R&D / PERSUASION · DOC CLASS — COUNTER-PATTERN LAB

The Counter-Pattern Lab

The Lab closes the loop between the Cognitive-Bias Bank and the field. We take a documented technique apart, design its inverse, bench-test it against the real pattern, trial it on live surfaces, and channel the proven counter-patterns into the public tools and the field. The goal is not cleverness — it is counter-patterns that hold up against the manipulation we actually face.

Doc Control QH-RND-0005 Revision Rev A Effective 2091-01-01 Owner R&D / Persuasion Review Cycle Annual Classification INTERNAL // QUANTUM-ZONE-TASK-FORCE EYES-ONLY

The Lab Program

⊕Design

Counter-Pattern Design

Controlled production of counter-patterns from the techniques in the Index. We isolate the mechanism a dark pattern exploits and design the smallest, most reversible inverse that holds it — turning one studied manipulation into a working antidote.

⊹Validate

Validation & Scoring

Every counter-pattern is scored against measurable, hard targets — reversibility (the user can always undo it), effectiveness (it actually holds the technique), and line-safety (it never crosses the "Lines We Won't Cross" doctrine). Counters are scored, not assumed.

🚀Deploy

Deployment Pipeline

Proven counter-patterns graduate from the bench to deployment. The best feed the public Unranked tools and the Detox & Recovery program, which puts them in front of real users in a later operational phase.

Closed loop. Techniques in, counter-patterns out. The Bias Bank preserves the mechanisms; the Lab inverts and tests them; deployment returns the antidote to the field. No counter-pattern ships without a validation record on file.

Counter-Pattern Pipeline

Every counter-pattern we ship passes through the same five-stage flow. Mechanism is drawn from the Cognitive-Bias Bank at specify and, for the counters cleared for release, exits into the Detox & Recovery program — which receives Lab output in a later phase rather than at first build.

Stage	Step	What happens	Gate
1	Specify	The target technique's exploited mechanism is isolated from the Index (sourced via the Cognitive-Bias Bank) and the counter-pattern's requirements are written.	Mechanism verified
2	Prototype	The lab builds candidate counter-patterns — the smallest reversible inverse that could hold the technique — over roughly a sprint.	Prototype builds
3	Bench-test	Each candidate is run against a captured instance of the real technique on an instrumented surface, and its effect is measured.	Holds the technique
4	Field-trial	Survivors are trialed on live surfaces and scored against the targets — reversibility, effectiveness, line-safety, relapse impact.	Metrics ≥ line target
5	Ship / Deploy	Cleared counter-patterns go to the public tools, partner integrations, or — for field-grade lines — into the deployment pipeline.	Validation record filed

Active Counter-Pattern Lines

Six counter-pattern lines are under active development this season. Full lineage and revision history for each is held in the Index.

Line	Targets	Validated for	Rev	Status
`CP-FRICTION-01`	Engagement & compulsion	Reversibility · effectiveness	R7	● Stable
`CP-EXIT-04`	Obstruction & sludge	One-click exit · line-safety	R5	● Stable
`CP-CONSENT-02`	Consent theater	Re-grounding · revocability	R4	◐ Validating
`CP-CALM-03`	Recommendation & feed	De-escalation · line-safety	R3	◷ Trial
`CP-JAMMER-01`	Targeting & segmentation	Profile noise · precision-loss	R2	◐ Validating
`CP-DEFAULT-09`	Sneaking & defaults	Default reset · reversibility	R6	✦ Ship-grade

Validation Metrics

Counter-patterns are advanced, held, or cut on hard numbers. The current targets and live readings are mirrored from the Metrics board and cross-checked against the Harm & Wellbeing Index.

Reversibility

Reversibility rate

Share of deployments the user can fully undo, and does undo without penalty when they choose — the front-line guarantee that a counter-pattern never becomes a hook of its own.

Target ≥ 99%Lab avg 99.4%

Relapse

Relapse impact

Change in user relapse rate after a counter-pattern is deployed, sampled monthly. Low and stable is the signal that the counter holds without leaving a worse hole behind.

Target < 2.0%Action ≥ 3.0%

Retention

Recovery retention

Share of users still helped — still off the loop — at follow-up. The single toughest test of whether a counter-pattern actually changed anything.

Target ≥ 85%Last cycle 88%

Effect

Effectiveness index

Composite of how reliably the counter-pattern holds its target technique against live deployments, drawn from Instrumentation Mesh telemetry.

Target ≥ 1.10Lab avg 1.04

Doctrine & ethics. The Lab at THE RESET serves the user, not the metric. We do not build counter-patterns that raise effectiveness at the cost of reversibility or that steer toward any belief, we retire a line the moment it crosses the doctrine, and every counter retains a reversibility floor so a fix never becomes a new trap. Every line is reviewed annually against the "Lines We Won't Cross" charter.

Restricted lines. The PL-4 Restricted counters — Generative Counter-Voice and the Sentinel Antidote tracked in Restricted Techniques — are not built in the open program. They fall under a separate, Steward-authorized restricted protocol with logged two-person handling and never enter the open design or deployment flow.