Giveaway Picker

Why a random picker is not enough

Every giveaway tool claims to be random. None of them can prove it to the people who lost. From the outside, a winner list on a screenshot is indistinguishable from a name the organiser simply liked — and in a community where the same few people keep winning, that suspicion does real damage.

This picker is built around that problem. The draw is reproducible: the same participants and the same seed always produce the same winners. That turns the result into something checkable instead of something to take on faith.

How to run a giveaway nobody can argue with

1. Announce the rules before you draw. How many winners, how many reserves, who is excluded, and when the entry list closes.
2. Publish the participant list. Post it in the topic, or leave it visible where people entered. The proof is only worth something if others have the same list.
3. Draw, and post the proof block. It contains the fingerprint of the list, the seed, and the winners.
4. Tell people they can check it. Anyone opens the Verify a draw tab, pastes the proof and the list, and sees for themselves.

If you want to go further, post the list fingerprint before you draw. Then you cannot quietly add or remove an entry afterwards — the fingerprint would no longer match.

What actually happens when you press Draw

No hand-waving, because the whole point is that you can check it:

• Names are trimmed, extra spaces collapsed, and duplicates merged case-insensitively. Ticket counts are read from Name x3 or Name, 3.
• The list is sorted alphabetically, then written as name|tickets, one per line. Sorting is what makes the order you pasted in irrelevant — otherwise two honest people with the same list would get different results.
• That text is hashed with SHA-256. This is the fingerprint in the proof.
• The fingerprint and the seed together seed a deterministic generator (xmur3 into sfc32).
• One ticket per entry goes into a pool, which is shuffled with Fisher–Yates — an unbiased shuffle, not a repeated random pick.
• Reading the shuffled pool in order and skipping people already drawn gives the winners, then the reserves. Extra tickets raise your chance of being drawn, but nobody can win twice.

Everything runs in your browser. The participant list never leaves your computer, and there is nothing stored on our side — which also means we could not tamper with a draw even if we wanted to.

What the proof does and does not show

It proves that these winners follow from that list and that seed. Change one character in a name, or one character of the seed, and the check fails.

It cannot prove the list was honest in the first place. If an organiser quietly drops someone before drawing, the proof is still internally valid — that is why publishing the list, or its fingerprint, before the draw matters. Nor can it prove the seed was not chosen after trying several: to rule that out, commit to the seed publicly before you draw, or use something nobody controls, such as a lottery number announced later that day.

What people use it for

• Game key, skin and in-game item giveaways
• Discord role, nitro and event rewards
• YouTube and Twitch comment or chat contests
• Forum raffles and community events
• Picking a random name for anything else — who goes first, who buys the pizza

Questions