To build a magic system for web fiction, start with its limits rather than its powers, give every use a real cost, and decide up front how much of the mechanics the reader is allowed to understand. A hard system with legible rules lets your protagonist win through cleverness and lets readers reason along; a soft system trades that away for mystery. In progression-centered web fiction the system is also the ladder the story climbs, so it has to keep generating new problems for hundreds of chapters, which means designing the constraints before the spectacle and keeping a searchable record so the rules you set in chapter three still hold in chapter three hundred.
What does it actually mean to build a magic system?
Building a magic system means defining the rules by which the impossible happens in your world, and the good ones are defined far more by what they forbid than by what they allow. This is the first thing that trips up newer authors, because the instinct is to start with the cool part, the fireball or the time-stop or the ability to read minds, and only later, usually when the plot has painted itself into a corner, to start bolting on the restrictions that keep that power from trivializing every scene. My feeling is that this is exactly backwards. A magic system is not a list of things characters can do; it is the shape of the space they are allowed to move in, and that shape is drawn by its edges. Brandon Sanderson, whose Mistborn (2006) built its Allomancy around a fixed set of metals each doing one specific thing, has argued for years in what he calls his Second Law of magic that limitations are more interesting than powers, and I think any author who has watched an overpowered protagonist drain the tension out of their own story will recognize why. The question that builds a system is never "what can this magic do," but "what does it cost, what can it not do, and who is shut out of it."
So the practical first move, before you name a single spell, is to write down the constraints. Decide the price of using the magic, whether that price is physical exhaustion, a finite resource, a moral cost, a social one, or simple risk of death. Decide the hard boundary, the thing the magic categorically cannot accomplish no matter how skilled the user, because a system with no ceiling is a system with no stakes. And decide the gate, meaning who gets to use it and why they and not everyone else, since that single choice generates most of your world's politics for free. Only once those three edges are fixed do you get to design the powers, and now every power you invent has to live inside the box you built, which is precisely what keeps it from breaking your story later.
Start with the cost, because a power with no price is not a story
The most important design decision in any magic system is what it costs to use, because cost is the thing that converts raw power into narrative tension. A protagonist who can throw fireball at no expense is a fire hose; a protagonist who can throw fireball but burns through a reserve that takes a day to refill, or who ages a little each time, or who cannot aim it when their hands shake with fear, is a character with choices to make. Ursula K. Le Guin understood this in A Wizard of Earthsea (1968), where every act of magic disturbs a cosmic equilibrium and the wise mage is the one who mostly declines to cast. Patrick Rothfuss built the sympathy magic of The Name of the Wind (2007) on something close to thermodynamics, so that a wizard drawing heat to light a candle can freeze the blood in his own arm if he sources the energy from his body. In both cases the cost is not a footnote; it is the reason the magic is dramatically interesting at all.
Web fiction sharpens this because of its length. A cost that feels severe in a single novel has to keep mattering across hundreds of chapters, so the most durable systems make the cost scale with the ambition, meaning the bigger the effect the steeper the price, which guarantees the tension survives even as the protagonist grows enormously strong. The authors I talk to who sustain a serial for years almost all describe some version of this: they did not just make their hero more powerful over time, they made the price of that power keep pace, so that a level-fifty problem still feels as dangerous to a level-fifty hero as the opening problem felt to the beginner. If you get nothing else right, get the cost right, because it is the single lever that keeps a long-running system from deflating.
Decide how much the reader is allowed to understand
The next decision is how legible your magic is to the reader, and this is the hard-versus-soft choice that quietly determines what kind of story you are able to tell. Sanderson's widely cited First Law, published on his website back in 2007, holds that an author's ability to solve conflict with magic is directly proportional to how well the reader understands that magic, and the logic is airtight once you sit with it. If your protagonist is going to win the climax by using the magic system cleverly, the reader has to already grasp the rules well enough to feel that the solution was earned rather than conjured, which means the system must be hard, explicit, and taught to the reader in advance. If instead the magic exists to create awe or dread, the way Gandalf's power does in The Lord of the Rings where we never learn its mechanics, then it can and probably should stay soft, because explaining it would only shrink the mystery without buying you anything.
