On Rehearing En Banc
GLICKMAN, Associate Judge:Two years ago, in United States Parole Comm’n v. Noble, 693 A.2d 1084, 1095 (D.C.1997), op. adopted, 711 A.2d 85 (D.C.1998) (en banc), this court resolved a decade-old dispute regarding “street time,” the colloquial term for time that a convicted offender spends serving his sentence while on parole. We held that a law enacted by Congress in 1932, providing for loss of accrued street time when parole is revoked, was not repealed by the Council of the District of Columbia in 1987, and remains in full force and effect. Our holding required the District of Columbia Department of Corrections to change its method of computing the amount of time that prisoners in its custody had left to serve on their sentences, because the Department had been allowing parole violators to retain credit for street time on the erroneous understanding that the Council had repealed the 1932 law by implication.
The present case presents a question that we reserved in Noble: whether our holding would apply retroactively. Appellants are three D.C.Code offenders whose paroles were revoked prior to our decision in Noble. In computing the time remaining on appellants’ sentences following their reincarceration, the Department of Corrections initially credited them with their street time. After our decision in Noble, however, the Department recomputed appellants’ remaining sentences by withdrawing credit for street time, thereby increasing the amount of time that appellants had left to serve.
Appellants contend that our holding in Noble announced a new rule of law which enhances the punishment imposed on D.C.Code offenders if they violate the terms of their parole. Claiming that there was widespread reliance in the District of Columbia on the pre-Noble understanding that parole revocation would not result in loss of street time, appellants invoke the equitable balancing test that this court adopted in Mendes v. Johnson, 389 A.2d 781 (D.C.1978) (en banc), to argue that Noble must be applied prospectively only. Appellants further argue that retroactive application of Noble to increase their sentences would not only be inequitable under Mendes, but would also violate the Ex Post Facto and Due Process Clauses of the Constitution.
A division of this court upheld the retroactive application of Noble to appellants, with one judge dissenting. We vacated the decision of the division and granted rehearing en banc, in part because recent retroactivity decisions of the Supreme Court have undermined the viability of Mendes.
We conclude that the Department of Corrections acted properly when it recomputed appellants’ sentences in accordance with Noble. In reaching that conclusion, *209we hold, first, that the retroactive application of Noble does not violate the Ex Post Facto or Due Process Clauses. Second, we hold that the time has come for us to jettison the retroactivity jurisprudence that we adopted in Mendes. Instead of using a balancing test, we follow the lead of the Supreme Court and adopt a firm rule of retroactivity for our rulings. Applying that rule in this case, we hold that Noble must be applied retroactively so long as it is Constitutional to do so.
BACKGROUND
This case has its genesis in a purported conflict between two statutes relating to street time. The first statute was enacted by Congress in 1932, and provides that a D.C.Code offender forfeits his accrued street time if he violates the conditions of his parole and is reincarcerated. The language of this statute is unequivocal: “If the order of parole shall be revoked, ... [t]he time a prisoner was on parole shall not be taken into account to diminish the time for which he was sentenced.” D.C.Code § 24-206(a) (1996). The revocation of parole results in the prolongation of the time that an offender serves on his sentence by the amount of street time that is lost.
The second statute was enacted fifty-five years later by the Council of the District of Columbia as part of the Good Time Credits Act of 1986 (“GTCA”), D.C. Law 6-218, § 5, 34 D.C.Reg. 484 (1987). This statute, which took effect in 1987, provides among other things that D.C.Code offenders get credit for their street time against the service of their sentences. The language of the statute is that “[ejvery person shall be given credit on the maximum and the minimum term of imprisonment for time spent in custody or on parole as a result of the offense for which the sentence was imposed.” D.C.Code § 24-431(a) (1996). Unlike § 24-206(a), however, the GTCA does not specifically address whether a D.C.Code offender forfeits his street time credit if his parole is revoked.
Disagreement over whether the street time forfeiture provision of § 24-206(a) remained in force surfaced almost immediately after the GTCA went into effect. The District of Columbia Corporation Counsel advised the Department of Corrections that in its opinion the GTCA implicitly repealed the provision of § 24-206(a) that required forfeiture of street time upon revocation of parole. See Noble, 693 A.2d at 1095. In reliance on the Corporation Counsel’s opinion, the Department of Corrections issued an order, and thereafter a formal regulation, providing that henceforth revocation of parole would not result in the loss of credit for street time toward service of the sentence for which parole had been granted. See 35 D.C.Reg. 1077, 1078 (1988).
