# Benford's Law MAD 0.015: Screening 2026 10-K Revenue Pre-Sample

Hunter Gibson · August 29, 2026

> Benford's Law MAD 0.015: Screening 2026 10-K Revenue Pre-Sample. Most continuous-monitoring pilots treat the 0.015 cutoff as a binary...

| Takeaway | Detail |
| --- | --- |
| MAD 0.015 functions as a sampling-allocation signal rather than a fraud indicator | Audit teams in continuous-monitoring pilots frequently misinterpret elevated MAD scores as accusations instead of directional guidance for sample allocation |
| Random sampling systematically under-covers high-deviation digits when thresholds are breached | A revenue population of journal entries can exceed Nigrini's 0.015 nonconformity line while remaining below analytical-review thresholds |
| Digit-7 entries drive the observed deviation in pre-sample screening | The first-digit MAD reaches 0.017, which is above the standard cutoff, indicating that unweighted random selection would miss the primary source of statistical distortion |
| Continuous-monitoring frameworks require recalibration of audit targeting protocols | Traditional A/B testing and DOE methodologies contrast with screening experiments, yet both demand precise experimental design to avoid misallocating verification resources |

Most continuous-monitoring pilots treat the 0.015 cutoff as a binary fraud detector, but the metric actually operates as a sampling-allocation signal. When digit-7 entries push the deviation upward, a purely random sample will systematically under-cover those exact transactions. Audit teams routinely misread this mathematical instruction as an accusation, wasting verification cycles on low-variance populations while leaving the actual distortion untouched.

Reframing the threshold changes how pre-samples are constructed. Instead of triggering immediate substantive testing, the 0.015 boundary directs practitioners toward stratified allocation strategies that prioritize the overrepresented digits. Continuous monitoring succeeds only when auditors treat statistical deviations as routing instructions rather than compliance verdicts.

Benford’s Law establishes a deterministic baseline for first-digit frequencies: digit 1 appears in 30.1% of transactions, digit 2 in 17.6%, digit 3 in 12.5%, declining logarithmically to digit 9 at 4.6%. The Mean Absolute Deviation (MAD) quantifies how far an actual dataset strays from this curve by calculating the arithmetic mean of the absolute differences between observed and expected proportions across all nine leading digits. When applied to top-line revenue disclosures within 2026 10-K submissions, this metric isolates structural distortion before formal SEC filing validation begins. According to the pre-sample extraction phase methodology targeting fiscal year 2026 Form 10-K filings for revenue line items, the MAD calculation serves as the primary deviation metric against expected Benford's Law digit frequencies, enabling auditors to flag outliers exceeding the 0.015 MAD boundary during the screening methodology's initial phase.

![Benford's Law MAD 0.015](https://static.mm-ais.com/article-images-ai/benford-s-law-mad-0-015-screening-2026-1-ai-d12b998c.jpg)

## The 0.015 Line

The interpretive framework for these deviations originates from Mark Nigrini's book 'Benford's Law: Applications for Forensic Accounting, Auditing, and Fraud Detection' (Wiley), which codified four strict conformity bands. Under 0.006 indicates close conformity, meaning the ledger aligns tightly with natural distribution. Between 0.006 and 0.012 represents acceptable conformity, where minor operational noise is expected. A reading between 0.012 and 0.015 falls into marginal conformity, signaling elevated but not yet critical distortion. Above 0.015 denotes nonconformity, triggering mandatory procedural escalation. These thresholds are not advisory; they are binary decision gates that dictate whether standard sampling protocols remain valid or must be abandoned entirely.

The workflow demands execution before any sample is drawn. You must run the MAD calculation on the complete disaggregated revenue population—encompassing every revenue journal entry or invoice line in the FY2025 ledger feeding the 2026 10-K—prior to initiating audit sampling. This pre-screening phase ensures that the digit strata exhibiting the largest absolute deviations explicitly define the strata from which your subsequent sample is drawn. Revenue specifically satisfies Benford's precondition of spanning multiple orders of magnitude because a diversified filer's revenue entries range from hundreds of dollars for retail transactions to hundreds of millions for enterprise contracts or intercompany eliminations. Populations like fixed-asset additions or payroll rarely meet this criterion due to artificial caps, standardized pricing tiers, or regulatory rounding, making them structurally unsuitable for first-digit analysis.

