Library
The Cores of Game Design · 5 of 12
The Cores of Game Design
ARG Design CRITICAL

Implementation Playbook — From Concept to Complete Prototype

development-cycle prototype planning iteration playtesting GDD

Key Principle

The Game Design Development Cycle has two phases: (1) Planning and Concepting — work in any order toward a playable build, and (2) Iterating and Improving — test, diagnose, and refine. The threshold between them is the complete prototype: at least one mechanic and one goal, making the game minimally playable and testable. (Chapter 2)

Why This Matters

Without a complete prototype, playtesting is impossible and the iterative cycle cannot begin. Teams that endlessly plan without producing something testable, or skip to polish before establishing a playable foundation, waste effort and lose motivation. The cycle anchors all development in player feedback rather than designer assumptions.

Good Examples

  • Complete prototype diagnostic: After playtesting, isolate which half failed. Mechanic fails but goal works? Fix the mechanic. Goal fails but mechanic works? Pivot the goal. Both fail? Return to concepting. (Chapter 2)
  • Emergent player goals: When players create their own goals during playtesting, treat this as a design signal. Consider pivoting the official goal to match emergent behavior rather than fighting it. (Chapter 2)
  • Family Tree balancing: Happiness threshold moved from 20 (too long) to 10 (too fast with lucky cards) to 15 (next test) — iterative economic balancing through playtesting. (Chapter 8)

Counterpoints

  • GDD-as-gospel anti-pattern: Template-driven GDDs accumulate speculative fields, become busywork, and the team stops reading them. "The information in the GDD should serve the project, not the other way around." (Chapter 4)
  • Software-as-protagonist trap: Programmer-background developers build engines and tools instead of reaching a complete prototype. "The software works like the mayonnaise on a sandwich: it helps to hold and enhance the other ingredients." (Chapter 4)
  • Perfectionism as release blocker: "Perfectionism is just an insecure feeling that stops you from understanding how careful and attentive you have been." External validation through playtesting, not self-judgment, is the antidote. (Chapter 17)

Key Quotes

"Thus, to playtest, it is necessary to have at least one mechanic and one goal." — Yvens R. Serpa, Chapter 2

"Planning itself is hard due to its nature: planning is nothing more than an attempt at predicting the future." — Yvens R. Serpa, Chapter 4

"From a practical standpoint, you should prefer improvements over additions. Improvements require less work and are more likely to strengthen the links between your game elements and the cores." — Yvens R. Serpa, Chapter 17

Rules of Thumb

  • Build the complete prototype before iterating — one mechanic + one goal is the minimum
  • Change fewer variables per iteration to keep feedback attributable
  • Assess planning feasibility on three dimensions: Proximity, Experience, Motivation — demotivation is the most corrosive
  • Use the halving-based point scale for estimation (0=1hr, 1=1day, 2=quarter-week, 4=half-week, 8=full week); always round up
  • GDD must be easy to access, easy to understand, easy to navigate — organize "certainly useful decisions," not "possibly useful decisions"
  • Integration of all game elements is the hardest task — do it early and continuously, not as a final step
  • Three-question completion checklist: (1) Has the game achieved its purpose? (2) What is still needed? (3) Is the fix an improvement or an addition?
  • Prefer improvements over additions in late-stage iteration

Related References