Comparisons
Choosing a prize-draw tool: the seven questions that actually discriminate
A draw your entrants can recompute themselves
Seed published before the draw, public drand randomness, shareable video. Free for entrants.
Create an accountVerify an existing drawPrize-draw tool comparisons line up features: entrant limits, CSV export, branding. None of those will help you the day an entrant writes "it was rigged". Here are the seven questions that will.
1. When is the entrant list frozen?
This is the most important question, and the least asked.
As long as the list can move after the draw, no other commitment holds: adding or removing an entry after seeing the result is enough to change the winner without contradicting anything published.
What you want to hear: "the list is sealed when entries close, and the seal's fingerprint is published at that moment." What isn't enough: "the list is stored in our database."
2. Where does the randomness come from?
An internal random generator is a black box: it yields a result you can neither predict nor reproduce, and certainly not verify.
What you want to hear: an external, public source, whose values are produced independently of the organizer and the tool. What isn't enough: "we use a cryptographically secure generator." Probably true, and unverifiable.
3. Did the randomness already exist when the list was frozen?
Subtle, and decisive. If the random value is already known when the list is sealed, whoever knows it can choose which entries to include to obtain the desired winner.
What you want to hear: the comparison is public and fits in two numbers — the randomness used belongs to a window that had not yet opened when the list was sealed.
4. Which rules exclude entries, and when are they published?
Every promotion excludes entries: duplicates, disposable addresses, unconfirmed sign-ups. The question isn't whether a filter exists, but when it is announced.
What you want to hear: the rules are public, identical for everyone, and applied before the list is frozen. What isn't enough: "we remove suspicious accounts." A discretionary filter applied after the draw is indistinguishable from choosing the winner.
5. Can an entrant verify they were in the pool?
Publishing the full list is out of the question — it's personal data. But an entrant must be able to confirm they were part of it.
What you want to hear: an individual proof of inclusion, demonstrating membership of the sealed list without revealing the other entrants.
6. Is the computation reproducible by a third party?
This is where everything above leads. With the published values, an outsider must be able to redo the computation and find the same winner.
What you want to hear: a documented method, and preferably a verifier independent of the organizer. What isn't enough: a "results" page hosted by the tool. It displays what the tool asserts, not what can be recomputed.
7. What happens to an unclaimed prize?
This happens more often than expected, and it's where many setups improvise.
What you want to hear: reassignment is a verifiable re-draw, run with the same values, in the same frame, traceable like the initial draw. What isn't enough: "we pick the next on the list." Which list, in what order, decided when?
The framework, to copy
| Question | Expected answer |
|---|---|
| When is the list frozen? | At closing, with published fingerprint |
| Where does randomness come from? | External, public source |
| Did it exist at sealing? | No, and publicly comparable |
| When are filters announced? | Before, public, identical for all |
| Individual proof of inclusion? | Yes, without revealing others |
| Reproducible computation? | Yes, by a third party |
| Unclaimed prize? | Verifiable re-draw |
A tool answering "yes" to all seven doesn't ask you to trust it. That is the only difference that matters, and no feature replaces it.
Key takeaways
Seven questions, one idea: what is frozen must be frozen before the randomness exists, and what is decided must be recomputable by someone else. Entrant limits, CSV export and visual branding never help you answer a challenge.
FAQ
Aren't these criteria excessive for a small promotion?
The stakes depend on the audience, not the prize. A modest promotion with a large following generates more disputes than a big confidential one. And these guarantees cost nothing in use: they are either built in or absent.
Can a tool answer "yes" to everything without being verifiable?
It can claim so. That's why the seventh question is the most useful: ask for the verifier. If it exists, everything else becomes checkable; if it doesn't, the rest is a declaration.
Do you need to understand cryptography to use these tools?
No. An organizer needs to know what is published and when; the computation is executed by a verifier. It's like accounting: you don't need to be an auditor to keep auditable books.
What if the tool already in place ticks nothing?
Don't change everything mid-campaign. But from the next one, publish a written commitment before entries open — filtering rules, selection method, treatment of an unclaimed prize. That's already half the way.