Proving your draw was fair

Last updated: August 12, 2026 · about 6 minutes

An accusation of rigging is almost impossible to disprove after the fact and almost trivial to answer in advance. The difference is four records, and gathering them takes about ten minutes spread across the giveaway.

The goal is not to convince a determined bad-faith critic — that is not achievable. It is to make the answer so immediate and so boring that the conversation ends in one reply, and everyone reading it can see who is being reasonable.

Record 1: the rules, published before entries open

Dated, public, and unedited. This is the foundation, because it converts every later question from a matter of opinion into a matter of fact. “Why did you disqualify accounts created that week?” has an easy answer if the rule was published on day one and a terrible one if it appeared after the draw.

Practical detail: platform edits are usually invisible and always suspicious. If you must correct something, post a new dated update rather than editing the original, and keep both. Screenshot your own rules post the day you publish it, so you have a copy that does not depend on the platform.

Record 2: the frozen entry list

At your published closing time, capture the entries in a form that cannot change: export the comments to a spreadsheet, screenshot the full thread, save the form responses. Note the exact time you captured it and the total count.

This is the record that gets used most often, and not for the reason people expect. The common dispute is not “you picked your friend” — it is “my entry was in there and you missed it.” Sometimes they are right, because platforms hide comments, filter them as spam, or collapse replies. A frozen list dated at closing time lets you check honestly and answer either way within a minute.

Keep the count. “Drawn from 412 valid entries” is a specific, checkable claim, and specificity is what reads as honest.

Record 3: the draw itself

Record the moment of selection. In order of strength:

• A screen recording, unedited, showing the full entry list on screen, the draw happening, and the result. This is the strongest option and takes thirty seconds. Phone screen recording is fine.
• A live draw on stream or in a story, with the list visible. Strong because it is unedited by definition, and it doubles as content.
• A screenshot of the result with the entry count visible. Weakest of the three, but far better than nothing.

Two habits to avoid, because both make everything above worthless. Do not draw several times and keep the result you liked — if you would not show every draw, the ones you hid are the story. And do not draw the backup winner before the first winner has actually missed their deadline; draw all placings in one recorded session from the same frozen list.

Record 4: the public announcement of the method

Say what you did when you say who won. One sentence:

That sentence pre-answers who, how many, when, with what, where the evidence is, and what happens next. Compare it with “Congratulations @name!”, which answers nothing and invites all of it.

How long to keep it, and why

Keep the four records until the prize has been delivered and accepted, plus a buffer — ninety days is a reasonable default for an ordinary giveaway. Challenges tend to arrive either immediately or weeks later, when someone notices the winner's account looks unusual.

There is a tension here worth naming. An entry list is personal data, and holding it forever is the wrong instinct: keep it as long as you need it for the promotion and delete it afterwards, as your own rules should say. Retain evidence and respect data minimisation — the two are compatible if you set an end date.

The credibility multiplier: get someone else involved

For anything with a meaningful prize, the strongest and cheapest upgrade is a second person. Have someone unconnected to you press the button, or draw live on a call with a guest. The reason it works is structural rather than technical: the accusation is always that you chose, and it evaporates when you demonstrably did not.

What to do when accused anyway

It will happen, and the first ninety seconds matter more than the evidence. Reply once, publicly, calmly, with the specifics and the link to the recording. Do not delete the accusation — deleting it is read as confirmation by everyone who saw it. Do not argue past that one reply; you have made your case and the audience can see it.

If the complaint turns out to be right — you missed an entry, the export dropped a page, you mixed up two similar usernames — say so immediately and fix it. Redraw publicly, or award a second prize, and explain what went wrong. Audiences are remarkably forgiving of a mistake admitted quickly and remarkably unforgiving of a defensive one.

The tools on this site cannot supply any of this for you. What they can do is be describable in a sentence, which is exactly the sentence you need. Why that matters is covered in how random is random.