HARDIMAN, Circuit Judge.

Roger Brown appeals an order of the District Court denying his petition for writ of habeas corpus. For the reasons that follow, we will vacate the order of the District Court and remand for discovery.

I.

Because we write exclusively for the parties, we will review only the facts essential to our decision.

While in prison for burglary, Brown was convicted of killing a man in the course of a previous robbery in the District of Columbia in 1978 and was sentenced in 1980. During his incarceration, Brown completed an apprenticeship as an electrician and received vocational training in building maintenance and construction, heating and air conditioning, and computers. He has no disciplinary infractions in the past 20 years, but he had five weapons and drug violations between 1981 and 1988.

In 2006, Brown's initial bid for parole was rejected by the United States Parole Commission. At his parole hearing, Brown refused to accept responsibility for his crime. In denying Brown parole, the Commission applied the current guidelines set forth in 28 C.F.R. § 2.80, rather than the guidelines that were in effect at the time of his offense. The guidelines in 1978 instructed the District of Columbia Board of Parole to consider:

(a) The offense, noting the nature of the violation, mitigating or aggravating circumstances and the activities and adjustment of the offender following arrest if on bond or in the community under any presentence type arrangement.

(b) Prior history of criminality noting the nature and pattern of any prior offenses as they may relate to the current circumstances.

(c) Personal and social history of the offender, including such factors as his family situation, educational development, socialization, marital history, employment history, use of leisure time and prior military experience, if any.

(d) Physical and emotional health and/or problems which may have played a role in the individual's socialization process, and efforts made to overcome any such problems.

(e) Institutional experience, including information as to the offender's overall general adjustment, his ability to handle interpersonal relationships, his behavior responses, his planning for himself, setting meaningful goals in areas of academic schooling, vocational education or training, involvements in self-improvement activity and therapy and his utilization of available resources to overcome recognized problems. Achievements in accomplishing goals and efforts put forth in any involvements in established programs to overcome problems are carefully evaluated.

(f) Community resources available to assist the offender with regard to his needs and problems, which will supplement treatment and training programs begun in the institution, and be available to assist the offender to further serve in his efforts to reintegrate himself back into the community and within his family unit as a productive useful individual.

D.C. MUN. REGS. tit. 9, § 105.1 (1972).

In 1987 the Board of Parole implemented a point-based scoring system to formalize the manner in which it exercised its discretion. Under the 1987 system, points were awarded in various categories which reflected the enumerated factors under the previous guidelines, and an inmate's total score determined his suitability for parole. The Board retained discretion in "unusual circumstances" to grant parole even when an inmate's score indicated that he should not be paroled. D.C. MUN. REGS. tit. 28, §§ 204.16-22 (1987).

In 1997 Congress abolished the District of Columbia Board of Parole and replaced it with the United States Parole Commission. The Commission promulgated its own regulations, which included a point-based scoring system. See 28 C.F.R. § 2.80. Unlike the 1987 system, in which an inmate's points dictated whether he was granted or denied parole, the current system uses an inmate's point score to determine the number of months an inmate should serve beyond his minimum sentence. § 2.80(h). Although an inmate is presumptively unsuitable for parole until he has served the time dictated by his point score, "[t]he Commission may, in unusual circumstances, grant or deny parole to a prisoner notwithstanding the guidelines." § 280(n)(1).

The current regulations are more detailed and uniform than the regulations in effect at the time of Brown's offense. An inmate's total score is calculated by adding points in the following categories: salient factor; current or prior violence; and death of victim. § 2.80(f). Salient factor points reflect the inmate's perceived risk of recidivism and are awarded based on his age at the time of the offense and the number and nature of his previous offenses. 28 C.F.R § 2.20. Violence points reflect the use of violence or a firearm in the current offense or in previous offenses. § 2.80(f). Death-of-victim points reflect murder and other specifically enumerated crimes involving a "high level" of violence, even if the victim survived. Id. After adding these points together, the inmate's base guideline range is calculated in months. § 2.80(h). This range may then be adjusted upward for disciplinary infractions while incarcerated, § 2.80(j), or downward for "superior program achievement," § 2.80(k). In denying Brown's request for parole, the Commission calculated a guideline range of 98 to 164 months past his initial parole eligibility date.

