Random Number Generator Wheel

Set your range

Type a minimum and a maximum, press Build the wheel, and every number in the range gets a slice. 1–20 is loaded to start with.

If you enter the numbers the wrong way round the tool swaps them rather than complaining. Ranges are capped at 200 numbers, because past that the slices are thinner than the text on them and the wheel stops being a wheel. For bigger ranges — a random number between 1 and 100,000 — you want a plain draw rather than a spin, and this is the wrong page for it.

Never repeat a number takes each number off the wheel once it comes up. That turns the tool into a draw without replacement: bingo calling, prize numbers, seat allocation, picking three different problems out of an exercise set.

Why a wheel instead of a number

A random number generator that prints "17" is faster. It is also completely unconvincing to anyone watching, which is the entire problem it needs to solve in a classroom, a livestream or a prize draw.

The wheel makes the range visible. Twenty people can see that there were twenty slices, that they were the same size, and that the pointer landed where it landed. Nobody has to trust a number that appeared out of nowhere.

That is the trade: you spend five seconds to buy credibility. When you are drawing alone and nobody needs convincing, use anything you like.

Common ranges

• 1–6 — a die. One slice per face, same odds as the real thing.
• 1–10 / 1–20 — the everyday classroom range: page numbers, question numbers, points.
• 1–100 — percentages, price-is-right style guessing games, random percentage discounts.
• 0–9 — a single digit, for building longer numbers one draw at a time.
• 1–52 — a card from a deck, if you map the numbers to cards yourself.
• Your own range — student IDs, seat numbers, raffle tickets. For actual raffle tickets the raffle wheel handles ticket ranges and prize tiers properly.

How the numbers are drawn

Exactly the same way as every other wheel on the site. The draw comes from crypto.getRandomValues(), not Math.random(), and it is taken before the wheel starts turning.

The one part worth explaining is the rejection step. Turning a 32-bit random number into "a number from 1 to 20" with a plain remainder would very slightly favour the low end, because 2³² is not divisible by 20. So any draw landing in the incomplete final block is thrown away and a fresh one is taken. The result window tells you how many draws were discarded — usually zero, occasionally one.

That is a small detail, and for picking a page number it does not matter. It matters for anything where somebody might reasonably audit the result, and it costs nothing to do properly.

The full explanation, with a 10,000-draw simulation you can run yourself →