We present a denominator-audit framework for Bell/CHSH statistics to disentangle mechanism effects from protocol-bookkeeping effects under a shared data context. The protocol compares strict and standard definitions in parallel and separates evaluation into an engineering gate (computability and basic data sufficiency) and a thesis gate (claim-level preregistered criteria). On the same NIST completeblind event stream, binary CHSH yields S_strict=2.336276 (window 0.0) and S_standard=2.839387 (window 15.0), i.e., Delta=+0.503111 from pairing-rule change alone. Bootstrap (n=2000) gives CI95_strict=[2.295151, 2.378669] and CI95_standard=[2.820420, 2.857413]; because 2sqrt(2)=2.828427 lies inside the standard CI, we do not claim Tsirelson violation. A simulated fallback run also shows that S_standard does not necessarily exceed S_strict, reinforcing protocol sensitivity. We report a method-level, boundary-aware conclusion: Bell/CHSH claims should be interpreted through protocol definitions, uncertainty intervals, and retained negative results, rather than peak statistics alone. Figures may be added in a subsequent version; current evidence is provided via tables and machine-readable result files.
We present a denominator-audit framework for Bell/CHSH statistics to disentangle mechanism effects from protocol-bookkeeping effects under a shared data context. The protocol compares strict and standard definitions in parallel and separates evaluation into an engineering gate (computability and basic data sufficiency) and a thesis gate (claim-level preregistered criteria). On the same NIST completeblind event stream, binary CHSH yields S_strict=2.336276 (window 0.0) and S_standard=2.839387 (window 15.0), i.e., Delta=+0.503111 from pairing-rule change alone. Bootstrap (n=2000) gives CI95_strict=[2.295151, 2.378669] and CI95_standard=[2.820420, 2.857413]; because 2sqrt(2)=2.828427 lies inside the standard CI, we do not claim Tsirelson violation. A simulated fallback run also shows that S_standard does not necessarily exceed S_strict, reinforcing protocol sensitivity. We report a method-level, boundary-aware conclusion: Bell/CHSH claims should be interpreted through protocol definitions, uncertainty intervals, and retained negative results, rather than peak statistics alone. Figures may be added in a subsequent version; current evidence is provided via tables and machine-readable result files.
Archive Notice (Deprecated): This record is kept for archival purposes and is no longer the current version. Please use the latest Bell Audit v5 record: https://zenodo.org/records/19904639 (DOI: https://doi.org/10.5281/zenodo.19904639). 留档提示(已过期): 本记录仅用于历史留档,已非当前版本。请优先访问最新版 Bell Audit v5:https://zenodo.org/records/19904639(DOI:https://doi.org/10.5281/zenodo.19904639)。 [EN] This paper presents a "Denominator Audit" of Bell/CHSH violations using NIST experimental data. Historically, violations of S > 2.0 are cited as proof of non-locality. Our audit reveals that these results are not intrinsic to the physical system but are "manufactured" through the selective manipulation of coincidence windows (Accounting Fraud). On the same NIST complete-blind stream, we compare a strict pairing tolerance (0.0) versus the Round-1 standard tolerance (15.0) in NIST grid-index units on the exported CSV t axis (not calibrated seconds in this deposit). The frozen binary-CHSH snapshot gives S_strict = 2.336276 vs S_standard = 2.839387 (Δ = +0.503111). A bootstrap CI for the standard branch includes 2\sqrt{2} (artifacts/reports/chsh_bootstrap_ci_standard15.json, ci_contains_tsirelson: true); we therefore discuss protocol sensitivity, not Tsirelson violation. Separately, the “30 cm truth” is a laboratory heuristic (about 1 ns ↔ ~30 cm in vacuum); it motivates coincidence bookkeeping sensitivity and must not be read as “parameter 15.0 equals 15 ns” here. Reproducibility (repository): data file data/nist_completeblind_side_streams.csv; run from repo root python scripts/explore/chsh_denominator_audit.py (writes artifacts/bell_window_scan_v1/). [中] 在同一组 NIST complete-blind 事件流上,我们在导出 CSV 的 t 网格上对比严苛配对容差(0.0)与第一轮标准容差(15.0),单位为 NIST 网格指数(本归档不将其标定为秒)。冻结的二元 CHSH 快照为 S_strict = 2.336276 与 S_standard = 2.839387(Δ = +0.503111)。标准分支的 bootstrap 置信区间覆盖 2\sqrt{2}(artifacts/reports/chsh_bootstrap_ci_standard15.json,ci_contains_tsirelson: true);因此本文强调协议敏感性,而非 Tsirelson 违背。另:“30 厘米的真相”是实验室启发(真空中约 1 纳秒 ↔ 约 30 厘米),用于提示符合会计的敏感性,不应被读作“参数 15.0 等于 15 纳秒”。 可复现(配套仓库): 数据 data/nist_completeblind_side_streams.csv;在仓库根目录运行 python scripts/explore/chsh_denominator_audit.py(输出写入 artifacts/bell_window_scan_v1/)。
