Sweeper's Gambit: Minesweeper x Chess

When I started working on Sweeper’s Gambit, my goal was simple: merge two games I found interesting and challenging—chess and Minesweeper—into something fresh and exciting. I’ve always been fascinated by chess’s strategic depth and Minesweeper’s logical problem-solving. Combining them seemed like a fun challenge, but it was also an opportunity to push my game design and development skills.

The concept for Sweeper’s Gambit came to me during my sophomore year. I wanted to create a game where players navigate a Minesweeper-style grid, but instead of just clicking tiles, they’d use chess piece movements to strategize. Imagine a Knight jumping over tiles or a Bishop sweeping diagonally across the board—it added a layer of strategy that felt new and exciting.

A large part of Sweeper’s Gambit—about 80–90%—was AI-generated using Amazon Q Developer. Here’s how I integrated it into my workflow:

• Code Generation: I described the core mechanics of the game, including how Minesweeper is traditionally played and the modifications I wanted to implement to blend chess-inspired movements. Amazon Q Developer generated a complete codebase, including the file structure, foundational logic, and essential features for these mechanics.

• Debugging Assistance: Whenever I encountered issues, I relied on the AI to suggest fixes or optimize existing code.

• Iterative Development: By iterating with AI-generated code, I could quickly test ideas and make changes based on what worked best in practice. It felt more like a conversation with a senior developer than an AI tool.

The rest of the project involved stitching the AI-generated code together with my customizations:

• HTML and CSS: To design the layout and styling of the game.

• JavaScript: For handling game logic, especially around integrating Minesweeper rules with chess-inspired movements (I had plans to add a chess game mode with AI and knew there was a stockfish module I could use in javascript).

AWS tools were essential for turning the prototype into a polished, playable game:

• AWS Amplify: Hosted and deployed the game with a seamless CI/CD pipeline.

After the initial build, I shared the game with family and friends. Their feedback helped me refine the AI-generated code and adjust the gameplay mechanics to make it more engaging.

I’ve used many LLMs for code generation, but Amazon Q Developer has been the most impressive so far. For this project, I hardly needed to tweak the generated code, which made the process incredibly smooth. However, I believe that for more complex projects, you might need to review and refine the code more thoroughly.

One key observation is that when working on larger or multifaceted projects, asking the model to implement multiple features simultaneously can reduce its efficiency. It’s better to strategize by breaking down the development process and tackling features one at a time.

Here’s how I approached it:

1. Leveraging ChatGPT for Prompt Refinement:

I started by providing ChatGPT with an overview of my project and describing a specific feature I wanted to implement. ChatGPT’s clarifying questions helped refine my understanding and ensure I was covering all necessary details.

2. Creating Detailed Prompts for Q Developer:

Once I was satisfied that ChatGPT had a complete understanding of the feature, I asked it to generate a detailed prompt. This prompt was then fed into Amazon Q Developer, ensuring that the generated code aligned closely with my project goals.

This method allowed me to combine the strengths of both tools, resulting in efficient and accurate code generation.

Lesson Learned: Playtesting is key. Listening to feedback made the game better than I could have imagined.

I plan to add elements of Minesweeper to chess to create a new chess variant.

This was a fun and rewarding experience, and I would highly recommend Amazon Q Developer to developers of all skill levels.

Play the Game: Sweeper’s Gambit