The United States Parole Commission promptly took issue with the Corporation Counsel’s statutory interpretation. See Noble, 693 A.2d at 1095-96. The Commission concluded that the Council did not repeal the street time forfeiture provision of § 24-206(a) by implication when it enacted the GTCA, but merely recognized the general rule that a sentence of imprisonment could be served on parole as well as in custody. Consistent with this conclusion, the Commission promulgated its own formal regulation to provide that D.C.Code prisoners under federal supervision would continue to lose credit for street time upon the revocation of their parole. See 28 C.P.R. § 2.65(i) (1999).
Because the Commission and the District disagreed over whether the GTCA repealed the street time forfeiture provision of § 24-206(a), offenders sentenced in the District to imprisonment were subject to disparate treatment upon revocation of *210their parole. The disparity depended on where the Attorney General chose to designate them to serve their sentences. Pursuant to D.C.Code § 24-425 (1996), the Attorney General has custody over all prisoners convicted in the District and has unfettered discretion to designate them to prisons maintained by the District of Columbia government or by the federal government. See District of Columbia v. Cooper, 483 A.2d 317, 322 (D.C.1984); Curry-Bey v. Jackson, 422 F.Supp. 926, 932 (D.D.C.1976). While District authorities supervise prisoners who are confined to D.C. correctional facilities, the Commission supervises parole of D.C. offenders housed at federal facilities.1 See D.C.Code §§ 24-206(b) and 24-209 (1996); Franklin v. Ridley, 635 A.2d 356, 357 n. 2 (D.C.1993); Goode v. Markley, 195 U.S.App. D.C. 391, 394, 603 F.2d 973, 976 (1979). Under § 24-209, the Commission must apply D.C. (rather than federal) parole law to these inmates. See Walker v. Luther, 830 F.2d 1208, 1217 (2d Cir.1987). See also D.C.Code § 24-1231(c) (2000 Supp.). In the case of the GTCA, however, the Commission applied its construction of that law rather than the Corporation Counsel’s construction. Hence in the years following the enactment of the GTCA, D.C.Code offenders supervised by the Commission continued to forfeit street time upon revocation of parole while locally supervised offenders did not.
Predictably, the Commission’s interpretation of the GTCA was soon challenged by a federally designated D.C.Code offender who was deprived of street time credit when his parole was revoked. Joseph Michael Tyler instituted this challenge by means of a habeas corpus petition in federal district court in Alaska after his parole was revoked in early 1988. The district court denied relief, and the Court of Appeals for the Ninth Circuit affirmed in Tyler v. United States, 929 F.2d 451 (9th Cir.1991). Invoking the “cardinal rule” of statutory construction that repeals by implication are disfavored,2 and specifically rejecting the D.C. Corporation Counsel’s statutory analysis as “cursory,” “ill-conceived” and “overly simplistic,” id. at 456, the Ninth Circuit held that “the GTCA did not impliedly repeal the longstanding requirement of section 24-206 that parole violators forfeit their street time.” Id. at 457. The court further held that Tyler was not entitled to retain credit for his street time merely because prisoners in the District did not lose street time upon revocation of their parole: “We cannot seriously entertain an argument that an erroneous statutory interpretation should be perpetuated simply because it would favor a prisoner who has not yet benefited from it.” Id.
Despite the decision in Tyler, the Corporation Counsel neither abandoned its interpretation of the GTCA nor sought a *211clarifying statutory amendment from the Council. See Noble, 693 A.2d at 1102-08. Instead, the Corporation Counsel continued to adhere to the view that the Council implicitly repealed the street time forfeiture provision of § 24-206(a) when it enacted the GTCA. For its part, the Department of Corrections also ignored Tyler and continued to credit prisoners in its custody with street time notwithstanding the revocation of their parole.
For several years after Tyler the issue of implied repeal of § 24-206(a) by the GTCA did not come before this court. In a few habeas corpus cases brought by federally designated prisoners, judges of the Superior Court agreed with the District that the GTCA abrogated the street time forfeiture provision of § 24-206(a). See, e.g., Beaty v. Ridley, No. SP 138-93 (D.C.Super.Ct. Jan. 26, 1993). None of those decisions, however, were appealed. Our opinions did contain dicta suggesting that we would agree with the District’s position. See Franklin v. Ridley, 635 A.2d 356, 358 (D.C.1993); Luck v. District of Columbia, 617 A.2d 509, 514 and 514 n. 6 (D.C.1992). The Luck court commented that the alternative construction of the GTCA approved in Tyler was “plausible” but “curious.” Id. However, the court also stated in Luck that we adhere strongly to the canon of statutory construction that repeals by implication are disfavored, and that we would deviate from that canon “only in exceptional cases.” Id.
