Launching JugPanda Studios — A Three-Year Retrospective

Three years ago, Coco Crunchers was just a simple educational game concept. Today, it stands as a full research platform — and JugPanda Studios is the home that grew around it. This post is the story of how a solo project became a studio, and what comes next. ## The Beginning In 2022, I set out to build an educational platformer that could actually measure whether players learned something. The idea was straightforward: a 2D game where a character named Coco (inspired by an English Bulldog) navigates levels that integrate math challenges into the gameplay itself. The first prototype was built in Python with tkinter. It was rough, but it worked well enough to test the core hypothesis: can a game improve math scores? ## Year 1: Proving the Concept The first year focused on one question — do players actually learn? With an initial group of participants, I ran pre- and post-tests around gameplay sessions. The results were encouraging: average scores improved from 83% to 92.5%, with statistical significance (p = 0.02459). That was the proof of concept. Games could teach, and we could measure it. ## Year 2: Understanding the Player Year 2 shifted focus from learning outcomes to player experience. I recruited participants across a wide age range (8 to 77) to study immersion — how differently do players of various ages engage with the same game? The technical stack evolved too. I migrated from tkinter to pygame (and later pygame-ce), built levels with the Tiled editor, and started collecting richer behavioral data. The game was becoming a proper research instrument. ## Year 3: AI Enters the Picture The final research year introduced an AI-powered hint system using OpenAI's GPT-4o. The question: does AI assistance improve learning outcomes or completion times? The results were clear — participants using the AI hintbot completed challenges 35.8% faster. The hint system provided contextual, adaptive guidance that met players where they were. On the technical side, this year also brought the biggest challenge: web deployment. The journey from local Python game to browser-playable experience went through Flask, Django, and finally Pygbag. Each attempt taught something new about the constraints of running Python in a browser. ## Building in Public Throughout all three years, I documented everything. Every pivot, every failed deployment attempt, every surprising data point. That philosophy — transparent development, honest about what works and what doesn't — became the foundation of JugPanda Studios. ## What's Next With Coco Crunchers documented and the studio infrastructure in place, the focus shifts forward. The full 3-year development timeline, research data, and playable game are available at CocoCrunchers.com. As for what comes next — new concepts in AI-driven gameplay and procedural storytelling are on the horizon. The research doesn't stop with one game. Follow along on the devlog for updates, or check out the source code on GitHub.