DESIGN.md

@@ -2415,7 +2415,7 @@ gated on the 9o5 landscape-flattening risk (§13 / `homemaker-py-xi7`) and its o
 Tests: `tests/test_collapse_global.py` ×6 (demand-set relabel, level hard constraint, c/o/s
 exclusion, no-op safety, keep-better/unmerged); 267 pass.
 
-## 18. Graded circulation-connectivity signal (`homemaker-py-qi6`) — DONE (negative)
+## 18. Graded circulation-connectivity signal (`homemaker-py-qi6`) — in progress
 
 **Motivation — the binary fail is flat.** After the §17 collapse, the residual fails on the
 harbor-house set are dominated by `level N not connected` (2 of the best layout's 12; also on
@@ -2458,39 +2458,13 @@ the gradient toward connected topologies.
 (env `HOMEMAKER_CONN_GRADE`, default OFF); the grade is read off the optimised tree, one extra
 native eval per child.
 
-**Build.** Signal, fitness wiring, CLI, and 9 tests landed (`tests/test_conn_grade.py`: pure-graph
+**Status / next.** Signal, fitness wiring, CLI, and 9 tests landed (`tests/test_conn_grade.py`:
-fraction contract, non-circ cells ignored, monotone under (dis)connection, and the score/fail-
+pure-graph fraction contract, non-circ cells ignored, monotone under (dis)connection, and the
-count-invariance of the flag). 276 tests pass.
+score/fail-count-invariance of the flag). The A/B — does the gradient actually pull evolve runs
-
+toward connected circulation and clear `not connected` fails — needs full-budget runs and is
-**A/B verdict (measured, 2026-07-22, qpk protocol, `experiments/run_qi6_ab.sh`) — NEGATIVE.**
+pending (short 60-eval smoke run confirms the plumbing only). If the graded key alone is
-Equal-budget `conn_grade` ON vs OFF, both arms finished with the standard finish-time `--collapse`
+insufficient, the follow-on is an insert/relocate-circulation mutation operator (mechanism (a),
-(94g), 4 workers, canonical `homemaker-fitness` re-score for the `.fails` breakdown:
+still `homemaker-py-qi6`) that now has a gradient to climb. 276 tests pass.
-
-- **harbor-house** (`init.dom`, budget 2500, seeds 1–3): **byte-identical output** in every seed
-  (dom, fail list, fitness all diff-clean ON vs OFF) — the secondary comparator key never fired,
-  i.e. the search trajectory never actually hit a tie at fail-count that the grade could break.
-  This is the programme §18 was motivated on (2 of 15 fails on the best layout are `not
-  connected`), and the signal moved nothing.
-- **programme-house** (`init.dom`, budget 3000, seeds 1–5): 3/5 seeds tie exactly (byte-identical
-  `.fails`); seeds 1 and 2 diverge to a **different topology** with one fewer total fail (8→7
-  each) — but the diff is entirely adjacency/crinkliness/width/access/size fails, not
-  connectivity. In all 4 seed-arms across both programmes where a `not connected` fail was
-  actually present (harbor 1&3, programme 3&4), the fail is **unchanged** in both arms — zero
-  cases of the grade clearing one.
-- **Conclusion: the grade does not do what §18 designed it to do.** It occasionally perturbs
-  tie-breaking among equal-fail-count neighbours (programme-house seeds 1/2), which can
-  incidentally shift the total fail count, but that perturbation never targets circulation
-  connectivity specifically — consistent with a comparator key that is either too weak relative
-  to the primary `(-n_fails, fitness)` keys to steer topology choice, or whose grade values are
-  rarely distinct enough between the actual neighbours the search compares to break a tie in the
-  intended direction.
-
-**Status / next.** Kept default OFF (already was). Mechanism (b) (graded proximity as a tertiary
-key) is falsified by this A/B, not just unconfirmed — do not re-attempt without a different
-mechanism. The remaining candidate from the original issue is mechanism (a): an explicit
-insert/relocate-circulation mutation/repair operator, which does not depend on the search
-stumbling onto a fail-count tie to act. Not started; low priority per DISCOVERED-FROM epic
-`homemaker-py-94g`'s framing (fitness fidelity, not search capability).
 