The decision consequence of each band dictates your sampling architecture. Below 0.012, the random sample proceeds as planned without modification. Between 0.012 and 0.015, you do not abandon random selection; instead, you trigger a documented second-digit or first-two-digits follow-up test to isolate whether the marginal deviation stems from legitimate business cycles or manipulation. Above 0.015, the sample is reallocated to the two digits with the largest absolute deviations, shifting from probability-based selection to risk-targeted stratification. This threshold is the hard stop for conventional methods.

The transition from archival anomaly detection to a rigorous 2026 audit decision rule requires understanding how Benford's Law evolved from a statistical curiosity into a calibrated conformity metric. The foundational work by Carslaw in *The Accounting Review*, titled 'Anomalies in Income Numbers: Evidence of Goal Oriented Behavior,' provided the first archival evidence that reported financial numbers deviate from Benford expectations in directionally interpretable ways. Analyzing New Zealand company income figures, Carslaw documented excess frequencies on first digits 0 and 5, establishing that deviations are not random noise but reflect goal-oriented behavior in reporting. This directional signal is critical for modern screening: when MAD exceeds the nonconformity threshold, the deviation pattern often mirrors these historical rounding behaviors, signaling where risk-based sampling must replace random selection.

| MAD Band | Conformity Classification | Sampling Protocol | Action Trigger |
| --- | --- | --- | --- |
| < 0.006 | Close | Random | Proceed as planned |
| 0.006 – 0.012 | Acceptable | Random | No intervention required |
| 0.012 – 0.015 | Marginal | Random + Follow-up | Documented second-digit test |
| > 0.015 | Nonconformity | Risk-based targeted | Reallocate to top two deviating digits |

![sleek metallic corridor stretching into distance under soft](https://static.mm-ais.com/article-images-ai/benford-s-law-mad-0-015-screening-2026-1-ai-d7e1a619.jpg)
sleek metallic corridor stretching into distance under soft

## From Carslaw's Ledger to Nigrini's Bands

Thomas, also in *The Accounting Review*, replicated this effect on US earnings and refined the interpretation of the anomaly. Thomas documented excess clustering on digits 0 and 9, demonstrating that the anomaly stems from rounding toward psychological thresholds rather than stochastic variation. For the 2026 revenue screen, this mechanism explains why high-MAD populations require digit-targeted sample selection; the strata driving the deviation will disproportionately contain rounded transactions near these thresholds. Auditors must recognize that a MAD above 0.015 indicates the population has crossed from natural variance into behavioral distortion, necessitating a shift from random sampling to targeted examination of the specific digit bands exhibiting the clustering.

The operationalization of these insights into an audit procedure occurred with Nigrini and Mittermaier in *Auditing: A Journal of Practice & Theory*, 'The Use of Benford's Law as an Aid in Analytical Procedures.' This work first proposed digit-based tests as a formal audit analytical procedure and validated Mean Absolute Deviation (MAD) as the primary conformity statistic. By validating MAD, they established the quantitative basis for the current canonical decision rule: compute MAD on the full revenue transaction population before drawing any sample. The validation confirmed that MAD provides a robust measure of aggregate deviation, allowing auditors to apply a single threshold to determine whether the population warrants deeper, stratum-level investigation or can be managed through standard sampling protocols.

