2255 Motion Statistics: Filing Volumes, Outcomes, and Timing

By Elizabeth Franklin-Best, Esq., and Christopher Zoukis, JD, MBA

How often are § 2255 motions granted?

Nobody publishes a national § 2255 grant rate — the Administrative Office of the U.S. Courts counts motions filed, not motions granted. Verified filings ran 24,837 in FY2016, 7,205 in FY2017, and 3,763 in FY2023. Any “success rate” you see quoted from those numbers is a misuse of them.

Key takeaways on § 2255 filing statistics

• The only reliable national § 2255 series is a filing count. The Administrative Office of the U.S. Courts reports motions to vacate sentence commenced, by nature of suit, in Judicial Business of the United States Courts. It does not publish a clean “granted” field.
• One retroactivity decision moved the numbers more than any other event on record. Filings rose to 24,837 in FY2016 — up 19,317, roughly 350% — after Welch v. United States, 578 U.S. 120 (2016), made Johnson v. United States, 576 U.S. 591 (2015), retroactive on collateral review.
• The spike unwound almost as fast. FY2017 fell to 7,205, down 17,632 (−71%). By FY2021 filings were 3,395, and FY2023 was 3,763.
• District-level swings were enormous. In FY2016 the Southern District of Florida alone rose by 966 filings to 1,233. In FY2017 it fell by 839, to 394.
• A filing count is not an outcome count. It tells you how many people asked. It tells you nothing about how many were granted, how many were dismissed on preliminary review under Rule 4, or how long any of them took.
• The AO changed what it itemizes. From FY2024 onward the district-courts narrative no longer separately breaks out motions to vacate sentence; it reports habeas corpus petitions and prisoner petitions instead. The filings did not stop — the itemization did, and later figures must come from Table C-2 directly.

Where does this data come from?

From one agency, under one statutory duty, published in one report series.

The Administrative Office of the U.S. Courts compiles federal caseload statistics because Congress told it to. Under 28 U.S.C. § 604(a)(2), the Director of the Administrative Office must “[e]xamine the state of the dockets of the courts; secure information as to the courts’ need of assistance; prepare and transmit semiannually to the chief judges of the circuits, statistical data and reports as to the business of the courts.” Section 604(a)(3) requires an annual report to the Judicial Conference containing that statistical data, and provides that “report, data and recommendations shall be public documents”; § 604(a)(4) requires copies to Congress and the Attorney General. Section 604(b) closes the loop: court clerical and administrative personnel “shall comply with all requests by the Director for information or statistical data as to the state of court dockets.”

The product of that duty is Judicial Business of the United States Courts, published annually for each fiscal year (October 1 through September 30). Its “U.S. District Courts” section carries a narrative describing changes in filings by case type, and its accompanying data tables carry the underlying counts. The companion series, Federal Judicial Caseload Statistics, publishes the same universe of district-court data on a twelve-month period ending March 31 rather than the fiscal year — which is why a figure from one series will not always match a figure from the other, and why this page uses the fiscal-year series throughout.

Within those publications, motions under § 2255 appear in the civil caseload as “motions to vacate sentence,” a nature-of-suit category. The detailed counts live in the caseload data tables, where Judicial Business Table C-2 reports civil cases commenced by nature of suit. Every number on this page is drawn from the Judicial Business series for fiscal years 2016 through 2025 and was transcribed from uscourts.gov, not from any secondary description of it.

One structural fact governs everything below. The AO’s job under § 604 is to describe the business of the courts — how much work came in and how much went out. It is not an outcome-reporting agency. That is why a filing count is what exists, and why an honest page about § 2255 statistics has to begin by saying what the number is not.

How many § 2255 motions are filed each year?

Here is the verified national series. Read the “Change” column before the “Filed” column — the swings are the story.

Table 1. Motions to vacate sentence commenced in U.S. district courts, FY2015–FY2023, with the driver the Administrative Office itself noted.

