonuses to the learn roll based on that kind of expenditure.

    And now this comment is longer than the original post, so I’ll stop yammering.

  8. Excellent points, as a player (in your setting) I do agree that the feeling is pretty unique and hits what you’re going for.

    At low levels I love it, I think your last comment about the useless slot is my actual fear.

    “rulings not rules” rules.

    It does make me wonder what I’d have to do (roleplaying wise) to make a intelligence 6 magic-user work as a character… I’m picturing an idiot-savant who has to spend ALOT of time with his master and really just has stick figures in his book for the hand signals.

  9. Accepting that I’m replying to a 9-year-old post at this point, the system you described here are similar to what I use for my personal D&D amalgam.

    1. Same regarding one check per spell per character level, but I do it with a 2d6 roll scaled against Intelligence (needing 8 or higher for 9-12, adjusted up or down from there in line with the typical B/X modifier-by-standard-deviations approach). This produces a more controlled distribution of results compared to a straight d20 stat roll, so magic users with higher Intelligence are substantially better at learning spells.
    2. For starting spells, I do Read Magic/Detect Magic automatically, plus 1d4+Intelligence modifier additional spells, selected randomly. The first two give a minimum baseline of utility along with the ability to make use of scrolls (either for casting or learning), while the others give some flavor/differentiation.
    3-5. Same as you’ve said.

    For learning new spells, the character is expected to go out in search of spellbooks/scrolls, whether that means cultivating partners to share spells with, hunting down renegade sorcerers to find their spellbooks, doing favors for established wizards in return for some scrolls, making unsavory deals with shady warlocks, etc. My exception to that is when I play with 2e-style specialist wizards, as they get a spell from their specialty school on gaining a level (this is one of their class perks). I’ve had some players complain about the lack of spell access, but it usually comes from a misalignment in expectations, and explaining to the player that I’m open to working with them to provide content that they’re interested in is usually enough to resolve it.

    I haven’t run OD&D, but I did eventually come to the conclusion that clerics should also maintain some sort of spellbook instead of just having access to everything they can pray for by default. The latter approach never sat well with me.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.