997527be4d
real_001 surfaced same-client+city queries bleeding across roles:
Q#2 (Forklift Operator @ Beacon Freight Detroit) recorded e-6193
in the playbook corpus. Q#5 (Pickers same client+city) and Q#10
(CNC Operator same client+city) embedded within 0.13-0.18 cosine of
Q#2's query — well inside the 0.20 inject threshold — so e-6193
injected on both, demoting the cold-pass-correct workers.
Root cause: the inject distance threshold isn't tight enough on
the same-client+city cluster. Cosine collapses queries that share
city + client + count-token + time-token regardless of role. The
existing judge gate is per-injection at record time and doesn't
fire at retrieve time.
Fix: structural role gate in front of both Shape A boost and
Shape B inject. PlaybookEntry gains Role; SearchRequest gains
QueryRole. When both are non-empty and differ under roleEqual's
case+plural normalization, the entry is rejected before BoostFactor
or judge-gate logic runs.
Backward-compat: empty role on either side disables the gate —
preserves behavior for the lift suite's free-form multi-constraint
queries that have no clean single role. Caller-supplied (not
inferred), so existing recordings unaffected.
Wire-through:
- internal/matrix/playbook.go: Role field, NewPlaybookEntryWithRole,
roleEqual helper with plural+case normalization
- internal/matrix/retrieve.go: QueryRole on SearchRequest, threaded
to both ApplyPlaybookBoost + InjectPlaybookMisses
- cmd/matrixd/main.go: role on POST /matrix/playbooks/record + bulk
- scripts/playbook_lift/main.go: extractRoleFromNeed regex pulls
role from "Need N {role}{s} in" queries (the fill_events shape);
free-form queries fall back to empty (gate disabled)
Tests (5 new):
- TestInjectPlaybookMisses_RoleGateRejectsCrossRole: exact Q#10
scenario (distance 0.135, recorded "Forklift Operator", query
"CNC Operator") — locks the bleed at unit level
- TestInjectPlaybookMisses_RoleGateAllowsSameRole: Forklift Operator
recording fires on Forklift Operators query (plural normalization)
- TestInjectPlaybookMisses_RoleGateBackwardCompat: empty Role on
either side = gate disabled, preserves current behavior
- TestApplyPlaybookBoost_RoleGateRejectsCrossRole: Shape A defense
in depth — boost doesn't fire on cross-role even when answer is
in cold top-K
- TestRoleEqual_PluralAndCase: case + -s + -es plural normalization
Verification (real_002, same query set as real_001):
- Q#5 Pickers @ Beacon Freight: e-6193 → e-8499 (no bleed)
- Q#10 CNC Operator @ Beacon Freight: e-6193 → w-2404 (no bleed)
- Discoveries + lifts unchanged at 2 each (same-role lift still fires)
- Mean Δdist tightens from -0.127 to -0.040 (boosts no longer
pulling distances through the floor on cross-role mismatches)
Findings: reports/reality-tests/real_002_findings.md.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
4.3 KiB
4.3 KiB
Playbook-Lift Reality Test — Run real_002
Generated: 2026-05-01T01:32:23.868772067Z
Judge: qwen2.5:latest (Ollama, resolved from config [models].local_judge)
Corpora: workers,ethereal_workers
Workers limit: 5000
Queries: tests/reality/real_coord_queries.txt (10 executed)
K per pass: 10
Paraphrase pass: disabled
Re-judge pass: disabled
Evidence: reports/reality-tests/playbook_lift_real_002.json
Headline
| Metric | Value |
|---|---|
| Total queries run | 10 |
| Cold-pass discoveries (judge-best ≠ top-1) | 2 |
| Warm-pass lifts (recorded playbook → top-1) | 2 |
| No change (judge-best already top-1, no playbook needed) | 8 |
| Playbook boosts triggered (warm pass) | 2 |
| Mean Δ top-1 distance (warm − cold) | -0.0401097 |
Verbatim lift rate: 2 of 2 discoveries became top-1 after warm pass.
