The PATTERN Risk Score

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

What is a PATTERN score and why does it matter?

PATTERN is the Bureau of Prisons’ recidivism risk tool. It sorts every sentenced federal prisoner into minimum, low, medium, or high risk. The level does not affect whether you earn First Step Act time credits — but minimum or low risk across the last two assessments is what lets the Bureau apply them to a halfway house, home confinement, or early supervised release.

Key takeaways on the PATTERN risk score

• PATTERN exists because 18 U.S.C. § 3632(a) required the Attorney General to build a risk and needs assessment system that classifies each prisoner as “minimum, low, medium, or high risk for recidivism.”
• The current version is PATTERN v1.3, and BOP publishes the scoring forms and the level thresholds openly. You do not have to guess at the inputs.
• Roughly two thirds of a typical male general score comes from items that can never change — age, criminal history points, offense of conviction, history of violence, history of escapes. The rest is earned back through programming, education, drug treatment, work, and clear conduct.
• Risk level is the gate on applying first step act time credits, not on earning them. Medium and high risk prisoners earn credits and cannot spend them without a Warden’s petition.
• You are entitled to reassessment. 18 U.S.C. § 3632(d)(5) requires reassessments “not less often than annually,” and more frequently for medium or high risk prisoners within five years of release.
• Only designated program completions reduce a score. BOP Program Statement 5405.01 (7 May 2026) — agency policy, not law — limits score reduction to programs identified in the FSA Approved Programs Guide.

What is the PATTERN risk score?

PATTERN stands for the Prisoner Assessment Tool Targeting Estimated Risk and Needs. It is an actuarial instrument: a fixed list of scored items, each with defined point values, summed to a total, with the total mapped to one of four levels. It is not a clinical judgment, and no individual staff member sets it.

It exists because Congress ordered it built. 18 U.S.C. § 3632(a) directed the Attorney General, in consultation with an Independent Review Committee, to “develop and release publicly on the Department of Justice website a risk and needs assessment system,” which “shall be used to — (1) determine the recidivism risk of each prisoner as part of the intake process, and classify each prisoner as having minimum, low, medium, or high risk for recidivism.”

That same subsection carries an instruction that gets overlooked and is worth holding onto. The System must reassign prisoners to programming based on revised determinations so as to ensure “all prisoners at each risk level have a meaningful opportunity to reduce their classification during the period of incarceration.” The statute does not treat the score as a verdict. It treats it as something the prisoner is entitled to a real chance to change.

PATTERN is one half of a two-part system. The Bureau’s newest program statement on the subject, PS 5405.01, describes the Risk and Needs Assessment System (RNAS) as comprising “the Prisoner Assessment Tool Targeting Estimated Risk and Needs (PATTERN) recidivism risk instrument and Standardized Prisoner Assessment for Reduction in Criminality (SPARC) criminogenic needs assessments.” PATTERN measures risk. SPARC-13 measures thirteen need areas — substance abuse, education, work, family, trauma, mental health, and so on — and drives which programs get recommended — including the substance-abuse programming that leads to rdap eligibility.

The two interact. SPARC-13 identifies what to work on; completing designated programming in those areas is what moves PATTERN.

The four levels, and the numbers behind them

BOP publishes the thresholds. The cut points for PATTERN v1.3 are:

Tool	Minimum	Low	Medium	High
Male — General	5 or less	6 to 39	40 to 54	55 or more
Male — Violent	7 or less	8 to 24	25 to 31	32 or more
Female — General	7 or less	8 to 38	39 to 52	53 or more
Female — Violent	1 or less	2 to 11	12 to 17	18 or more

Source: Federal Bureau of Prisons, “Cut Points Used for PATTERN v. 1.3.”

Two things are worth noticing about that table. There are separate male and female instruments, scored differently on the same items. And each person receives two scores — a general recidivism score and a violent recidivism score — with different thresholds. A person can sit at low on one and medium on the other, and the consequences of that are not intuitive. Ask which score is driving a classification before assuming you know.

Notice also how wide “low” is: on the male general tool it runs from 6 to 39, a 34-point band. A person at 38 and a person at 7 are both “low,” and the practical difference between them is enormous in terms of how much cushion they have before a single incident report moves them.

Which PATTERN items can change, and which cannot

This is the part of the subject nobody writes about, and it is the only part a person in custody can act on.

BOP publishes the scoring instrument itself. The following is drawn from the Bureau’s own Male PATTERN Tool, version 1.3; the female instrument uses the same architecture with different weights.