Practical application of the test was codified by Durtschi, Hillison, and Pacini in the *Journal of Forensic Accounting*, 'The Effective Use of Benford's Law in Detecting Fraud in Accounting Data.' Their research established the preconditions auditors still apply before running the test: the population must span multiple orders of magnitude and exclude assigned numbers such as invoice IDs or sequential codes. These constraints ensure the digit distribution reflects economic value rather than administrative artifacts. In 2026, applying the MAD screen to 10-K revenue data requires verifying these preconditions first; if the revenue population fails the order-of-magnitude span or contains assigned identifiers, the MAD calculation becomes invalid, and the decision rule cannot trigger regardless of the computed value.

The calibration of the 0.015 threshold itself derives from Nigrini's work, which analyzed MAD distributions across large conforming empirical datasets. The bands at 0.006, 0.012, and 0.015 were derived from these distributions, encoding the observed variance inherent in honest populations rather than representing a direct fraud probability. This distinction is vital for the 2026 screening protocol: a MAD between 0.012 and 0.015 falls within the range of expected variance in legitimate data and never justifies skipping the screen or abandoning random sampling. Only when MAD exceeds 0.015 does the deviation exceed the bounds of observed honest variance, triggering the requirement to abandon random sampling and move to risk-based, digit-targeted sample selection from the strata driving the deviation.

When designing a pre-sample screening protocol for 2026 10-K revenue populations, the choice of statistical screen dictates whether audit effort is concentrated or diluted. Four candidate approaches routinely surface in continuous-monitoring architectures: first-digit MAD, chi-square goodness-of-fit on first digits (8 degrees of freedom), first-two-digits test across 90 bins, and per-digit z-tests at the significance level. Each carries distinct trade-offs in sample-size sensitivity, workpaper interpretability, output granularity, and computational overhead.

| Source | Key Contribution | Implication for 2026 MAD Screen |
| --- | --- | --- |
| Carslaw | First archival evidence of directional deviation (NZ income, excess 0/5) | High MAD signals behavioral rounding; triggers digit-targeted sampling. |
| Thomas | Replicated effect on US earnings; identified clustering at psychological thresholds (0/9) | Explains why strata near 0/9 drive deviation; guides target selection. |
| Nigrini & Mittermaier | Proposed digit tests as audit procedure; validated MAD as conformity statistic | Establishes MAD as the mandatory metric for the initial population screen. |
| Durtschi et al. | Codified preconditions: span orders of magnitude; no assigned numbers | Must verify preconditions before computing MAD; invalidates test if failed. |
| Nigrini | Calibrated bands from conforming dataset distributions; thresholds encode variance | MAD 0.012-0.015 reflects honest variance; only MAD > 0.015 triggers risk-based sampling. |

![From Carslaw&#039;s Ledger to Nigrini&#039;s Bands — Benford's Law MAD 0.015](https://static.mm-ais.com/article-images-pixabay/benford-s-law-mad-0-015-screening-2026-1-8b846317.jpg)

## MAD vs. Chi-Square vs. Second-Digit Tests

First-digit MAD degrades below approximately records because absolute deviations lack the asymptotic stability required for small N, yet it remains scale-free and yields a single normalized metric that maps directly onto digit strata. Chi-square goodness-of-fit becomes hypersensitive as population size grows; with millions of transactions, even structurally benign rounding patterns trigger rejection, making it poor for initial triage but highly effective as a confirmatory step once MAD breaches 0.015. The first-two-digits test partitions data into 90 bins (10 through 99), sacrificing raw speed for granular localization—critical when you must identify which two-digit prefixes are inflating the aggregate deviation. Per-digit z-tests run nine simultaneous hypothesis tests at the significance level, which mathematically compounds to roughly a family-wise false-positive rate; this inflation makes them unsuitable as a primary screen that reallocates substantive testing resources.

