Map the Level Route, Then Give Each Segment a New Decision
Break the level into situations players can perceive instead of forcing every level into the same number of segments. Fill in these fields for each segment:
| Item | Design question |
|---|---|
| Entry state | What have players learned, acquired, or discovered? |
| New decision | Does this segment add a judgment, tradeoff, or rule combination? |
| Supporting situation | What space, encounter, or constraint makes the decision happen? |
| Player action | How do players apply a learned rule or test a new one? |
| World response | What outcome becomes visible and changes what follows? |
| End marker | What observable state shows that the segment is complete? |
| Next recombination | How will an old approach meet a conflict or new condition? |
A segment may teach one rule or test a combination players have already learned. Practice, recovery and atmosphere may use familiar actions. Consider merging or deleting repetition when it supports neither mastery nor a change in pressure, experience or later state.
Let Players Cause the First Result Themselves
A rule explanation cannot replace action. When a new rule first appears, give players a situation in which they can see the prerequisite, try the action, and read the result. Keep failure feedback close to its cause so players know what to change next time.
After the teaching segment, the next segment does not need to repeat the same pop-up. Change the space, objective, or rules in effect so players must recall the lesson themselves. Larger values may change resource or risk judgments, or merely add time. Check their actual effect before deciding to recombine rules.
Original Design Example: The Shadowlight Theater's Missing Set Piece
This original paper level has not been built or playtested. The player restores a theater using shadow lamps: one lamp projects a single puppet silhouette, while two illuminating the same puppet produce a split. A screen mechanism recognizes a single shadow covering its pattern only when the screen is secured. Recognition latches its door open until level reset. Fixed screens start secured; movable screens need the player to engage a locating pin. These are fictional game rules, not claims about real theater equipment.
The goal is returning the Moon Boat set piece from the storeroom to the backstage lift. Players move puppets, switch lamps and move or secure screens. Both hands pushing the boat cannot simultaneously adjust a puppet or lamp. This route is not a universal segment template.
| Segment | Situation, action and result |
|---|---|
| Empty Stage | One lamp and a secured screen. Move the puppet before the circular pattern and switch on the lamp; recognition latches the side door open. Passing through ends the segment |
| Two-Lamp Corridor | Two beams create a split. Turn off the distant lamp or move the puppet out of one beam, retaining the near lamp’s aligned single shadow. The door opens and releases the next screen track |
| Moving Screen | Keep puppet and near lamp fixed. Move the unsecured screen until its pattern overlaps the single shadow, then engage the pin; recognition latches the storeroom door open |
| Return | The boat is found but the return cargo door remains closed. Arrange that door’s lamp and puppet first, then push the boat through with both hands. Reaching the lift lights its number plate and ends the level |
The third segment does not assume that moving a puppet with a screen guarantees shadow alignment. The relationship among lamp, puppet and projection surface must be readable in implementation. Here the shadow stays fixed while the screen moves into alignment and is secured. The return keeps earlier doors open and prepares a new cargo door before transport, without simultaneous puppet manipulation.
Next observe whether players understand splitting, see the secured screen and encounter a cargo door that truly stays open. Paper rules do not establish projection display, control feel or pacing.
Use the Default-Action Test to Find Filler
Write down the easiest default action for each segment, such as always leaving the near lamp on, always taking the closest door, or always attacking the nearest target. Then ask:
- Can players pass without reading the segment's new information?
- Do rules learned earlier still change the outcome here?
- After failing, will players understand why the default action was wrong?
In the Shadowlight example, if players can open the door in the Two-Lamp Corridor while leaving both lamps on, the splitting rule is not actually part of the level. Fix how the rule affects the outcome or remove the segment. Do not pile on unrelated timers, enemies, or currencies just to block a default action. First make the rule promised by the segment genuinely change the available opportunities.
Carry Each Segment's Outcome Into the Next
Carry open doors, learned rules and the released track into later segments rather than resetting each new location to blank state.
Doors remain open after recognition, preserving an earlier route when the player leaves a puppet. The third segment checks screen securing; the return prepares the cargo door before pushing. The lit plate means the set piece reached the lift, not that the entire performance is complete.
A linear route can include spatial reading, execution judgments and reused rules without forced branches. Allocate teaching, variation, practice and recovery to the project’s pacing.
Remove a Segment to Test Whether It Has a Distinct Purpose
Temporarily remove one segment and connect the segments before and after it. Ask:
- Can players still understand and complete the next segment?
- Did the removed segment teach or recombine a rule that a later segment uses?
- Did removing it eliminate a state, space, or consequence?
If nothing is lost, removing the segment is usually cleaner than adding a reward. If deleting a teaching segment leaves players unprepared for what follows, keep one sufficiently clear first use. If two segments ask players to make the same judgment, combine them into one more complete situation. A paper route establishes only the level's logic. Observe actual play time, cognitive load, and enjoyment in candidate builds.
Common Problems and Fixes
| Problem | Fix |
|---|---|
| Values or quantities rise without changing decisions or experience | Give each segment a new judgment or rule combination |
| Teaching consists only of a pop-up, so players never cause the result themselves | Arrange an action, response, and end marker |
| A new rule disappears after one use | Bring it back later under a new condition |
| Removing a segment has no effect on anything that follows | Merge or remove it, or add a genuine state handoff |
| New systems keep being added to prevent repetition | Recombine existing rules, spaces, and objectives first |
| A paper route is presented as proof of smooth pacing | State that it is untested, then observe timing and comprehension after building it |
FAQ
Does a level always have to get harder?
No. Variation, practice, recovery and closure can alternate. Check each segment’s contribution to mastery, pressure, experience or later play without requiring rising difficulty or a new rule every time.
Does a linear level leave players with no choices?
A linear route can still require players to observe, judge, and act. Whether a level needs multiple routes depends on the intended experience. A puzzle with one solution does not need false branches added for appearances.
Do it with the skill
In a project that already has SOURCE_BIBLE.md, a selected CONCEPT.md, and PRODUCT_BRIEF.md, use /game-world-design (or $game-world-design in Codex) to write a level route. For each segment, list its entry state, new decision, player action, world response, end marker, and the next recombination, then run the default-action and deletion tests. Deliver the design only; do not build it or claim that playtesting is complete.
Read the method: Method used in this article