The issue was finally joined in a lawsuit commenced in 1995 in the United States District Court for the District of Columbia by a federally designated prisoner named Matthew Noble. The district court granted Noble habeas corpus relief, holding that the GTCA impliedly repealed § 24-206(a) and required the Commission to credit Noble for time he spent on parole for a D.C. offense even though his parole had been revoked. See Noble v. United States Parrole Comm’n, 887 F.Supp. 11 (D.D.C.1995). On appeal, however, the United States Court of Appeals for the District of Columbia Circuit certified the question of implied repeal of § 24-206(a) to this court for resolution. See Noble v. United States Parole Comm’n, 317 U.S.App. D.C. 304, 82 F.3d 1108 (1996). The D.C. Circuit found that there was no controlling precedent in our decisions and that we had “sent mixed signals” regarding the implied repeal question in Luck and Franklin. Id., 317 U.S.App. D.C. at 307-08, 82 F.3d at 1111-12.
A division of this court answered the certified question in United States Parole Comm’n v. Noble, 693 A.2d 1084 (D.C.1997). The following year the en banc court adopted the division’s opinion. See 711 A.2d 85 (D.C.1998) (en banc). Agreeing with the Ninth Circuit’s 1991 decision in Tyler, we held that the Council did not repeal § 24-206(a) by implication when it enacted the GTCA, and that the District had therefore erred in crediting prisoners with street time after their parole had been revoked. Noble, 693 A.2d at 1095.
Because this court was addressing only the narrow question of law certified to it by the D.C. Circuit, we were not called upon in Noble to decide whether our ruling would apply “retroactively.” We expressed no opinion on that issue, though we recognized that prisoners might claim to have relied upon the Department of Corrections regulation which credited them with street time despite the revocation of their parole. Flagging the question of whether Matthew Noble himself had a legitimate reliance interest in that regulation, however, we observed that this was not a case of unconstitutional ex post facto legislation, because the legislature had not “given credits toward completion of a sentence and then taken them away” retroac*212tively. Id. at 1104. We also noted that the Ninth Circuit issued its decision in Tyler over two years before Noble’s parole was revoked. Id.
The Department of Corrections undertook to comply with our en banc ruling in Noble and the street time forfeiture requirement of § 24-206(a) by recalculating the sentences of all parole violators who were still in custody. The District exercised its enforcement discretion not to reincarcerate inmates who had been released from custody (either on re-parole or because their sentences had expired) based on erroneously awarded street time credit. Nonetheless, as appellants emphasize, the practical effect of the Department’s belated decision to rescind street time in accordance with § 24-206(a) was to add months or years to the sentences of upwards of one thousand former parolees who were still subject to the Department’s authority.3
Meanwhile, the D.C. Circuit remanded Matthew Noble’s habeas corpus petition to the district court for resolution in light of our decision. The district court held that Noble could not complain about the application of our ruling to him, because as a federal prisoner he had no legitimate expectation otherwise, since the Commission had never credited D.C.Code offenders in federal custody with street time following revocation of their parole. Accordingly the district court denied Noble’s habeas corpus petition. See Noble v. United States Parole Comm’n, 32 F.Supp.2d 11 (D.D.C.1998). In dictum, however, the court opined that the District ought not to apply our ruling in Noble retroactively to inmates under its supervision, since those persons did have “the right to rely on the District of Columbia’s prior interpretation of its own laws.” Id. at 14.
On appeal once again to the D.C. Circuit, Noble argued on equal protection grounds that our ruling in his case should not be applied retroactively to him, because it was not being applied retroactively to former D.C. prisoners whose sentences had expired “early” after the Department of Corrections mistakenly credited them with street time despite the revocation of their parole. The appeals court rejected this argument even assuming that Noble was similarly situated to D.C. prisoners, “because the difficulty of rearresting inmates who have already been released would provide a rational basis for the disparate treatment.” Noble v. United States Parole Comm’n, 338 U.S.App. D.C. 362, 364, 194 F.3d 152, 155 (1999). Furthermore, the *213court stated, “[n]either authority nor common sense support the proposition that if the government erroneously confers a benefit on some people, then other people have a Constitutional right to receive the same ■windfall.” Id. The court affirmed the denial of Noble’s habeas petition.
