Chapter 37 — Key Takeaways

1. Scope is a cube, not a line. It has three axes: content volume, system depth, and polish bar. You do not cut scope by cutting along a single axis. A small game can be deep. A large game can be shallow. A shippable scope is one where all three axes are compatible with the time, money, and team you actually have — not with the ones you wish you had.

2. Everything takes 3x as long as you think. Hofstadter's Law, compounded by the nature of creative and technical work. Teams estimate the happy path; reality delivers the average path, which includes every edge case, refactor, and dependency. Multiply your initial estimates by three as a matter of private planning discipline. You will not overshoot.

3. Build a vertical slice before you commit to full production. A vertical slice is one polished piece of the game at final shippable quality. If the slice is fun, the game will probably be fun. If the slice is not fun, the game will not be fun, and you should learn that cheaply instead of expensively. Half-Life 2's sandbag demo and Breath of the Wild's Great Plateau are the canonical examples.

4. Scope by mechanics, not by content. One deep mechanic carries a game; ten shallow mechanics carry nothing. Braid, Baba Is You, Celeste, and Downwell are each one mechanic explored exhaustively. Pick your mechanic. Explore its implications. Do not add mechanics because you are worried the game is thin — the answer to a thin game is depth, not breadth.

5. Maintain a living Must-Have / Nice-to-Have / Cut list. Every feature lives in one tier. Must-Have means the game does not ship without this; each one must justify its tier with a single sentence. Nice-to-Have is the queue of quality upgrades. Cut is final — do not reserve "cut for now" escape hatches; cut means gone. Review the list weekly. Cut something every month.

6. The true cost of a feature is implementation plus integration plus testing plus balancing plus UI plus teaching plus forever-maintenance. Implementation is about 10% of a feature's real cost. When someone says "I can build this in a weekend," they mean the first 10%. The other 90% shows up over the remaining development and the entire lifespan of the game. Features are debts. The cheapest feature is the one you never start.

7. Polish takes 2-3x implementation time. Budget for it, or the polish won't happen. Projects that do not schedule polish die in the polish phase. If a feature is not worth polishing, cut it — a half-polished feature is worse than no feature at all. The 80/20 rule applies: the features players interact with most, and the ones that set the quality bar, deserve the deepest polish. Everything else can be adequate.

8. By the end of week two, you must have a playable prototype. If you don't, the project is probably in trouble — either the scope or the tech stack is wrong, and that diagnostic is cheaper to act on now than later. The playable prototype transforms the project from speculation into production. Everything else follows from having hands on the thing.

9. Scope creep comes from many sources. Resist all of them. Yourself (notebook for next game, not this one). Your team (every proposal goes through the list). Your publisher (trade new features for cut ones, or reject). Playtesters (distinguish feedback that fixes problems from feedback that wishes for more). "Wouldn't it be cool if..." (cool is not a tier; run it through the list). The discipline of saying no is the designer's most important shipping skill.

10. Cut your darlings. Especially the ones you love most. The feature you love is often the feature slowing everything else. Stephen King's advice applies: what you cherish is what distorts the work around it. Cutting feels like grief because it is grief. Give cuts a small ceremony — a sentence-long obituary in your design journal — and then move on. BioShock 2's cut multiplayer saga, Final Fantasy XV's cut chapters, Stardew Valley's countless cut features all made the shipped games possible.

11. Re-scope mid-project when the signals say so. Re-scoping is a tool, not a failure. When milestones slip repeatedly, when morale drops, when "one more feature" becomes chronic, when playtesting stops because the build is too broken — re-scope. Demote Must-Haves, cut Nice-to-Haves, accept a longer schedule or a smaller game. No Man's Sky needed a re-scoping meeting before launch it never held. Hollow Knight held one and delayed accordingly. The outcomes reflect the choice.

12. Honest timelines are humbling. Plan for reality. Solo indies ship meaningful first games in 1-3 years (often 3+). Small teams take 2-4. AA takes 3-5. AAA takes 5-7+. If your plan is meaningfully shorter than these benchmarks, you are not on the short end of variance — you are underestimating. Lengthen the plan or shrink the game. Those are your two real options.

13. Ship small, ship often. The finishing is the skill. Terry Cavanagh's catalog, Ludum Dare culture, Daniel Cook's "one game a year" — the habit of finishing many small games builds shipping skill better than any amount of work on a single big project. The last 20% of development is where the shipping skill lives; you only develop it by reaching the last 20% repeatedly. A designer who has shipped five small games has skills a designer half-way through one big game does not.

14. Underpromise pre-launch. Overdeliver post-launch. The asymmetry is permanent. A feature you promised and cut costs more trust than a feature you never promised costs at all. A feature you never promised but ship for free six months later earns more goodwill than a feature you promised and shipped on time. Design your launch scope conservatively. Save the surprises for post-launch. No Man's Sky's launch disaster and eight-year recovery teach this lesson at a scale few games will match.

15. Your first game will be bad. Ship it anyway. The bad first game is the foundation of the good second game. The designer who has shipped, even once, has skills the designer who has not cannot acquire any other way. Cross the threshold. Cut what you have to. Save what you must. Ship.

One Sentence If You Remember Nothing Else

Cut enough that the remaining game can actually be finished, polished, and shipped — because a shipped small game is infinitely more valuable than an unshipped large one.