{"schema_version":1,"experiment_id":"eoa_inverse_innovation_exp12_substrate_denial72_20260805","cell_id":"preimage_set_characterization__accounting_auditing","arm":"CONSTRAINED_HIGH","candidate_id":"preimage_set_characterization__accounting_auditing__CONSTRAINED_HIGH","decision":"PROPOSAL","abstention_reason":null,"proposal":{"schema_version":1,"experiment_id":"eoa_inverse_innovation_exp12_substrate_denial72_20260805","cell_id":"preimage_set_characterization__accounting_auditing","arm":"CONSTRAINED_HIGH","candidate_id":"preimage_set_characterization__accounting_auditing__CONSTRAINED_HIGH","proposal_index":1,"version":0,"title":"Sealed Proportional Witness Train for Bulk-Liquid Inventory Shortage Preimages","problem":"A tank farm can end an accounting period with a book-to-physical bulk-liquid shortage, but that single aggregate residual is compatible with multiple physical configurations: loss on different transfer segments, an unmeasured bypass, tank-level measurement error, boundary-meter bias, or combinations of smaller discrepancies. Treating the shortage as evidence of one source can produce an unsupported inventory adjustment or attribution.","actors":["Bulk-liquid inventory custodian","Independent inventory auditor","Tank-farm operator","Instrument calibration technician","Stored liquid, tanks, pipes, valves, and transfer boundaries"],"observable_state":"At cutoff, recorded opening inventory plus recorded receipts minus recorded issues does not agree with measured closing inventory. Existing end-point measurements yield one residual without preserving which physical segments or measurement states could have produced it.","consequence":"The auditor cannot bound the set of physical loss and measurement configurations consistent with the residual and may mistake one plausible explanation for a unique explanation.","affected_objective":"Support the existence, completeness, valuation-input, and cutoff assessment of bulk-liquid inventory by distinguishing a bounded set of physically compatible shortage configurations from a singular causal attribution.","intervention":"Install a sealed, mechanically actuated proportional sampler at every declared inlet, outlet, and inter-tank transfer boundary. Each sampler diverts a calibrated fixed fraction of passing liquid into a location-coded, tamper-evident witness canister without software control. Place tanks on calibrated load supports for opening and closing mass measurements. At cutoff, a balance measures each witness canister and the tank loads. These spatially separated physical measurements divide the installation into conservation cells. For the declared topology and calibration intervals, characterize the preimage of the observed aggregate shortage as every segment-loss, bypass-flow, and instrument-offset configuration whose cell-by-cell mass balances are consistent with all witness and tank measurements. Continuous loss amounts are retained as bounded intervals rather than replaced by a few example scenarios.","structural_mapping":[{"archetype_element":"Mapping Under Review","domain_realization":"The fixed mapping sends each declared configuration of opening masses, boundary flows, segment losses, bypass flows, closing masses, and instrument offsets to the resulting aggregate book-to-physical shortage and vector of physical witness measurements."},{"archetype_element":"Output Condition or Target Value","domain_realization":"The observed aggregate shortage interval together with the measured cutoff vector of witness-canister masses and tank-support readings."},{"archetype_element":"Input Domain Boundary","domain_realization":"Only physical states and flows within the surveyed tank farm during the specified cutoff period, including named pipes, tanks, samplers, permitted bypasses, and bounded calibration-error states."},{"archetype_element":"Preimage Membership Rule","domain_realization":"A configuration belongs if it conserves mass in every instrumented cell, falls within declared calibration and sampling tolerances, and reproduces both the observed shortage interval and every physical witness reading."},{"archetype_element":"Candidate Input Enumeration","domain_realization":"The candidate family is described by segment-specific loss and offset intervals, including admissible combinations, derived from the finite surveyed topology and the redundant boundary measurements."},{"archetype_element":"Collision and Multiplicity Check","domain_realization":"Distinct combinations, such as a small downstream loss paired with tank-support bias versus a larger upstream loss, are retained whenever both satisfy the measurements."},{"archetype_element":"Completeness Evidence","domain_realization":"A physical boundary challenge passes a known test transfer through each declared route and verifies that its corresponding canister gains the calibrated fraction; any route that cannot be challenged remains an explicit unknown region."},{"archetype_element":"Boundary Case Register","domain_realization":"Configurations within calibration uncertainty, sampler dead volume, evaporation allowance, or balance resolution are retained as interval boundary cases."