Creating A Magic System: “The Case for Levelless Spells”

For some context, this post is the result of brainstorming about an upcoming campaign which I’m calling Wizard School for the time being. The concept, which I described as “cocaine Hogwarts in a failing Byzantine empire,” was picked by my players from among four campaign pitches.

Since I intend to run an OSR campaign where every PC is a wizard, I had to do a dive into OSR magic systems to ‘see what’s out there’. When reviewing the most relevant codices and grimoires, I was immediately taken by the concept of levelless spells as seen in places like the GLOG or Wonder and Wickedness. It was something I wanted to include in my game for reasons I’ll dive into here. This post should provide some answers to the question: why use a levelless spell magic system?

To briefly establish some terms: leveled spells are spells which are organized into a hierarchy by power level, while levelless spells are spells which are not organized into a hierarchy. 

To compare the two, let’s think about why our spells have levels in the first place. 

We should consider that the preference for leveled spells is an appeal to tradition. After all, this is how things have been done since the three little brown books! The appeal to tradition is a logical fallacy where bad ideas are sheltered from scrutiny by their long history, but even if this is true to some degree, leveled spells have some obvious advantages. For starters, their long history means that there is an extensive, time-tested, curated list of spells for spellcasters to use in their games. From the effort standpoint this is a massive advantage to the referee. I could decide I don’t need any new spells, new system, new anything. It’s all right there for me.

If, however, we want to create new content, leveled spells provide ease of design. The designer has a ready model for future spells, whether they are modifications of existing spells or entirely original. Even for a completely homemade magic system, it’s a simple matter to create a scale for magical effects. Given that magic doesn’t exist, the scale is by default subjective. Once you have your scale, you just slide spells into appropriately leveled slots. This is a much easier environment in which to attempt game balance

As the fighter does more damage, gains a second attack, or learns the Cleave Maneuver, the game designer wants a comparable and equivalent way to progress the spellcaster. We can be confident that spells created in this framework will be balanced because we are giving them hierarchy. Leveled spells make it easy to categorize any effect (and intensity of effect) into a nice, neat box. It makes it easier to match a magical effect to a non-magical analog. It also works on unique and seemingly non-scaleable effects like invisibility, telepathy, or clairvoyance. Hierarchy produces tidiness, which is generally useful in game design and in life.

Hierarchy gives us a straightforward way to control access. This is the big one, the main reason we use leveled spells. Simply put, there are powerful magical effects that should, from the view of both the storyteller and the game designer, be withheld from novice wizards. Slowly expanding access to better spell effects is how casters progress, and progress is the key to any story. It’s why most people like RPGs. Just as the fighter gets stronger or quicker, the caster gets new and more powerful spells. It matches what we’ve learned to expect from magical fiction, where the master can access mystical energies that the apprentice can only dream of. Without controlling access, we have no structure, no progress. We end up playing Calvinball. For most people, most of the time, this is not fun. 

Image: Andre Bdois

“But master, why can’t I cast Major Illusion?” 

“Because you aren’t level 5 yet, boy. And when you’re level 5, they give you one of these floppy hats.” 

So we admit that we need to control access. Our previous definition of levelless spells doesn’t include access and is therefore unlikely to stand up to scrutiny. A magic system with no concern for access will suffer outside of truly low-magic campaigns and one-shots, and then only because these reduce exposure to our bad magic system. 

So let’s redefine our levelless spells with four defining characteristics:

  1. Spells exist alongside other spells without an explicit hierarchy.
  2. Spells have a comparable, low-powered magical effect when cast.
  3. The magical effects of spells can scale on some axis. 
  4. Access to further effects on the axis is controlled by some game mechanics.

All four must be present to be a viable alternative to leveled spells. 

#1 is definitional. Without it, we don’t have what we want. 

Without #2, you are simply creating a magical system that demands hierarchy. If every spell has wildly different effects and raw power but no categorization, you have no way to control access. Apprentice wizards cast Time Stop and Master wizards struggle with Color Spray. We chose to make our baseline effect “low-powered” because that works for our goal of progression, but also because it grounds our magic system. Without it, earth-shattering fireballs are everywhere and you’re back playing Calvinball.

Without #3, all spells are shit and every magician is simply lighting a match flame underneath the foot of their enemy like some sorcerous Kevin McCallister. Again, for a low-magic campaign this is fine but most campaigns that feature magic, it isn’t.

#4 seems loaded but we already control access in many ways outside of level. Spell points is an obvious and common method. Spell rarity is another. We aspire to something more interesting but it’s already being done.

In reviewing our new definition, it’s clear that a more accurate name for this concept would be scalable spells rather than levelless spells. While we can still control access using various narrative and mechanical methods, the ability to control scaling rather than being the right level becomes our primary way of controlling access. 

“Whoa whoa whoa,” you say. “Aren’t you just replacing spell levels with some other arbitrary mechanic?”

In response, let’s consider what advantages scalable spells have over leveled spells. 

The primary narrative advantage is that scalable spells more closely match our intuitions. Whether spellcasting starts as a gift or talent, genetic or bestowed, it functions in our game as a skill that can be developed. Spells that scale in power as the caster gets stronger match our intuitions about how spellcasting should work. If I learn the Firebolt Spell and later become a much stronger wizard, I shouldn’t have to discover some scroll for, or have someone teach me, an entirely new spell to produce a Fireball. It should be a natural progression of my skills. It’s something I should be able to figure out for myself, even if that process is dangerous.

Our intuitions are also that magic is messy. This is a key feature of nearly all magical fiction that we consume. It may be a matter of taste, but most people feel that magic should be, narratively speaking, unpredictable, difficult to classify or pin down, wonderful and amazing. A magical system where all spells are already classified and organized loses much of what makes it magical: it becomes science, just with different clothes. There is nothing unexpected. There is nothing left to discover.

Consider the compass. Without an understanding of magnetism or planetary poles, the compass is an incredible magical artifact. It just knows which way is north. You, the reader, understand the science behind it and probably haven’t thought about the compass in many months or years. There is no sense of wonder remaining about the compass.

We have come to accept the ludonarrative dissonance that occurs when every spell that exists in our TTRPG appears in a list in the back of the players’ materials. Newly learned spells can not produce wonder because they aren’t new. We know them all. That classic curated list of spells from D&D’s antiquity suddenly seems like an albatross around the neck.

“Ah, but this is a red herring,” you say. “How do we play a fantasy game without defining its spells? I need to know what’s in the Spell Section of the Rule Book.” The in-depth answer is another, upcoming blog post, but the fastest answer is to hide information from your players and give your scalable spells room to expand along their axis. A Fireball spell could scale downward to a matchflame-type cantrip without much mechanical consideration, while the Light spell could scale upward to a blinding offensive weapon that uses more of whatever spell resource we decide on. 

To elaborate on the last point, scalable spells have the capacity to scale on non-obvious axes. Combine this ability with the idea of hiding the upper ranges of the spell until cast and you get a magic system that, while completely under control of the gamemaster, appears to the player as something worth discovering, with spells that can mutate in unexpected ways. Take the example of a hypothetical Frost spell, which places a painful cold snap around a target, reducing their agility. The player knows what this does mechanically and can always rely on it, but they don’t know what will happen if they upcast it with more spell resource. Perhaps they expect it to freeze a larger area but it actually freezes a single target more strongly, or causes a small blizzard, or creates ice armor, or raises a platform of ice. These are all effects we can define mechanically, in advance or not, but won’t be known to the player without in-game experimentation. 

To briefly expand on other methods of access control, because these answers are incredibly important, you can include one or more of the following: a free-form magic system, hide unknown spells from the players, hide the upper and lower ends of the spell axis until they are used, leave the upper and lower ends of the spell axis undefined and improvise them at the table, include spell research rules for players to create or ‘rediscover’ new spells. I will argue that these ideas are much more viable in a scalable spell system and none of them reduce crunch if you don’t want them to. You can probably think of a few systems right now that include some of these ideas. 

