Definition #
Layer 3’s fifth stress-test tool in the Operator Decision Tree. Runs the aggregated read against the priors declared at Layer 0 (Declaration), asking whether the picture the operator has assembled is what the evidence actually shows or what the prior expected to find.
Family #
Child of [The Decision Architecture]. Sibling stress-test tool to [The Five Questions], [Structural Scale], [The Five Stakeholder Read], and The Decoy Option inside Layer 3 (Aggregation). Runs against the antagonist [Informal Instrumentation]. Cannot function without [Declaration] naming the defendant first.
Why Behind the Thinking #
The prosecution runs three moves in fixed order. First, where the evidence confirms the prior — note it, do not weight it, because confirmation is not proof. Second, where the evidence contradicts the prior — this is the signal, not a problem to explain away. Third, strip everything that is not evidence: the preference, the historical scorecard, the disruption lean, the feel. The three named defendants are confirmation bias (weighting historical success), [Golden Rule Bias] (substituting operator preference for market evidence), and [Disruption Bias] (overcorrecting toward novelty). The prosecution standard does not care which direction the bias points — it cares only whether the evidence was read honestly. Untrained reps compound bias because every surviving decision gets credited to the reps and every failure gets attributed to circumstance; trained reps become pattern-testing instead of pattern-recognizing, because the reps are prosecuted rather than trusted blind. Formalization does not diminish experience — it makes experience legible.
Pairs With #
[Declaration], [Informal Instrumentation], [Golden Rule Bias], [Disruption Bias], [The Operator Decision Tree], [The Five Stakeholder Read], [Structural Scale], [The Five Questions], [The Decoy Option], [Operator R&D]
Placement #
Perspective. Manuscript section 1.DA.7, Layer 3 of the Decision Architecture.