PATTERN v1.3 scored items: static versus dynamic

Item	Static or dynamic	Direction	What it means in practice
Current age	Static (but improves with time)	Points fall as age rises — 35 points under 26, 0 points over 60 on the male general tool	The single heaviest item. It cannot be acted on, only waited out
Walsh Act conviction	Static	Fixed	Determined by the conviction
Violent offense (PATTERN designation)	Static	Fixed	Determined by the offense of conviction
Criminal history points	Static	Fixed — up to 40 points on the male general tool	Carried from the presentence report
History of escapes	Static, decays	Weight falls as the escape recedes in time	Recency matters; nothing else does
History of violence	Static, decays	Weight falls with time since the incident, and with severity	Same structure — recency and seriousness
Education score	Dynamic	−1 enrolled in GED; −2 high school diploma or GED earned	Enrolling helps immediately; finishing helps more
Drug program status	Dynamic	−2 non-residential drug program complete; −4 RDAP complete; −6 assessed as no need	The largest single point swing available
All incident reports (120 months)	Dynamic	0 to 3 points, rising with the number	Every incident report costs
Serious incident reports (120 months)	Dynamic	0 to 3 points, rising with the number	Scored separately from and on top of the item above
Time since last incident report	Dynamic	3 points under 3 months, falling to 0 at 12+ months or none	A clean year erases this item entirely
FRP refuse status	Dynamic	+2 on the male general tool if refusing	Refusing the financial responsibility program costs points and separately stops time-credit earning
Programs completed	Dynamic	Reduces the score	Only designated completions count — see below

Source: Federal Bureau of Prisons, “Male PATTERN Risk Scoring,” version 1.3. Point values shown are from the male general and violent columns of that form; the female instrument scores the same items on a different scale.

The structure of that table is the message. Age and criminal history are, on the male general tool, capable of contributing 75 points between them — well past the 55-point high-risk threshold on their own. The dynamic items are worth a good deal less. That asymmetry is real, and pretending otherwise sets families up for disappointment.

What it means practically: a person with heavy criminal history points and a young age has a long way to travel and should start immediately, because the dynamic items accumulate slowly and the reassessment schedule is what it is. A person already at the low end of “low” can lose that position quickly — a single incident report scores on three separate items at once, because it counts in the incident-report tally, potentially in the serious-incident tally, and resets the time-since-last-incident item to its maximum.

“Only designated program completions may reduce risk scores”

The most important operational sentence about lowering a PATTERN score is in PS 5405.01 (7 May 2026), the Bureau’s current First Step Act program statement — an internal agency policy document, not law:

“Participation in EBRRs and PAs is strongly encouraged and contributes to meaningful personal growth. Only designated program completions may reduce risk scores; these are identified in the FSA Approved Programs Guide and listed on the FSA pages of the Bureau’s public and intranet sites with additional scoring details.”

The same document is equally direct about what does not move: “other factors, such as the offense of conviction and criminal history remain the same despite program completions.”

So the honest advice is narrower than “take every class.” It is: find out which programs are designated as score-reducing in the FSA Approved Programs Guide, find out which of them address a need area on your own SPARC-13 assessment, and get on those lists. Programming outside that intersection may be worth doing for its own sake — it is not going to move the number.

The Bureau’s own First Step Act FAQ also states the prioritization rule, which cuts against intuition: “Priority for participation in recidivism reduction programs will be given to medium-risk and high-risk prisoners, with priority access to productive activities given to minimum-risk and low-risk prisoners.” Someone at high risk who wants a seat in an intensive program is, on paper, near the front of that queue — which matters, because those completions are what earn first step act time credits.

Applied Insight — Christopher Zoukis, JD, MBA, Managing Director: The drug program item is the largest single lever on the instrument and it is routinely misread. On the male general tool, completing RDAP is worth four points off; being assessed as having no substance abuse need is worth six. Those are different things, and the second is not something you achieve — it is something the assessment finds. The practical point is that if a substance abuse need is coded and the person has not been placed in treatment, that gap is costing points on PATTERN, blocking a § 3621(e) benefit, and leaving a need area open on SPARC-13 all at once. It is worth chasing as a single problem rather than three.

How PATTERN controls First Step Act time credits

Risk level does two separate jobs in the time-credit system, and conflating them is the most common mistake we see.

It does not gate earning the base rate

18 U.S.C. § 3632(d)(4)(A)(i) gives every eligible prisoner in earning status “10 days of time credits for every 30 days of successful participation.” That clause says nothing about risk level. A person at high PATTERN risk who is participating in recommended programming earns at that rate.

