RNG Integrity Studies

A series of empirical studies that collect thousands of rounds via browser automation and run rigorous statistical pipelines to determine whether commercial RNG games show any exploitable structure. Spoiler: they don't — and proving it cleanly is the point.

Stack

Engineering highlights

• Sporty Hero crash game: 15-dimension analysis — distribution fitting (5 models), independence (ACF, Ljung-Box), extreme-value (GEV/GPD), and volatility clustering (ARCH/GARCH).
• Confirmed a log-normal model via Kolmogorov-Smirnov (p=0.49), quantified the ~10.7% house edge, and proved zero exploitable edge via Kelly analysis.
• Virtual-soccer study: 51K matches via iframe-aware scraping, 13 statistical tests, calibration analysis pinning a 5.0% edge with <1pp pricing error.
• Reusable pattern: Playwright collector → SQLite (WAL) → analysis pipeline → reproducible report.