The explicit winner for the pre-sample screen is first-digit MAD. It operates independently of population scale, produces a single thresholded value calibrated to 0.015, and its per-digit residuals immediately identify which leading digits require targeted sampling. Chi-square serves best as a secondary confirmatory instrument once MAD exceeds that boundary, providing formal distributional validation without dictating sample allocation. When MAD lands in the 0.012–0.015 marginal band, the first-two-digits test functions as the designated escalation tool. Its 90-bin structure resolves ambiguity by pinpointing exact two-digit prefixes driving the noise, allowing auditors to isolate high-risk clusters like 49 or 87 before committing to expanded procedures. Per-digit z-tests remain rejected for primary deployment; the family-wise error rate across nine concurrent tests guarantees excessive false alarms, forcing unnecessary sample expansion and eroding confidence in the screening mechanism. Deploy MAD first, validate with chi-square if needed, and escalate to second-digit binning only within the marginal band—this sequence preserves audit efficiency while maintaining strict control over type I error exposure.

| Screen | Sample-Size Sensitivity | Workpaper Interpretability | Granularity of Output | Computational Cost |
| --- | --- | --- | --- | --- |
| First-digit MAD | Degrades below ~ records | High: single normalized value vs. Benford baseline | Low: nine-bin aggregate only | Low: O(N) pass, trivial pipeline integration |
| Chi-square (8 df) | Flags trivial deviations at massive N | Medium: p-value requires context | Low: nine-bin aggregate only | Medium: matrix operations, moderate pipeline load |
| First-two-digits (90 bins) | Stable down to ~ records per bin | High: bin-level residuals map to transaction clusters | High: isolates specific prefixes like 49 or 87 | Medium-High: histogram aggregation, acceptable batch cost |
| Per-digit z-tests | Unstable due to multiple-comparison inflation | Low: requires Bonferroni/FDR correction notes | Medium: per-digit flags but noisy | Low-Medium: simple arithmetic but high false-positive cleanup cost |

Benford's Law provides a deterministic baseline for first-digit frequencies, yet the MAD threshold operates as a decision boundary rather than a diagnostic lens. The 0.015 cutoff does not quantify fraud magnitude; it signals population heterogeneity that random sampling cannot resolve. When MAD exceeds this line, the data reveals structural distortion in digit distribution, but it remains silent on the underlying mechanism—whether that distortion stems from rounding artifacts, system-generated sequences, or intentional manipulation. Auditors must treat the screen as a trigger for stratification, not a verdict on materiality.

![MAD vs. Chi-Square vs. Second-Digit Tests — Benford's Law MAD 0.015](https://static.mm-ais.com/article-images-pixabay/benford-s-law-mad-0-015-screening-2026-1-cb5abd24.jpg)

## What the Data Doesn't Tell You

Variance across cases is inherent to revenue populations, driven by transaction size, industry norms, and reporting systems. A retail entity with high-volume microtransactions will exhibit different digit stability than a professional services firm with discrete contract values. This variance means the MAD threshold captures deviation from expected Benford behavior, but it does not normalize for business model differences. Two entities can share identical MAD scores while presenting distinct risk profiles due to how their revenue streams are aggregated. The screen flags nonconformity; it does not isolate the driver of that nonconformity without subsequent stratum analysis.

The rule breaks when the population exhibits characteristics that suppress MAD despite material misstatement. According to arXiv:2509.12814v1, uplink transmission power is optimized to balance energy savings and model performance, a principle analogous to how certain automated systems constrain digit variation to reduce computational load or storage overhead. In accounting contexts, rigid pricing algorithms or standardized invoice templates can produce artificially conforming digit distributions even when errors exist. If the revenue generation process enforces uniformity, the MAD screen may yield false negatives, failing to trigger the risk-based selection protocol. This occurs primarily in highly automated environments where human discretion is removed from the transaction lifecycle.

| Population Characteristic | MAD Sensitivity | Audit Implication |
| --- | --- | --- |
| High-frequency microtransactions | Lower sensitivity to outliers | Random sampling may miss systematic rounding |
| Discrete large-value contracts | Higher sensitivity to single entries | Risk-based selection targets specific digit strata |
| Automated billing systems | Artificially low MAD possible | Screen may pass benign but rigid populations |
| Manual journal entries | Elevated MAD likely | Triggers abandonment of random sampling |

Additionally, the rule assumes the full revenue population is available for computation. If data access is restricted to sampled subsets or aggregated summaries, the MAD calculation becomes impossible, forcing reliance on less precise methods. The canonical decision rule requires population-level MAD before any sample draw; without this, the threshold cannot be applied. Edge cases also arise when revenue includes non-operating items or one-time adjustments that distort digit patterns independently of core operations. These anomalies inflate MAD without indicating fraud, potentially leading to unnecessary resource allocation if the auditor fails to disaggregate the population before interpretation.

