Skill Families and How They Grow

Authors rarely publish one skill. They publish a family, and the family's shape tells you more about the author's process than any single score.

Grouping skills by the first segment of their slug reveals 37 families of three or more skills, covering 214 of 591 evaluated entries:

Families score better than singletons

The ten largest families average between 6.97 and 9.14, and most sit above the directory average of 8.53. Authors who build a family tend to have a system — shared conventions, a consistent section structure, a common way of describing triggers.

baoyu (21 skills, 9.11) and linkedin (13, 9.14) are the clearest examples: every member follows the same template, and the template scores well. alphaear (9, 6.97) is the exception, and it is the family that supplies five of the eighteen skills below 7.0.

The spread inside a family is the interesting number

A family average hides its own range. Within the video-related families, scores run from 5.7 to 9.7 — a 4.0-point spread under one naming convention. The baoyu family is tighter, but still spans roughly a point.

That spread is what you would expect from a system applied unevenly: the early skills in a family are often the strongest, because they were the reason the family existed, and the later additions are variations that did not need to exist.

What this means for readers

A family name is not a quality guarantee. It tells you the skills share conventions, which is useful for consistency, not for picking one. Read the individual score.

The largest family is not the best. baoyu and linkedin are close, but threejs at 12 skills averages 8.05 — below several smaller families.

What this means for authors

Building a family is a real advantage: shared conventions raise the structural floor, and a reader who likes one skill is likely to try the next. The risk is the tail — the 15th skill in a family is usually weaker than the first, and publishing it dilutes the average without adding much.

The data suggests a rule: grow a family only when the new skill would score at or above the family average. By that test, several of the families in this directory should be smaller than they are.