CB-WP-0050 T02: F31 falsified — the module is unreachable, then real

GateAttackPolicy ranks ATTACK above GROUND at the stress gate and
delegates everything else. A PROBE, not a better bot: it wins 12/100
where greedy wins 60/100 under the same module, and the panel's banner
says so, because a column that looks like a policy comparison will be
read as one.

(a) THE MODULE IS UNREACHABLE IN THE PRINTED GAME. Selected alone it
still never fires: peak Stress never exceeds 2, so the gate never bites
and no ATTACK is ever gated. A module can be unreachable because of an
aspect it does not name -- attack_relief needs a problem_stress module
before it can act at all, and nothing in its own declaration says so.

(b) ONCE REACHABLE IT IS REAL. Holding the probe fixed and varying only
the module: group wins 12->21, 19->28, 50->57 at 3/4/6p, and DARVO arms
299->221, 294->213, 334->164. Under flat pressure the same shape appears
in DARVO alone: h1 arms 200 against flat_any_open's 300/400/600.

Stated as the conditional it is: GIVEN seats that attack at the gate.
Not advice to attack, and not evidence against F17.

AND IT CHANGES WHAT WE TOLD GROUND-GAME. CB-EV-0030 rejected H1 with
policies that never attacked, so H1-B was inert for every game behind
that verdict. The rejection stands on flat pressure alone -- wins are 0
either way -- but "H1-B does nothing" was never established and is now
known to be false. We owe them that correction.

Sensitivity: vary only the ATTACK rank. At 10 (greedy) and +55
(module-aware) the module never fires; at 110 it fires in every game.
Nothing about the module changed -- only whether any seat gave it a turn.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

workplans/CB-WP-0050-every-configuration-faceted-by-aspect.md

@@ -84,16 +84,60 @@ The numbers in such a row are real for the modules that did fire. Saying
 which one did not is what stops the row being read as a verdict on all of
 them.
 
+## Task: F31's falsifier
+
+```task
+id: CB-WP-0050-T02
+status: done
+priority: high
+```
+
+`GateAttackPolicy` ranks ATTACK above GROUND at the stress gate and
+delegates everything else. **A probe, not a better bot** — its win rate
+is not a recommendation, and the panel's banner says so, because a
+column that looks like a policy comparison will be read as one.
+
+**Controls:**
+- **it actually attacks, and at the gate** — asserted on real games, not
+  on the ranking: a rank of 110 proves an intention where the panel needs
+  an occurrence;
+- **blind** (ADR-0023);
+- **the probe's row decides "never fired"**, not the others — a module is
+  unreachable only if the seat built to reach it could not reach it
+  either.
+
+**Done 2026-08-08.** F31 answered, in two parts.
+
+**(a) The module is unreachable in the printed game.** Selected *alone*
+it still never fires: peak Stress never exceeds 2, so the gate never
+bites and no ATTACK is ever gated. **A module can be unreachable because
+of an aspect it does not name** — `attack_relief` needs a
+`problem_stress` module before it can act at all, and nothing in its own
+declaration says so.
+
+**(b) Once reachable it is real.** Holding the probe fixed and varying
+only the module: group wins +9 / +9 / +7 at 3/4/6p, and DARVO arming down
+by a quarter to a half.
+
+Stated as the conditional it is: *given seats that attack at the gate*.
+The probe wins 12/100 where greedy wins 60/100 under the same module, so
+this is not advice to attack and not evidence against F17.
+
+**And it changes what we told ground-game about H1.** CB-EV-0030 rejected
+H1 with policies that never attacked, so H1-B was inert for every game
+behind that verdict. The rejection stands on flat pressure alone — wins
+are 0 either way — but *"H1-B does nothing"* was never established and is
+now known to be false.
+
 ## Not done here
 
 - **The panel measures SHARED GROUND only.** Mode × module is a
   three-dimensional sweep and nothing yet says the modules behave the
   same under the competitive modes — where, per CB-EV-0033, *who wins*
   moves even when group success does not.
-- **F31 has no falsifier running.** A seat that attacks at the gate would
-  settle whether `attack_relief.self_soothe_ge4` is real or merely
-  unreachable; until one exists, no claim about that module is supported
-  in either direction.
+- **ground-game has not been told about F31.** It bears on a verdict we
+  already reported to them (H1's rejection), and a correction we owe is
+  worth more than one we volunteer.
 - **The scope term is inert at round-one positions** (CB-WP-0049 T04), so
   `problem_stress.scoped`'s measured effect is driven by later rounds.
   The panel does not report per-round effect, and that is where the