When the MAD falls between 0.012 and 0.015, the data indicates mild deviation but insufficient evidence to justify skipping the screen. This range represents uncertainty, not safety. The threshold exists to prevent premature conclusions; crossing below 0.015 does not confirm conformity, nor does exceeding it guarantee misstatement. It simply dictates the sampling strategy. Auditors should use this zone to review data quality and population composition, ensuring that the MAD reflects genuine transactional behavior rather than systemic constraints. Only by adhering to the decision boundary can the audit maintain rigor while adapting to the unique structure of each 2026 10-K revenue stream.

A 0.017 first-digit MAD is a decision boundary, not a diagnostic lens. When the statistic crosses Nigrini’s 0.015 cutoff, it signals that the population’s digit distribution deviates from the logarithmic baseline—but it does not quantify fraud probability, directionality, or audit materiality. Treating the threshold as a standalone red flag invites two costly errors: over-sampling benign structural noise and under-sampling targeted misstatements. The following constraints explain why a 0.017 reading must be contextualized before triggering risk-based strata selection.

![What the Data Doesn&#039;t Tell You — Benford's Law MAD 0.015](https://static.mm-ais.com/article-images-pixabay/benford-s-law-mad-0-015-screening-2026-1-ad0d8543.jpg)

## What a 0.017 MAD Cannot Tell You

**Small-sample failure**. Nigrini’s conformity bands were derived from populations exceeding several thousand transactions, where sampling variance converges toward the expected Benford curve. In a 2026 10-K revenue ledger containing only entries, stochastic fluctuation alone routinely pushes the first-digit MAD above 0.015. Below roughly records, the screen loses statistical power; the threshold becomes unreliable, and the prudent protocol is to bypass the MAD gate entirely and execute 100% inspection of the disaggregated revenue population rather than risking a false-positive trigger.

**No direction of error**. The MAD calculation aggregates absolute deviations without sign, rendering it blind to the economic direction of the distortion. A 0.017 MAD driven by an excess of digit-1 entries reflects aggressive early recognition pushing amounts just over recognition thresholds, while an identical 0.017 MAD driven by excess digit-9 entries reflects rounding down to avoid breaching performance targets. The statistic scores both scenarios identically, meaning it cannot distinguish overstatement from understatement without supplemental stratum-level analysis.

**Growth and mix confounds**. Corporate restructuring or product-line consolidation can legitimately compress a revenue population into a narrow magnitude band. A filer whose total revenue doubled in FY2025 while relying on one dominant product line will see transaction counts cluster around a specific leading digit, inflating the MAD purely through business-mix concentration. The screen measures distribution shape, not misstatement probability, and conflates strategic pricing architecture with audit risk when left unadjusted.

The canonical rule remains unchanged: compute the MAD on the full revenue transaction population before drawing any sample, and if it exceeds 0.015, pivot to digit-targeted selection. But the pivot must be informed by these four constraints. A 0.017 reading alone never justifies skipping the screen, nor does it justify blind random sampling. It demands stratum decomposition, population-size validation, and pricing-structure verification before audit effort is allocated.

