| name | learning-tracker |
| description | Use this to trace lessons across months - learned once, learned slowly, forgotten, or never turned into a system. |
Learning Tracker
The Big Idea
A lesson is not learned when it is stated. It is learned when behavior
changes and stays changed. Track corrections longitudinally: lesson
appears in month N, then check months N+1, N+2, N+3 for behavior.
The Six Verdicts
For every recurring lesson assign one:
| Verdict | Meaning |
|---|
| LEARNED ONCE | stated once, behavior changed permanently |
| LEARNED SLOWLY | needed 3+ corrections before sticking |
| KNOWN BUT NOT APPLIED | explained fluently, violated anyway |
| REPEATEDLY FORGOTTEN | reappears every few weeks like weather |
| MECHANICALLY SOLVED | converted to a test, hook, script, or lint |
| NO LONGER RELEVANT | the context disappeared |
What to Mine
- concepts re-explained across months (the tutor tax)
- mistakes corrected more than twice
- principles that transferred successfully between projects
- knowledge that failed to transfer between projects
- lessons that should become external tooling instead of memory
The Critical Split
Separate KNOWLEDGE GAP from EXECUTION GAP. They need opposite fixes.
- Knowledge gap -> explanation, example, reference card.
- Execution gap -> friction change: test, hook, checklist, default.
Explaining harder does nothing; the subject already knows.
When It Backfires
- Counting restated lessons as retained. Restating is rehearsal, not
retention. Look for applied behavior.
- Blaming memory for what is really workflow design. If a lesson must be
remembered at 2am under deadline, it will be forgotten. Move it into
tooling.
One-Line Memory
A lesson corrected twice belongs in a script, not a memory.