A complete attributed proof that ordinary conservative chip-firing HALTING is NP-complete under deterministic polynomial-time many-one reductions on finite loopless simple strongly connected digraphs with nonnegative binary configurations. Hardness holds with polynomial numerical total initial chips and a specified root whose deletion leaves an acyclic graph with no path of two arcs. The main construction permits antiparallel arcs; a separately proved oriented construction forbids them. The proof includes actual legal checkpoints, conservation, schedule-independent least action, polynomial output bounds, all preprocessing cases and a polynomial-bit NP certificate. The literal modular-bank formula and parallel primed-toppler mechanism are credited to Holroyd, Levine and Winkler, Abelian logic gates, arXiv:1511.00422v2, Lemma 4.1. The prime Boolean and local clause-product encoding is credited to Stockmeyer and Meyer, STOC 1973, Theorem 6.1, and its reconstruction by Gregor Gramlich alone, MFCS 2003, Fact 3. The allowed-predicate intersection is the complement of the classical invalid-variable or failed-clause union; extra local nonBoolean clause-complement residues are already variable-invalid. Neither ingredient nor avoiding global-period expansion is claimed as new. NP membership, capacity and the root-last viewpoint are prior Farrell-Levine/Farrell methods, and long simple-graph clocks are prior Eriksson. Full-result, conservative simple/oriented realization and remaining method priority, and preexisting worldwide openness, remain UNKNOWN after a bounded audit. AI tools were used extensively in solving, verification, source comparison, programming and writing. This is an unrefereed preprint, without external human peer review, journal acceptance or formal proof-assistant certification. No firstness is claimed. The support ZIP contains the complete TeX source, exact authored historical snapshots, bibliographic hashes/read scopes and fresh standard-library finite diagnostics with meaningful faults. No third-party source bodies are redistributed, no omitted historical program is claimed replayed, and finite tests do not certify the unbounded proof or the literature.
No takes yet. Share an insight, caveat, or question.
Kriebel Alec (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: