casino Random Number Generator star

How the Numbers Are Actually Generated

Not all "random" is equal. Math.random(), available in every browser, is a fast pseudo-random generator — deterministic under the hood and, in principle, predictable if someone knows its internal state, which makes it unsuitable for anything security-sensitive like draws with real stakes. This tool instead uses crypto.getRandomValues(), the cryptographically secure random generator built into the browser, which draws from the operating system entropy sources and is not predictable even in principle.

Having a secure random source solves only half the problem, though. Mapping a random value into a specific range (say, 1-100) naively — by taking random % 100 — introduces a subtle but real modulo bias: because the raw output range of the generator usually is not an exact multiple of 100, some remainders end up very slightly more likely than others. The bias is small, but it is real and it compounds across many draws — exactly the kind of thing that should not exist in a supposedly fair raffle.

The fix is rejection sampling: instead of forcing every raw value into range via modulo, out-of-range raw values are discarded and a new one is drawn until it fits cleanly into an exact multiple of the target range. This guarantees a truly uniform distribution — every number in the range has exactly equal probability — at the cost of occasionally needing an extra draw or two, which is imperceptible at the speeds involved.

Choose a minimum, a maximum and how many numbers are needed; tick unique to prevent repeats, which is what raffles, prize draws, sampling and games need when the same entry should not win twice. Everything runs on the local device — no number is transmitted or stored anywhere, and refreshing the page produces a completely fresh draw.

Things Worth Knowing

• Modulo bias is invisible in a handful of draws but becomes statistically detectable over thousands — rejection sampling removes it entirely rather than just shrinking it.
• crypto.getRandomValues() being "cryptographically secure" means its output cannot be predicted even by someone who has seen previous outputs — a property Math.random() explicitly does not have and was never designed to have.
• The range is inclusive on both ends: a 1-10 draw can produce both 1 and 10, not just the values strictly between them.
• For an auditable raffle, number participants 1 to N, set the range accordingly, enable unique, and take a screenshot of the result, since nothing is stored server-side after the draw.