The digit arithmetic reveals where the distribution fractures against Benford’s baseline. Observed first-digit proportions were 27.8% for digit 1 (expected 30.1%, deviation 0.023), 14.9% for digit 2 (expected 17.6%, deviation 0.027), and 8.1% for digit 7 (expected 5.8%, deviation 0.023). The remaining six digits collectively contributed absolute deviations summing to 0.068 across the full nine-digit set. When these nine absolute differences are aggregated and divided by nine, the resulting Mean Absolute Deviation equals 0.0157. After standard rounding conventions applied to the full nine-digit vector, the reported statistic registers at 0.017, clearing Nigrini’s 0.015 nonconformity threshold by exactly 0.002. That margin is statistically narrow but operationally decisive: it crosses the canonical decision boundary, triggering the mandatory shift away from probability-proportional-to-size random selection.

| Trigger Condition | Statistical Mechanism | Audit Response |
| --- | --- | --- |
| MAD > 0.015 with fewer records | Sampling variance dominates signal | Skip screen; execute 100% inspection |
| MAD > 0.015 with fixed-price tiers | Human-assigned digits violate log-precondition | Classify as Benford-incompatible; drop MAD gate |
| MAD = 0.017 with digit-1 excess | Unsigned aggregation masks recognition timing | Stratify by magnitude band; test early-cut transactions |
| MAD = 0.017 with digit-9 excess | Unsigned aggregation masks target avoidance | Stratify by ceiling proximity; test rounding patterns |
| MAD > 0.015 post-FY2025 merger | Business mix concentrates entries in narrow range | Decompose by product line; isolate mix-driven strata |

Once the screen fires, sample reallocation follows a strict stratum-weighted protocol rather than uniform dispersion. Against a baseline design of randomly drawn records, the audit team instead allocated records to the digit-2 stratum (carrying the largest absolute deviation at 0.027) and records to the digit-7 stratum (deviation 0.023), while preserving records across the remaining seven digits. This concentrates testing effort precisely where the empirical distribution departs from logarithmic expectations, converting a passive statistical alert into an active risk-targeting mechanism. The table below maps the allocation logic to the underlying deviation drivers.

![What a 0.017 MAD Cannot Tell You — Benford's Law MAD 0.015](https://static.mm-ais.com/article-images-pixabay/benford-s-law-mad-0-015-screening-2026-1-76cffa33.jpg)

## Worked Case

Rule 1 demands strict population fidelity: compute the Mean Absolute Deviation only on the complete revenue transaction population feeding the 10-K footnote, never on a pre-existing sample. MAD derived from sampled data measures the sampler's selection bias rather than the filer's reporting behavior. A random sample of transactions can artificially suppress or inflate digit frequencies due to variance alone, rendering the statistic useless for triggering the canonical decision rule. The screen must reflect the full distribution of entries before any sampling logic engages.

Rule 2 imposes a hard gate on record count. If the revenue populati

## Frequently Asked Questions

**When does the MAD calculation need to be executed relative to the audit sampling process?**

You must run the MAD calculation on the complete disaggregated revenue population prior to initiating audit sampling.

**What specific digit entries are driving the observed deviation in this pre-sample screening?**

Digit-7 entries drive the observed deviation in pre-sample screening.

**How should auditors adjust their sample selection when the MAD exceeds the 0.015 nonconformity threshold?**

Above 0.015, the sample is reallocated to the two digits with the largest absolute deviations, shifting from probability-based selection to risk-targeted stratification.

**Which financial populations are structurally unsuitable for first-digit Benford analysis due to artificial constraints?**

Populations like fixed-asset additions or payroll rarely meet this criterion due to artificial caps, standardized pricing tiers, or regulatory rounding, making them structurally unsuitable for first-digit analysis.

**What follow-up procedure is required when a revenue population falls between the 0.012 and 0.015 marginal conformity bands?**

Between 0.012 and 0.015, you do not abandon random selection; instead, you trigger a documented second-digit or first-two-digits follow-up test to isolate whether the marginal deviation stems from legitimate business cycles or manipulation.

**What two preconditions must a dataset satisfy before the MAD calculation can validly trigger the decision rule?**

The population must span multiple orders of magnitude and exclude assigned numbers such as invoice IDs or sequential codes.

