The question every bad-paper veteran asks first is the one nobody answers with a number: what are my chances? The boards don't publish win rates in any form a veteran can use, and the lawyers who quote success rates are quoting their own. In July 2025 the Government Accountability Office
did the count, across all six boards, for the cases that make up more than half of all upgrade applications: the ones involving PTSD, other mental health conditions, TBI, or sexual assault, where the boards are required to apply "liberal consideration." The numbers are the closest thing to an honest answer that exists.
The rates
Full grants, January 2018 through March 2024, in liberal-consideration cases,
as reported by Stars and Stripes from the GAO's data: the Army Discharge Review Board granted 29% in full. The Naval Discharge Review Board, which also hears Marine Corps cases, granted 27%. The Army Board for Correction of Military Records, 26%. Then the drop: the Board for Correction of Naval Records granted 13%, the Air Force Board for Correction of Military Records 10%, and the Air Force Discharge Review Board 8%.
Across the six boards, counting partial grants, the GAO's range was 18% to 49%. The boards closed 21,817 of these cases in the period: 10,237 Army, 9,941 Navy and Marine Corps, 1,639 Air Force.
Read those numbers twice. A soldier and an airman with the same diagnosis, the same misconduct, and the same evidence face a three-and-a-half-fold difference in their odds depending on which branch's seal is on the DD-214. That's not what the law says should happen.
The Kurta memo and the cases that enforced it set one standard for every board.
Why the numbers spread
The GAO's explanation is unglamorous. The boards applied the same guidance differently in two specific places.
The first was VA documentation. The Kurta memo tells boards that a VA finding of service connection for a mental health condition is strong evidence that the condition existed in service and may have contributed to the misconduct. Some boards treated a VA rating decision as close to dispositive on that point. Others treated it as one piece of paper among many and discounted it because the VA's standard is different from the board's. Same memo, opposite weight.
The second was testimony about sexual assault and harassment. The memo says a veteran's own statement can establish that the event occurred, because these events are underreported and the records are usually silent. Some boards took the statement as sufficient. Others looked for corroboration the memo doesn't require and denied when they didn't find it.
Layer on the rest of what the GAO found, decisions that didn't explain how the four Kurta questions were answered, boards with no required time frames, nearly half of decisions never posted to the public reading room, and a staffing decline against a more complex caseload, and the spread stops being a mystery.
What the GAO told the Pentagon to do, and what the Pentagon said
Nine recommendations: evaluate how the boards are applying the guidance and consider a single joint board; monitor the service boards periodically; set required time frames for the DRBs and the appeal board; post time estimates on the websites and in letters to applicants; answer the four Kurta questions explicitly in every decision; and fix the reading room. The Defense Department concurred with three, partially concurred with one, and did not concur with five, including the joint board, the monitoring, and the required time frames. All nine remain open.
The one the Department rejected that would have changed your odds most is the explicit-answers requirement. A decision that has to state, in writing, whether you had a condition, whether it existed in service, whether it excuses the misconduct, and whether it outweighs it, is a decision that can be checked. A decision that doesn't is the kind
a federal court sends back, and the kind you get more often from the boards at the bottom of the list.
How to use the numbers
Three ways.
Pick the board with the rates in mind. Within 15 years of discharge you can go to the DRB; after that, or for anything the DRB can't change, you go to the BCMR or BCNR. For an Army veteran inside the window, the DRB's 29% beats the ABCMR's 26% by a little and runs faster. For a sailor or Marine inside the window, the NDRB's 27% is double the BCNR's 13%, which makes the DRB the first stop almost every time and makes blowing the 15-year window expensive. For an airman, both boards are hard, and
the packet has to do more of the work.
Build the record against the two failure points. If you have a VA rating for the condition, lead with it and say what the Kurta memo says it proves, in those words, so a board that's inclined to discount it has to explain why. If your case involves sexual assault or harassment, your statement is evidence under the memo; say so, and add whatever corroboration exists anyway, because some boards will look for it.
Ask for the hearing. The GAO didn't break out hearing versus records-review rates, but
we've written about why a personal appearance tends to help, and in a system where boards are deciding the same facts differently, putting a face and a voice on the record is one of the few variables you control.

The number the GAO didn't count
Most veterans with bad paper never apply. The CRS put VA applications from less-than-honorable veterans at about 10,000 in 2023, against a population of hundreds of thousands. An 8% grant rate is a bad number. A 0% rate, which is what not filing gets you, is worse, and the boards with the lowest rates are the ones where a well-built case stands out most.
Where we come in
We're a veteran-led firm, and discharge upgrades across every branch's boards are core work for us. We know which board reads a VA rating as proof and which one needs to be told, and we build the case for the board you're in front of. If you've been sitting on an application because you didn't know your odds, now you do.
Reach out.
