Featured Case Study · Cursor Vibe Jam 2026 · Solo精选案例 · Cursor Vibe Jam 2026 · 独立
tarduck — extraction shooter
Three.js · Browser · Risk/Reward Loop · Solo · 30 Days
A compact extraction shooter built to answer one design question: can a browser-scale prototype still create the pressure of "one more room, or get out now"? I designed the risk/reward loop and progression surface, and directed the build end to end. 一个紧凑的撤离射击原型，用来回答一个设计问题：浏览器体量的项目，能不能仍然制造“再深入一间，还是现在撤离”的压力？我端到端负责风险/回报循环与成长界面，并主导整个开发。
Design Goal设计目标
Compress extraction-shooter tension into a short browser session: readable combat, meaningful inventory pressure, and a clear decision between greed and safety.
把撤离射击的张力压缩进一次短浏览器体验：战斗要易读，背包要有压力，玩家要清楚地在贪心和安全之间做选择。
Player Loop玩家循环
Enter raid → scout through fog → fight for loot → manage limited grid space → choose extract timing → convert loot into upgrades for the next run.
进入战局 → 在迷雾中探索 → 为战利品战斗 → 管理有限网格背包 → 选择撤离时机 → 把收益转化为下一局升级。
🔧 The Build🔧 程序实现
Built in Three.js + Vite with fully procedural geometry — no asset pipeline. The renderer uses physically based materials with bloom and ACES tone mapping for a cartoon-grim look, plus an in-engine lil-gui panel for live tuning. The real work was the direction layer: breaking systems into clear specs, reviewing and integrating each change, and keeping a coherent architecture across 57 commits in 30 days.
用 Three.js + Vite 构建，几何体全程序化生成，没有资源管线。渲染使用 PBR 材质、bloom 和 ACES 色调映射，形成偏卡通但冷峻的视觉质感，并内置 lil-gui 面板实时调参。真正的工作在“指挥层”：把系统拆成清晰的规格、审阅并整合每次改动，并在 30 天 57 次提交里维持架构一致。
🎯 The Design🎯 设计思路
The core is a risk/reward extraction loop: push deeper for better loot, or extract early and bank what you have. I designed the supporting systems around that tension:
核心是一套风险/回报的撤离循环：深入拿更好的战利品，还是早撤保住已有收益。围绕这个张力，我设计了这些支撑系统：
- 5 weapons and 5 enemy types with body-part hitbox damage, armor & durability
- Grid inventory · 12-talent progression tree · 8 crafting recipes
- 3 extract points · fog-of-war · wave spawning · difficulty scaling
- 5 种武器、5 种敌人，含部位 hitbox 伤害、装甲与耐久
- 网格背包 · 12 天赋成长树 · 8 个合成配方
- 3 个撤离点 · 战争迷雾 · 波次生成 · 难度缩放
What I Tuned我调了什么
The key tuning problem was not adding more content, but making the run state readable enough for players to make informed risks. I used fog density, enemy waves, extract placement, inventory pressure, and upgrade costs as the main pacing levers.
关键调优问题不是继续加内容，而是让玩家足够读懂当前局势，从而愿意承担风险。我主要用迷雾密度、敌人波次、撤离点位置、背包压力和升级成本来控制节奏。