Fiscal year	Motions to vacate sentence filed	Change from prior FY	Driver noted by the AO	Source
2015	~5,520 (derived — see note)	—	Baseline	Derived from the FY2016 report’s stated increase
2016	24,837	+19,317 (≈ +350%)	Welch v. United States made Johnson retroactive	Judicial Business FY2016
2017	7,205	−17,632 (−71%)	The Welch surge unwinds	Judicial Business FY2017
2018	5,342	−1,863 (−26%)	None noted beyond continued decline	Judicial Business FY2018
2019	5,877	+535 (+10%)	None noted	Judicial Business FY2019
2020	6,834	+957 (+16%)	None noted	Judicial Business FY2020
2021	3,395	−3,439 (−50%)	None noted	Judicial Business FY2021
2022	~3,392 (derived — see note)	—	None noted	Derived from the FY2023 report’s stated increase
2023	3,763	+371 (+11%)	None noted	Judicial Business FY2023

Source: Administrative Office of the U.S. Courts, Judicial Business of the United States Courts, “U.S. District Courts” section, fiscal years as noted; nature-of-suit basis (Judicial Business Table C-2). These are filings, not grants. Bolded years are figures reported by the AO. FY2015 and FY2022 are derived by subtracting the following year’s reported change from that year’s reported total, and are labeled as derived rather than reported. Because FY2022 is derived, the FY2023 percentage change is computed against a derived baseline. “Driver noted by the AO” records only causes the AO itself identified; where it identified none, this page does not supply one.

The arithmetic, shown

A maintained data asset should show its work, because a series assembled across nine annual reports is exactly where transcription errors hide. Every year-over-year transition from FY2016 through FY2021 reconciles: prior year plus the reported change equals the reported current-year total.

• FY2015 5,520 + 19,317 = 24,837 (FY2016)
• FY2016 24,837 − 17,632 = 7,205 (FY2017)
• FY2017 7,205 − 1,863 = 5,342 (FY2018)
• FY2018 5,342 + 535 = 5,877 (FY2019)
• FY2019 5,877 + 957 = 6,834 (FY2020)
• FY2020 6,834 − 3,439 = 3,395 (FY2021)
• FY2023 3,763 − 371 = 3,392 (FY2022, derived)

Two years carry an asterisk for a reason. FY2015 and FY2022 are not figures the AO stated as totals in the reports consulted; they are back-calculated from the following year’s stated change. They are internally consistent and useful for trend context, and they are labeled derived so that nobody re-publishes them as reported figures.

How often are § 2255 motions granted?

No one publishes that number nationally, and nothing on this page answers it. This is the single most important sentence on the page, and it is worth being blunt about why.

The Administrative Office’s series counts cases commenced. There is no companion field in the Judicial Business district-courts narrative that reports how many motions to vacate sentence were granted. Terminations are reported, but a termination is not a grant: the same column captures motions dismissed on preliminary review, motions denied on the merits, motions dismissed as untimely, motions transferred, motions withdrawn, and motions granted, without distinguishing among them.

So when you see a page confidently reporting that “§ 2255 motions succeed X% of the time,” ask where the numerator came from. If the answer is the AO series, there is no numerator — only a denominator. Presenting a filing count as a success rate is not a rounding error; it is a different statistic entirely.

What can be said honestly is structural, and it is not encouraging. Federal collateral review is built as a series of screens, each of which disposes of motions before any merits question is reached:

• Preliminary review. Under Rule 4, Rules Governing Section 2255 Proceedings, the assigned judge “must promptly examine” the motion and “must dismiss” it if “it plainly appears from the motion, any attached exhibits, and the record of prior proceedings that the moving party is not entitled to relief.” A large share of motions end here, before the government files anything.
• The one-year limitations period. Section 2255(f) runs from the latest of four trigger dates, and untimeliness disposes of motions without any look at the claim. See 2255 statute of limitations.
• Procedural default. A claim available on direct appeal and not raised there is barred absent cause and prejudice or actual innocence. See procedural default.
• The successive-motion gate. A second or successive motion requires authorization from a court of appeals panel on one of two narrow grounds. See second or successive 2255.
• The merits standards themselves. The most common § 2255 claim is ineffective assistance of counsel, and it is measured against a deliberately demanding two-prong test. See Strickland v. Washington.

None of that produces a percentage. It does explain why practitioners describe relief as uncommon without being able to quantify it — and why the responsible way to assess your odds is a record-specific evaluation rather than a national average. Which claims actually succeed, and why, belongs to grounds for a 2255 motion; whether a lawyer is worth retaining for the attempt is addressed at federal 2255 motion lawyer.

Applied Insight — Elizabeth Franklin-Best, Esq., Principal Attorney: When a family asks us “what are the odds,” the honest answer is that the national number they are hoping for does not exist — and that even if it did, it would be close to useless for their case. A § 2255 built on a documented conflict of interest, a suppressed report, or a plea offer that was never conveyed sits in a different universe from a motion re-arguing sufficiency of the evidence. What we can give is an assessment of the specific record: what the transcripts show, what the file is missing, whether the claim survives the procedural screens, and whether it presents a factual dispute the court cannot resolve on paper. That last question — whether the case can get to a hearing — is usually a better predictor than any published statistic.

Why did filings spike in FY2016, and why did they collapse after?

Because of two Supreme Court decisions, one year apart, and this is the clearest natural experiment in federal post-conviction law.

In Johnson v. United States, 576 U.S. 591 (2015), the Court held that “imposing an increased sentence under the residual clause of the Armed Career Criminal Act violates the Constitution’s guarantee of due process,” overruling James and Sykes. The residual clause had swept in any prior offense that “otherwise involves conduct that presents a serious potential risk of physical injury to another,” and it had been used for years to trigger the Act’s fifteen-year mandatory minimum. Striking it down potentially affected every federal prisoner whose sentence rested on it.

But a new constitutional rule does not automatically reach a conviction that is already final. That question was answered the following term in Welch v. United States, 578 U.S. 120 (2016): “Johnson is thus a substantive decision and so has retroactive effect under Teague in cases on collateral review.” The Court reasoned that “[b]y striking down the residual clause as void for vagueness, Johnson changed the substantive reach of the Armed Career Criminal Act, altering ‘the range of conduct or the class of persons that the [Act] punishes.'”

Welch was decided April 18, 2016 — inside fiscal year 2016. What followed is in the table: 24,837 motions to vacate sentence filed that fiscal year, an increase of 19,317 over the prior year, roughly 350%. Nothing else in the verified series comes close.

The collapse has the same explanation running in reverse. A retroactivity ruling creates a one-time, bounded population: the prisoners whose sentences rest on the invalidated provision, filing within the one-year window that § 2255(f)(3) opens from the date the right was initially recognized. Once that cohort filed, it was gone. FY2017 fell to 7,205 — down 17,632, or 71% — and the AO’s own narrative attributes the decline to the Welch surge unwinding. FY2018 fell again to 5,342, and by FY2021 the series reached 3,395, roughly a seventh of the FY2016 peak.

Read as a whole, the series is the most concrete available evidence for a proposition the rest of this section argues in prose: retroactivity is the single most powerful variable in federal post-conviction law. One decision about whether a new rule reaches final convictions moved national filings by nearly 20,000 in a year. If a recent Supreme Court decision looks like it helps your case, the retroactivity question is the one to answer first — see retroactive Supreme Court decisions. It also explains why the § 2255(f)(3) clock runs from the date the right was initially recognized, not from the later decision holding it retroactive, a distinction that has cost many filers their claim.

Which districts see the most filings?

The national totals conceal district-level swings that are proportionally far larger, and the verified detail below shows where the Welch wave broke hardest.

Table 2. Verified district-level movement in motions to vacate sentence, by fiscal year.

