Random number from 1 to 1000
A thousand numbers, for giveaways, raffles and spot checks.
Between 1 and 1000
No number drawn yet.
Tap Draw, or press Space
Range
Last draws
None yet. Your draws line up here, newest first.
- Free, no sign-up
- The result is drawn before anything moves
- Nothing leaves your browser
What it does
Tap Draw for a whole number from 1 to 1000, each with a 1 in 1000 chance. Draw up to 50 at once with No repeats for a fair sample, such as which rows of a spreadsheet to check by hand. Your browser’s secure random generator decides every number before the reels turn.
How to use it
- Tap Draw for one number from 1 to 1000.
- For a sample, set How many up to 50 and turn on No repeats and Sort.
- Tap Copy and paste the list into your spreadsheet.
What people use it for
Giveaways
Number the comments or entries in the order they came in, then draw one per prize with No repeats so nobody wins twice.
Spot checks
Auditing 1000 invoices or products? Draw 30 different row numbers, turn on Sort, and check those.
Guess the number
A thousand sounds hard, but with higher or lower hints 10 guesses always find it.
Odds from 1 to 1000
| From 1 to 1,000 | Odds |
|---|---|
| Any one number | 1 in 1,000 (0.1%) |
| An even number | 500 in 1,000 (50%) |
| The top ten percent (901 to 1,000) | 100 in 1,000 (10%) |
| A repeat within the first 38 draws | 50.9% |
| A given number never comes up in 1,000 draws | 36.8% |
| Draws to see every number at least once, on average | 7,485 |
The birthday problem bites hard here. With a thousand numbers to choose from, a repeat is more likely than not by the 38th draw. Anything with prizes needs No repeats on.
How it picks
The way the RandALL app does it: the number first, the reels after.
- Your range is read. Whole numbers up to 15 digits, negative or positive. A range typed backwards is turned round, so 100 to 1 means 1 to 100. Both ends can come up.
- The number is drawn first. For a range of n numbers, a whole number from 0 to n minus 1 is drawn with
crypto.getRandomValuesand added to your minimum. Rejection sampling throws away the few raw 32-bit values that would give some numbers an extra share, so each one has exactly a 1 in n chance. - Wide ranges get more bits. One 32-bit draw covers 4,294,967,296 numbers. For a wider range, two draws are joined into a number with just enough bits for it, and anything past the end is drawn again. No remainders, no rounding.
- No repeats is a fair set, then a shuffle. Floyd’s algorithm picks numbers that all differ, with every possible set equally likely, and a shuffle makes every order equally likely too.
- Then the reels turn. Each reel steps through the digits before its own, slowing as it goes, and they stop left to right, as in the app. They only show a number that is already decided.
Tested. In 100,000 test draws each from 1 to 10, 1 to 6 and minus 3 to 3, every number came up evenly (chi-square test). A range three times wider than 32 bits and the full 15-digit range split evenly into slices, odds and evens, and last digits. With No repeats no draw ever repeated, and all 20 ordered pairs from 1 to 5 and all 20 sets of three from 1 to 6 came up evenly.
More random tools
Questions
How do I pick a random sample from 1 to 1000?
Set How many to the sample size, up to 50, turn on No repeats and Sort, then tap Draw and Copy. Every set of numbers is equally likely to be picked.
What are the odds of the same number coming up twice?
Twice in a row, 1 in 1000. Somewhere in a run of draws, much higher: by the 38th draw a repeat is more likely than not (50.9%).
Can I go higher than 1000?
Yes. Set any maximum up to 15 digits on the random number generator, and every number still has the same chance.
Take it with you
RandALL for iPhone has 25 tools in one app: wheels and teams, real 3D dice, random alarms, Finger Picker and more. It works offline.
Your first 24 hours are free, with every tool. After that, subscribe or buy the lifetime unlock once. Family Sharing included.
Get it on the App StoreiPhone only for now. These web tools are yours to keep using.
