In a 2015 wire-fraud trial, a judge let the defense argue the numbers were innocent using a 1938 law of how often digits start with 1.

In a December 2015 ruling, a federal judge in New Mexico let a wire-fraud defendant's expert use Benford's Law — a 1938 observation about how often numbers begin with the digit 1 — to argue the data looked innocent.

Here is the thing most people never learn about numbers in the wild: in large, naturally occurring sets, the first digit is far more likely to be a 1 than a 9. About 30% of the time the leading digit is 1; a 9 shows up as the first digit only about 5% of the time. That lopsided pattern has a name — Benford's Law — and in a 2015 federal courtroom in Albuquerque, it became a defense argument.

In *United States v. Channon*, Matthew and Brandi Channon stood accused of wire fraud and conspiracy. Their expert, forensic accountant Janet McHard, proposed to testify that the financial data at issue conformed to Benford's Law — that the digits looked the way honest numbers tend to look, not the way cooked books do. The Government did not want the jury to hear it. On November 2015 it filed a Consolidated Motion to Exclude (Doc. 235), arguing McHard's method wasn't reliable enough to be evidence.

On December 2, 2015, U.S. District Judge Judith C. Herrera disagreed. In a Memorandum Opinion and Order (Doc. 262), she found McHard's Benford's Law testimony sufficiently reliable under the Daubert standard and Federal Rule of Evidence 702, and denied the motion. The jury would get to hear it.

Why a law about digits ends up in fraud cases

The pattern was first noticed in 1881 by astronomer Simon Newcomb, who observed that the early pages of logarithm tables — the ones for numbers starting with 1 — were far more worn than the later pages. People were looking up numbers beginning with 1 much more often. It was the physicist Frank Benford who, in 1938, gathered tens of thousands of figures from river areas to street addresses to molecular weights and formalized the distribution. The law now carries his name.

The forensic hook is simple. Honest financial data — invoices, expense reports, reimbursements — tends to follow Benford's distribution. People who fabricate numbers usually don't. A fraudster inventing dollar amounts tends to spread the leading digits too evenly, or lean on favorites, producing a profile that deviates from what Benford predicts. Auditors and the IRS have used exactly this to flag statements worth a second look.

What makes *Channon* notable is the direction the tool was pointed. Benford's Law is typically the prosecutor's magnifying glass — a reason to suspect numbers are fake. Here it ran the other way. The defense wanted to say: look at the digits, they behave normally, this is what clean data looks like. Judge Herrera's ruling meant the same statistical lens auditors use to catch fraud could be argued, in open court, as evidence that no fraud occurred.

How the case turned out

Admitting the testimony did not win the trial. The jury convicted. On appeal, the Tenth Circuit upheld the convictions in a decision issued January 31, 2018 (*United States v. Channon*, 881 F.3d 806), where the court described Janet McHard plainly as "Defendants' expert, a forensic accountant." The Benford's analysis got in front of the jury; the jury was not persuaded.

That outcome points at the seam still worth arguing. A judge ruling testimony *admissible* under Daubert is only saying it clears a reliability threshold — that it's grounded enough in accepted method to be heard, not that it's correct. Benford's Law is genuinely powerful at scale, but its reliability depends on the data: it works best on sets that span several orders of magnitude and are free of artificial floors, caps, or assigned numbers. Whether any particular spreadsheet is the kind of data the law even applies to is a judgment call — and that call is what a jury, not a formula, is left to make.

So the fact holds in two layers. Benford's Law earned its place in the Channons' trial. And the deeper lesson — that a tool built to catch liars can be turned to defend them, with the same math pointing both ways — is the part still open to dispute.