🔀Shuffle any list into an honest random order.

✓ Tested & verified Updated: Aug 6, 2026 How the randomness works

List randomizer

Do the dishes
Take out the bins
Walk the dog
Water the plants
Cook dinner
Clean the bathroom

Press shuffle — every possible order is equally likely

Your list — one item per line

Try also

What is the right way to shuffle a list?

The list randomizer runs Fisher–Yates, and essentially nothing else. Walk backwards through the list; at each position, swap the item with one chosen at random from the positions at or before it. Do that once and every possible ordering of the list is exactly as likely as every other — a guarantee, not an approximation.

Published by Fisher and Yates in 1938, recast into its modern in-place form by Durstenfeld in 1964 and canonised by Knuth in The Art of Computer Programming, it is the shuffle that gets it right. The alternatives do not: sorting a list by random comparisons produces a measurably lopsided distribution, and swapping every item with any random position skews towards the orderings a short-cut favours. Both look shuffled. Neither is.

Six items have 720 possible orderings, ten have 3,628,800, and the shuffle picks between them evenly. Paste a rota, a set of slots or a playlist, press the button and read off the order — nothing is stored and nothing leaves the page.

The number of possible orderings grows faster than almost anything else in everyday arithmetic, which is why a shuffled list of any real length is effectively unique.

Items	Possible orderings
3	6
5	120
6	720
8	40,320
10	3,628,800
15	over 1.3 trillion

Fisher–Yates picks one of those evenly, in a single pass, however long the list is.

Three jobs that look similar and are not: a running order for everyone is this page, one winner from many is the random name picker, and an even split into groups is the team generator. For the reasoning behind all three, see how random is random.

Frequently asked questions

Why does the shuffle algorithm matter?

Because the obvious approaches are wrong in ways you cannot see. Sorting with a random comparator gives certain orderings noticeably more often than others, and the bias grows with the list length. Fisher–Yates is uniform by construction and runs in a single pass. For a chore rota nobody will notice; for a draw with something at stake, it is the difference between fair and merely mixed.

Is this the same as shuffling music?

It is the honest version. Most music apps do not use a pure shuffle — they nudge results to spread artists out or to favour tracks you skip less, because a true shuffle plays the same artist three times in a row often enough to generate complaints. This page does no such nudging, which is exactly why it can be used for a draw.

Can I shuffle numbers or a spreadsheet column?

Paste anything that arrives as one item per line, and that includes a column copied from a spreadsheet. Blank lines are ignored. If you want random numbers rather than a shuffled list of your own, the random number generator with the unique option draws them directly.

How many items will it handle?

Hundreds without complaint — Fisher–Yates does one pass over the list, so the work grows in a straight line rather than exploding. The limiting factor is reading the result; past a couple of hundred lines, copy the output somewhere you can scroll it comfortably rather than squinting at a phone.

Will I get the same order twice?

Essentially never, for a list of any size. Ten items have 3,628,800 possible orderings and twenty have more orderings than there are seconds since the universe began, so repeats are not something you need to guard against. Short lists are different: three items have only six orderings, so repetition is expected and unremarkable.

Can I share or save the shuffled order?

Your input list is carried in the page address, so the link restores your items — but pressing shuffle again produces a fresh order, since results are not stored. To share the actual order you generated, copy or screenshot it. Nothing is written to a server, so the page cannot recover a shuffle you have already replaced.

How the randomness works

The list is shuffled with Fisher–Yates: starting from the last item, each position is swapped with one chosen uniformly from the positions at or before it, using indices drawn from crypto.getRandomValues with rejection sampling to avoid modulo bias. One pass produces an ordering drawn evenly from all n! possibilities — 720 for six items, over 3.6 million for ten. Blank lines are dropped before shuffling and duplicate lines are kept as distinct items. The list is mirrored into the page URL so it can be bookmarked or shared; the shuffled result itself is not stored anywhere.

Sources:

Why the order itself is a decision

Shuffling a list is the least glamorous randomiser here and the one most likely to be used for something that matters. Presentation order at a conference, the rotation for who does the school run, the sequence in which candidates are interviewed — all of these are decisions where the order itself has consequences and where the appearance of impartiality is half the point.

The practical advice is to shuffle once and publish. A shuffle you ran three times before liking the result is not a shuffle, and the fact that nobody can see the earlier attempts is precisely what makes it worth avoiding. If the order needs to be seen to be fair, generate it in front of the people it affects, or paste the result into a shared thread immediately.

Order effects are real, incidentally. Going first in an interview round or last in a talent show measurably changes how you are judged — which is an argument for randomising the order rather than letting it be chosen, and for saying openly that you did.