Testing Absoluteness: A ruler for the coherence of records in the friend test, its scaling, and a protocol for its run
I take the friend test of Measuring Absoluteness, with an odd number n of laboratories, where outcomes that hold absolutely score at most n − 1 and quantum theory reaches n + 1. I ask what a measured score proves about the coherence of the friends’ records, and how an imperfect or modified experiment should be read. Under quantum theory I prove five results. (1) The joint-coherence law holds on the commuting algebra formed from record and later measurement, and does not depend on which measurement is called the record. (2) A score above G, the best score with one incoherent record, forces every pair of records to have joint coherence at least (S − G)/(2(n + 1 − G)), and this bound is exact. (3) With misreport rate ε and state drift μ, everything holds for the observed score less nμ + 2n(n + 1)ε, and neither term can be reduced. (4) A law that dephases records with visibilities v_k multiplies pair coherence by v_i v_j, so a measured score bounds any such law. (5) The full friend score has a coherence threshold fixed at 1 − 2^(−1/2) of its maximum at every n. I then ran a registered six-qubit protocol on IBM’s ibm_fez, with 176,000 shots. The corrected score was 3.155, above both the classical limit of 2 and the every-pair threshold of 2.83, which certifies joint coherence of at least 0.14 for every pair. Dephasing the records lowered the score along the predicted curve. Repeats on ibm_kingston and ibm_marrakesh showed violation and the same relations, but missed the every-pair threshold because of larger readout budgets. The results confirm quantum theory; no new physics is claimed. Included: the paper (PDF and Word), the Lean 4 companion (exit 0, 44 axiom lines, run on the author’s machine) with its run script and log, the numerical check script (29 checks), the registered run plan, the run and analysis scripts, and the raw counts and reports from all four IBM jobs. Submitted for technical scrutiny, not belief.
Authors
- Aron Reid
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-21
- DOI
- https://doi.org/10.5281/zenodo.24412057
- Primary Topic
- Quantum Mechanics and Applications
- Type
- preprint