Elsewhere, another federally designated D.C.Code offender named James F. Johnson challenged the retroactive application of our decision in Noble as a violation of due process rather than equal protection, on the theory that Noble effected an unforeseeable, after-the-fact increase in the statutorily authorized degree of punishment for D.C. offenses. In Johnson v. Kindt, 158 F.3d 1060 (10th Cir.1998), the court of appeals rejected that challenge. The court agreed that retroactive operation of an unforeseeable judicial enlargement of a criminal statute would violate due process by depriving defendants of fair warning of the penal consequences of their acts. See id. at 1068. However, the court stated, the test of unforeseeability for purposes of due process is whether the judicial construction of the statute is “unexpected and indefensible by reference to the law which had been expressed prior to the conduct at issue.” Id. (quoting Bouie v. City of Columbia, 378 U.S. 347, 354, 84 S.Ct. 1697, 12 L.Ed.2d 894 (1964)). “Unforeseeable judicial decisions include expansion of a statute narrow and precise on its face beyond [its] terms, ...; the overruling of precedent, ...; or when ‘an in— depth inquiry by a dedicated and educated student of [the relevant] law would have revealed nothing to foreshadow the [controlling court] opinion,’...” Johnson, 158 F.3d at 1063 (citations omitted). In light of the early disagreement between federal and District authorities over the proper construction of the GTCA, and the Ninth Circuit’s decision in Tyler holding that § 24-206(a) remained in effect, the Tenth Circuit stated that it was “foreseeable that the Commission’s view would ultimately prevail.” Id. In addition, the court said, our decision in Noble “did not expand a narrow and precise legislative provision or overrule controlling precedent. Moreover, elementary legal research would have revealed the principle that repeals by implication are disfavored.” Id. The Tenth Circuit accordingly found no due process bar to the retroactive application of our holding in Noble that the GTCA did not repeal the street time forfeiture provision of § 24-206(a).
Like Matthew Noble and James F. Johnson, appellants in the present case were reimprisoned after violating the conditions of their parole during the interregnum between the enactment of the GTCA in 1987 and the decision in Noble a decade later.4 However, unlike Noble and John*214son, who were in federal custody, appellants were in the custody and under the supervision of local agencies, the D.C. Department of Corrections and the D.C. Board of Parole. When their parole was revoked, appellants were officially informed by the Department of Corrections that, in accordance with its administrative regulation and the then prevailing local understanding of the GTCA and § 24-206(a), they would retain credit against their sentences for street time accrued while on parole prior to revocation.5 But in response to our decision in Noble, the Department of Corrections recomputed appellants’ sentences by subtracting their street time credit.
Seeking to limit the resultant lengthening of the time that they would spend serving their sentences, appellants filed petitions for writs of habeas corpus in the Superior Court challenging the District’s retroactive application of Noble to them. Those petitions were denied, and on consolidated appeal a division of this court affirmed with one judge dissenting. We vacated the decision of the division and granted rehearing en banc, in order to reconsider the propriety of applying Noble to appellants and others similarly situated in conjunction with a reconsideration of the basic principles governing retroactive application of decisions of this court that announce “new” rules of law.
DISCUSSION
Emphasizing that Noble invalidated a formal regulation of the Department of Corrections and reversed the understanding, in the District of Columbia at least, that revocation of parole entails no loss of street time credit, appellants urge us to declare our holding in Noble to be purely prospective in effect, i.e., not applicable to persons who committed their offenses before the issuance of our decision. Appellants argue that in light of widespread reliance on the District’s pre-Noble interpretation of the GTCA, any retroactive application of Noble’s “new rule of law” would be unfair and contrary to this court’s equity-based retroactivity jurisprudence. Appellants also argue that since the practical impact of Noble was to increase the potential punishment for commission of a criminal offense, any retroactive application of that decision would violate the Constitutional prohibition against ex post facto laws and amount to a denial of due process of law.
Appellants advocate these positions with admirable force and imagination, but *215in the end we are not persuaded. Taking appellants’ arguments in reverse order, which we think analytically useful, we conclude first that nothing in the Constitution requires that our ruling in Noble be limited to purely prospective application. The Constitutional proscription of ex post facto laws applies only to retrospective legislation. That proscription therefore has no bearing on this case, inasmuch as the street time forfeiture provision of § 24-206(a) went into effect long before appellants committed their offenses and was never repealed. Appellants cannot base an ex post facto claim on the existence of an administrative regulation that purported to preserve street time credit after revocation of parole, because insofar as that regulation was contrary to the express command of § 24-206(a), it was invalid.
The Due Process Clause does not bar retroactive enforcement of Noble either, because our decision in that case, even if contrary to the District’s position and not expected by appellants or others in the District, was not unforeseeable. Rather, in rejecting the argument that the GTCA impliedly repealed the street time forfeiture provision of § 24-206(a), Noble employed accepted principles of statutory interpretation and approved the same construction of the GTCA that the United States Parole Commission adopted when that legislation first took effect, and that the Ninth Circuit endorsed in 1991. In the decade between the enactment of the GTCA and our decision in Noble, it was foreseeable to every D.C.Code offender that revocation of parole might result in loss of street time pursuant to § 24-206(a) notwithstanding the GTCA, depending on whether the offender was designated to the custody of federal or District authorities, or on whether the federal or the District view of the effect of the GTCA ultimately prevailed.
