Stack Tower Project Report
1. The Concept
Stack Tower is a lightweight, web-based arcade stacking game designed around quick timing, simple controls, and satisfying visual feedback. I chose a tower-stacking game for this one-day sprint because it is small enough to build in a single file, but still complex enough to test a multi-AI workflow with real gameplay logic, animation, UI design, scoring, saved progress, and responsive layout.
2. The Stack
The Stack: HTML5, Vanilla CSS, and JavaScript.
The AI Assist: ChatGPT handled the initial product planning and generated the technical handoff document. OpenAI Codex was then used for the coding phase, taking the markdown handoff and turning it into a playable single-file browser game with canvas gameplay, home-screen UI, level progression, localStorage saving, and visual polish.
3. The 1 Day Sprint:
Morning (Hours 1–3): We started by defining the core game loop: a moving block travels across the screen, the player taps or presses a key to drop it, the overlap is calculated, and any missed part is sliced away. The planning phase focused on keeping the project simple enough for one file while still including the expected arcade features: score, best score, restart flow, sound toggle, mobile support, and satisfying feedback.
Afternoon (Hours 4–6): The coding phase began with the core canvas game. Codex built the basic stacking mechanics, block trimming, falling pieces, camera movement, score tracking, and localStorage best score. After that, the project expanded into a premium arcade interface with a home screen, Level Mode, Endless Mode, high score panels, settings, and a 20-level progression system. A lot of time went into keeping the gameplay logic intact while improving the UI.
The Finish Line (Hour 7-8): The final stretch was mostly polish and bug fixing. The first versions had layout problems, oversized fonts, overlapping panels, unresponsive sections, and several menu buttons that looked clickable but did nothing. We spent the last phase improving responsiveness, wiring every major button, replacing unfinished overlays with reusable premium dialogs, improving the in-game HUD, upgrading the background and blocks, and checking that the JavaScript still parsed correctly.
4. The Roadblocks
Roadblocks: The biggest roadblock was UI consistency. The gameplay became playable quickly, but the visual presentation needed many iterations before it felt like a polished indie arcade game. Another issue was feature creep: the project started as a simple endless prototype, then grew into home screens, level selection, settings, stats, responsive design, and 20 unique levels. The hardest part was adding those features without rewriting the original gameplay or breaking the single-file requirement.
5. Key Takeaways
What was the part that worked the best? The core gameplay implementation. Codex was able to turn the handoff into a functional stacking game with collision detection, scoring, animation, and saved progress very quickly.
What was the part that worked the worst? Visual precision and responsive UI. Matching premium mobile game references required repeated prompts and manual refinement because the first layouts had oversized typography, broken spacing, inactive buttons, and overlay issues.
My final thought: If I had one more day, I would separate the code into modules, add a proper asset pipeline, and run visual QA on multiple device sizes. If I could redo this project, I would lock the UI direction earlier with a smaller reference set so the gameplay, level system, and menu design could evolve together instead of being redesigned several times near the end.