Your Magic System Needs Three Sentences, Not a Bible
My protagonist picked a lock in chapter twenty-two of a draft I was writing in the winter of 2023, and she should not have been able to do it.
Mireille had been locked in a grain cellar by the people who were supposed to be her allies, and I needed her out before the scene turned into a hundred pages of waiting. She had a small, unglamorous magic -- she could feel the shape of any object she had touched before, the way you can find a light switch in a dark room you know well. It was a nice, specific little power. I had never once, in twenty-one chapters, established that it could do anything to a lock. So I wrote that it could. The bolt clicked. She was free. I kept writing, pleased with myself for solving the scene so cleanly, and did not notice what I had actually done until a friend read the draft three weeks later and asked, not unkindly, "wait, since when can she do that?"
Since I needed her out of a cellar, is the honest answer. That is not magic. That is a plot mechanic wearing a costume, and readers can smell the difference from a hundred pages away even when they cannot name what tipped them off.
The fix took me an afternoon, and it taught me more about magic-system design than the two years of worldbuilding notebooks I had filled before it. I did not need a bible. I needed three sentences, written down before I ever got Mireille near that cellar: what her power could do, what it cost her to use it, and what it flatly could not touch no matter how badly the plot wanted it to. Once I had those three sentences, the lock stayed locked, and the scene got better, because now she had to get out some other way -- a way that actually cost her something.
Sanderson's Laws Are for Readers, Not Spreadsheets
Brandon Sanderson's First Law of Magic gets quoted constantly, and I think it gets misread almost as often: the ability of magic to resolve conflict is directly proportional to how well the reader understands it. Writers hear this and conclude they need an exhaustive, textbook-grade system before they can put a single spell on the page. That is backward. The law is not a mandate to build more. It is a mandate to reveal what you have built, clearly, before you spend it.
Here is what I mean in practice. You do not need to know, in chapter one, how magic interacts with iron, or whether it can be taught to the deaf, or what the seventh-century schism between two magical orders was actually about. You need to know, and the reader needs to have felt, the specific rule your climax is going to use. If your protagonist is going to save herself in chapter thirty by exploiting the fact that her magic cannot lie, that rule needs to be on the page well before chapter thirty -- ideally demonstrated in a small, low-stakes moment where the reader can watch it work and file it away. Sanderson's own Mistborn opens with Allomancy's core mechanic -- burning metals for specific effects -- established in the first fifty pages, in scenes that matter for their own sake, long before the climax needs the reader to already trust the rule. The law is about sequencing a promise, not front-loading an encyclopedia.
I have watched writers -- and I include the version of myself who filled two years of notebooks -- treat "understand the magic" as a synonym for "answer every possible question about the magic." Those are different tasks, and only one of them helps the reader. A reader does not need the metaphysics. A reader needs the load-bearing rule, delivered early enough and clearly enough that its use later feels earned instead of convenient.
The One Rule That Powers All of Earthsea
Ursula K. Le Guin built the opposite kind of system, and it is worth sitting with how little of it there actually is. In Earthsea, to know the true name of a thing is to have power over it. That is close to the entire system. She does not explain where the true names come from, how the Old Speech relates to the dragons' native tongue in any rigorous historical sense, or what percentage of the population can be trained in it. What she gives us is a single, elegant rule and its consequences, worked all the way through: wizards keep their own true names secret because to speak your name aloud is to hand someone your leash. Ged's arc in A Wizard of Earthsea is legible entirely through that one rule -- the shadow he releases, the pursuit, the eventual reckoning where he names the thing chasing him and, in naming it, understands that it is himself.
I do not think Le Guin's restraint was a limitation. I think it was the discipline. One rule, applied with total consistency, produces more dramatic pressure than seven half-explained ones, because the reader can actually hold one rule in their head and watch every consequence land. A system with fifteen moving parts asks the reader to do bookkeeping. A system with one moving part asks the reader to feel something, which is the only job magic ever actually had.
The practical version of this for your own draft: before you add a second rule to your system, ask whether the story needs it or whether you are just curious about it. Curiosity is not a reason. A scene that cannot resolve without the second rule is a reason.
Where Jemisin Puts the Cost of Orogeny
N.K. Jemisin's The Broken Earth trilogy is, underneath its withheld, elliptical prose, one of the most rigorously rule-bound magic systems I have read in the genre -- and she almost never explains it directly. Orogeny, the ability to sense and control seismic energy, comes with a precise mechanism (drawing heat and motion from the surrounding area, at a radius that scales with the orogene's skill and the size of the quake being stilled) and a precise social cost (orogenes are legally property, trained by Guardians, some of them wired directly into node stations to keep the earth calm at the price of their own minds). You could write a tidy paragraph explaining all of that. Jemisin never does. She shows you a child's hand, and the box she is kept in, and lets the horror of the system arrive through the body of one specific person before it ever arrives through exposition.
