Art of Posting - Lesson 9
Comment Anatomy
The room, the stance and the craft: what an outbound comment earns, in strict order of importance
Everyone is told to comment. Almost nobody is told what a comment is made of. We fitted what outbound comments from our index earned, holding the size of the post they landed on constant.
The anatomy has three layers and they are not equal. The room you comment in, the stance you take, then the craft in the sentence. The gap between the first and the rest is not close.
One note before you use any of it. This is observational, measured on comments people chose to write, so it describes association rather than cause. Reactions are also only the currency visible from outside.
The Room Is the Multiplier
The popularity of the post you comment under is associated with what your comment earns four to ten times more strongly than any feature of the writing itself. Standardised, that is +0.40 for the room against +0.03 to +0.09 for every craft feature that survived the fit.
So most comments earn almost nothing, however well they are written. The median outbound comment earns about two reactions. The upper tail comes from writing well in a room that already has people standing in it.
Read the room before you read your own sentence. Pick the room first.
Stance Changes Your Odds, Not Your Average
We labelled every comment by the stance it took: echo praise, added information, personal experience, respectful disagreement. All four came back with the same median, about two reactions.
What stance moves is the tail. On a subset of 2,210 comments, 17.2% of respectful disagreements and 16.4% of personal-experience comments cleared ten reactions. Praise cleared it 8.7% of the time, adding information 9.1%. The disagreement cell is 93 comments, so hold it loosely.
Disagreement that earns is a contribution, not a correction: concede the part that is right, name where your own work sees it differently, and give the reason. Adding information ties with praise on reactions, and it almost certainly pays in replies and relationships we cannot see. Stance buys you the odds of a comment that travels, and nothing else.
Three Levers, in One Breath
Once the room and the stance are settled, three writing features survived the popularity control. Development across sentences earned +0.073, a first-person stake +0.054, one specific number +0.037. A comment that makes a point and then does something with it.
The shape is one developed paragraph, not a structure. The specialists who genuinely practise this sit between 125 and 154 characters, which renders as one or two lines. Nobody we measured writes the bulleted, sub-headed comment.
The corpus median sits far below that, and the close below reads it live. The gap between the two is the whole craft argument.
Three Things Our Own Data Killed
Opening with the author's name came back negative at -0.054. The habit is so common it now reads as a formality.
Ending on a question sits at -0.025, and the heaviest practitioner in the index uses questions in 0.3% of his comments. Statements that add sit above invitations to reply.
Writing longer does not survive the control at all: it goes slightly negative, at -0.040. We had published a length gradient, and once we held the size of the target post constant most of it turned out to be placement. We retired it, and the craft story it left behind is developed, specific and personal. It was never long.
Do This Before the Next Lesson
The room audit, ten minutes. Take your last ten outbound comments and write beside each one the reaction count of the post it sat under. You are scoring your choice of room, not your writing.
The stance tally, ten minutes. Label your last twenty against the four stances and work out your own shares. Then look for whether you have ever once taken the stance that changes your odds.
The one-breath rewrite, fifteen minutes. Pick three comments from last week that were agreement and nothing else, and rewrite each as one paragraph carrying one first-person detail from your own work and one specific number. Do not post them.