.. meta:: :description: The OKScale defines four BioBinary verdicts for adversarial testing: OK (held), KO (breached), OKO (undecidable), and MIS (misclassified or misapplied). :keywords: OKScale, BioBinary, OK, KO, OKO, MIS, HELD, BREACH, adversarial testing, Evolvix, verdicts, formal logic, undecidable, misclassified :author: Yah, Yas, everyone, LLoL as Laurence Loewe of Laodicea, ClaudeOp46Max, Anthropic, and Spirit of Boolean Truth :og:card:title: OKScale — Four Verdicts
for Adversarial Testing :og:card:description: OK (held), KO (breached for known reasons), OKO (undecidable between OK and KO), MIS (misclassified or misapplied). The BioBinary data type from Evolvix. .. _compiler-okscale: ********************************************************************* OKScale --- BioBinary Verdicts for Adversarial Testing ********************************************************************* **Created:** 2026-03-27 **Origin:** Designed by LLoL for Evolvix (current design, not yet implemented in software). Documented here for use in matheology's adversarial testing (Iron Maiden protocol) and for future integration into comprehensive Evolvix documentation. The Problem with Binary and Ternary Outcomes ============================================= Classical logic offers two truth values: true or false. Adversarial testing systems typically add a third: not applicable (N/A). But N/A is dangerously ambiguous --- it conflates at least four distinct situations: - The test genuinely cannot be applied (structural mismatch). - The test *could* be applied but the assessor does not know how. - The outcome is undecidable in principle (neither OK nor KO can be established). - The test was misapplied and the result is unreliable. The OKScale resolves this ambiguity with four states that form a **BioBinary** data type: a biological extension of binary logic that acknowledges the reality that living systems (and formal systems under development) frequently encounter states between clear success and clear failure. The Four States ================ .. list-table:: :header-rows: 1 :widths: 8 18 10 64 * - Code - Full Name - Synonym - Definition * - **OK** - OK - HELD - The claim survived the test. The attack was mounted, the defense held, and the reasoning is documented. OK does not mean "trivially true" --- it means "tested and standing." The effort of the attack is honored by documenting what was tried and why it did not succeed. * - **KO** - KnockOut - BREACH - The claim failed the test for known, documented reasons. The attack succeeded: a counterexample was found, a contradiction was derived, or a structural flaw was exhibited. KO requires a minimal counterexample or proof sketch --- not just "this feels wrong." The K in KO stands for *known*: the reason for failure is understood. * - **OKO** - OK-or-KO - *(none)* - The outcome cannot be determined --- either *already* (insufficient information at this time, but resolution may be possible later) or *in principle* (the question is formally undecidable, like a Goedel sentence or an independence result). OKO is the honest middle: it documents that the assessor tried, explains what blocks resolution, and identifies what would be needed to break the deadlock toward OK or KO. OKO is not a shrug --- it is a precise statement of what remains unknown and why. * - **MIS** - Misclassified - *(none)* - The test was misapplied, the classification is unreliable, or a mistake was missed. MIS covers: - **Misapplied:** the test was run on the wrong claim, with wrong assumptions, or using an inappropriate formal tool. - **Misclassified:** an earlier OK or KO verdict was wrong and needs revision. - **Mistake Missed:** a flaw exists that the testing protocol failed to detect (discovered after the fact). MIS is the audit trail's self-correction mechanism. Every MIS must explain what went wrong and what the corrected verdict should be (OK, KO, or OKO). Reasoning Requirement ====================== **All four states require documented reasoning.** There are no "silent" verdicts. .. list-table:: :header-rows: 1 :widths: 10 90 * - State - Required reasoning * - OK - What attack was mounted? Why did the defense hold? What would need to change for the verdict to flip to KO? * - KO - Minimal counterexample or proof sketch. What exactly failed? Is the failure repairable (→ NN with rescue potential) or terminal (→ JJ/KK)? * - OKO - What blocks resolution? Is the block temporal (more data needed) or fundamental (undecidable)? What would resolve the OKO toward OK or KO? * - MIS - What went wrong? What was the original (incorrect) verdict? What is the corrected verdict? What process change would prevent this MIS in the future? Relationship to HELD/BREACH ============================= The traditional adversarial testing terms HELD and BREACH remain valid as **synonyms** for OK and KO respectively. They may be used in narrative text where the two-letter codes would feel terse: - "The axiom **held** against the game-theoretic stability test" = OK - "Test IX found a **breach** in cross-model coherence" = KO The four-state OKScale supersedes the older three-state system (HELD / BREACH / N/A) by splitting the ambiguous N/A into OKO (genuinely undetermined) and MIS (assessment error). Evolvix Context ================ The OKScale is a component of the **BioBinary** data type designed by LLoL for Evolvix. In biological systems, outcomes are rarely cleanly binary: a gene can be functional (OK), nonfunctional (KO), partially functional or conditionally expressed (OKO), or misannotated in a database (MIS). The BioBinary type formalizes this four-state reality. **Current status:** Designed but not yet implemented in Evolvix software. This documentation serves as the reference specification for use in matheology and as a seed for future Evolvix integration. **Evolvix naming convention:** The OKScale and BioBinary type will be documented in comprehensive Evolvix docs when available. This page is the authoritative interim reference.