Test your system coherence
Join the waitlist to test SagaScope: constraint checks and long-term coherence.
Quick answer (if you only read this)
A coherent magic system is easy to define and hard to maintain.
Definition:
- Hard magic = precise enough rules that the reader can understand consequences.
- Soft magic system = mysterious effects whose coherence comes mainly from the story, tone and emotional logic.
Maintenance:
In both cases, you must keep a magic contract: what’s possible, what’s impossible, and what consequences follow.
Then you preserve coherence by verifying (1) rules, (2) costs, and (3) narrative promises across chapters.
This guide gives you a full method: from hard vs soft design to a promises-vs-proofs checklist, plus examples from famous books.
Hard vs soft magic: what it really means
“Hard” and “soft” aren’t only about detail level.
They describe the relationship between:
- what your magic makes the reader believe;
- and what your narrative actually proves over time.
Hard magic: when the reader can anticipate
A Hard magic system isn’t necessarily “scientific.”
It’s a magic where:
- constraints are clear;
- costs are identifiable;
- limits don’t change just because the plot needs a convenient solution.
The reader doesn’t have to know everything by chapter 2.
But they must feel that every use follows stable logic.
Soft magic: when the reader follows an emotional logic
Soft magic isn’t “rule-less.”
It’s consistent through:
- narrative intention;
- narrative consequences;
- limits revealed gradually.
Soft systems work best when:
- characters understand magic through lived experience;
- your plot turns “rules” into stakes (fear, hope, sacrifice, obsession).
Soft magic usually fails not on “detail,” but on “promise.”
If an effect suddenly becomes available “because the plot needs it,” coherence breaks.
The real core: promises vs proofs
A coherent magic system keeps a promise.
So you must connect:
- Promise: what your magic leads the reader to believe (capability, intention, limits).
- Proof: what the novel later demonstrates (cost, consequence, constraint).
Most incoherences come from a mismatch:
- you delivered a promise;
- you never built the proof;
- or the proof changed without explanation.
This promise/proof logic applies to any magic:
- when magic “works,” it must have consequences;
- when magic “fails,” there must be stable reasons.
To connect this to your structure, you can also read:
- how to structure a compelling plot
- magic system coherence checklist
- designing a coherent magic system (short article)
Model your system: hard anchors + soft levers
Think of your system as having zones:
- Hard anchors: what must never move.
- Soft levers: what can vary, as long as the story stays coherent.
What is a hard anchor?
A hard anchor answers a question you don’t want to change:
- what cost makes magic rare or dangerous?
- what condition prevents certain uses?
- what type of consequence always follows magical actions?
- who can (or can’t) manipulate the system, and why?
Examples of anchors:
- a ritual requires a time that can’t be compressed;
- an invocation consumes a physical resource;
- an enchantment changes memory at the price of trauma;
- a safety rule prevents certain categories of effects.
What is a soft lever?
A soft lever is where you can be more poetic:
- the “shape” of effects;
- metaphorical ritual details;
- culture’s interface with magic;
- variations tied to character personality;
- incomplete ritual knowledge.
Soft levers work as long as you:
- don’t break the anchors;
- pay the consequence;
- don’t introduce “free bypasses.”
Build a magic contract without overdetailing
You want the reader to understand logic without reading an encyclopedia.
The secret: you don’t need to explain every rule in prose.
First you document the fixed facts (costs, limits, conditions, consequences).
Then you write only what serves the scene:
- an action;
- a cost;
- a consequence;
- a decision.
Magic becomes “legible” because each scene is a payment.
Table: what to write vs what to keep as a contract
| In the text | As coherence contract |
|---|---|
| What the character does | observable limits and conditions |
| Immediate results | stable costs and consequences |
| Conflict (stakes) | rules that prevent free workarounds |
| Emotional impact | allowed variations (soft levers) |
This table avoids a common trap:
Confusing “knowing magic” with “explaining magic.”
Checklist: hard vs soft in 3 levels
To prevent contradictions, use a checklist with three levels: system, scene, plot.
Level 1: System (what must not move)
Verify:
- Limits: what’s impossible, and why?
- Costs: what does magic always demand?
- Conditions: what triggers or blocks an effect?
- Consequences: what happens afterward, not only during the act?
Level 2: Scene (what must pay off)
Verify:
- Did I show the cost (not only declare “there’s a price”)?
- Is any workaround justified (by a rule, learning, or a resource)?
- Do my characters act as if they truly follow the rule?
Level 3: Plot (the long-term promise)
Verify:
- Does the system change only when the story requires it (with justification)?
- Do magic revelations arrive at the right moment?
- Did I build cumulative evidence (not “free exceptions”)?
If you run these three levels at the right cadence, your system stays stable.