It does gate the extra five days

Clause (ii) awards “an additional 5 days of time credits for every 30 days” to a prisoner “determined by the Bureau of Prisons to be at a minimum or low risk for recidivating, who, over 2 consecutive assessments, has not increased their risk of recidivism.”

The implementing regulation states the second half of that test differently. 28 C.F.R. § 523.42(c)(2) requires that the prisoner “(i) Is determined by the Bureau to be at a minimum or low risk for recidivating; and (ii) Has maintained a consistent minimum or low risk of recidivism over the most recent two consecutive risk and needs assessments conducted by the Bureau.”

“Has not increased their risk” and “has maintained a consistent minimum or low risk” are different requirements, and the difference bites in exactly one situation: a person reassessed downward from medium to low. Under the statute’s language they have not increased their risk. Under the regulation’s language they have not maintained a consistent minimum or low risk across both assessments. We flag the divergence and quote both. It is the kind of thing worth raising in a BOP sentence computation dispute rather than assuming away.

It absolutely gates applying credits

This is where the score becomes the whole ballgame. 28 C.F.R. § 523.44(c) permits application of earned credits toward prerelease custody only when the prisoner has “(1) Maintained a minimum or low recidivism risk through his or her last two risk and needs assessments; or (2) Had a petition to be transferred to prerelease custody or supervised release approved by the Warden.”

And § 523.44(d) adds a further requirement for early transfer to supervised release: minimum or low risk “through his or her last risk and needs assessment.”

The consequence is stark. A medium-risk prisoner can accumulate a full year of credits and be unable to use a day of them. The credits are not lost — they sit in the balance — but they buy nothing until the level changes or a Warden’s petition succeeds. What they would otherwise buy is an earlier residential reentry center or federal home confinement placement, and up to a year off the transfer to supervised release. The placement itself runs to 12 months in a halfway house under 18 U.S.C. § 3624(c)(1), or 180 days or 10 percent of the sentence, whichever is shorter, on home confinement under § 3624(c)(2) — which is a great deal of time to be unable to reach because of a risk level.

Bureau policy also confirms that the projection itself is withheld. PS 5410.01 CN-2 states: “Medium and High PATTERN risk inmates may earn FTC, but will not receive an estimated FSA PRD.” Families searching for a projected First Step Act date that does not exist are usually running into this rule, not a clerical failure.

How PATTERN level changes what the Bureau may do

PATTERN level at the relevant assessments	Earn the base 10 days?	Earn the extra 5?	Apply credits to prerelease custody?	Apply credits to early supervised release?	Estimated FSA projected release date issued?
Minimum, two consecutive assessments	Yes	Yes	Yes	Yes, if the other § 523.44(d) conditions are met	Yes
Low, two consecutive assessments	Yes	Yes	Yes	Yes, if the other § 523.44(d) conditions are met	Yes
Minimum or low, but not yet two consecutive	Yes	Not until the condition is met	Only on a request submitted at program review to the Warden	Not yet	Yes
Medium	Yes	No	Only on a Warden-approved petition under § 523.44(c)(2)	Only after the above	No
High	Yes	No	Only on a Warden-approved petition under § 523.44(c)(2)	Only after the above	No

Sources: 18 U.S.C. § 3632(d)(4)(A); 28 C.F.R. § 523.42(c), § 523.44(c)–(d); BOP Program Statement 5410.01 CN-2 (10 March 2023) — agency policy, not law.

The Warden-petition route exists, and BOP policy tells you how narrow it is. PS 5410.01 CN-2 directs the Warden to consult the Regional Director, and lists people ordinarily considered inappropriate: those at medium or high risk; those with a current or prior offense in the Bureau’s Categorization of Offenses program statement; those with an unwaived Public Safety Factor; those with a history of supervision non-compliance; and those found to have committed 100- or 200-level prohibited acts during the current term. Medium and high applicants must also show clear conduct for at least three years from the date of the request and completion of at least one residential evidence-based recidivism reduction program in the past five years. The request itself goes in on a BP-A0148, Inmate Request to Staff, at a regularly scheduled program review.

When is PATTERN reassessed?

The statute sets a floor. 18 U.S.C. § 3632(d)(5) provides that a prisoner “who successfully participates in evidence-based recidivism reduction programming or productive activities shall receive periodic risk reassessments not less often than annually, and a prisoner determined to be at a medium or high risk of recidivating and who has less than 5 years until his or her projected release date shall receive more frequent risk reassessments.”

BOP implements that through the program review cycle. PS 5405.01 provides that “[i]nmate risk scores are reassessed in accordance with the Program Statement Unit Management and Inmate Program Review,” and that during those reviews “the Unit Management team utilizes the Insight application to reassess the recidivism risk of each inmate.” The policy identifies what usually drives a change: “Ordinarily, changes are a result of the inmate’s age, completed programs, and incident reports.”

