What the 386-Case Study Teaches About How Brady Material Gets Lost

The largest Brady database ever assembled tells a story most defenders already suspect but rarely see in numbers. When you study 386 confirmed Brady violations from 2004 to 2022, the pattern isn't a rogue prosecutor hiding a folder in a drawer. It's a system that loses evidence before anyone makes a decision about it.

According to the study by Professor Jennifer Mason McAward, 45% of the suppressed evidence could have impeached a critical prosecution witness. Another 34% of cases involved withheld material that both supported innocence and undermined the prosecution's case theory or a witness's credibility. Impeachment is where the volume lives.

The more important number for your practice is this one: in 51% of successful claims, the Brady material never reached the line prosecutor at all. Not buried. Never delivered.

That reframes the problem. The "prosecution team" under Brady is not just the trial ADA. It includes police officers, crime labs, fire marshals, and federal agencies. When a detective's bodycam contradicts his report and nobody forwards the footage, the ADA can be genuinely surprised at trial while your client sits in prison. Professor McAward describes the team as "just a collection of parts that do not work well together."

Where the breakdowns happen, per the same research:

• Law enforcement holds 31% of violations, and in 43% of the good-faith ones, inadequate training and resources were at fault.
• Caseloads make oversight nearly impossible: Cook County prosecutors carry 300 or more open cases at a time, and Fort Worth misdemeanor prosecutors handle 1,200 to 1,500 matters each.
• Large offices coordinate with sprawling agency networks. The Los Angeles prosecutor's office works with nearly 100 law enforcement agencies.
• Training is thin. Professor Adam Gershowitz reports that most prosecutors receive minimal Brady training, and many offices run a "once and done" session at onboarding.

So stop hunting for bad faith as your opening move. Assume systemic failure first, and treat every case as a candidate for accidental suppression. That means a repeatable discovery-review checklist rather than an ad hoc read-through of the dump.

Practically, that shifts your target list. Start with impeachment material, because that's where suppressed evidence concentrates. Prior inconsistent statements, officer discipline findings, and report-versus-bodycam discrepancies are the highest-yield items to chase. A structured Brady list research workflow for officer and witness records maps directly onto this finding.

It also means serving your demands on the whole team, not just the ADA who shows up at the calendar call. Name the lab, the fire marshal, the federal task force. Document what you asked each of them for and when. If the ADA never saw the evidence, your log of who was asked and who went silent becomes the record you argue from.

Evidence Copilot supports that kind of disciplined review on the defense side. When discovery arrives as bodycam, jail calls, and scanned reports, the platform surfaces candidate inconsistencies as unreviewed review signals before you even ask, and every AI answer cites the exact second of video or page of the PDF it came from. You still make the judgment calls. The system just makes sure the gaps don't hide.

The Legal Standard You're Actually Arguing Under: Brady, Bagley, and Kyles

Most defenders can recite Brady v. Maryland, but fewer can state the standard they'll actually be arguing under in 2026. That gap matters, because the doctrine has shifted since 1963 and the empirical reality on the ground has shifted even faster.

Start with the basics. A Brady violation occurs when the prosecution suppresses evidence that is both favorable to the defense and material to guilt or punishment. Two later cases reshaped how you litigate it. Under United States v. Bagley, no pretrial request is required: the prosecution has a constitutional duty to disclose all material, favorable information whether or not the defense asks for it. And the duty is breached regardless of whether the withholding was intentional or unintentional, which means you don't need to prove bad faith to win the motion.

Then there's Kyles v. Whitley, which governs how you frame materiality. Under Kyles, all undisclosed information must be considered collectively, not item by item. The standard is a "reasonable probability" that the outcome would have differed, and the defendant merely must show the withheld information can reasonably be taken to put the whole case in a different light. That's not a sufficiency-of-evidence test. Practically, it means you should keep a running log of every suspected non-disclosure, even the small ones, so you can aggregate them at motion or trial instead of arguing each scrap in isolation.

The enforcement reality is thinner than the doctrine. According to the Innocence Project, via the Baltimore Sun, only five prosecutors have been disbarred for a Brady violation nationwide since the rule was established, even as withheld-evidence cases have cost Maryland millions of dollars. You cannot count on professional discipline to surface violations for you. Contemporaneous documentation of what you requested, when, and from whom is the practical remedy.

Here's the part that should reshape your workflow: the data shows suppression is usually a team failure, not a lone prosecutor's bad act. In a study of 386 Brady violations from 2004 to 2022, in 51% of successful claims the Brady material never made its way to the line prosecutor at all. The "prosecution team" extends beyond the DA's office to police agencies, crime labs, fire marshals, and federal partners. Your discovery demands and follow-ups should name those entities explicitly.