For a more detailed checklist, see:
Examples from real books (hard, soft, and blended)
Mistborn (Brandon Sanderson): hard magic oriented toward consequences
What this kind of fantasy nails:
- alchemy logic;
- implicit costs;
- difficulty of “solving” without paying.
Why it matters for writers:
- if effects become too easy, tension disappears;
- when costs are stable, scenes feel credible.
Dune (Frank Herbert): structured logic serving politics
Even when it isn’t “magic” in the fantasy sense, the idea overlaps:
- implicit rules that guide usage;
- system coherence;
- consequences that spill into politics.
Track:
- who can use it;
- when they can use it;
- and what it costs at narrative scale.
Harry Potter: soft magic with graded anchors
The series does two things at once:
- it leaves space for mystery;
- but it gradually anchors certain “laws” (spells, limits, effects).
Inspiration:
- you can be soft, but you must lock in evidence over time;
- readers accept the unknown when the system feels framed.
The Name of the Wind (Patrick Rothfuss): soft magic with coherence via intention
What’s most useful for your writing:
- magic isn’t always explained; it’s lived;
- the plot pays emotion and consequences, not nomenclature.
Use “editor AI” support to keep system coherence
Goal: don’t ask AI to “invent your magic.”
Goal: ask for verifications and resolution options.
1) Promises vs proofs diagnostic (questions)
Ask:
- What did I make the reader believe about this magic (promise)?
- What did I actually show in-scene (proof)?
- Where is the most dangerous gap (the one that breaks tension or enables free bypass)?
- What is the least invasive fix (cost, constraint, consequence, bridge scene)?
2) Free workarounds diagnostic (questions)
Ask:
- Which workarounds does my story currently allow “for free”?
- If a workaround exists, what cost/constraint makes it paid (and coherent)?
- Is this workaround becoming a new rule? If yes, what limits must I lock?
3) Hard anchor calibration (questions)
Ask:
- Which constraints must remain stable no matter what?
- Which areas can remain flexible without breaking the promise (soft levers)?
- Where would a rule change be acceptable (with proof), and where would it feel like cheating?
These patterns preserve coherence, without dumping creative responsibility onto the tool.
Table: hard vs soft in one line each
| Approach | What it makes easier | What it requires |
|---|---|---|
| Hard magic | anticipate consequences | lock costs/limits |
| Soft magic | create living mystery | pay emotions and narrative constraints |
Common failures (and how to repair them without rewriting 100 pages)
Failure 1: the exception that becomes a rule “silently”
Symptom:
- one scene uses an exception (artifact, rare ritual, emotional state);
- then the exception keeps showing up as if it’s normal.
Repair:
- either lock the exception (conditions + cost + limits),
- or remove it and redesign the scene that relied on it.
Failure 2: the cost exists… but changes nothing
Symptom:
- you say “magic has a price”;
- but the price never changes decisions.
Repair:
- pay it with a visible consequence (fatigue, debt, injury, loss, stigma),
- or pay it socially (reputation, prohibition, blackmail, dependency).
Failure 3: free bypass of a constraint
Symptom:
- the reader sees an obvious workaround;
- your characters don’t act like they see it.
Repair:
- either the workaround is impossible (hard anchor),
- or it’s possible but paid (cost/risk),
- or it changes the world (political/social consequence) and you must commit to it.
Failure 4: rule revelation arrives too late
Symptom:
- a scene resolves “by magic”;
- the reader only understands the rule afterward.
Repair:
- move proof earlier (a micro-proof, partial demonstration),
- or make the resolution more ambiguous (soft)—but pay emotion and consequence.
Workshop: “promise → proof → consequence” in 20 minutes
Pick one central power/ritual.
- Promise (1 sentence): what the reader believes is possible.
- Proof (2 scenes): where you show cost/limits.
- Consequence (1 scene): how the magic changes an important decision.
If you have no consequence, your magic is probably decoration.
Workshop: write a hard-magic scene without a lecture
Goal: make logic legible without exposition.
Inside the scene, ensure you have:
- intention (what the character wants);
- constraint (what blocks them);
- cost (what forces a choice);
- consequence (what changes afterward).
You can stay soft on “how it looks” as long as these four elements are visible.
Workshop: calibrate soft magic (mystery without cheating)
Soft magic remains coherent if:
- effects have stable intention;
- consequences are regular;
- limits lock in through accumulated proof.
Quick checklist:
- Does the effect serve emotion (fear, desire, sacrifice)?
- Does the effect create a problem as often as it solves one?
- Does every “miracle” pay something (even symbolically)?