II.

After being denied parole, Brown filed a timely federal habeas petition, arguing that application of the current parole guidelines violates the Ex Post Facto Clause because it significantly increased his risk of serving a longer sentence than he would have served under the guidelines in effect at the time of his offense. The District Court denied the petition and Brown filed this timely appeal.

We have appellate jurisdiction pursuant to 28 U.S.C. § 1291 and our review of a habeas petition dismissed without a hearing is plenary. Richardson v. Pa. Bd. of Probation Parole, 423 F.3d 282, 287 n. 3 (3d Cir. 2005).

The gravamen of Brown's appeal is that the current parole guidelines emphasize different factors and are objectively harsher for inmates in his shoes than the guidelines in effect at the time of his offense. The Government views the current guidelines as nothing more than a procedural mechanism for implementing the same substantive criteria that were relevant under the prior guidelines, but with more detailed specificity and less open-ended discretion.

III.

The Ex Post Facto Clause applies to "a statute or policy change which 'alters the definition of criminal conduct or increases the penalty by which a crime is punishable.'" Richardson, 423 F.3d at 287 (quoting Calif. Dept. of Corr. v. Morales, 514 U.S. 499, 506 n. 3, 115 S.Ct. 1597, 131 L.Ed.2d 588 (1995)). "Retroactive changes in laws governing parole of prisoners, in some instances, may be violative of [the Ex Post Facto Clause]." Garner v. Jones, 529 U.S. 244, 250, 120 S.Ct. 1362, 146 L.Ed.2d 236 (2000). However, "not every retroactive procedural change creating a risk of affecting an inmate's terms or conditions of confinement is prohibited." Id. (citing Morales, 514 U.S. at 508-09, 115 S.Ct. 1597).

In Morales, the Supreme Court rejected an ex post facto challenge to the retroactive application of a California statute that decreased the frequency of periodic parole review, noting that the Ex Post Facto Clause should not be invoked to "micro-manage[] . . . legislative adjustments to parole and sentencing procedures." Morales, 514 U.S. at 508, 115 S.Ct. 1597. But the Court also stated that an inmate challenging the retroactive application of new parole guidelines need not show that he would have been immediately paroled under previous guidelines — only "that the measure of punishment itself has changed." Id. at 510 n. 6, 115 S.Ct. 1597. Accordingly, "[t]he question of what legislative adjustments will be held to be of sufficient moment to transgress the constitutional prohibition must be a matter of degree." Id. at 509, 115 S.Ct. 1597 (emphasis in original) (quotation omitted).

In Garner, the Supreme Court considered a Georgia statute similar to the California law at issue in Morales. The Court held that although the statute did not necessarily violate the Ex Post Facto Clause, the lower court erred by ignoring the parole board's internal policy statements interpreting the statute's effect. Therefore, the inmate was entitled to "adequate discovery" to demonstrate that the new guidelines posed a "significant risk of prolonging [his] incarceration." Garner, 529 U.S. at 251, 255-57, 120 S.Ct. 1362. The Court also cautioned that "[t]he States must have due flexibility in formulating parole procedures and addressing problems associated with confinement and release." Id. at 252, 120 S.Ct. 1362.

Following Morales and Garner, we have had occasion to consider ex post facto challenges to changes to parole guidelines. See Mickens-Thomas v. Vaughn, 321 F.3d 374 (3d Cir. 2003) (finding ex post facto violation where change in parole guidelines emphasized a new criterion that was solely responsible for inmate's continued incarceration); Richardson, 423 F.3d at 293 (distinguishing Mickens-Thomas as "an exceptional case" and rejecting ex post facto challenge where inmate failed to proffer "at least some evidence of disadvantage to warrant habeas relief) (emphasis in original).