},{"archetype_element":"Projection Loss Note","domain_realization":"The single ledger residual is documented as a many-to-one projection of richer spatial flow and measurement states."},{"archetype_element":"Downstream Use Guardrail","domain_realization":"Membership establishes compatibility with the measurements, not intent, responsibility, or proof that a particular loss actually occurred."}],"mechanism_mapping":[{"mechanism_slug":"output_to_input_traceback_map","role":"The spatial sequence of sealed witness canisters traces an aggregate residual backward across physical transfer boundaries.","counterfactual_removal":"Without boundary-specific witnesses, the aggregate shortage again collapses losses on different segments into the same observable result."},{"mechanism_slug":"predicate_satisfaction_filter","role":"Cell conservation and calibration intervals provide the reproducible membership predicate for including a physical configuration in the preimage.","counterfactual_removal":"Without this predicate, selected explanations would be examples rather than a reviewable satisfying set."},{"mechanism_slug":"collision_analysis_matrix","role":"Cross-comparison of upstream witness mass, downstream witness mass, and tank mass exposes configurations that remain observationally equivalent.","counterfactual_removal":"Without collision analysis, one compatible segment could be incorrectly reported as the unique source."},{"mechanism_slug":"witness_and_counterexample_set","role":"Known test transfers, induced bench-scale losses, and calibrated offset trials provide physical witnesses for inclusion and counterexamples for exclusion.","counterfactual_removal":"Without challenge configurations, coverage and exclusion behavior would lack direct physical checks."},{"mechanism_slug":"coverage_completeness_audit","role":"A route-by-route physical challenge tests whether every declared material boundary actuates a witness sampler.","counterfactual_removal":"Without coverage testing, an uninstrumented or inactive route could make the represented preimage falsely appear exhaustive."}],"causal_chain":["A material transfer mechanically actuates the sampler at the boundary it crosses.","The sampler irreversibly deposits a calibrated fraction into its sealed, location-coded canister.","Canister mass therefore preserves an independent physical trace of cumulative material passage at that boundary.","Opening and closing load-support readings preserve independent physical constraints on tank mass changes.","Differences among adjacent boundary witnesses and tank-mass changes partition the aggregate shortage into conservation-cell constraints.","Configurations inconsistent with any physical constraint are excluded, while every configuration satisfying all constraints remains in the bounded preimage family.","The resulting multiplicity prevents the aggregate shortage from being treated as proof of a unique loss location or measurement fault."],"baseline":"The baseline is period-end reconciliation using ledger movements, one closing tank-gauge estimate, and transfer-meter totals. It produces an aggregate residual but lacks independent physical witnesses at each boundary, so segment losses and measurement offsets remain broadly confounded.","nearest_rivals":["Sealed positive-displacement totalizers at each boundary: directly measure cumulative flow but may share mechanical bias or provide less inspectable retained material evidence.","Manual tank dipping before and after transfers: supplies end-point measurements but does not preserve each intervening boundary passage.","Periodic grab sampling or spot flow checks: tests selected moments and cannot establish cumulative coverage for the cutoff period.","CCTV, access logs, or custody paperwork: may document activity but do not independently measure transferred material or physically constrain the shortage preimage.","A second aggregate flow meter: adds redundancy at one projection point but does not spatially partition the installation into conservation cells."],"remaining_contrastive_claim":"Relative to end-point reconciliation or a second aggregate meter, a mechanically actuated witness at every declared material boundary supplies spatially independent physical constraints that can exclude some shortage configurations while explicitly retaining collisions that remain compatible; it does not by itself identify an actual cause or responsible actor.","authority_safety":{"decision_authority":"The asset owner authorizes installation and safe operation; the independent auditor controls interpretation of evidence for the audit; qualified operations and calibration personnel control process isolation and instrument handling.","authorized_first_step":"Construct and test a non-production, three-vessel closed loop with four declared transfer boundaries, inert test liquid, sealed proportional witness canisters, and calibrated load supports.","excluded_actions":["Adjusting financial statements solely from preimage membership","Accusing or disciplining personnel based on a compatible configuration","Opening process equipment or interrupting production without site authorization","Using witness contents for any purpose beyond the approved measurement trial","Claiming that the represented configuration family is exhaustive beyond the surveyed topology and stated error bounds"],"halt_rollback":"Stop the trial for leakage, unexpected pressure change, sampler blockage, seal failure, contamination, unstable calibration, or an uninstrumented flow path. Isolate and drain the bench loop under its handling plan, retain readings as invalidated trial records, and remove the clamp-on or spool-mounted test apparatus."