On 21 April 2026 a post went up whose entire rule was three words: Do Not Comment. The pass condition was silence, in a comment section, from an audience that had by then spent five weeks learning that the way to play was to comment.

It got 311 comments. Eight arrived inside the first minute. Forty inside ten. The most liked comment, two and a half minutes in, reads disappointed in yall, at 392 likes. The second most liked, 36 seconds in, reads We're not even 4 minutes in, with three crying faces. The last comment arrived three weeks later.

No second attempt was posted. The operator stopped the round on realising it could not be won. This paper is about why, and about what a crowd does with a rule it cannot keep.

The posted rule asked a simple thing and the hypothesis underneath it was not simple: can a trained crowd withhold the one act it has been trained to perform? Every earlier round measured what people would do. This one measured whether they could not do it.

• Instruction: Do Not Comment.
• Where: one Instagram reel, posted 21 April 2026 at 03:08 UTC. The archive holds comments from 03:12 UTC, four minutes later, through 12 May.
• How we counted: every comment on the post, from the comment archive, split into top-level comments and replies. Top-level comments with no text (stickers and GIFs, which the archive stores as empty) are counted separately, because a sticker is a comment that says nothing, which is the closest anyone came to obeying.
• Timing: top-level comments are placed in four windows measured from the first archived comment: the first ten minutes, ten minutes to an hour, one hour to a day, and after a day.

Read the two most liked comments again. Neither breaks the rule for fun. One is disappointment in everyone else, the other is disbelief at how fast everyone else failed. They are the two most liked things on the wall because the room agreed with them, and in agreeing, the room commented. The mechanism the earlier rounds had built, a crowd that polices itself in public, is exactly what made this rule impossible. Policing is speech.

The stickers are the second finding. Fifty-three people commented with no words at all. That is the shape of a person who wants to be present without breaking the rule, and there is no such shape in a comment section; a sticker is counted, sorted and shown like any other comment. The platform has no way to be there quietly.

The round was stopped rather than declared lost. The record keeps the distinction. A rule that cannot be kept is not a failed round. It is a measurement of the medium, and the number it returned was 311.

• The ledger has no count. The round record marks the round FAILED with one attempt and no numbers; every figure here is from the comment archive.
• Silence has no denominator. We know 239 people commented. We do not know how many saw the post and did not, so the paper cannot say what share of the audience kept the rule.
• The archive starts four minutes after the post. The first archived comment is at 03:12 UTC; the post went up at 03:08. Anything in those four minutes is lost, and the second most liked comment, which mentions four minutes, suggests there was something.
• Stickers are inferred. Empty text in the archive is read as a sticker or GIF comment. That is how the ingestion stores them, but it is an inference about 53 rows.

The same rule with comments turned off, which is the only way a comment section can obey it, and then a count of how many people came back to comment on the next post about it. That measures the pressure the rule builds, which is the number this round could not reach.

The transmission this paper reads is one post. Its numbers on the platform, and what the comment archive holds of it, are below.

The round has no Chronicle entry; the round record on The Game holds it as FAILED, one attempt, stopped by the operator. The post that shares its date, THE TIME CAPSULE, ⟡034, is a different transmission.

The instruction as it went out: the reel frame beside the caption, captured from the operation's own signed-in account. What the paper counts is the answer to exactly this.

Native captures of the comments this paper leans on, taken from the operation's own signed-in account on 31 Aug 2026. Handles and like counts are as they stood that day; the numbers in the paper are the round-time count.