In Richardson, we noted that "[t]he ex post facto inquiry has two prongs: (1) whether there was a change in the law or policy which has been given retrospective effect, and (2) whether the offender was disadvantaged by the change." Richardson, 423 F.3d at 287-88. To violate the second prong, "a retroactive change in the law or policy must create a 'sufficient risk of increasing the measure of punishment attached to the covered crimes'; a 'speculative and attenuated possibility of . . . increasing the measure of punishment' is not enough." Id. at 288 (quoting Morales, 514 U.S. at 509, 115 S.Ct. 1597).

IV.

In applying the controlling law to his case, Brown claims three aspects of the new parole guidelines constitute substantive changes that materially affect his chances for release. We consider each aspect in turn.

A.

Brown argues that the implementation of a point-based scoring system with specific guideline ranges necessarily violates the Ex Post Facto Clause. We disagree.

Although the current guidelines "mandate" specific periods of additional time for various reasons ( e.g., death of a victim and possession of contraband while incarcerated), these considerations would have been evaluated by the Board of Parole under the prior guidelines. See D.C. MUN. REGS. tit. 9, §§ 105.1(a), (e) (1972) (directing the Board to consider the nature of the underlying offense and the inmate's institutional experience). Moreover, the Parole Commission retains discretion under the current system to depart from the guidelines. 28 C.F.R § 2.80(n)(1). See Garner, 529 U.S. at 253, 120 S.Ct. 1362 (noting relevance of discretion in ex post facto analysis).

Thus, the fact that the point-based scoring system of the current guidelines results in a presumptive denial of parole in Brown's case sheds no light on what would have been his fate under the guidelines in effect in 1978. Accordingly, Brown's broadside challenge to the current system must fail.

B.

Brown also claims that the current guidelines are more detrimental to him than the prior guidelines in several particular respects. The current guidelines assess six points for the violent death of his victim while the guidelines implemented in 1987 would have assessed only one point. He also notes that after being denied parole in 2006, he must wait five years for another parole hearing whereas the prior guidelines allowed for annual rehearings. Finally, Brown argues that his "remote" disciplinary infractions, which resulted in additional time under the current guidelines, would have been utterly irrelevant under the old guidelines, because "an offender . . . who had forfeited good time restored was not penalized and could be paroled." Id. at 23.

The constitutional relevance of the frequency of rehearings is questionable given Morales and Garner, especially since Brown was presumptively unsuitable for parole for at least another eight years according to the Parole Commission's calculations. And contrary to Brown's assertion, the language of the 1978 guidelines does not specifically prohibit consideration of his remote disciplinary infractions. See D.C. MUN. REGS tit. 9, § 105.2 (1972). But the purported discrepancy between the weight given to the violent death of Brown's victim under the old guidelines and under the current guidelines may support his argument because it might pose a significant risk of prolonging Brown's incarceration.

The six points assessed for the violent death of his victim under the current guidelines added between nine and twelve years to Brown's guideline range. See § 2.80(h). Although a guideline range by itself does not necessarily indicate a significant risk of prolonged incarceration, the length of this range is striking, especially considering the relative lack of emphasis given to the nature of the underlying offense in the prior guidelines. Compare § 105.1(a) (single reference to "the nature of the violation, mitigating or aggravating circumstances"), with §§ 105.1(c)-(f) (all emphasizing post-incarceration factors or personal/background characteristics of the offender).

The scoring system implemented in 1987 — which, under District of Columbia law, implemented no significant substantive changes to the 1978 guidelines, Davis v. Henderson, 652 A.2d 634, 636-37 (D.C. 1994) — makes this range appear even more disparate. Brown would have received at most five points at his initial parole hearing under the 1987 guidelines, with an automatic rehearing every subsequent subsequent year. D.C. MUN. REGS. tit. 28, § 200, Appx. 2-1 (1987). Assuming continued good behavior and achievement in institutional programs, this score would have been reduced by one point at each successive rehearing. § 200, Appx. 2-2. Under this system, Brown would have been presumptively suitable for parole at his second annual rehearing when his score fell to three. Id.