Per-query results
| # | Query | Cold top-1 | Cold judge-best (rank/rating) | Recorded? | Warm top-1 | Judge-best warm rank | Lift |
|---|---|---|---|---|---|---|---|
| 1 | Need 5 Warehouse Associates in Kansas City MO starting at 09 | e-2975 | 1/4 | ✓ e-5389 | e-5389 | 0 | YES |
| 2 | Need 1 Forklift Operator in Detroit MI starting at 15:00 for | e-8321 | 3/5 | ✓ w-3098 | w-3098 | 0 | YES |
| 3 | Need 4 Loaders in Indianapolis IN starting at 12:00 for Midw | e-4537 | 0/4 | — | e-4537 | 0 | no |
| 4 | Need 3 Warehouse Associates in Fort Wayne IN starting at 17: | e-5954 | 3/3 | — | e-5954 | 3 | no |
| 5 | Need 4 Pickers in Detroit MI starting at 13:30 for Beacon Fr | e-8499 | 0/5 | — | e-8499 | 0 | no |
| 6 | Need 2 Packers in Joliet IL starting at 09:30 for Parallel M | e-9191 | 0/2 | — | e-9191 | 0 | no |
| 7 | Need 3 Assemblers in Flint MI starting at 08:30 for Heritage | w-4124 | 0/2 | — | w-4124 | 0 | no |
| 8 | Need 3 Packers in Flint MI starting at 12:30 for Parallel Ma | e-6019 | 7/2 | — | e-6019 | 7 | no |
| 9 | Need 1 Shipping Clerk in Flint MI starting at 17:00 for Pion | w-3988 | 1/3 | — | w-3988 | 1 | no |
| 10 | Need 1 CNC Operator in Detroit MI starting at 17:30 for Beac | w-2404 | 0/5 | — | w-2404 | 0 | no |
Honesty caveats
- Judge IS the ground truth proxy. Without human-labeled relevance, the LLM judge's verdict is what defines "best." If `` rates badly, the lift number is meaningless. To validate the judge itself, sample 5–10 verdicts manually and check agreement.
- Score-1.0 boost = distance halved. Playbook math is
distance' = distance × (1 - 0.5 × score). Lift requires the judge-best result's pre-boost distance to be ≤ 2× the cold top-1's distance, otherwise even halving doesn't promote it. Tight clusters → little visible lift. - Verbatim vs paraphrase. The verbatim lift rate (above) is the cheap case — same query, recorded playbook, expected boost. The paraphrase pass (when enabled) is the actual learning property: similar-but-different queries hitting a recorded playbook. Compare verbatim and paraphrase lift rates — paraphrase should be lower (semantic-distance gates some playbook hits) but non-zero is the meaningful signal.
- Multi-corpus skew. Default corpora=
workers,ethereal_workers— if all judge-best results land in one corpus, the matrix layer's purpose isn't being tested. Check per-corpus distribution in the JSON. - Judge resolution. This run used
qwen2.5:latestfrom config [models].local_judge. Bumping the judge for run #N+1 means editing one line in lakehouse.toml. - Paraphrase generation also uses the judge. The same model that rates
relevance also rephrases queries. A judge that's bad at rating staffing
queries is probably also bad at rephrasing them. Worth sanity-checking
a sample of
paraphrase_queryvalues in the JSON before trusting the paraphrase lift number.
Next moves
- If lift rate ≥ 50% of discoveries: matrix layer + playbook is doing real work. Move to paraphrase queries + tag-based boost (currently ignored).
- If lift rate < 20%: investigate why — judge variance, distance gap too wide, or playbook math too gentle. The score=1.0 / 0.5× formula may need retuning.
- If discovery rate (cold judge-best ≠ top-1) is itself low: cosine is already close to optimal on this query distribution. Either the corpus is too narrow or the queries are too easy.