RatLabA Boar Research
Closed recordas of what’s still open →

Decision ·

Treat v0.2 as burnt (dev); a final number needs v0.3 written outside the Claude family

From the record

Context
Retrieval scorecard: coverage is the largest failure bucket (49 %); recall@5 0.333
Why
plaintext in private git history + 4 runs + a targeted fix
What followed
v0.3 spec written 10-07; items not written by 10-10
Reversible
partly
Status
kept
Alternatives considered
  • rewrite git history
  • keep using v0.2 as held-out

Linked facts · 2

  • Treat v0.2 as burnt (dev); a final number needs v0.3 written outside the Claude family — why: plaintext in private git history + 4 runs + a targeted fix — outcome: v0.3 spec written 10-07; items not written by 10-10Treat v0.2 as burnt (dev); a final number needs v0.3 written outside the Claude family — why: plaintext in private git history + 4 runs + a targeted fix — outcome: v0.3 spec written 10-07; items not written by 10-10Date2026-10-06Kinda record, not a measurement (set, status, judge, n, seeds: not recorded)LaterPartly superseded: v0.2 stays burnt (dev); the "needs a v0.3" path was dropped on 2026-10-18 and replaced by the v0.4 ruler set.Sourcenot in the public repositories yetOpen the data file
  • Treat v0.2 as burnt → dev; final number needs a v0.3 written outside the Claude family (why: plaintext in git history, 4 runs, Claude-written items)Treat v0.2 as burnt → dev; final number needs a v0.3 written outside the Claude family (why: plaintext in git history, 4 runs, Claude-written items)Date2026-10-06Kinda record, not a measurement (set, status, judge, n, seeds: not recorded)Caveatcitation fix: D5 is l.57LaterPartly superseded: v0.2 stays burnt (dev); the "needs a v0.3" path was dropped on 2026-10-18 and replaced by the v0.4 ruler set.Sourceresearch record reports/SET-STATUS.md:14Open the data file