Two timing rules from elsewhere in the system matter as well.

The initial assessment is supposed to happen within 28 days. PS 5410.01 CN-2 states that “[o]rdinarily, the Initial Classification is completed within 28 days of the inmate’s arrival, but completing the PATTERN tool requires the inmate’s sentence computation to be completed” — and if the computation is incomplete, the classification is delayed and the reason annotated on the Program Review Report. That dependency is a common source of delay, and it is documented.

For people already in prerelease custody, reassessment changes shape. Under the February 2023 change notice to PS 5410.01, those “not subject to regularly scheduled program review, will not receive reassessments. For inmates in prerelease placement, reassessments will be completed automatically on a monthly basis.” And the same document freezes the level at transfer: once transferred to prerelease custody, “the inmate will maintain the recidivism risk level the inmate had at the time of the transfer, unless the inmate benefits from a lower recidivism risk level based on the passage of time or the inmate’s actions result in a higher risk rating.”

The Bureau is behind on its own deadlines, and GAO measured it

In GAO-26-107268, published 27 January 2026, the Government Accountability Office reported that “BOP did not conduct all assessments within required time frames (28 days for initial and 90 or 180 days for reassessments) for various reasons, including technology issues.” Specifically, “BOP conducted initial risk assessments within required time frames for about 75 percent of the 57,902 incarcerated people who entered a BOP facility from June 1, 2022, to March 30, 2024,” and completed 69 to 95 percent of that cohort’s needs assessments on time depending on the need area.

GAO also found the underlying programming data unreliable: “BOP does not have accurate program data because, for example, staff used different methods to record when an incarcerated person declined to participate in a recommended program,” and “GAO also found inaccuracies in program participation data, which BOP officials attributed to data entry errors.”

If your PATTERN reassessment is late, or a completed program is not reflected in it, you are not imagining a problem that does not exist. GAO has documented both at scale, and BOP concurred with all six of GAO’s recommendations.

What has DOJ itself said about PATTERN?

The First Step Act required the Department of Justice to review and validate the system annually and to release the results publicly. It has done that four times. BOP publishes all four on its First Step Act reports page:

• 2020 Review and Revalidation (published January 2021)
• 2021 Review and Revalidation (published December 2021)
• 2022 Review and Revalidation (published March 2023)
• 2023 Review and Revalidation (published August 2024)

A separate 2023 Review and Validation of SPARC-13 followed in September 2024.

As of 2 September 2026, the most recent PATTERN revalidation published by the Bureau is the August 2024 report. No 2024 or 2025 revalidation appears on that page. The Attorney General’s separate § 3634 annual report series likewise stops at the June 2024 update. Some of that is explained by statute: GAO reported in January 2026 that the § 3634 reporting requirement “expired in 2025, and absent congressional actions, DOJ no longer has to submit a report to Congress,” and recommended that Congress consider amending 18 U.S.C. § 3634 to extend it.

The substance of the most recent review is not comfortable reading for the tool. The National Institute of Justice’s own abstract for the August 2024 report (NCJ 309264) states:

“While the findings continue to indicate PATTERN’s accuracy across the five racial and ethnic groups analyzed, there remains evidence that the instruments predict differently across those groups, including overprediction of risk of Black, Hispanic, and Asian males and females, relative to White individuals, on the general recidivism tools.”

That is the Department’s own assessment, in its own words, in the report the statute required it to produce. NIJ has since marked the publication page with an archival notice indicating it is no longer being updated.

We state this because it is true and sourced, not as an argument that any individual score is wrong. A documented pattern of overprediction across a population is not, by itself, a claim about a particular person’s file. But it is relevant context for anyone being told that the number is objective and therefore beyond argument.

Can you challenge a PATTERN score?

You can challenge the inputs. Challenging the judgment is much harder, and the reason is statutory.

What is worth challenging

PATTERN is arithmetic performed on recorded facts. If a recorded fact is wrong, the arithmetic is wrong — and because the level gates the application of credits, an error in a scored item is an error about liberty, corrected through the bop sentence computation channels. The items that produce genuine, correctable errors:

Criminal history points. These come from the presentence report, the same document that later drives rdap eligibility. If the presentence report miscounted, PATTERN inherits the error, and it inherits it as one of the heaviest items on the instrument.

Violent offense designation. BOP publishes its PATTERN Violent Offense Codes. Whether an offense of conviction falls inside that list is a checkable question, not a matter of impression.