This paper synthesizes two published audit lines into a single reproducible framework for Bell and GHZ interpretation under analysis-rule sensitivity. Using public records and artifact-linked evidence, we show that Bell/CHSH on the same NIST event stream shifts from S_strict=2.336276 to S_standard=2.839387 (Delta=+0.503111) under protocol change, while GHZ search under the current audited pipeline remains far from F=4 (best fine-stage F=0.085396). We formalize a ledger-style audit protocol that treats pairing windows, denominator policy, post-selection, and uncertainty as first-class reported variables. We further introduce a singles-aware diagnostics layer (C_norm = coincidences / sqrt(singles_A * singles_B)) as a testable bridge for normalization-path analysis, while explicitly avoiding ontology-level promotion from current evidence. The contribution is a publication-ready, boundary-aware audit standard with claim-to-artifact traceability across both lines. Related records: Bell DOI https://doi.org/10.5281/zenodo.19763028 ; GHZ DOI https://doi.org/10.5281/zenodo.19763473 ; Code repository: https://github.com/tomnattle/chain-explosion-model
This deposit provides a professional Bell audit rewrite based on the Apr-29 v4 evidence package and is designed for reviewer-level traceability. Key quantitative findings on a fixed event stream: Binarization shift: S_raw(continuous) = 1.117683 vs S_raw(quant_2) = 2.297779. Semantic premium staircase: same_index = 2.329417, external_bucket_all = 2.775687 (+0.446270), event_anchor_nearest = 2.834670 (+0.058983). Closure checks: same_index_not_near_2p82 = True, pure_bucket_in_2p8_zone = True, anchor_asymmetry_small = True (delta_abs = 0.001356), edge_sensitivity_bounded = True. Experimental principle and method: Principle: CHSH headline values can be altered by representation and inclusion rules even when the underlying row/event stream is unchanged. Method chain: same data -> quantization choice -> pairing semantics -> denominator inclusion policy -> CHSH report. Scope: methodological reproducibility and protocol sensitivity audit. Not claimed: ontology-final conclusion or direct falsification of all competing theories. Primary manuscripts: papers_final/05_Bell_Audit/version_01/【0】_Bell_Audit/files/01_Bell_Audit_Professional_CN.md papers_final/05_Bell_Audit/version_01/【0】_Bell_Audit/files/02_Bell_Audit_Professional_EN.md Reproducibility Run from repository root: python scripts/explore/nist_same_index_quantization_sweep_v4.py python scripts/explore/nist_unified_semantics_audit_v4.py python scripts/explore/nist_revival_20pct_closure_v4.py
[EN] This deposit summarizes two frozen workflows in one methodological bundle. (1) Bell / CHSH (real public data): Binary CHSH on the NIST complete-blind CSV. Pairing uses tolerances in NIST grid-index units on the exported t axis (not calibrated seconds in this repository). Frozen snapshot: battle_results/nist_completeblind_2015-09-19/battle_result.json (S_strict = 2.336276 at tolerance 0.0 vs S_standard = 2.839387 at 15.0; pair counts 136632 vs 148670). Reproduce the window scan via python scripts/explore/chsh_denominator_audit.py. (2) GHZ-style statistic (simulation, V10.4 only): The GHZ leg is not tied to ghz_loop_explosion_v19.py. It is anchored to scripts/explore/ghz_medium_v10/v10_4_real_cost_curve.py, which emits a computed_curve (type in V10_4_REAL_COST_CURVE.meta.json): a fixed RNG phase sample (medium_phase_states), soft amplitude gating with parameter gate_k, reported F_gated, mean simultaneous-hit retention R_gated, and a matched-retention random control (F_random_mean, F_random_std). Canonical outputs live under artifacts/ghz_medium_v10/ (V10_4_REAL_COST_CURVE.png, .csv, .meta.json, V10_4_REAL_COST_CURVE_REPORT.md). Scope statement: This is methodological reproducibility documentation — not a claim that either pipeline certifies a specific hardware Bell or GHZ experiment. [中] 本归档将两条冻结工作流合并为一条方法论叙述。 (1)贝尔 / CHSH(真实公开数据): 在 NIST complete-blind CSV 上做二元 CHSH。配对容差取导出列 t 上的 NIST 网格指数单位(本仓库不标定为秒)。冻结快照:battle_results/nist_completeblind_2015-09-19/battle_result.json(容差 0.0 时 S_strict = 2.336276,15.0 时 S_standard = 2.839387;配对数 136632 / 148670)。窗口扫描复现:python scripts/explore/chsh_denominator_audit.py。 (2)GHZ 型统计量(仿真,仅 V10.4): GHZ 侧不再绑定 ghz_loop_explosion_v19.py。唯一锚点为 scripts/explore/ghz_medium_v10/v10_4_real_cost_curve.py,产出 computed_curve(见 V10_4_REAL_COST_CURVE.meta.json 中 type):固定 RNG 相位样本(medium_phase_states)、参数为 gate_k 的软振幅门控、报告 F_gated、同时触中平均保留 R_gated,以及同保留率随机对照(F_random_mean、F_random_std)。规范产物位于 artifacts/ghz_medium_v10/(V10_4_REAL_COST_CURVE.png、.csv、.meta.json、V10_4_REAL_COST_CURVE_REPORT.md)。 范围声明: 本文为方法论与可复现性文档——不声称任一脚本对某一具体硬件贝尔或 GHZ 实验作出鉴定。 Reproducibility (repository): Bell: python scripts/explore/chsh_denominator_audit.py (needs data/nist_completeblind_side_streams.csv). GHZ V10.4: python scripts/explore/ghz_medium_v10/v10_4_real_cost_curve.py. 可复现(配套仓库): Bell:python scripts/explore/chsh_denominator_audit.py(需 data/nist_completeblind_side_streams.csv)。GHZ V10.4:python scripts/explore/ghz_medium_v10/v10_4_real_cost_curve.py。
Paper 32B analyzes the public Weihs Alice/Bob archive with a focus on separating timing diagnostics from setting-resolved Bell-test statistics. The bluesine data are used as a timing-acceptance lane, showing a finite timing continuum with a central peak near zero offset, shoulders near plus or minus 10 nanoseconds, and a flat far-offset baseline. As the coincidence half-window widens from 3 to 5 to 7 nanoseconds, total coincidences increase while aggregate parity decreases, showing the expected throughput-versus-contrast tradeoff. The paper keeps this timing analysis separate from Bell-sector claims because bluesine does not provide a recoverable settings timeline. For the CHSH benchmark, the paper uses the longdist32 Coincidence Evaluation report, which includes setting-resolved correlators. The reported value is S = 2.702 plus or minus 0.078, which is well above the classical bound of 2 and below the Tsirelson bound of about 2.828. Raw longdist32 timetag scans are used only as timing and ingestion sanity checks, not as an independent reconstruction of CHSH. The result is a reproducible audit framework that separates timing-continuum validation, report-based CHSH benchmarking, and future raw-event reconstruction. The paper does not claim a new Bell bound, a proof of Holosphere theory, or evidence for superdeterminism; it establishes a disciplined baseline for later setting-resolved and theory-specific tests.
This paper introduces a denominator-recovery audit framework for high-correlation claims in a GHZ three-body model, aimed at separating mechanism gain from post-selection bookkeeping gain. Within a unified pipeline, we compare `none` and `energy_weighted` denominator modes and run a two-stage search (90 deg coarse + 2 deg fine), while jointly auditing `F` and `coincidence_rate` trade-offs. Under the current configuration (`samples=80000`, `numba_cpu`, fixed gating), the best fine-stage candidate reaches only `F=0.085396`, remaining far from the target `F=4` with `|err|=3.914604`. Target-attainment decomposition consistently indicates correlation-shape mismatch as the dominant bottleneck rather than coincidence sparsity. Robustness indicators (bootstrap and seed sweep) do not show convergence toward 4 either. We therefore report a method-level conclusion: within the tested model family and protocol, no evidence supports an `F->4` transition. Strong GHZ-style claims should be evaluated through joint auditing of denominator logic, sample-inclusion rules, and retained failure cases. The contribution is a reproducible, ledger-style audit workflow for GHZ correlation studies. Code repository: https://github.com/tomnattle/chain-explosion-modelThis version adds explicit disclosure of the post-selection cost-benefit regime: while F can approach 4 under aggressive filtering, the retained sample ratio drops to about 25% (about 75% discarded). The update includes the corresponding cost-benefit curve and supporting artifacts.