Most web fiction, and nearly all of progression fantasy and LitRPG, sits firmly on the hard end, and for a genre-specific reason. These readers came to reason along with the protagonist, to see the stat block or the skill tree and work out the optimal build before the character does, and a soft system denies them the puzzle they showed up for. He Who Fights with Monsters by Shirtaloon lays its essence-and-affliction rules bare so the reader can follow every combat calculation; Mother of Learning by the author known as nobody103 is admired precisely because its magic is tight enough that fans reconstruct the rules from evidence like a detective story. That said, you do not have to make everything legible at once. A strong approach, and the one I would usually recommend for a serial, is to run a hard core the reader learns early and keep a soft frontier at the edges, some deeper layer of the system still shrouded, so you always have a legitimate mystery to unveil two hundred chapters in without ever contradicting the rules you have already taught.
In web fiction, the system is also the ladder
In progression-centered web fiction the magic system is doing a second job that a standalone-novel system never has to do, which is to serve as the story's advancement ladder, and that changes how you build it. A Mistborn reader does not need Allomancy to keep escalating forever, because the book ends; a progression fantasy reader is going to climb your system for years, so it has to be designed to keep generating fresh problems at every rung rather than just bigger numbers. This is the difference between a system that scales and a system that merely inflates. Will Wight's Cradle works because each advancement stage, from Foundation up through the sacred arts, does not simply grant more madra but reframes what kind of conflict the protagonist can even engage in, so the story at the top looks nothing like the story at the bottom. The cultivation genre it draws on has done the same thing for a very long time, tying each tier of qi to a new category of ability and a new class of enemy.
The trap, then, is a progression system whose only variable is strength, because that runs out of story long before a multi-year serial does. If levelling up just means the same fights with larger numbers, readers feel the treadmill even when they cannot name it, which is a big part of why some LitRPG protagonists start to feel hollow a hundred chapters in. The fix is to build change into the ladder itself, so that every tier the hero reaches alters the questions they face and not merely the answers they can afford. I have written before about why so much of world-building is better discovered as you go than planned in advance, and the magic system is the one deliberate exception to that rule: the cultures, the maps, the history can all accrete around your story chapter by chapter, but the load-bearing system has to be engineered up front, because it is the part you can least afford to contradict later.
Keep the rules consistent across hundreds of chapters
The final and least glamorous skill in building a magic system is keeping it consistent once it is live, which in web fiction is genuinely hard because you are publishing in order, in public, with no chance to quietly fix an earlier contradiction after readers have seen it. No author reliably remembers, two hundred chapters and two years later, the exact cost they assigned a particular spell or the name they gave a minor technique in an early arc, and the moment the rules drift the comment section will catch it, because the readers of this genre keep better notes than most authors do. The only real defense is to write the system down and keep it somewhere you can search, so that consistency becomes a lookup instead of a feat of memory. This is exactly the kind of work IlorisNovel's AI world builder and wiki are meant to carry, holding your system's rules, named abilities, and their costs as entities you can reference while you draft rather than as things you hope you remember, which is also part of the broader craft of world-building that survives a long serial.
None of this is a formula, and everything above can be, and has been, successfully contradicted by an author with the instinct to break the rule on purpose. But if you are staring at a blank page trying to invent something impossible, the reliable path is to resist starting with the spectacle. Draw the edges first, price every use, decide how much you are willing to explain, and make sure the system can keep asking new questions long after the hero has grown strong, and you will have built something a reader can think inside of rather than merely watch. That, far more than any particular cool power, is what turns a magic system into the thing people argue about between your updates.
by Jacob Tam · July 6, 2026
I run IlorisNovel, a platform for web fiction writers. If you want to see what building a world here feels like, you can try the editor and world builder with no account.