History of violence and escapes. Both are scored by recency and severity. A stale entry scored as though it were recent, or a minor incident scored as serious, is a specific and fixable error.

Incident reports. If a report was expunged on appeal, or the DHO finding was overturned, the item should reflect that. It does not always. And because a single incident report touches three separate scored items, an expungement that never propagated can be worth several points.

Program completions that were never entered. GAO found this problem is systemic. A completion that exists on paper but not in the system is doing nothing for the score.

Drug program status. Coded as a need with no treatment placement, or a completed program not credited — both are common, and both are worth several points.

How the challenge is made

Through the BOP administrative remedy program at 28 C.F.R. Part 542. The Bureau’s new First Step Act program statement says so explicitly: PS 5405.01 provides that “[i]nmates have the option to utilize the Program Statement Administrative Remedy Program for review of issues or concerns related to this program statement, as outlined in 28 C.F.R. § 542.10 and subsequent sections.”

In practice, the sequence starts before the formal remedy: an Inmate Request to Staff (BP-A0148) to the unit team, at or before a program review, identifying the specific item and the specific error, with the supporting document attached. Most correctable errors are corrected there, and doing it that way preserves the informal-resolution step the formal remedy requires anyway.

If that fails, the formal sequence runs BP-9 to the Warden within 20 calendar days of the event under 28 C.F.R. § 542.14, BP-10 to the Regional Director within 20 days of the Warden’s signature, and BP-11 to the General Counsel within 30 days of the Regional Director’s, under § 542.15. Our page on bop sentence computation covers the deadlines and the traps in detail.

Why the judgment itself is nearly unreviewable

18 U.S.C. § 3625 provides that the Administrative Procedure Act’s provisions at 5 U.S.C. §§ 554, 555 and 701 through 706 “do not apply to the making of any determination, decision, or order under this subchapter.” That removes the ordinary arbitrary-and-capricious challenge to an individual Bureau determination.

The practical translation: an argument that the Bureau weighed things badly, or that a reasonable assessor would have reached a different level, has almost no purchase. An argument that the Bureau scored a fact incorrectly, or applied its own published instrument incorrectly, or violated its own regulation, is a different kind of argument and survives in a way the first does not. Frame accordingly.

Applied Insight — Elizabeth Franklin-Best, Esq., Principal Attorney: The winning frame in this area is almost never “the score is unfair.” It is “the score is wrong, and here is the document that proves it.” A PATTERN challenge that arrives with the presentence report page showing the correct criminal history count, or the DHO expungement that never propagated, is a different filing from one that argues about the tool. Section 3625 does a great deal of work against the second kind and very little against the first. And every one of these has to be preserved through the administrative remedy process before a court will look at it — which is why the twenty-day clock in § 542.14 matters more than almost anything else in the file.

Working with a federal prison early release lawyer

A PATTERN problem is worth attention out of proportion to how technical it sounds, because the level is the gate on everything downstream: the extra five days of credit, the ability to apply credits at all, the residential reentry center referral, the federal home confinement placement, and, under the Bureau’s May 2026 camp initiative, the custody move that now follows a placement date — and, at the far end, the early termination of supervised release motion that a documented reduction in risk supports.

What we do on these files is unglamorous and effective: pull the scoring documentation, compare each scored item against the underlying record, identify the discrepancies, and pursue them on the timeline the regulations allow. Where the score is right and the level is simply high, the work shifts to the programming plan and, where appropriate, the Warden’s petition route.

Elizabeth Franklin-Best, P.C. is a boutique federal criminal defense and appellate firm founded in 2019, with offices in Columbia and Mount Pleasant, South Carolina, serving clients nationwide on an intentionally small caseload. Elizabeth Franklin-Best is admitted to the U.S. Supreme Court and twelve of the thirteen U.S. Courts of Appeals — every regional circuit (1st–11th) plus the D.C. Circuit. Christopher Zoukis, JD, MBA, leads the firm’s federal prison consulting work and is the author of the Federal Prison Handbook and the Directory of Federal Prisons. We do not promise outcomes, and we will tell you when a score is simply correct.

Schedule a consultation

If a PATTERN level is blocking the use of credits that have already been earned, or a scored item does not match the record, we can review it.

Schedule a consultation with Elizabeth Franklin-Best, P.C. Our practice is nationwide.

For the wider picture — how the credits, the placements, and the court-based routes fit together — start at our federal prison early release lawyer hub. For the credits themselves, see first step act time credits; for the separate stream that PATTERN does not touch, see good conduct time.

Frequently Asked Questions About the PATTERN Risk Score

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