Fiscal year	District	Movement	Ending count
2016	S.D. Fla.	Largest growth nationally — up 966	1,233
2016	S.D. Tex.	Second largest — up 624	866
2017	S.D. Fla.	Largest decline nationally — down 839	394
2017	E.D. Pa.	Down 726	—
2018	D.P.R.	Largest reduction — down 194	75
2019	M.D.N.C.	Largest increase — up 188	341
2020	D. Md.	Largest growth — up 319	462
2021	D. Md.	Down 379	83
2021	S.D. Fla.	Down 173	147
2023	M.D.N.C.	Greatest rise — up 56	191

Source: Administrative Office of the U.S. Courts, Judicial Business of the United States Courts, “U.S. District Courts” section, fiscal years as noted. Filings, not grants. A dash means the AO narrative reported the change without the ending count.

The Southern District of Florida is the clearest illustration of the whole phenomenon. It gained 966 filings in FY2016, reaching 1,233 — and then lost 839 in FY2017, falling to 394. That is one district absorbing, and then shedding, a caseload larger than most districts see in a decade. The District of Maryland shows the same shape three years later on a smaller scale: up 319 to 462 in FY2020, then down 379 to 83 in FY2021.

Two cautions before anyone reads local strategy into this. First, filing volume tracks the size and composition of a district’s criminal docket — districts that charged large numbers of Armed Career Criminal Act cases had large numbers of Johnson claims. It does not track how receptive a district’s bench is. Second, these are the movements the AO narrative itself singled out as the largest for the year; they are not a ranked list of the busiest districts, and this page does not present them as one.

Applied Insight — Christopher Zoukis, JD, MBA, Managing Director: The most useful thing in this data for a family is not the national trend — it is the FY2016 spike, because it shows what actually moves outcomes at scale. Twenty thousand additional filings in a single year did not happen because prisoners became more motivated. They happened because a legal category changed and a large group of people suddenly had a claim they could not have brought the year before. That is the thing to watch for: not “is a § 2255 worth it in general,” but “has anything changed that gives this specific sentence a claim it did not have.” Most years, for most people, the answer is no. When the answer is yes, the window is one year from the date the right was recognized, and it closes quietly.

What this data does not tell you

Being precise about the limits is what makes the rest of the page usable. Six things this series does not measure:

1. Outcomes. No grant count, no denial count, no rate. The AO reports terminations, but terminations lump together summary dismissals, merits denials, untimeliness dismissals, transfers, withdrawals, and grants.
2. Time to disposition. The verified source for this page carries no median-interval figure for motions to vacate sentence, so this page prints none. How long a § 2255 actually takes, stage by stage, is treated at what happens after filing a 2255, which owns that question.
3. How many motions were summarily dismissed versus set for hearing. The AO does not break this out. The legal standard that governs it — a court must hold a hearing unless the motion and record “conclusively show that the prisoner is entitled to no relief” — is at 2255 evidentiary hearing.
4. Appeal rates or certificate-of-appealability grant rates. Not published in this series. The standard a movant must meet to appeal a denial is at certificate of appealability.
5. What kind of claim was raised. A nature-of-suit code records that a motion to vacate was filed. It does not record whether the claim was ineffective assistance, a suppressed-evidence claim under Brady, a jurisdictional attack, or a claim of actual innocence.
6. Anything about a specific case. A national aggregate is not a prediction. Two motions filed in the same district in the same week can have entirely different prospects depending on the record, the claim, and whether the movant clears the procedural screens.

There is also a boundary worth naming, because it produces double-counting in careless analyses. Motions to vacate sentence under § 2255 are counted separately from habeas corpus petitions under §§ 2241 and 2254. A federal prisoner challenging the execution of a sentence — sentence computation, credits, program eligibility — files under § 2241 in the district of confinement, and that filing lands in a different category. See 28 U.S.C. § 2241.

What happened to the data after FY2023?

The itemization changed. The filings did not stop.

