Understanding Random Number Generators

By the Riskr team · 6 min read

Every digital game of chance — a slot reel, a Plinko drop, a crash point, a shuffled deck — rests on the same hidden engine: a random number generator, or RNG. It's the thing that decides outcomes when there's no physical wheel or dice to do the job. Understanding what an RNG is (and isn't) clears up a lot of confusion about fairness, streaks, and whether a game is "rigged."

What an RNG actually does

At its simplest, an RNG produces numbers that have no predictable pattern — each one statistically independent of the last. Give it a call and it hands back a value, say between 0 and 1. That single unpredictable number is the raw material; the game then translates it into something you can see, like a reel position or a multiplier.

Pseudo-random vs. true random

There are two broad flavors. A true random generator draws from physical chaos — electronic noise, radioactive decay, atmospheric static — sources that are genuinely unpredictable. A pseudo-random generator (a PRNG) is a clever algorithm: feed it a starting value and it spits out a long stream of numbers that look random and pass statistical tests, even though they're produced by deterministic math.

Almost all software uses PRNGs, because they're fast, repeatable, and — when built with cryptographic-grade algorithms — practically impossible to predict without the starting value. "Pseudo" isn't an insult; a good PRNG is indistinguishable from true randomness for all practical purposes.

The seed

That starting value is called the seed. A PRNG is fully determined by its seed: the same seed always yields the exact same sequence of "random" numbers. That sounds like a bug, but it's actually the foundation of one of the most important fairness mechanisms in online gaming — because a result that can be reproduced from a seed can also be verified.

From a number to an outcome

Games map the RNG's raw output onto their own outcomes. The mapping is where the design lives:

Same fair, uniform random number — but the slices aren't equal. That's a crucial point we'll come back to: the RNG can be perfectly unbiased while the mapping still carves out an edge for the house.

"Provably fair" systems

Many sites use a commit-reveal scheme to prove they didn't tamper with results. Before a round, the operator generates a server seed and publishes a scrambled fingerprint of it (a cryptographic hash) — committing to a value it can no longer change. Your client often contributes a seed too. After the round, the operator reveals the original server seed, and you can check it matches the fingerprint and recompute the outcome yourself. If everything lines up, the result was locked in before any bets and wasn't altered afterward.

Random is not the same as rigged

People often suspect a game is "rigged" after a brutal losing streak. But a fair RNG has no memory. It doesn't know you've lost five in a row, doesn't owe you a win, and won't punish you for winning. Each draw is independent. A long streak — good or bad — is exactly what genuine randomness produces from time to time; it's clustering, not conspiracy.

So keep two ideas separate. Fairness is about whether the RNG is unbiased and untampered — verifiable, in a provably fair system. The edge is about how the payouts are structured, and it's deliberately tilted toward the house. A game can be 100% fair and still be one you'll lose on over time. Both things are true at once, and they're not in conflict.