Revision control tables (extremely practical)
Table 1: “Rule → cost → consequence”
Fill this for 5 rules max (more than that tends to become overload).
| Rule | Immediate cost | Return consequence | Risk if forgotten |
|---|---|---|---|
| Example: seal | pain + memory loss | debt to the healer | free bypass |
| Example: vision | exhaustion | paranoia / isolation | prophecy becomes too easy |
Table 2: “What the reader believes” (promises)
| Promise delivered | Proof shown | To pay later via… |
|---|---|---|
| “Magic is rare” | limited resource | conflict, economy, prohibitions |
| “Spells have a price” | on-page cost | moral choice, injury, debt |
These tables do one thing: prevent contradictions you only notice at chapter 40.
Genre lens (what breaks tension depending on niche)
Epic fantasy
Main risk:
- power becomes a universal solution.
Antidote:
- multiply costs + political consequences (who controls it, who bans it, who exploits it).
Dark fantasy
Main risk:
- magic feels scary but never truly costs.
Antidote:
- make the cost intimate (trauma, dependency, identity loss) and recurring.
Romantasy
Main risk:
- magic accelerates relationship flips without proof.
Antidote:
- align magic with emotion: each use changes the relationship (debt, trust, danger).
SF with “psi powers”
Main risk:
- mental abilities with no logistical limits (range, delay, fatigue).
Antidote:
- treat the ability like tech: range, latency, energy cost, proof moments.
Extra case study: a “rule drift” you can fix with one bridge scene
Scenario:
- Early chapters establish: “Teleportation requires a blood price.”
- Mid-book, the Protagonist teleports twice with no visible cost.
Reader reaction:
- tension collapses (“so why not teleport all the time?”)
Three clean fixes:
- Pay the cost: show the blood price affecting a later choice (weakness, risk, scarcity).
- Lock an exception: a temporary buffer reduces cost, but creates a different cost (debt, corruption, limited charges).
- Change the goal: solve the scene socially/logistically, not magically.
The key: keep the contract stable, even when you compress the explanation.
Workshop: build “proof moments” so readers trust your system
Pick 3 moments where the system is tested, not merely used.
For each moment, answer:
- What does the character want?
- Which limit blocks them?
- What cost must be paid to push through?
- What consequence returns later?
If you place these proof moments early, you can stay concise later.
Mini checklist: hard anchors you should never violate
These are the anchors that usually destroy a system when they drift:
- cost (if cost disappears, power becomes free)
- limit (if limits expand silently, stakes collapse)
- channel (who can do it and why)
- scarcity (resource/time/energy)
- consequence (what the world does in response)
Conclusion: the system is a contract, not a backdrop
A coherent magic system:
- makes choices believable;
- increases conflict impact;
- prevents “free solutions” that destroy tension.
Hard or soft, your goal stays the same:
keep promises with proofs.
FAQ
Hard or soft magic: which should I choose?
Choose based on your narrative promise. Hard maximizes verifiability; soft supports emotional mystery. In both cases, keep a contract: limits + costs + consequences.
How do I avoid inconsistencies while writing?
Verify your rules, costs and proofs at the right cadence. Tie magic to consequences (scenes) and long-term plot impact (story).
How do I reduce the risk of overdetailing?
Document fixed facts in your narrative base, then write what the scene must pay: action, cost, result.
Do I need to define everything before writing?
No. Start with anchors, then refine. Each new rule must be compatible with what your novel has already proven.
Keep this framework handy
Get SagaScope early access + writer templates — one email, no account required.
Ready to structure your novel?
Join the SagaScope waitlist. Founder perk: locked pricing + AI credits at launch.
Frequently asked questions
- Hard or soft magic—what should I choose?
- Choose based on your narrative promise. Hard magic is easier to verify; soft magic supports mystery and emotional effects. In both cases you need a contract: what’s possible, what’s not, and what always has a cost.
- Why do magic systems become inconsistent while writing?
- Usually because key rules shift or get forgotten: a constraint stops being applied, a limit becomes suddenly bypassable, or your story stops paying the consequences. The fix is a coherence checklist and verification at the right cadence.
- How do I avoid overdetailing magic (info-dumping)?
- Document fixed facts in your narrative base, then write only what serves the scene: an action, a cost, a result, a decision. The reader should feel logic without reading an encyclopedia.
- Do I need to define everything before writing?
- No. Start with anchors (major limits + costs), then refine as scenes require it. The key is that new rules remain compatible with what your novel has already proven.
SagaScope Team
The team behind SagaScope, passionate about writing and technology.
Similar articles
Complete guide: worldbuilding with AI without overdetailing your novel (Fantasy/SF)
Worldbuilding with AI without overdetailing: how to dose details, lock useful rules, and keep a coherent universe.
Writing fantasy/SF with AI: worldbuilding, magic systems and chronology
Guide to write fantasy or science-fiction with AI: coherent worldbuilding, explicit magic/tech constraints, reliable chronology, and long-form coherence with.
Why SagaScope doesn't generate text (and why that's better)
Why SagaScope doesn't ghost-write your novel: long-form coherence, voice control, and an « editor » AI instead of prose generation.
