diff --git a/docs/claude/stage-decision-record-S2-2026-11-15.md b/docs/claude/stage-decision-record-S2-2026-11-15.md index d53b1c813..aedea53ec 100644 --- a/docs/claude/stage-decision-record-S2-2026-11-15.md +++ b/docs/claude/stage-decision-record-S2-2026-11-15.md @@ -58,11 +58,11 @@ S3-A runs as approved. In parallel, the route-A falsification tests (O-vol oracl - A stronger distance weighting f = D_w^2 (20.2 %, against 9.1 % for D_w^1), for this graph approximation and clock only. - Volume-consistent remapping of R2: no net gain (-0.09 %, CI includes 0). -**Measured references (oracles; not deployable, not floors):** true full-domain order plus training clock gives 1.48 %. The LOG time map is within 1 pp of linear for 411/624 parts, with a tail up to 34 pp. Part 996 shows that the last-filled timing can sit in the post-switch-over phase. +**Measured references (oracles; not deployable, not floors):** true full-domain order plus training clock gives 1.48 %. The LOG time map is within 1 pp of linear for 411/624 parts, with a tail up to 34 pp. Part 996 shows only that whole-field scoring is sensitive to the end point chosen; it does not attribute the late labels to the post-switch-over controller. (Narrowed 2026-11-15 after MeltAstra's correction. In Slack I twice wrote 'what is left is mostly ordering' and 'last-filled timing is governed by the post-switch-over phase'; neither follows from the evidence. The R2-vs-oracle difference cannot be split into an ordering share, and part 996 is an end-point sensitivity case only.) **Untested:** O1 (region or branch organisation with interfaces); flow-competition (branch-split) ordering; continuous eikonal versus Dijkstra; process, gate and material variation (impossible on public data); acceptance targets (asked through TRON, not yet answered); whether neural models do better with far more data. -**Next route: the filling order from flow competition, S3-B(a) then (b).** CPU only, no training without a separate contract. +**Next route: the filling order from flow competition, S3-B(a) then (b).** Rationale: the true-order oracle is low and the flow-competition mechanism is untested. This does not rest on a claim that R2's remaining error is mostly ordering. CPU only, no training without a separate contract. - (a) A graph filling prior (Mold's definition; conductance g_ij = dual-volume/L_ij^2 x D_w^2, fixed before the run). Scored like the naive batch against B0_DW1 and R2. - (b) Only if (a) passes: its order as a single extra R2 feature, using the six-region table against R2.