How to Run a Fair Raffle or Door-Prize Draw

A raffle goes wrong in one of two ways: the draw is biased, or people believe it is biased. Both are fixed the same way — make the mechanics boringly explicit. A fair door-prize draw is just a ticket pool, a shuffle, and a stated rule about repeat winners, and every one of those parts can be written down before the first name is pulled.

The ticket pool is the whole model

Every raffle question reduces to one object: the pool of tickets. One name on one slip is the simplest pool — everyone has exactly one chance per draw. The classic door-prize twist is weighted entries: a person who bought five tabs of tickets has five slips in the drum, and is therefore exactly five times as likely to be drawn. That is all "weighting" ever means in a raffle — more copies of your name in the pool, never a bigger pool for you alone.

The digital version works the same way. Paste each entry as its own line and duplicates naturally stack the odds; a list of Ada, Ada, Ben gives Ada two chances in three. What matters for fairness is that the mapping from entries to tickets is written down — "one line per ticket sold" — so nobody suspects the list was padded after the fact.

Shuffle, then deal in prize order

With the pool set, the fair procedure is a full shuffle followed by dealing: randomize the ticket pool once, and the first prize goes to the first drawn ticket, the second prize to the next, and so on. Dealing from one shuffle and drawing one-at-a-time are mathematically equivalent — both give every ticket the same chance at every prize — but dealing has an audit advantage: the entire outcome is determined by a single shuffle you can re-run.

Prize order deserves its own decision. Because the first line receives the first winner, put the grandest prize on line one and let the values descend. Reversing the order by accident — handing the big prize to the "third" winner — is the most common raffle dispute there is.

The repeat-win rule changes everyone's odds

Before drawing, state whether one person may win more than once. If repeat wins are off, the moment someone is drawn every remaining ticket bearing their name leaves the pool — their losing tickets cannot win a second prize, and everyone else's relative chances rise slightly. If repeat wins are on, only the single drawn ticket is removed, so a heavy ticket-holder can sweep the table.

Neither rule is more fair; they are different games. Community events usually want the prizes spread across people, which means no repeats. Fundraisers that sold tickets by the sheet sometimes reward volume and allow repeats. The unfair outcome is not choosing either rule — it is choosing implicitly and explaining it after someone wins twice.

Reproducibility: the seed as a receipt

A seeded draw is a draw with a receipt. The seed drives the random number generator deterministically, so the same entries, prizes, and seed always produce the same winners. Leave the seed blank and every draw is fresh; set it to, say, 42 and announce it before drawing, and you can re-run the exact result later if anyone asks — the digital equivalent of keeping the drum and tickets in the back room.

The generator underneath uses the same Park–Miller style seeded pseudorandom generator described in how pseudorandom numbers work: good enough to be unpredictable in practice, deterministic enough to be audited. That combination — unpredictable to participants, reproducible to the organizer — is precisely what a trustworthy draw needs.

Running the draw in public

Three habits make a draw unarguable. First, freeze the entry list before the event and show the count: "217 tickets, 34 people". Second, state the repeat-win rule and the prize order in the same breath. Third, publish the winners with the undrawn pool — everyone who did not win should be able to find their own name among the losers, which is the strongest possible evidence that the list was real.

For very large draws the mechanics do not change, only the bookkeeping: the same shuffle handles thousands of entries. What stays constant is the promise that the same inputs always give the same winners.

Try it yourself

The Raffle Draw Generator implements the whole procedure: one line per ticket, prizes dealt in listed order, a one-win-per-person switch, and an optional seed for a reproducible result. For a plain shuffled list without prize assignment, the Randomizer is the lighter tool; for weighted "pick one of my options" draws — lunch, chores, draft order — the Pick For Me decision maker adds per-option weights on top of the same engine.