That ordering is the whole lesson. The cost of your magic should be visible on a character before it is explained in a rule. If a reader can recite your magic system's cost as a fact but has never watched it actually take something from someone they care about, the cost is decoration. Jemisin, who spent years working as a counseling psychologist before she wrote full time, builds pain the way a clinician documents it -- specific, embodied, never abstract. Essun flinching at a particular kind of touch tells you more about what orogeny costs than a page of rules ever could.
I try to run this check on my own systems now: can I point to one scene, with one character, where the cost of the power actually lands as loss? If the honest answer is no, the cost is not a cost yet. It is a warning label nobody has read.
Borrow the Writers' Room Whiteboard
The reference that changed how I actually build these systems did not come from fantasy at all. It came from watching an old interview with Vince Gilligan describing how the Breaking Bad writers' room worked. The show ran on a strict internal logic -- actions have consequences, consequences compound, nothing gets undone for convenience -- and the room enforced it with a physical whiteboard tracking cause and effect scene by scene: if this happens, then this becomes true, which forces this decision, which closes off that option permanently. It is not a fantasy magic system. It is the same discipline wearing different clothes, and I think it is the single most useful tool nobody tells fantasy writers to steal.
A magic system is a cause-and-effect machine before it is anything else. Fire magic costs stamina, therefore a mage who has cast three times today is a different person in a fight than one who has not cast at all, therefore your fourth-act ambush only works if you have kept honest track of who has spent what. The writers' room whiteboard exists because television cannot afford to forget its own rules across forty episodes and several writers. Your magic system cannot afford to forget its own rules across four hundred pages and eighteen months of drafting, and the failure mode is identical: a reader notices the rule broke before you do, because you were the one who invented it and they are the one paying close attention.
I now keep an actual flowchart for any rule with more than one consequence -- this cost enables this ability, which forecloses this option, which is the fact my climax needs to be true. It takes twenty minutes to draw and it has caught three continuity errors I would not have found by rereading, because rereading tests whether the prose sounds right, and a flowchart tests whether the logic actually holds.
Three Sentences to Write Before Chapter One
So here is the method, stripped down to what I actually do now, before I let a character near their first spell:
- What can it do? One sentence, specific enough that you could not confuse it with any other magic in any other book. Not "she can manipulate energy." "She can feel the shape of anything she has touched before, at any distance, as long as she is still touching something that touched it."
- What does it cost? One sentence, and the cost has to be visible on the character, not just true on paper. Not "magic drains the user." "Every time she reaches past a mile, she loses the memory of something small from that same day -- she will not know, until later, what she traded."
- What can it not do, no matter what the plot needs? One sentence, and this is the one that saves you from your own Mireille moment. Not "it has limits." "It cannot open a lock it has never touched, and it cannot be used on a person who does not want to be found."
That third sentence is the one almost every writer skips, and it is the one that would have kept me from writing a cheap escape scene in the winter of 2023. A limit you write down before you need it is a rule. A limit you invent after you have already broken it is an excuse.
Let the Rest of the System Grow Later
None of this means the rest of the system does not matter. It means the rest of it does not matter yet. Once those three sentences exist, everything else -- the history of who first discovered the ability, the political faction that wants to regulate it, the seventh-century schism I mentioned earlier and am not going to build until a scene actually needs it -- can wait until a specific chapter asks a specific question. I keep a running note for exactly this: not a magic-system bible written in advance, but a place where the answer to "wait, can she also do this?" gets recorded the moment I actually need it, so that by chapter thirty I have not quietly contradicted chapter four. The three sentences from the start stay the spine. Everything else gets added the way Jemisin's world reveals itself -- one true, specific fact at a time, exactly when the story is ready to carry it, which is also the argument at the center of the piece I wrote about worldbuilding's one real job. If you want a fuller structure for tracking a system once it does start growing -- the origin, the cost ledger, the social fallout -- the template I use for that is built around the same three questions, just with more room to grow into.
The grain cellar scene got rewritten that afternoon in 2023. Mireille did not pick the lock. She sat with her back against it, waited for the one guard who trusted her to check on her, and used the only thing her magic had ever been able to do -- feel the shape of something she had touched before -- to find the seam in his coat where he kept the key, because she had brushed against him three chapters earlier and some part of her had filed it away without either of us noticing. It cost her the memory of her sister's face for one whole day. It was a worse scene to write and a much better scene to read, and it only existed because the lock had a rule I had refused to break, even when breaking it would have been easier.
Plotiar keeps a running Lore entry for every rule your world grows, alongside a flowchart for tracing what each rule costs and forecloses -- so the three sentences you write today stay the spine of a system that can grow for four hundred pages without contradicting itself. Start a project for free and write the sentence your climax is going to need.