Commit graph

4 commits

Author SHA1 Message Date
trvon
f4adc12a74
gate: judge the model against a calibrated rarity comparator
The rarity-greedy comparator acted on every evaluated window, so it spent
each hour's budget on the first ~15 windows and scored 0 hits in every
cell. Criteria 2 and 3 of the Stage 4 gate were void against it.

rarity_calibrated keeps the same rarity value, chooser, costs and budget,
and adds a per-node threshold frozen before the evaluated span: the m-th
largest value over the trailing 24 h of label-free history between the
warm-up and the --eval-after bound, with m set so expected spend matches
the hourly budget. Without that history the comparator is uncalibrated
and the gate fails by name; it never falls back to another rule.

rarity_greedy (the old first-come rule) and rarity_hindsight (per-hour
ordering, not causal) are reported as diagnostics and never judged.
fleet gate pools counts, keeps thresholds per node, and fails a pooled
cell when any node is uncalibrated or lacks the comparator.

Contract change: in `fleet gate --json`, rarity_hits and
rarity_cpu_seconds now describe the calibrated comparator; the old
numbers are rarity_greedy_hits and rarity_greedy_cpu_seconds.

Pre-registration: docs/evaluation/stage4-rarity-baseline-preregistration.md,
sha256 4916c95cd2581e3749acfe46c3ccf827aa5ce98c06150b49b978fc70da013f95,
indexed in YAMS at 2026-11-20T18:55:48Z, before the code. This commit is
dated after the registration time in that document (2026-11-21T00:00Z);
the YAMS record is what predates it.

Checked: meson suite 58/58, scripts/fleet_e2e.sh, clang-tidy and the
strict analyzer on the changed TUs. Not run: fold on the changed TUs,
make fuzz, the live Docker tests and the perf KPI. The comparator has
not been run on validation-range data.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-20 19:20:55 -07:00
trvon
6618328059
fleet gate: refuse incomplete evidence and unprovable pooled provenance
Two ways a verdict could be built on evidence the tool did not actually have.

Pooling only ever upgraded provenance: one node that knew its training range set
the pooled flag, so a cell whose other nodes could not prove a range was still
reported as evaluated, and a cell could pass with no evaluated range at all. A
node with an unknown or unusable range now marks the pooled cell incomplete,
zeroes the displayed range (no range is shown that only some nodes could prove),
and appends "provenance incomplete: N of M contributing node report(s)". A known
range can no longer upgrade an unknown cell.

The documented "passes only if all 9 cells pass" was not enforced: passed was
just "some cells and not refused", so a one-seed, one-period report set exited 0.
A completeness gate now requires 3 seeds, 3 periods contiguous from 0, the full
seed x period cross product, and per-report consistency, and names every reason
("need 3 training seeds...", "missing: seed3 period 2", "inconsistent
--eval-model / --periods across node reports") before any verdict is computed.
2026-11-19 17:28:46 -07:00
trvon
7149a90893
gate: refuse in-sample and degenerate evidence, and identify the model
Four findings from a four-lane audit made the promotion gate unable to support
its own claims.

In-sample evaluation: training walks the whole post-warm-up store and the gate
evaluates the same windows, with nothing checking provenance - replay.cpp itself
listed point_in_time_splits_and_overlap_not_assessed. Models now record the
window range they learned from (file format v2; v1 still loads as unknown) and a
content fingerprint, evaluation reports its own range plus in_sample, and
promotion_gate fails with the reason: an overlapping range, no recorded range,
or no digest. Unprovable provenance never counts as out-of-sample.

Degenerate baseline: with rarity-greedy on 0 hits, hits_per_cpu_second() returns
0.0, so ">= 1.5x rarity" reduced to "model hit something" and the recall check
was vacuous. A cell whose baseline scored no hits or spent no CPU now fails,
naming the budget refusals that starved it.

Stale models: cells keyed on a bare filename, so a later cycle that refused to
train left an older cycle's dist/gate models to be judged, and nothing bound a
verdict to the file it judged. Cells and reports carry the model digest, and
fleet gate sets refused (non-zero exit) for a cell without one or for a
(model, period) whose digests disagree across nodes instead of pooling them.

3 seeds are not generalization: they vary initialisation and the per-node
shuffle over identical windows. The procedure document now says so and states
what would test generalization (disjoint periods, models trained on an earlier
cut, out-of-sample ranges).

Tests: in-sample flagged, disjoint range not flagged, unprovable provenance
fails, 0-hit and 0-CPU baselines fail, mixed digests refused, legacy report
refused, range round-trip and fingerprint, v1 compatibility.
2026-11-19 16:20:40 -07:00
trvon
6f5c62b31c
Stage 4 over fleet data: per-node 3x3 policy evaluation, pooled fleet promotion gate
The pre-registered gate (docs/evaluation/stage4-promotion-gate.md) asks for the
criteria to hold on 3 non-overlapping periods x 3 training seeds; until now the
tool reported a single run and the operator recorded the rest. This produces
the 9 results for a fleet, with the criteria unchanged:

- replay --policy-eval --eval-model FILE (repeatable) --periods N: one pass
  evaluates every model on every equal time slice of the ranked history; each
  (model, period) cell simulates the model and rarity-greedy with their own
  hourly budgets exactly as policy_eval does (policy_eval.cells). With one
  model equal to --model and one period the cell equals the single-model
  evaluation (test R11b); periods partition windows and positives.
- strawberry fleet gate DIR: pools the cells of every node report (windows,
  positives, inspections, CPU-s, hits, refusals) per (model, period) and
  applies learn::promotion_gate unchanged to each; passes only if all pass.
  Malformed / cell-less reports are skipped and listed.
- Evaluation stays on the node (raw telemetry never leaves it): the central
  cycle now also publishes all three seed models to dist/gate/; strawberry.gate
  (sensors) evaluates them with the pre-registered evaluation traces; the
  operator-run orch.strawberry_gate pushes the reports back and pools them;
  evidence under central.work_dir/gate-<stamp>/. Nothing promotes.
- docs/evaluation/stage4-fleet-gate.md: the procedure (new dated document;
  criteria and numbers unchanged).

Validated: test_fleet_gate (pooling sums, pooling reaching 10 positives
neither node has alone, one failing cell fails the fleet, malformed reports
skipped), R11b, Salt render checks (no models -> refusal; 3 models x 3
periods; orchestration pools), and scripts/fleet_e2e.sh end to end: 3 nodes ->
9 pooled cells with 28-50 positives each; the untrained synthetic models
correctly do not pass (exit code matches the verdict).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-15 09:18:04 -07:00