When "Probably" Is All You've Got

The prosecution in the Kouri Richins trial called George Matthew Throckmorton to the stand on Day 10. He's a forensic document examiner with 30 years of experience, 1,500 cases, FBI Academy training at Quantico, and 48 prior courtroom appearances in his specialty. On paper, he looks like exactly the kind of witness you want delivering the goods on a forgery charge.

And he did his job. He compared 86 known signatures and 24 sets of initials from Eric Richins against the questioned signature on a life insurance application. He walked the jury through it with a visual comparison exhibit, stepping down from the witness stand to point out the differences letter by letter. One-stroke signature on the questioned document versus consistent pen lifts in every known sample. Second E starting at the baseline instead of the top. R beginning from the bottom rather than the staff.

His conclusion: no evidence Eric authored the signature. It appeared to be a simulated forgery written by someone who was familiar with Eric's handwriting and tried to copy, draw, or duplicate it.

That sounds devastating for the defense. Until you look at the actual words he used.

The Gap Between "Probably" and "Beyond a Reasonable Doubt"

Throckmorton didn't eliminate the signature. He didn't reach "strong probability." He landed at "probably did not." That's the middle of his own field's standardized scale of conclusions. Not the strongest opinion. Not the second strongest. The third.

Why? Because he was working from a PDF. Not the original document. Not even a photocopy. A PDF, which he himself described as worse than a photocopy for his purposes. PDFs pixelize when you zoom in. Fine details disappear. The finer characteristics he relies on to make definitive calls get lost in digital compression.

So the prosecution's forgery evidence comes from an expert who, by his own methodology, couldn't reach his own field's standard for certainty because the evidence he was given wasn't good enough.

Now think about what the jury has to do with that. They don't convict on "probably." They convict on "beyond a reasonable doubt." That's the highest standard in the American legal system. It exists specifically to prevent the government from taking someone's freedom based on likelihood instead of proof.

The Certification Problem

Nester went somewhere the prosecution clearly hoped she wouldn't. She asked Throckmorton about certification.

There are multiple accredited organizations that certify forensic document examiners. They test candidates. They conduct peer review. They run oral boards. They provide a layer of independent verification that an examiner's work meets professional standards. Throckmorton listed off several of these organizations by name. He knows they exist. He knows what they do.

He has never pursued certification from any of them. Not in 30 years.

That doesn't mean he's incompetent. It doesn't mean his analysis is wrong. Certification isn't required to testify. The judge already ruled he's qualified as an expert. But when the defense stands up in closing arguments and says "the state's handwriting expert had three decades to get independently verified and never once did it," that's a data point the jury gets to weigh. And it's the kind of thing that sticks.

The Science Itself Is Under Fire

Nester didn't stop at the man. She went after the field.

The 2009 National Research Council report concluded that the scientific basis for handwriting comparison needs to be strengthened. It found limited research quantifying the reliability and replicability of the practices used by trained document examiners. The PCAST report raised concerns about validity and reliability. A NIST expert working group found multiple problems with the reliability of handwriting analysis.

These aren't defense attorney talking points. These are federal scientific advisory bodies. And Throckmorton acknowledged them on the stand.

He also confirmed the error rate. Trained, fully qualified forensic document examiners get it wrong approximately 4% of the time. Four out of every hundred cases. That's not zero. And when you're asking a jury to convict someone of forgery, which is one of the charges Kouri Richins faces, the jury has every right to ask whether this expert's opinion falls in the 96% or the 4%.

On redirect, the prosecution landed one clean point: PCAST, the 2016 report, did not specifically address forensic document examination. That's fair. But the NRC report did. And the NIST working group did. And the error rate studies did. One report not mentioning your field doesn't erase three others that did.

What He Can't Tell the Jury

There's a hole in this testimony that Nester walked right through. Throckmorton can tell the jury the signature probably wasn't Eric's. He cannot tell them who signed it.

In a simulated forgery, the writer gives up their own handwriting habits to imitate someone else's. That's the whole point. Copy, draw, duplicate. And in doing so, they hide their own identity. The examiner can exclude the person it's supposed to be, but the actual author's characteristics are buried under the imitation.

So the prosecution has an expert who says Eric probably didn't sign it, but can't say Kouri did. Or anyone else, for that matter.

And then Nester planted a seed that could grow into something bigger. She asked Throckmorton whether he had any knowledge of whether Kouri had permission from Eric to sign his name. He didn't. It wasn't part of his examination. He was never asked.

If the defense can establish that Kouri routinely signed Eric's name on documents with his knowledge, a signature that doesn't match Eric's handwriting stops being evidence of forgery and starts being evidence of a wife doing paperwork. That question is now in the jury's mind, and Throckmorton's testimony can't answer it.

What This Means for the Forgery Charge

The prosecution is nearing the end of its case. Throckmorton is likely the only forensic document examiner the jury will hear from on the state's behalf. This is the evidence they have on forgery.

An uncertified expert. Working from a PDF. Reaching a "probably did not" conclusion. Unable to identify the actual author. With no knowledge of whether the defendant had permission to sign. In a field where federal scientific bodies have questioned the reliability of the methodology.

Is that enough?

I'm not the jury. Neither are you. But here's what I know from watching trials for years, and from growing up watching my father fight cases where the system tried to convict on less than it should: "probably" is a dangerous word when freedom is on the line. The burden of proof exists for a reason. It doesn't bend because the prosecution's evidence wasn't as strong as they needed it to be.

The state has to prove forgery beyond a reasonable doubt. Not "probably." Beyond a reasonable doubt. The jury will decide whether Throckmorton's testimony clears that bar. But after what Nester did on cross, at least they'll know what questions to ask before they get there.

▶ WATCH THE FULL TESTIMONY Expert Says Eric Richins' Life Insurance Signature Was a Simulated Forgery