R024 — Unused Input Driver

Rule-Reference • — • 3 min read

Audience: Technical Reviewers • Auditors • Model Developers
Last Reviewed: July 2026
Updated: Version 1.0

Executive Summary

R024 (Unused Input Driver) flags a numeric, non-formula cell on a worksheet already identified as a model's input layer that no formula anywhere in the model actually references. An input that cannot be traced to any output is either a stale assumption that was never connected, or a driver that was intended to feed a calculation but was silently missed. Confidence is deliberately kept low, since the engine only resolves single-cell references and cannot see a driver consumed only through a range formula or an Excel data table.

Key Takeaways

✓ R024 flags a numeric input-sheet cell that no formula in the model references, meaning it cannot influence any model output.
✓ Confidence is deliberately low, since the engine's reference resolution only sees single-cell precedents — a driver consumed only through a range formula, such as a SUM over a range including it, would appear unreferenced even though it is genuinely used.
✓ Cells inside an Excel Data Table region, and contiguous columns of binary 0/1 toggle values, are excluded, since both patterns are legitimate structures this rule would otherwise misclassify.
✓ An unreferenced input is not automatically a defect — it may be a placeholder, a retained historical assumption, or genuinely stale; the finding asks a reviewer to confirm which.

Purpose¶

R024 flags a numeric, non-formula cell on a worksheet already identified as the model's input layer that no formula anywhere else in the model references — an assumption with no traceable path to any model output.

Detection Logic¶

Every numeric, non-formula cell on an identified input sheet is checked against the set of every cell referenced as a precedent by any formula in the workbook. A cell not found in that referenced set is a candidate finding, subject to two exclusions: cells recorded as part of an Excel Data Table region are excluded (they are Data Table inputs or outputs, not free-standing unused drivers), and contiguous columns of five or more cells holding only 0 or 1 values are excluded as a recognized binary toggle-array pattern rather than a genuinely unreferenced driver. Confidence is set low throughout, since the engine's reference resolution only resolves single-cell precedents — a driver consumed only through a range reference (such as being included in a SUM(A1:A20)) would appear unreferenced under this specific check even though it is genuinely in use.

Why It Matters¶

An input cell that no formula in the model actually references cannot influence any output, regardless of what value it holds. This can indicate a stale assumption that was once connected but is no longer, or a driver the model's author intended to wire into a calculation but never did — in either case, the model's displayed input does not reflect what actually drives its results, which can mislead a reader who changes the input expecting to see an effect.

Evidence Produced¶

A triggered R024 finding's evidence identifies the specific sheet and cell address of the unreferenced input, its current value, and a description stating that the cell is not referenced by any formula in the model and cannot influence model outputs.

Severity¶

Medium. Weight 6.5. R024 is not one of the five critical-override rules. Confidence is always low.

Example¶

Trigger. An input sheet holds a numeric "Contingency Reserve %" value in a cell that no formula anywhere in the model references — perhaps a driver added during an earlier model version and later disconnected. R024 flags this cell.

Non-trigger — range-reference limitation. The same cell, but genuinely consumed via =SUM(B5:B20) elsewhere in the model, where the cell falls inside that range. Because the engine's reference resolution in this version only resolves single-cell precedents, this cell would still appear unreferenced under R024's check — a known, disclosed limitation of the current detection rather than a claim that the cell is actually unused.

Recommended Remediation¶

Confirm whether the input cell represents an assumption that is still active. If it is stale, remove it or clearly mark it as retained for historical reference only. If it was intended to feed a calculation, identify where the connection was lost and restore the reference.

R024 shares the Model Governance taxonomy category with the other Model Governance rules published in the Rule Reference — see the Rule Taxonomy page for the full category membership.

No existing Knowledge Centre technical guide addresses unused input drivers specifically as of this writing; this section will be updated with a link in the same edit that publishes one.

Version History¶

No version-specific changes to R024's severity, weight, or category are recorded as of engine version 2.1.0. This section will be updated in the same edit as any future change to the rule's declared attributes.

How OXXON tests thisRun a free structural check with FMAE

Frequently Asked Questions

What does FMAE rule R024 check for?

A numeric value on a worksheet identified as the model's input layer that no formula anywhere in the model references — meaning the assumption it represents has no path to influence any model output.

Why is R024's confidence always low?

Because the engine's reference resolution in the current version only resolves single-cell precedents. A driver consumed only through a range formula (a SUM over a range that happens to include it) or an Excel data table would appear unreferenced under this check even though it genuinely is used — low confidence reflects that real limitation honestly rather than overstating certainty.

Does R024 flag every unreferenced numeric cell on an input sheet?

No. Cells within an Excel Data Table region are excluded, since Data Table inputs and outputs have their own distinct mechanism the reference check does not see. Contiguous columns of five or more binary 0/1 values are also excluded, since this pattern is a common toggle-array structure rather than a genuine unused driver.

FMAE Rule Taxonomy

Every rule in the FMAE structural rule pack declares a category attribute at the point it is defined in source — this is not a classification imposed on the rules afterward for documentation purposes, it is the classification the engine itself uses. Six categories cover all 26 rules — Structural (18 rules), Assumptions Governance (1), Integrity Controls (2), Structural Hygiene (1), Aggregation Logic (1), and Model Governance (3). This page publishes that taxonomy as the FMAE equivalent of a control catalog's classification scheme, cross-linked to the Rule Reference page for each member rule.

Finding Data Model & Confidence

Every FMAE finding carries the same fixed structure — which rule triggered it and that rule's identity at the time it ran, its evidence, a narrative explanation, and a confidence label of high, medium, or low. Confidence is a deterministic categorical label derived from how strongly a row's cells agree on a pattern, not a machine-learning probability estimate. Stating this plainly matters, since describing FMAE's confidence as AI-derived or probabilistic would overstate what the engine actually does.