 ## 19. Geometry/topology repair for shape-intrinsic fails (`homemaker-py-7fm`) — DONE (negative)
 

experiments/run_qi6_ab.sh

@@ -1,60 +0,0 @@
-#!/usr/bin/env bash
-# qi6 A/B (DESIGN.md §18): does the graded circulation-connectivity signal
-# actually pull full-budget evolve runs toward connected circulation and clear
-# 'level N not connected' fails, or is the secondary-comparator gradient too
-# weak to matter (mirrors the qpk §20 protocol: equal-budget ON vs OFF, both
-# arms finished with the standard finish-time --collapse (94g) so the
-# comparison is apples-to-apples on the final collapsed score)?
-#
-# Authoritative metrics: total fail count and 'not connected' fail count, read
-# from the .fails file homemaker-fitness writes (evolve.py itself only prints
-# n_fails, not the fail list). Each run appends one TSV row so partial results
-# survive an interrupt.
-#
-# Usage: experiments/run_qi6_ab.sh
-set -u
-cd "$(dirname "$0")/.."
-
-WORKERS=4
-OUT=scratch/qi6_ab; mkdir -p "$OUT"
-TSV=scratch/qi6_ab_results.tsv
-[ -f "$TSV" ] || printf 'programme\tseed\tconn_grade\tbudget\tfails\tnot_connected\tfitness\telapsed_s\n' > "$TSV"
-
-run() {  # programme seed conn_grade(0|1) budget
-  local prog="$1" seed="$2" cg="$3" budget="$4"
-  local tag="cg${cg}"
-  local dom="$OUT/${prog}_${tag}_s${seed}.dom"
-  local log="$OUT/${prog}_${tag}_s${seed}.log"
-  local flag="--no-conn-grade"; [ "$cg" = 1 ] && flag="--conn-grade"
-  echo ">>> $prog seed=$seed conn_grade=$cg budget=$budget"
-  local t0; t0=$(date +%s)
-  homemaker-evolve "examples/$prog/init.dom" \
-    --budget "$budget" --workers "$WORKERS" --seed "$seed" \
-    $flag --output "$dom" > "$log" 2>&1
-  local t1; t1=$(date +%s)
-  local fitness fails notconn
-  fitness=$(sed -n 's/^best *: \([0-9.e+-]*\) .*/\1/p' "$log")
-  fails=$(sed -n 's/^best *: [0-9.e+-]* (\([0-9]*\) fails).*/\1/p' "$log")
-  # re-score with the canonical scorer to get the .fails breakdown (evolve.py
-  # only prints the count, not the list); dom lives outside examples/$prog so
-  # pass an absolute path while cd'd there for patterns.config resolution
-  ( cd "examples/$prog" && homemaker-fitness "$(realpath "../../$dom")" > /dev/null 2>&1 )
-  notconn=0
-  if [ -f "${dom}.fails" ]; then
-    notconn=$(grep -c 'not connected' "${dom}.fails")
-  fi
-  printf '%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s\n' \
-    "$prog" "$seed" "$cg" "$budget" "${fails:-ERR}" "$notconn" "${fitness:-ERR}" "$((t1-t0))" >> "$TSV"
-  echo "    -> ${fails:-ERR} fails (${notconn} not-connected), fitness=${fitness:-ERR}, $((t1-t0))s"
-}
-
-# harbor-house: budget 2500, seeds 1-3 (qpk protocol)
-for seed in 1 2 3; do run harbor-house "$seed" 0 2500; done
-for seed in 1 2 3; do run harbor-house "$seed" 1 2500; done
-
-# programme-house: budget 3000, seeds 1-5 (qpk protocol)
-for seed in 1 2 3 4 5; do run programme-house "$seed" 0 3000; done
-for seed in 1 2 3 4 5; do run programme-house "$seed" 1 3000; done
-
-echo "=== qi6 conn_grade A/B complete ==="
-column -t -s $'\t' "$TSV"