Ranking signals

Every list has a sort. The sort is a design decision made by people with their own goals, and it decides which tokens a stranger ever sees. Understanding a list means understanding what its sort selects for and what it therefore hides.

These notes take the sorting rules apart: what turnover measures, what a percentage-change column rewards, why recency lists behave the way they do, and why a narrative can travel through a feed that no numeric signal would have surfaced.

Scope Sorting rules, published filters and the observable behaviour of ranked lists
Method Quote the documentation where it exists; otherwise state what is visible and mark the gap
Refused Reverse-engineered algorithm claims presented as fact, and any promise a signal can be forced

3 field notes in this section

Sorting rules, published filters and the difference between a measurement a surface states openly and a mechanism people infer from watching it behave.

02Comparison

Attention is not volume

Two different measurements that people treat as one. What turnover counts, what attention would have to count, and the four cases where the two numbers pull apart.

Read the field note
03Analysis

Narrative and why it carries

A narrative is a compression format. Why some token stories travel through a group chat unchanged and others need a paragraph of explanation and therefore never move at all.

Read the field note

The other two halves

A signal only matters because of the surface it sorts and the people who react to the result. Both have their own section.

Surfaces

The screens, feeds and lists a token can appear on, what each one is built to do, and which of them a launch team can observe directly rather than guess at.

Open surfaces

Communities

Chats, holder bases and the people who keep talking after the first day. How groups form, what makes them durable, and where honest growth stops being possible.

Open communities