For the aforementioned reasons, we agree .with Brown that the discrepancy between the guidelines constitutes more than conjecture that "the practical effect of the [new guidelines] was to change the weight that [this particular factor) was given in the parole calculation." Richardson, 423 F.3d at 289 (citing Mickens-Thomas, 321 F.3d at 387).

C.

Finally, as another ex post facto problem, Brown notes "the new scoring system's emphasis on criminal history rather than rehabilitation, physical and emotional health, and institutional experience and achievements." We agree that this is a facial discrepancy that seems to shift the overall emphasis of the parole inquiry.

The guidelines in effect at the time of Brown's offense listed six criteria to guide the Board of Parole in exercising its discretion. The first two criteria related to the underlying offense and the offender's prior criminal history. §§ 105.19(a), (b). Another criterion related to the offender's "institutional experience." § 105.1(e). The other criteria related to: the "[p]ersonal and social history of the offender, including such factors as his family situation, educational development, socialization, marital history, employment history, use of leisure time and prior military experience, if any," § 105.1(c); the offender's "physical and emotional health and/or problems which may have played a role in [his] socialization process, and efforts made to overcome any such problems," § 105.1(d); and the "[c]ommunity resources available to assist the offender with regard to his needs and problems," § 105.1(f).

The current guidelines make no provision for any of the criteria referenced in §§ 105.1(c), (d), and (f). Potentially sympathetic consideration of the mental, physical, and social problems plaguing an offender has been completely eliminated. Likewise, any progress made by the offender in confronting these problems and resources available to help the offender on the outside are no longer a factor. Whereas the prior guidelines emphasized the internal and external problems that might have driven an offender to commit a crime and his efforts to overcome them, the current guidelines focus almost exclusively on the severity of the underlying crime and the offender's criminal history. See 28 C.F.R §§ 2.20, 2.80.

Although a parole agency or state "may learn from experience" and adjust its parole criteria accordingly, that "does not mean that those who were sentenced at an earlier juncture may now be more severely resentenced in the light of newly-found wisdom. This is precisely what the Ex Post Facto Clause prohibits." Mickens-Thomas, 321 F.3d at 387. The discrepancy between the emphasis placed on an offender's mental, physical, and social characteristics by the prior guidelines and the lack of consideration given to those factors by the current guidelines indicates a substantive shift that might subject Brown to a more severe sentence.

V.

In sum, the additional weight given to the violent death of the victim and the apparent de-emphasis of Brown's mental, physical, and social characteristics collectively indicate that retroactive application of the current parole guidelines might pose a significant risk of prolonging Brown's incarceration. We acknowledge that while finding a potential ex post facto violation in retroactive application of the current guidelines for reparole, the D.C. Circuit equated the current and 1987 guidelines for parole in terms of "rehabilitative focus." Fletcher v. Reilly, 433 F.3d 867, 869 (D.C. Cir. 2006).

Because parole was not at issue in Fletcher, however, the D.C. Circuit had no reason to engage in a detailed comparison between the current guidelines and those in effect in 1978. Thus, although we find the legal analysis in Fletcher helpful, its dictum that the current guidelines for parole essentially mirror the 1987 guidelines is unpersuasive in Brown's case in light of our detailed explanation of the differences between the current guidelines and the 1978 guidelines. Here, it appears that the "substantive criteria for parole release [may] have changed" with the current guidelines. Mickens-Thomas, 321 F.3d at 392.

In Richardson we denied an evidentiary hearing where the habeas petitioner "provided no evidence, and for that matter . . . proffered no allegation, that a 'significant risk' of increased punishment was created" by application of new parole guidelines. Richardson, 423 F.3d at 293. In this appeal, Brown has pointed to specific disparities between the prior and current guidelines that could provide the basis for habeas relief. Therefore, he is entitled to "a searching comparison of the old and new [parole] regimes in order to determine whether the [] Parole Commission's application of the federal . . . regulations . . . created a significant risk that he will be subjected to a lengthier incarceration than he would have been if the Commission had adhered to the rules and practices of the D.C. Board." Fletcher, 433 F.3d at 879.

For the foregoing reasons, we will vacate the order of the District Court and remand the case for further proceedings consistent with this opinion.