Evidence Coverage and the Search for Unrecorded Liabilities: Data, Code and Analytical Figures

Complete data and computational companion to Evidence Coverage and the Search for Unrecorded Liabilities. The deposit contains a fixed synthetic accounts-payable case, the full evidence-packet library and truth key, 10,000 synthetic posted entries, a source-frame manifest, deterministic Python calculations, seven numerical result tables, three analytical figures with executable source code and publication-quality exports, and Online Appendix S1 containing the proofs and engagement-partner review workpaper. The analysis examines the identification boundary of procedures restricted to recorded inputs, the value of acquiring evidence from independently selected frames, asymmetric losses, perceived source coverage, and the complementarity or substitutability of sources under aggregation and materiality thresholds. All case amounts, dates, companies and documents are synthetic. There are no human-participant data, observed client records, fitted behavioral estimates or random simulation draws. Reproduction: extract AJPT_Completeness_Data_Code_v1.0.0.zip and run python reproduce.py using Python 3.10 or later. The core computation uses only the standard library and requires no external data, accounts, downloads or expenditure. A successful run produces results/verification.json with status PASS. All seven CSV outputs and the verification report reproduced exactly in the release check; all input SHA-256 hashes matched. Plotting dependencies and commands are documented in README.txt, with pinned versions supplied in requirements-figures.txt. The complete archive is self-contained. The three input files, data dictionary and appendix are also available separately for direct inspection. DATA_DICTIONARY.csv defines every tabular field and the case JSON structures. SHA256SUMS.txt inside the archive and UPLOAD_SHA256SUMS.txt verify file integrity. Probabilities of acquiring decisive evidence are distinguished from probabilities of correct judgment. Version 1.0.0. Creator: Craig Wright, University of Exeter Business School; ORCID 0000-0001-9374-0507. Funding: None. Financial conflicts: None. Nonfinancial conflicts: None. AI DisclosureAI assistance was used in searches, selecting material, summarising papers, red teaming, and checking code and validation. All figures were created using AI assistance.