Random Number Generator

Genuinely Random Numbers, Any Range, Any Quantity

Humans are famously bad at producing randomness — asked for a number between 1 and 100, we cluster around 37 and 73, avoid round numbers and refuse to repeat — and equally bad at trusting whoever picked the “random” winner. A proper generator solves both problems. This one draws numbers from your browser's cryptographically secure random source within any range you set, one at a time or in batches, with or without duplicates, instantly and locally. Raffles, decisions, sampling, games, classroom picks: the number is fair because no human chose it.

How to Use It

1. Set the range — minimum and maximum, both inclusive. Any integers work, negative numbers included.
2. Choose how many numbers to draw.
3. For multiple draws, decide on duplicates: allowed (independent rolls, like dice) or unique (drawing from a hat without replacement, like lottery balls and raffle winners).
4. Click Generate. Results appear large and copyable; generate again for a fresh draw.
5. Shortcut presets cover the classics: coin flip (1–2), die (1–6), d20, 1–100.

The Distinction That Matters: With or Without Replacement

The duplicates toggle is the mathematically meaningful choice. With replacement, every draw is independent — rolling five dice can absolutely produce three fours, and a fair generator sometimes will (streaks are a property of true randomness, not a defect). Without replacement, each result is removed from the pool — the right model for picking three different winners from fifty tickets, assigning unique numbers to participants, or sampling distinct items from a list. Choosing the wrong mode is the most common randomization error in informal raffles; having it as an explicit switch prevents it.

Fairness You Can Defend

Two properties make a draw defensible to the people affected by it. First, the randomness source: this tool uses the Web Crypto API — the same generator trusted for cryptographic keys — rather than weaker arithmetic generators, and it applies rejection sampling so every value in your range is exactly equally likely (naive modulo arithmetic subtly favors low numbers; this implementation avoids that bias). Second, the procedure: announce the rules before generating — range, number of winners, duplicates mode — and generate in view of the audience or on screen-share. The combination of unbiased source and pre-committed procedure is what makes “the computer picked it” a legitimate answer.

Beyond Raffles

• Decisions: number the options, generate, obey — surprisingly effective against choice paralysis, and the flash of disappointment when the “wrong” option wins tells you what you actually wanted.
• Sampling: pick unbiased subsets of rows, tickets or participants for spot-checks and QA.
• Games: lost dice are no obstacle; the d6 and d20 presets stand in faithfully.
• Practice data: quick numeric fixtures for testing and teaching.

Generation is local, instant and unlogged — nobody, including this site, knows what numbers you drew. For random secrets rather than numbers, use the companion Password Generator, built on the same cryptographic source.

To close with the most common real event: a social media giveaway. Number the entrants in order of entry, announce that order publicly, set the range here, draw once on a screen recording, and post the clip. The whole procedure takes two minutes and leaves nothing to argue about — which, for anyone who has moderated the comments after an opaque winner announcement, is worth far more than the two minutes.