RANDOM NUMBERS: THE GUIDE

"Say a number from 1 to 10." Sounds easy — but humans are terrible randomness generators: we say 7 far more than we should, we avoid the extremes, and we never repeat the number that just came up. That's why, when the outcome matters — a draw, a raffle, an allocation — you should delegate to a real generator. This guide explains, without unnecessary jargon, what a random number is, where it comes from, and how to use it properly.

What "random" actually means

A process is random when it meets two conditions: unpredictability (nobody can know the next result, even knowing all the previous ones) and uniformity (if it's "1 to 100", every number has exactly a 1-in-100 chance). Mind the nuance: random doesn't mean "nicely spread out". Twenty dice rolls can contain three sixes in a row — in fact, it would be suspicious if that never happened. Real randomness has streaks; human-invented "randomness" doesn't. That's exactly how teachers catch students who fake their experiment data.

Physical randomness: dice, drums and coins

For centuries, randomness was manufactured with physics: dice, lottery drums, shuffled decks, coins. They work because the result depends chaotically on details impossible to control (force, spin, bounces). Well built, they're excellent generators — which is why lotteries still use drums: not because they beat a computer, but because everyone can see nobody is cheating.

Computer randomness: the PRNG trick

A computer is a deterministic machine: it does exactly what it's told, which is the exact opposite of randomness. Its solution is the pseudo-random number generator (PRNG): an algorithm that starts from an initial value (the seed) and produces a sequence that passes every statistical test of randomness, even though it's technically computed. To make even the seed unpredictable, modern systems feed it real-world "noise": timing micro-variations, movements, hardware events. The practical upshot: numbers indistinguishable from pure chance for any everyday use.

Are the numbers on this site good enough?

Yes — for everything everyday. Our random number generator (and the rest of the tools: name picker, wheel, dice…) uses your own browser's random generator with a uniform distribution: every number or name gets identical treatment, no memory of previous results, and no way to manipulate the outcome — everything runs on your device, no servers involved. For draws among friends, classroom raffles, giveaways, allocations and games, that's more than enough (and far more reliable than any human "thinking of a number"). Only regulated contexts — official lotteries, real-money gambling, cryptography — require specifically certified and audited generators.

Practical everyday uses

• Raffles with numbered tickets: generate a number between 1 and N in front of everyone. Rules first, result final — the fair draw guide has the full protocol.
• Picking without a list: 30 numbered students? One number from 1 to 30 and nobody can complain.
• Allocations and turns: highest number starts; or roll dice, which is the same thing with more theater.
• Sampling and spot checks: to audit 5 orders out of 200 without bias, let 5 random numbers choose.
• Home bingo: our bingo caller is exactly this: numbers 1 to 90, no repeats, called one at a time.

One important non-use: passwords. Browser randomness would be up to the task, but a password also needs secure management (length, storage, no reuse) — use a password manager; that's its job.

The gambler's fallacy (and other myths)

Myth number one: "red has come up five times, black is due now". False — the roulette wheel has no memory, and the next spin's probability is identical to the first's. That's the gambler's fallacy, and it bankrupted so many people at the Monte Carlo casino in 1913 (26 blacks in a row, with half the room betting red "because it was due") that it carries the alternative name Monte Carlo fallacy. Its close cousins: lottery "hot numbers" (coming up a lot lately raises nothing), "overdue numbers" (years without appearing raises nothing either), and the lucky charm of the day. In uniform randomness, every draw starts from zero. The only real strategy is understanding the odds… or not betting the money.

Mini-glossary

• Uniform: every possible outcome has the same probability.
• Seed: the initial value from which a PRNG derives its whole sequence.
• Without replacement: each result can only come up once (like bingo). With replacement, every draw is independent (like dice).
• Bias: any deviation from uniformity — a loaded die, a bent coin, a human "choosing at random".