Draw several entries from a list with no entry appearing twice. This is sampling without replacement, done with a partial Fisher–Yates shuffle so every subset is equally likely — the right tool for shortlists, winners and reserves.
Settings
Recent results
Generated locally in your browser — your settings and results never leave this page.
With replacement, or without
Two operations get called “picking at random” and they are not the same. Drawing with replacement puts each selection back before the next, so the same entry can come up repeatedly and each draw is independent. Drawing without replacement sets each selection aside, so every result is distinct and each draw changes the odds for the rest.
Choosing the wrong one produces results that are wrong in a way nobody notices at first. A prize draw with replacement can award both prizes to the same person. A rota with replacement can put the same name on Monday and Tuesday while somebody else does nothing all week.
Why the method matters
The obvious implementation is to draw at random and try again if you get something already chosen. It gives correct results, and it degrades badly. Selecting 95 items from 100 means the last few draws are rejecting almost every attempt, and the time taken grows without bound as the list empties.
A partial Fisher–Yates shuffle avoids the problem. Walk through the list, swap each position with a randomly chosen later one, and stop after you have as many as you need. Every possible subset comes out with equal probability, the work is proportional to the number drawn, and there is no retry loop to spin.
The distinction rarely matters for a list of ten names. It matters a great deal when the list is thousands of entries and you are drawing most of them.
Draw order carries information
Results appear in the order they were drawn, and that order is often meaningful. First out is first prize. Beyond the winners, the next few are reserves in the order they should be contacted if someone declines.
Drawing winners and reserves in one operation is better practice than running a second draw later, because the reserve list is fixed at the same moment as the winners. A separate draw held after somebody drops out gives the organiser a chance — real or perceived — to influence who replaces them.
The half that is left over
Showing the entries that were not selected turns a single draw into a workable process. Assign three people to a task and the remainder are the pool for the next one. Pick a group for the first activity and the rest form the second group. Work through a list in stages and you always know what remains.
It is also a check on the draw itself: the two halves together should account for every entry exactly once, and seeing them side by side makes it obvious if something has gone missing.
Refusing impossible requests
Asking for twenty distinct entries from a list of twelve has no valid answer. The tool says so rather than returning twelve and letting you assume you got twenty. Silent truncation is the kind of failure that produces a shortlist which looks complete and is not, and the mistake usually surfaces long after the draw.
Privacy
Your list is processed in your browser and is never uploaded, stored or placed in the page link.
Drawing in stages
Because the tool shows what remains as well as what was drawn, a single list can be worked through in stages. Draw three for the first task, paste the remainder back in, draw three for the second, and so on until the list is empty.
That is more useful than it sounds for anything organised in rounds — allocating people to sessions, dividing a class into activity groups, or assigning a backlog across a team. Each stage is a fair draw from whoever is genuinely still available.
It also keeps an audit trail if you keep the intermediate lists, because each stage records who was in the pool at the time. Questions about why somebody ended up in a particular group are then answerable.
How to use the Random Picker Without Repeats
- Paste your list, one entry per line.
- Set how many entries to draw — every one will be different.
- Pick, then copy the winners or the remaining entries.
Frequently asked questions
What does "without repeats" actually mean?
Each entry can be selected at most once. This is sampling without replacement: imagine drawing names from a hat and not putting them back. It is the right behaviour for winners, shortlists and rotas, where selecting the same person twice would be a mistake rather than a result.
How is this different from the standard item picker?
The standard picker lets you choose. This one is built around the no-repeat case: the option is always on, the presets are about drawing subsets, and the output shows both what was drawn and what remains.
Is every possible combination equally likely?
Yes. The tool uses a partial Fisher–Yates shuffle, which produces each possible subset with equal probability. A naive approach — drawing at random and retrying when you hit a duplicate — gives the same result but slows down badly as the list empties; the shuffle avoids that entirely.
What if I ask for more entries than the list holds?
The tool refuses and says how many are available. Silently returning fewer would be worse: you would get a short list that looked complete.
Why does it show me what was not selected?
Because that is frequently the useful half. Drawing three people for a task leaves everybody else available for the next one, and a rota that works through a list needs to know who is left.
Can I draw the whole list?
Yes — set the count to the length of the list and you get every entry in a random order, which is the same thing as a shuffle. That is a legitimate way to produce a full running order.
Does the order of the results mean anything?
Yes. Results come out in draw order, so the first is the first drawn. That matters when the positions differ — first prize and second prize, or a winner followed by reserves in the order they should be approached.
