What question did this study set out to answer?

February 9, 2026Open Access

Testing Event Clustering in a Biblical Timeline Version 2: Formal Sensor Construction and Preregistered Validation

Key Points

To analyze and validate temporal event clustering in the biblical narrative using a formal sensor approach.
Utilized a rules-based extraction pipeline to analyze events from the American Standard Version (1901).
Conducted 10,000,000 Monte Carlo simulations to assess peak-to-background ratios and significance levels.
Performed secondary tests to detect concentrations during the Exilic/post-Exilic period and evaluated change-point hypotheses.
Identified 5,100 dated events, with the dominant bin showing a peak-to-background ratio of R = 26.005.
Monte Carlo analysis showed zero exceedances (p < 3×10⁻⁷), affirming the significance of event clustering.
Detected a broad concentration of events between -1650 and -425 CE, passing statistical tests for significance.

Abstract

Update — Preprint Version 2 (February 7, 2026) Summary. This release presents a preregistered confirmatory analysis of temporal event clustering in the biblical narrative. A frozen, rules-based extraction pipeline (Sensor A1, superseding the exploratory v1 analysis) yields 5, 100 dated events from the American Standard Version (1901), spanning –4025 to +100 CE across 165 fixed 25-year bins. The dominant bin (25–50 CE, midpoint 37. 5 CE) contains 698 events with a peak-to-background ratio R = 26. 005; across 10, 000, 000 uniform-timing Monte Carlo simulations, zero exceedances were observed (p < 3×10⁻⁷). Secondary preregistered tests detect a broad Exilic/post-Exilic concentration (–1650 to –425 CE) and demonstrate specificity through the correct failure of a change-point hypothesis. All extraction code, input data, exclusion lists, and outputs are archived with SHA-256 verification for full reproducibility. What's new since v1 Sensor formalization: Deterministic, rules-based extraction (extractorᵥ1. 2. 1 + rulesᵥ1. 2. 1. json) ; documented exclusions; fixed binning; preregistered null models. Scale-up: From 109 events (v1) to 5, 100 events (v2) with 303 exclusions documented. Reproducibility: Added repobundleA1Sensor. zip packaging all inputs, scripts, outputs, environment specs, and hashes. Program framing: This manuscript reports Module A1 (instrument validation) within a larger research program; system-identification modules follow separately. Preregistered confirmatory results (high level) H1 Peak location: Midpoint in [0, 75) CE → Pass (37. 5 CE). H2 Dominance: R ≥ 10 → Pass (R = 26. 005). H3 Extremity: p < 0. 001 (uniform-timing Monte Carlo) → Pass (0/10, 000, 000 exceedances; p < 3×10⁻⁷). H4 Overdispersion: Negative binomial ≫ Poisson (LRT, p < 10⁻⁷, BH-FDR) → Pass. H5 Temporal scan: –1650 to –425 CE concentration, p < 6×10⁻⁴ (BH-FDR) → Pass. H6 Change-points (PELT + permutation): p = 1. 0 → Not supported (specificity). Robustness (H7–H9): Peak persists under bin shift, point vs. interval assignment, and interval-weighted counts. Data, code, and reproducibility Corpus: American Standard Version (1901), treated as narrated events only. Dating: Mid-year representation from documented chronology sources; priority rules preregistered. Determinism: Fixed seeds; identical inputs → identical outputs (hash-verifiable). Update — Preprint Version 2 (February 7, 2026) Summary. This release presents a preregistered confirmatory analysis of temporal event clustering in the biblical narrative. A frozen, rules-based extraction pipeline (Sensor A1, superseding the exploratory v1 analysis) yields 5, 100 dated events from the American Standard Version (1901), spanning –4025 to +100 CE across 165 fixed 25-year bins. The dominant bin (25–50 CE, midpoint 37. 5 CE) contains 698 events with a peak-to-background ratio R = 26. 005; across 10, 000, 000 uniform-timing Monte Carlo simulations, zero exceedances were observed (p < 3×10⁻⁷). Secondary preregistered tests detect a broad Exilic/post-Exilic concentration (–1650 to –425 CE) and demonstrate specificity through the correct failure of a change-point hypothesis. All extraction code, input data, exclusion lists, and outputs are archived with SHA-256 verification for full reproducibility. What's new since v1 Sensor formalization: Deterministic, rules-based extraction (extractorᵥ1. 2. 1 + rulesᵥ1. 2. 1. json) ; documented exclusions; fixed binning; preregistered null models. Scale-up: From 109 events (v1) to 5, 100 events (v2) with 303 exclusions documented. Reproducibility: Added repobundleA1Sensor. zip packaging all inputs, scripts, outputs, environment specs, and hashes. Program framing: This manuscript reports Module A1 (instrument validation) within a larger research program; system-identification modules follow separately. Preregistered confirmatory results (high level) H1 Peak location: Midpoint in [0, 75) CE → Pass (37. 5 CE). H2 Dominance: R ≥ 10 → Pass (R = 26. 005). H3 Extremity: p < 0. 001 (uniform-timing Monte Carlo) → Pass (0/10, 000, 000 exceedances; p < 3×10⁻⁷). H4 Overdispersion: Negative binomial ≫ Poisson (LRT, p < 10⁻⁷, BH-FDR) → Pass. H5 Temporal scan: –1650 to –425 CE concentration, p < 6×10⁻⁴ (BH-FDR) → Pass. H6 Change-points (PELT + permutation): p = 1. 0 → Not supported (specificity). Robustness (H7–H9): Peak persists under bin shift, point vs. interval assignment, and interval-weighted counts. Data, code, and reproducibility Corpus: American Standard Version (1901), treated as narrated events only. Dating: Mid-year representation from documented chronology sources; priority rules preregistered. Determinism: Fixed seeds; identical inputs → identical outputs (hash-verifiable). AI-assisted development disclosure: AI tools assisted with code drafting/editing and administrative text; all logic and decisions are author-defined; all code reviewed and executed by the author. Files in this release (included as computational notebook in last update-minus the preprint) TestingEventClusteringBiblicalTimelineVersion 2. pdf — V2 preprint manuscript eventsₘergedwithᵧearscombo. csv — all 5, 100 dated events including exclusion flags exclusionsconfirmatory. csv — 303 preregistered exclusions bincounts₂5y. csv — 165-bin event counts familyₐₛummary. csv — Family A hypothesis outcomes familybₛummary. csv — Family B hypothesis outcomes (BH-FDR adjusted) analysisᵣeport. txt — complete numerical results log ColabCode. py — end-to-end analysis implementation AI-assisted development disclosure: AI tools assisted with code drafting/editing and administrative text; all logic and decisions are author-defined; all code reviewed and executed by the author. Status Extraction + preregistered confirmatory analysis complete (manuscript finalized for v2) ; subsequent modules (temporal geometry, dynamics, mechanism) to be reported separately. Project / preregistration Project updates (OSF): https: //osf. io/as8mq Preregistration (stats): https: //osf. io/vqzm4/ Preregistration (extractor & rules): https: //osf. io/8xquy/ How to cite McCain, W. (2025). Event Distribution in the Biblical Narrative (−4025 to 100 CE): Quantifying Temporal Clustering with Preregistered Methods (Version v2). Zenodo. https: //doi. org/10. 5281/zenodo. 18065787

Ask AI

Helpful

Bookmark

View Full Paper