From FY2024 onward, the Administrative Office’s district-courts narrative no longer separately breaks out “motions to vacate sentence.” The narrative reports habeas corpus petitions and prisoner petitions instead. That is a reporting-format decision, not a disappearance of the underlying data: motions under § 2255 continue to be filed, docketed, and counted by nature of suit. What changed is which figures the annual narrative highlights.

The practical consequence for anyone maintaining or citing this series is specific. Figures for FY2024 and later must be taken from Judicial Business Table C-2 directly, where civil cases commenced by nature of suit are reported, rather than from the narrative summary. This page will not carry a post-FY2023 figure until it has been read out of that table and cross-checked, because a maintained data asset that guesses once is not a data asset.

That is also the update rule for this page generally: it is rebuilt when the Administrative Office publishes the next Judicial Business report and the motion-to-vacate figures can be confirmed against Table C-2. No figure is added on the strength of a press release, a news summary, or another law firm’s page.

How does this compare to compassionate release?

The two cannot be compared numerically, and understanding why is more useful than any figure would be.

The § 2255 series is a filing count maintained by the Administrative Office under its § 604 duty to report the business of the courts. It counts cases commenced by nature of suit and has no outcome field. Compassionate release under 18 U.S.C. § 3582(c)(1)(A) is tracked differently: the United States Sentencing Commission data reports are the body of work that examines motions for sentence reduction, because the Commission’s statutory mission is sentencing policy, and outcomes are the point of the exercise. See the Commission’s published research at ussc.gov.

So a sentence like “§ 2255 motions succeed less often than compassionate release motions” is comparing an agency that publishes outcomes against an agency that does not — and reaching a conclusion the § 2255 data cannot support in either direction. This page will not make that comparison, and readers should be skeptical of pages that do.

There is one genuinely important connection between the two remedies, and it is legal rather than statistical. Fernandez v. United States, 608 U.S. 411, 146 S. Ct. 1292 (2026), holds that a prisoner attacking the validity of a conviction must proceed under § 2255 rather than § 3582 — and a pending compassionate release motion does not toll the § 2255 one-year period. People who spent 2025 and 2026 litigating conviction-validity arguments in the wrong statute are a real and growing population, and their § 2255 clocks kept running. That situation, and what to do about it, is covered at 2255 after compassionate release denial.

Why work with Elizabeth Franklin-Best, P.C.

Elizabeth Franklin-Best founded the firm in 2019 and practices federal criminal appeals and post-conviction relief nationwide, on an intentionally small caseload. She has handled more than 330 federal proceedings, over 100 of them appeals, with sixteen published decisions as counsel of record, and is recognized by Best Lawyers in America 2026 for Appellate Practice and by Chambers USA 2026 for Litigation: White-Collar Crime & Government Investigations. She is admitted to the U.S. Supreme Court and twelve of the thirteen U.S. Courts of Appeals, and is the author of Reversing Your Criminal Conviction: Develop Your Winning Strategy (2023).

We publish this page because the alternative — telling people relief is “rare” or “possible” and leaving it there — is not information. What a national filing series cannot do is evaluate a specific record, and that is the work: reading the transcripts and the plea colloquy, finding what the trial file is missing, testing the claim against the limitations period and the procedural-default rules, and deciding whether the facts can be pleaded specifically enough to require a hearing. If you are weighing whether to start, the mechanics of doing it correctly are at how to file a 2255 motion, and the restrictions AEDPA imposes on every one of these filings are at AEDPA.

Get a record-specific assessment

A national average cannot tell you whether your motion is worth filing. A review of your record can. Schedule a consultation or call (843) 620-1100. Bring the criminal judgment, the docket sheet, the plea agreement or trial transcripts if you have them, and the date of your last appellate ruling.

Frequently Asked Questions About § 2255 Success Rates

Reviewed for legal accuracy by Elizabeth Franklin-Best, Esq., Principal Attorney·September 2026