METHOD-001

Methodological contract

Research methodology

METHOD-001 defines how Coran Codex moves from a computational question to a documented public research record.

A numerical pattern is an observation to investigate. It is not treated as proof of supernatural or divine origin.

Scope

Methodological contract

Every Coran Codex study must identify its dataset, analytical view, counting rules, transformations, statistical procedures and limitations clearly enough for the analysis to be inspected and reproduced.

MethodMETHOD-001

DatasetDATA-001

Corpus viewsRAW / BODY

Research recordReproducible

DATA-001 defines the authoritative corpus identity and exact counting baseline. METHOD-001 defines how questions are tested against that corpus.

Research process

From question to public record

Define

State the research question, dataset, analytical view, counting unit, transformation and intended comparison.

Test

Execute the documented measurement and, when appropriate, compare it with a defined null model or reference distribution.

Challenge

Test alternative explanations, sensitivity to analytical choices, multiple-testing effects and relevant prior art.

Publish

Record methods, results, limitations, interpretation, revisions and negative findings without hiding unsuccessful tests.

Study classification

Exploratory and Confirmatory research

Exploratory work identifies questions and possible patterns; confirmatory work fixes the hypothesis and decision rules before the relevant analysis. These labels do not substitute for evidence.

Exploratory analysis searches for structure, generates hypotheses and identifies candidate relationships worth investigating.

Because the pattern may have been noticed after inspecting the data, exploratory significance values must not be presented as though the hypothesis had been specified in advance.

Confirmatory analysis fixes the hypothesis, measurement, comparison, decision rule and relevant analytical choices before evaluating the target result.

A confirmatory label is used only when the analysis genuinely satisfies that separation.

Statistical testing

Testing patterns rather than admiring them

Null model

When a probability claim is made, the study must state what alternative outcomes are considered possible under the null model and why that model is relevant to the question.

Test statistic

The measured quantity must be stated explicitly before interpreting its extremeness. A statistic selected only because it produced an unusual value is treated as exploratory.

Multiple testing

Searches across many formulas, moduli, verse groups, letter sets, parameters or alternative definitions increase the opportunity to find apparently unusual results by chance.

Relevant search breadth must therefore be reported and accounted for when statistical significance is interpreted.

Simulation and permutation

Monte Carlo or permutation procedures may be used when an analytic probability is unsuitable, provided the randomized unit, preserved structure, number of iterations and resulting uncertainty are documented.

A small p-value does not by itself establish importance, causation, intentional design or supernatural origin.

Controls

Robustness

A result becomes more informative when it survives reasonable tests designed to make it fail.

Corpus view

Compare RAW and BODY when the distinction could materially affect the result.

Counting rules

Test whether the finding depends on a particular segmentation, normalization or character treatment.

Duplicates and dependence

Repeated verses, repeated phrases and correlated observations must not silently be treated as independent evidence.

Alternative specifications

Reasonable nearby definitions should be examined when changing them could reverse the conclusion.

At evidence level E3, the result survives relevant alternative counting rules, corpus views, and methodological controls.

Novelty

Prior-art investigation

Statistical support and novelty are separate questions. A finding can be statistically robust while already known, or potentially novel while still weakly supported.

Novelty labels

Not Checked: No meaningful prior-art investigation has yet been completed.
Investigating: The literature and public record are still being examined.
Potentially Novel: No earlier equivalent formulation has yet been located within the documented search.
Previously Known: Earlier publication or documentation of the relevant result has been identified.
Not Applicable: Novelty is not a meaningful classification for the record.

Prior-art searches are necessarily limited by available sources, languages, indexing and search terms. “Potentially Novel” is therefore not proof that no earlier discovery exists.

Evidence scale

E0 to E4

E0: ObservationA pattern has been identified and documented.
E1: Structured patternThe relationship is reproducible under the stated corpus and counting rules.
E2: Statistical supportA documented statistical model provides support beyond the raw observation.
E3: Robust across controlsthe result survives relevant alternative counting rules, corpus views, and methodological controls.
E4: Independently reproducibleAn independent reproduction confirms the result using adequately documented methods.

Evidence level, scientific status and novelty status are independent dimensions and must not be collapsed into a single label.

Scientific status

A result can change status

ObservationInterestingStatistically SignificantRobustReproducedRejectedUnder Review

Status records the current scientific assessment. It may change when stronger controls, new evidence, corrected calculations or independent reproduction become available.

Interpretation

What the result does not prove

Each important result separates direct measurement from statistical interpretation and broader claims.

Failure policy

Negative results remain visible

Negative and rejected results remain part of the research record.

A hypothesis is not removed merely because a stronger control destroys the apparent pattern. Rejections, failed replications and corrected interpretations are valuable because they document which approaches did not survive scrutiny.

Where practical, revisions preserve enough history to distinguish the original observation from the later reason for changing its status.

Reproducibility

Minimum study record

A published computational study should expose enough information for another researcher to understand what was measured and repeat the analysis.

Dataset: DATA-001 or another explicitly identified dataset.
Analytical view: RAW, BODY, or another precisely documented transformation.
Question: The hypothesis or exploratory question being investigated.
Counting rule: The exact definition of every measured unit.
Statistical method: Null model, statistic, simulation or test procedure where relevant.
Search breadth: Relevant multiple testing, parameter exploration or post-hoc selection.
Controls: The robustness checks that were performed and their outcomes.
Prior art: The scope and limitations of the investigation into earlier work.
Revisions: Material changes to calculations, interpretation or scientific status.

Record history

Revision history

2026-09-15: METHOD-001 baseline established: study classification, statistical testing, multiple-testing policy, robustness, prior-art review, evidence levels, negative-result policy and reproducibility requirements.