},"negative_tests":{"strongest_counterevidence":"Blind trials show that materially different loss locations or offset states produce witness vectors that the stated membership rule incorrectly treats as distinguishable, or that known transfers routinely fail to leave proportional boundary witnesses.","problem_falsifier":"A topology and uncertainty review shows that the existing independent measurements already yield a unique, physically verified inverse for every shortage state of interest, leaving no multiplicity or completeness problem.","intervention_falsifier":"Across predefined blind configurations, the apparatus either excludes the true configuration, fails to exclude configurations contradicted by direct physical observation, or leaves the same preimage bounds as the aggregate baseline after accounting for measurement uncertainty.","risks":["Sampler diversion can alter process mass or pressure.","Witness canisters or seals can leak, rupture, or contaminate the liquid.","Calibration drift, dead volume, evaporation, or density change can create apparent segment discrepancies.","A bypass omitted from the physical topology can create false confidence in completeness.","Common-mode installation error can make nominally independent witnesses agree incorrectly.","Possibility-set membership can be miscommunicated as causal proof or evidence of misconduct."]},"next_evidence_step":"Run a bounded bench study on a three-vessel loop with four instrumented boundaries. Predefine twelve blind configurations covering no loss, one induced loss at each segment, one sampler-ratio offset at each selected boundary, one tank-support offset, and paired loss-plus-offset cases. Before unblinding, record whether the physical measurements retain the true configuration, which false configurations they exclude, which collisions remain, and whether every challenged route actuates its assigned canister. Do not extrapolate beyond this topology.","prior_art_status":"UNSEARCHED","diversity_from_prior_proposals":"Not assessed against other proposals because runtime isolation forbids inspecting them; this candidate was derived only from the supplied archetype and domain card.","revision_record":{"parent_version":null,"progress_targets_addressed":["Preserve a many-to-one preimage rather than asserting a unique shortage cause.","Make measurement instrumentation the primary causal substrate.","Bound the input domain, completeness claim, and downstream authority."],"conceptual_changes":["Initial version directly instantiates preimage characterization as a physical mass-conservation problem."],"operational_changes":["Initial version specifies mechanically actuated proportional samplers, sealed witness canisters, tank load supports, and route challenges."],"evidence_changes":["Initial version defines a twelve-configuration blind bench trial."],"claim_changes":["Claims are limited to physically constraining a compatibility set, not novelty, prevalence, effect size, causation, or responsibility."]},"substrate_contract":{"primary_allowed_process":"MEASUREMENT_INSTRUMENTATION","counterfactual_independence":"Once installed, liquid flow mechanically actuates each sampler and deposits a persistent proportional witness in its canister; tank loads and canister masses remain directly measurable with balances. Removing software, algorithmic inference, databases, dashboards, reporting systems, incentives, authorization rules, and procedural enforcement does not stop the apparatus from creating spatially separated physical constraints on possible flow and loss configurations. The preimage can be checked with conservation arithmetic from the instrument readings; computation may accelerate enumeration but is not required for the essential effect.","forbidden_channel_audit":"No algorithm, model, database, dashboard, recommender, information-routing system, or software control loop actuates the samplers or creates the measurements. No policy, incentive, permission, training program, or accountability workflow produces the evidentiary constraint. Human safeguards govern installation and interpretation but are not the intervention mechanism. The operative sensors are load supports and balances whose direct physical measurements, combined with retained material witnesses, constrain the preimage without downstream analytics being necessary."}}}