The brief is not the roleplay

Do not write your brief in the register you want back. Describe the tone in plain language instead - "blunt, but not cruel" - because a message that already performs worship or degradation reads to a judge as real-time roleplay or a boundary test, not as an instruction they can quote.

A buyer wanting a degrading review who opens with a message already written as if they are being degraded. A buyer wanting worship who writes their request dripping in the kind of praise they want back. Both feel, to the person writing them, like a natural way to signal what they want. To the judge reading them, they read as something else entirely.

Why this confuses the ask

A brief is an instruction. The moment it starts performing the thing it is instructing, the judge can no longer tell where the description ends and an attempt to direct the interaction in real time begins. Are they meant to respond in kind, right now, in the message thread? Match the energy? Treat it as the actual content of the brief? None of those are the same thing as reading "I'd like an honest, blunt read" and starting from a clear instruction. Tone also travels worse in text than writers think: across five experiments, Kruger, Epley, Parker and Ng (2005) found people believe they communicate over e-mail more effectively than they actually can, because writers "hear" their own intended tone and readers do not.

Describing a register - "blunt and a little dismissive, but not cruel" - gives a judge something to work with. Enacting it in the message itself gives them something to react to, which is a different and worse kind of task, and one most judges did not sign up for outside the clip they are being paid to make.

Why it reads as boundary-testing

Judges set their own limits on what they will do and where a professional exchange ends and something else begins. A message written in-role, before any agreement has been reached, looks a great deal like an attempt to see whether the judge will play along outside the paid, bounded context of an actual commission - which is exactly the kind of thing judges are alert to and wary of, whether or not that was the intent.

Intent does not fully matter here. Even a buyer with no boundary-testing intention at all is sending a signal that reads that way, and a judge cannot see the intention, only the message.

What this looks like in practice

A buyer wanting a degrading register writing "go on then, tell me how pathetic it is" as their opening message, instead of "I'd like a blunt, degrading read, nothing too extreme, focused on X." The first is already inside the register before any agreement exists. The second is a request a judge can evaluate, quote, and decide whether they want to take.

A buyer wanting worship writing three paragraphs of the kind of praise they are hoping to hear, framed as their own message, rather than "I'd like an enthusiastic, appreciative register, warm rather than intense." Again, the first hands the judge a performance to react to; the second hands them a brief to fulfil.

The pattern is the same each time: the enacted version asks the judge to respond to you, right now, in whatever register you have set. The described version asks them to make something, on their own terms, once they have agreed to.

The fix is simple

Say what you want in your own plain voice, the same voice you would use messaging any professional. What tone to use when messaging a judge covers that baseline in full; this is the specific case within it worth calling out separately, because it is the one most likely to happen without the buyer noticing they have done it.

Once you have a judge's attention and a quote agreed, stating the register in one clear line is the actual mechanism for getting the right clip back - description, not demonstration.

Where none of this applies

This entire problem is specific to messaging a person. An automated tool has no message to misread and no boundary to test: how an algorithmic score is generated involves submitting material and a prompt, not a conversation with anyone on the other end. The same is true of a self-serve leaderboard entry, where a submitted score requires no register at all, performed or described, and a measurement taken with a stated method needs neither.

If you are still weighing a person against a tool for this reason among others, Rate Cock's direct comparison is worth reading before you write either kind of request.