Mazocarta: Rust-powered deckbuilder uses deterministic seeds for automated playtesting
A procedural deckbuilder in Rust with automated balance probes and QR-based multiplayer
Mazocarta, created by Timothy C. Cogan, is a seeded procedural tactical deckbuilder implemented in Rust and compiled to WebAssembly for browser play (also natively executable). Its primary contribution isn’t a new genre but a reference artifact for instrumented game development: a single deterministic rules engine that serves interactive play, native command-line simulation, automated end-to-end tests, save/load fixtures, and local-area multiplayer via QR-mediated WebRTC pairing. The paper describes its architecture, deterministic run model, and reproducible balance probes. An evaluation over 1,000 seeds produced a single-player autoplay win rate of 36.1% and two-player rate of 34.9%—not final balance metrics, but repeatable probes for regression detection.
This approach brings software engineering rigor—continuous integration, automated regression checks, and deterministic simulation—to game design. Developers can commit balance changes and instantly verify their impact through the same shared production core. Mazocarta is open-source, positioning it as a practical tool for both game developers and researchers studying game design as a software engineering discipline. The paper is 9 pages with 4 figures and code available.
- Built in Rust, compiled to WebAssembly for browser play and natively for simulation
- One deterministic rules engine powers interactive play, CLI simulation, automated tests, save/load, and local multiplayer
- Over 1,000 deterministic seeds: single-player win rate 36.1%, two-player 34.9% as repeatable balance probes
Why It Matters
Brings software engineering best practices (automated testing, deterministic validation) into iterative game development workflows