• Impeachment evidence dominates: 45% of suppressed evidence could have impeached a critical prosecution witness.
• In 34% of cases, withheld evidence both supported innocence and undermined the prosecution's case theory or witness credibility.
• 49% of violations occurred in homicide cases, where the stakes and the bad-faith rate are highest.

That's why the highest-yield review targets are officer and witness credibility records: prior inconsistent statements, discipline findings, and report-versus-bodycam discrepancies. A structured Brady list research workflow for officer and witness records maps directly onto those findings, and tools like Evidence Copilot exist to make that kind of grounded, citation-backed review repeatable across a discovery dump. The doctrine is on your side. The system's enforcement isn't, so build the record yourself.

Build the Defense-Side Brady File: Impeachment Records First

Start with the credibility records, not the exculpatory file. The empirical case for this ordering is straightforward: in the largest Brady violation database ever assembled (386 violations from 2004 to 2022), 45% of suppressed evidence could have impeached a critical prosecution witness, and another 34% both supported innocence and undermined the prosecution's case theory or witness credibility. The study's authors found impeachment material dominates the data, so that's where your review hours should go first.

Practically, that means your Brady file opens with three categories for every officer and key witness on the state's list:

• Prior inconsistent statements, including earlier police reports, grand jury testimony, and statements to other agencies that diverge from the trial narrative.
• Discipline findings and documented dishonesty, which is the substance of official Brady lists and the target of any Giglio motion you plan to file.
• Report-versus-bodycam discrepancies, where the written narrative and the recorded events tell different stories about consent, the stop's duration, or what was said.

That third category is where you can generate your own impeachment material without waiting for anyone to hand it over. When the report says the defendant consented to a search and the bodycam captures hesitation, a timestamped transcript of the footage gives you a quotable, citable contradiction. Tools like Evidence Copilot surface these candidate inconsistencies automatically on ingest, comparing each new file against earlier evidence in the case, though every signal stays flagged for attorney review before it becomes anything more.

The second structural decision is who you serve. In 51% of successful Brady claims, the material never reached the line prosecutor at all, because the "prosecution team" extends well past the trial ADA to police agencies, crime labs, fire marshals, and federal partners. Professor Jennifer Mason McAward describes that team as "just a collection of parts that do not work well together," which is exactly why a demand addressed only to the assigned prosecutor leaves the biggest gaps untouched. Name each entity and each custodian in your demand, and keep proof of what each was asked for and when.

Third, log every suspected non-disclosure as it happens, even the ones that look trivial. Under Kyles v. Whitley, all undisclosed information must be weighed collectively, not item by item, against a "reasonable probability" standard. A running log of small gaps, dated and documented, is what lets you aggregate them into a collective-materiality argument at motion or trial. Itemized, each gap looks dismissible. Together, they can put the whole case in a different light.

Don't count on discipline to do this work for you. Only five prosecutors have been disbarred for a Brady violation nationwide since the rule was established, so the contemporaneous record you build is the practical remedy, not a professional-conduct board.

A Repeatable Review Checklist for Every Case

You cannot catch a Brady violation if you do not know what is missing. The prosecution team is a "collection of parts that do not work well together," according to peer-reviewed research.

In 51% of successful Brady claims, the material never reached the line prosecutor. It got lost in the gap between the police department, the crime lab, and the DA’s office. You cannot rely on the state to self-report these gaps.

Treat every case as a candidate for accidental suppression. You need a repeatable review routine that flags inconsistencies before you build your motion. This workflow ensures you catch the 45% of cases where suppressed evidence could have impeached a critical witness.

Here is a practical checklist for your discovery review:

• Map the prosecution team. Identify every agency involved (police, state lab, federal partners). In 31% of violations, law enforcement was responsible for withholding evidence. Serve demands on each entity explicitly.
• Cross-reference reports and bodycam. Compare the narrative in the police report against the bodycam audio. Look for timing discrepancies, omitted actions, or altered quotes. This is where impeachment material often hides.
• Log every non-disclosure. Keep a running log of missing items. Under Kyles v. Whitley, you must weigh all undisclosed information collectively. Small gaps add up to materiality.
• Verify witness credibility records. Check for prior inconsistent statements or discipline findings. With 45% of suppressed evidence involving witness impeachment, this is your highest-yield target.

Running this manually is tedious. You are scrubbing hours of video and reading dozens of PDFs. A grounded evidence workspace automates the grunt work. It generates timestamped transcripts and flags report-vs-bodycam inconsistencies automatically.

Evidence Copilot helps you build this record. The system compares new uploads against existing evidence and surfaces candidate inconsistencies as review signals. You get a case brief that lists key moments and discrepancies, ready for your attorney review.

You can see how this workflow handles officer records in the Brady list research workflow. It maps the specific data points needed for impeachment motions.

If you want to see how this fits into your current practice, book a 20-minute walkthrough. We will show you the checklist in action without any pressure.