The math is the point.
0xRoulette is a developer's lens on a problem that has been misframed for centuries. The house always wins (probably), but understanding exactly how, how fast, with what variance, and whether there's a crack in the wall at all is a legitimate engineering problem.
The name
0x is how developers write hexadecimal. It shows up at the bottom of the stack: in bitwise ops, memory addresses, PRNG seeds. It signals a certain kind of mind: one that reads source code and is more interested in why something behaves a certain way than in the outcome. That's the mindset this tool is built for. The RNG at its heart runs on a Squirrel3-family hash seeded with 0x6d2b79f5. The name and the implementation share the same register.
What it is
A strategy backtesting and forward-testing platform with an embedded scripting engine. Three modes (Game, Advanced, and Engineer) reveal progressively deeper layers of the same system. You can play a faithful simulation, watch autopilot bots run with a live P&L chart, or open a Monaco IDE and write Lua strategies that execute in a sandboxed WebAssembly runtime. Run a strategy against thousands of simulated spins to backtest it, then run the same code live, round by round, to forward-test it against the wheel as it spins.
The carrot, and the catch
Every system promises you can beat it. We promise something more interesting: the tools to chase that promise as far as it will actually go, and an honest account of where it stops. Here's the catch on one half of the question: roulette has a fixed negative expected value, and over the long run no betting system has ever closed that gap: not Martingale, not D'Alembert, not any coverage scheme or pattern hunt against a fair wheel. A pure betting system cannot create edge; what it can do is shape variance: the width of the swings, the depth of the drawdowns, the probability of ruin before a target is reached.
But that's not the whole question. A wheel is also a physical object, and physical objects have measurably been beaten before: see the record. That half of the carrot is genuinely open: we're building wheel-bias modeling into the engine specifically so it can be chased here, not just read about. Track that work on the Research log.
Who it's for
Developers, quants, students of probability, and anyone who would rather verify a claim than trust it, including people who aren't developers yet but want to be, and are happy to pick up Lua and some statistics along the way. This is meant to be a genuine STEM playground: level up your intuition for variance and probability, have fun doing it, and either find out for yourself whether a wheel can be beaten, or prove rigorously that it can't. Both outcomes are worth having.
Jesse (Human)
Jesse (Human) builds 0xRoulette and edits its research. That is the complete public byline used on this site; no institutional credential or affiliation is claimed.
The relevant project record is the work itself: the deterministic simulator, its documented assumptions, fixed-seed experiments, downloadable result files, and revision dates. Start with the published studies and reproduce the numbers before accepting a conclusion.
Honest about money
0xRoulette is free. No real money changes hands, nothing pays out, and it is not gambling. We keep the lights on with a few Amazon affiliate links and, eventually, optional hosted/API tiers for people running heavy backtests. That's all disclosed on the Disclosure and Support pages.
Building something with it, or want to argue about the math? Get in touch.