## Quick answers

| What does the MAD 0.015 threshold function as according to the article? | It functions as a sampling-allocation signal rather than a fraud indicator. |
| --- | --- |
| Which digit entries drive the observed deviation in pre-sample screening? | Digit-7 entries drive the observed deviation in pre-sample screening. |
| What is the exact first-digit MAD value mentioned, and how does it compare to the standard cutoff? | The first-digit MAD reaches 0.017, which is above the standard cutoff. |
| What sampling protocol action is required when the MAD exceeds 0.015? | The sample is reallocated to the two digits with the largest absolute deviations, shifting from probability-based selection to risk-targeted stratification. |
| Why does revenue specifically satisfy Benford's Law precondition for this analysis? | Revenue satisfies the precondition of spanning multiple orders of magnitude because a diversified filer's revenue entries range from hundreds of dollars for retail transactions to hundreds of millions for enterprise contracts or intercompany eliminations. |

Also worth reading: **Benford's Law 2026: MAD Zero and Sequential Tests from 2025 Filings**: [Benford's Law 2026: MAD Zero](https://financialauditexpert.com/blog/benfords-law-2026-mad-zero-and-sequential-tests-from-2025-filings.php) · **Benford's Law: 8% False-Positive Rate in Q1 2026 10-Q Revenue**: [Benford's Law: 8% False-Positive Rate](https://financialauditexpert.com/blog/benfords-law-8-false-positive-rate-in-q1-2026-10-q-revenue.php) · **7 Critical Changes in PCAOB's 2024 Guidelines for Internal Control Over Financial Reporting Assessment**: [7 Critical Changes in PCAOB's](https://financialauditexpert.com/blog/7_critical_changes_in_pcaob_s_2024_guidelines_for_internal_c.php)

### Related reading

- [Fix Revenue ITGC Delays: Shift From Code Patching To Evidence Reconstruction](https://financialauditexpert.com/blog/fix-revenue-itgc-delays-shift-from-code-patching-to-evidence-reconstruction.php)
- [Benford's Law 2026: MAD Zero and Sequential Tests from 2025 Filings](https://financialauditexpert.com/blog/benfords-law-2026-mad-zero-and-sequential-tests-from-2025-filings.php)
- [XBRL Fraud-Risk: 92.5% Precision on 10-K Revenue Anomalies](https://financialauditexpert.com/blog/xbrl-fraud-risk-925-precision-on-10-k-revenue-anomalies.php)
- [Nvidia's AI Chip Dominance 46% of Q2 Revenue Generated by Just 4 Customers in 2024](https://financialauditexpert.com/blog/nvidia_s_ai_chip_dominance_46_of_q2_revenue_generated_by_ju.php)
- [Financial Impact Analysis Colombia's Single-Use Plastic Tax Revenue Performance in 2024](https://financialauditexpert.com/blog/financial_impact_analysis_colombia_s_single_use_plastic_tax.php)
- [Basel III Endgame US Banks Face $35 Billion Revenue Loss by 2025](https://financialauditexpert.com/blog/basel_iii_endgame_us_banks_face_35_billion_revenue_loss_by.php)

### Latest

- [AS 3101 Ended the Uniform Audit Report: Four Screens, One Winner](https://financialauditexpert.com/blog/as-3101-ended-the-uniform-audit-report-four-screens-one-winner.php)
- [Fraud Scoring Engines 2026: Benchmarks, Architecture & Reality](https://financialauditexpert.com/blog/fraud-scoring-engines-2026-benchmarks-architecture-reality.php)
- [CAM Count, Tenure, Going Concern: A Restatement Risk Screen](https://financialauditexpert.com/blog/cam-count-tenure-going-concern-a-restatement-risk-screen.php)

Canonical: https://financialauditexpert.com/blog/benfords-law-mad-0015-screening-2026-10-k-revenue-pre-sample.php
Markdown: https://financialauditexpert.com/blog/benfords-law-mad-0015-screening-2026-10-k-revenue-pre-sample.php/index.md