Running a Ballot for Scarce Tickets
Ten times more applicants than tickets, and every losing entrant is a person you know.
A ballot for oversubscribed tickets is the only draw here where most participants lose and most of them know the organiser personally, which makes the waitlist design and the group-application rule matter more than the randomness.
A ballot is the only draw on this site where most participants lose, and where most of them know the organiser personally. That combination is what makes the design choices matter more than the randomness.
The three decisions that determine whether people accept the result are how group applications work, how the waitlist works, and whether previous years count.
Group applications are the hardest part
People want to attend with their friends, and a ballot that allocates individually will split groups constantly. A ballot that allocates groups atomically gives larger groups a different chance than individuals, which is its own unfairness.
Neither answer is wrong and the choice must be made and published before applications open. Allocating by group means a group of six either all get in or none do; allocating individually means a group of six may get three places and have to decide what to do with them.
The middle option that works well in practice is to cap group size — applications for up to four, allocated atomically — which bounds the distortion while preserving most of the point of applying together. Whatever you choose, say it in the announcement.
- Atomic group allocation: groups stay together, larger groups distort the odds.
- Individual allocation: odds are clean, groups get split constantly.
- Capped groups: bounds the distortion, keeps most of the benefit. The common compromise.
The waitlist is part of the ballot, not an afterthought
Places are always returned, and how they are reallocated is the second most contested part of a ballot. The clean answer is to draw the full ordering at the outset — not just the winners — so the waitlist is a ranked list produced by the same draw.
That makes reallocation automatic and non-discretionary. When a place is returned it goes to the next name, and nobody has to decide anything, which removes the accusation that returned places went to friends.
It also means the whole ballot is one operation with one record. Publishing the ordering, or a commitment to it, covers both the initial allocation and every subsequent return.
One draw, two outputs
- 500 applications, 120 places.
- Shuffle all 500 once. Positions 1–120 get places; 121–500 are the waitlist in order.
- A returned place goes to position 121, then 122. No decisions, no discretion.
Whether previous years should count
Weighting by how many years someone has been unsuccessful is common and it is a real change to the draw rather than a tie-break. Someone with three unsuccessful years at weight four has four times the chance of someone applying for the first time, and that is a policy decision with winners and losers.
It is defensible — it converts a repeated lottery into something closer to a queue — and it is unfair to new applicants in a way that compounds. A popular event with heavy loyalty weighting becomes progressively harder to enter for the first time, which may or may not be what you want.
If you do it, publish the weights as odds rather than as a policy. "Applicants unsuccessful in the last three years receive four entries" is a policy; "a three-year applicant has a 4.8% chance and a first-time applicant has 1.2%" is the same thing stated in a form people can evaluate.
Announcing a result where most people lost
The mistake is announcing the winners. In a ballot most participants are not winners, and an announcement addressed to the successful minority reads badly to everyone else.
Tell everyone their own result directly and at the same time, and publish the aggregate — applications received, places available, success rate. A stated 24% success rate is information; silence invites the assumption that it was higher for people who knew someone.
And publish the method with the numbers. A ballot where the odds and the mechanism were both stated in advance produces complaints about the odds, which is a much better conversation than complaints about the process.
Frequently asked questions
Should groups be allocated together?
Decide and publish before applications open. Atomic allocation keeps groups together and distorts the odds; individual allocation is cleaner and splits groups constantly.
What is the common compromise?
Cap group size — applications for up to four, allocated atomically. It bounds the distortion while preserving most of the point.
How should the waitlist work?
Draw the full ordering at the outset. Positions beyond the place count are the waitlist in order, so returns are automatic and non-discretionary.
Why does that matter?
Because a returned place allocated by decision invites the accusation that it went to a friend. A pre-drawn ordering removes the decision entirely.
Should previous unsuccessful years count?
It is defensible and it is a policy choice, not a tie-break. It converts a repeated lottery into something closer to a queue, at the cost of new applicants.
How should weighting be published?
As odds rather than as a policy. '4.8% versus 1.2%' is evaluable in a way that 'four entries versus one' is not.
How should I announce the result?
Tell everyone their own result at the same time, and publish the aggregate — applications, places, success rate. Announcing only the winners reads badly to the majority.
Why publish the success rate?
Because silence invites the assumption that it was higher for people who knew someone. A stated 24% is information.
Tools for this job
Raffle & Giveaway Picker
Paste your entrant list and draw a fair, verifiable winner — built for streamers.
Weighted Random Picker
Give each entry its own odds — and see the exact probability table before you draw.
Random Sample From a List
Draw a fixed-size sample without replacement — for audits, spot checks, and surveys.
Background reading
How Weighted Random Selection Works
Cumulative sums, one float, and why duplicating entries is the wrong way to do it.
Giveaway Fairness Expectations
What entrants, platforms, and regulators generally expect from a prize draw. General information.
Sampling With and Without Replacement
The single decision that changes every number downstream of your draw.