Returning to the point about our intuitions, doing so has the advantage of promoting verisimilitude, the appearance of being true or real. It means that if we accept the fantasy premises provided, magic is operating in a believable or realistic fashion. It keeps us invested in the fiction. This is why we’re considering our intuitions in the first place. For some players, removing spell levels will help them feel more like a wizard when roleplaying. Unless you’re running a shounen-inspired campaign, the concept of spell levels probably doesn’t exist in the fiction of your world. There are spells that I’m strong enough to cast and there are spells that I’m not, spells that I know or that I don’t. When players discuss magic, in-character, at the table, they never mention their own level or the level of spells. They simply say “I can help. I know how to cast [spell].”

As a final example for promoting verisimilitude, think of the confusion around character levels and spell levels that always happens when teaching the most common TTRPG, 5th Edition D&D. “I can’t cast Fireball? But I’m level 3, it’s level 3, what’s the problem here?” Nothing takes the player out of the fiction faster. Almost no RPGs coordinate character and spell level. I’m interested in other examples but the only one that comes to mind is Knave. Scalable spells neatly sidestep this and many other issues.

Consider that removing spell levels might remove player bias towards certain effects. If my spell system places the spell Speak to Wood as a 1 on my 1 to 10 power scale, my player may never consider it, learn it, or prepare it. But there are situations (“The king has been murdered in the throne room!”) where Speak to Wood is the best spell they know (“I interrogate the throne.”). Decoupling levels and spells lets the players decide for themselves what spells are “good” and invites them to imagine a spell’s possibilities rather than passing them over for the next level spell. 

The majority of leveled spell systems cause frequent decision paralysis when a player is in the spotlight. A 5th level spell is often just a hammer in search of a 5th level nail. Players will reserve their highest leveled spells for the optimal situation and are often unwilling to consider using a high-level spell in a low-stakes situation. And why should they? It’s a waste of their resources, even when a creative and rewarding use of that spell is right in front of them. In contrast, a scalable spell starts in the basic form and can be powered up to what the player needs in the moment. This makes the caster’s toolbox more versatile, which reduces decision paralysis and increases problem solving opportunities. I may never cast Earthquake as I fear not having that spell when I need it most, but if Earthquake is simply the Move Earth spell upcast by several steps, I won’t fear using Move Earth to shift the ground from under my enemy’s feet. 

This is the first point I’m making that explicitly references leveled spell slots. Somewhat paradoxically, I think the idea of spell slots makes some narrative sense: spells are so dangerous and complex that the caster needs to prepare them beforehand, and can only commit a certain level of time and effort to that process each day. Spell slots represent the efficiency of the wizard at this prep process, which increases as they progress. Unfortunately, this preparation aspect is an afterthought at every table, and in any event can be recreated in our new scaling magic system. 

Finally, scalable spells should reduce spell redundancy. Low-level iconic spells like Healing Word or Burning Hands are eventually outclassed by higher leveled but similar spells like Mass Healing Word or Fireball. It’s obvious how a straightforward effect like healing or producing fire can be expanded on predictable axes, ad infinitum, to produce a dozen different spells, which is what we see. This has the practical effect of cluttering the player’s in- and out-of-game spellbook with near duplicate spells, and dulls the excitement of gaining a new spell. If a player learns a spell that’s simply Existing Spell They Have Plus Four Damage, or at Plus Four Area Units, or Plus Four Targets, that’s not exciting. It’s boring. We should be focusing our creativity on new spells with new effects rather than refurbishing existing ones.

The obvious downside to this system is that deviating from the common leveled magic system is hard work which takes up time that a lot of game masters don’t have. I hope to do a lot of this work for you over the coming weeks and months but if you are interested in an OSR-compatible scaling magic system right now, you only need to check out the GLOG. It skews low-magic but it has massive community created spell lists and wizard classes, and it’s entirely free. When you return to this blog in the future, you should see a hybrid magic system and spell list that uses magic-dice as spell fuel but a tiered spell effect list similar to Dungeon Crawl Classics. This system should provide a levelless spell system appropriate for high magic campaigns.