We also conclude that equitable considerations such as those advanced by appellants do not justify prospective-only operation of Noble. Taking this occasion to reconsider our retroactivity jurisprudence and conform it to that of the Supreme Court, we adopt today a firm rule of retro-activity for our decisions expounding District of Columbia law. Our holding in Noble was a definitive statement of that law, and of what the Council did and did hot do when it enacted the GTCA. In the absence of any due process or other Constitutional impediment to retroactive application of our holding, we have no warrant to exempt the parties before us from the law that governed them.
A. Constitutional Challenges to Retroactivity
1. Prohibition Against Ex Post Facto Laws
In response to our decision in Noble, the Department of Corrections stopped abiding by its 1988 regulation which preserved street time credit following revocation of parole, and started applying the rule that prisoners forfeited their street time when their parole was revoked. Appellants argue that the 1988 regulation was the law in effect when they committed their offenses, and that therefore the Department’s retroactive application of a new rule which made their punishment more onerous by depriving them of street time credit infringed the Constitutional prohibition against ex post facto laws.
Retroactive application of a law that imposes a greater punishment than the law in effect when the crime was committed is forbidden by the Ex Post Facto *216clauses of the Constitution.6 See Lynce v. Mathis, 519 U.S. 433, 439-41, 117 S.Ct. 891, 187 L.Ed.2d 63 (1997); Weaver v. Graham, 450 U.S. 24, 28-29, 101 S.Ct. 960, 67 L.Ed.2d 17 (1981). A statute retroactively increasing the penalties imposed upon revocation of parole would fall within the prohibition, see Johnson v. United States, 529 U.S. 694, 701, 120 S.Ct. 1795, 146 L.Ed.2d 727 (2000). And since administrative regulations that are validly promulgated pursuant to statutory authority have the force and effect of statutes, see Dankman v. District of Columbia Bd. of Elections and Ethics, 443 A.2d 507, 513 (D.C.1981) (en banc), we agree with appellants that a new regulation which enhances punishment beyond what was formerly authorized by a valid penal regulation is within the ex post facto prohibition.
These principles are of no avail to appellants, however, because our holding in Noble means that the regulation promulgated by the Department of Corrections in 1988 was not a valid regulation. When this court interpreted the GTCA in Noble, we did not “undo” its repeal of the street time forfeiture provision of § 24-206(a); rather, we declared that the GTCA never effected a repeal of that provision, and that § 24-206(a) continued in full force and effect after the GTCA was enacted. See Rivers v. Roadway Express, Inc., 511 U.S. 298, 312-13, 114 S.Ct. 1510, 128 L.Ed.2d 274 (1994) (“A judicial construction of a statute is an authoritative statement of what the statute meant before as well as after the decision of the case giving rise to that construction.”); see also United States v. McKie, 315 U.S.App. D.C. 367, 371,73 F.3d 1149, 1153 (1996) (“a decision interpreting a statute does not change the statute but rather interprets the law as enacted by the legislature”). The 1988 regulation was, therefore, invalid from its inception because it was directly contrary to the governing statutory language, namely, the provision in § 24-206(a) that required forfeiture of street time upon revocation of parole. The Department of Corrections had no authority to abrogate § 24-206(a), and the Department’s correction of its erroneous interpretation of that law was not the same thing as a change in the law. The corrective action therefore did not run afoul of the ex post facto taboo. “[A]n agency misinterpretation of a statute cannot support an ex post facto claim.... ‘The ex post facto clause of the Constitution does not give [appellants] a vested right in such an erroneous interpretation.’” Caballery v. United States Parole Comm’n, 673 F.2d 43, 47 (2d Cir.1982) (quoting Mileham v. Simmons, 588 F.2d 1279,1280 (9th Cir.1979)).7
2. Due Process
Appellants have a second string to their Constitutional bow, and with it they aim their arrow directly at this court’s decision in Noble. Citing Bouie v. City of Columbia, 378 U.S. 347, 84 S.Ct. 1697, 12 L.Ed.2d 894 (1964), appellants argue that our construction of the GTCA in Noble was so unexpected and contrary to the prevailing view that it would offend due process *217to apply our holding retroactively. In Bouie the Supreme Court overturned convictions under a state criminal trespass statute because the state supreme court’s “unforeseeable and retroactive” expansion of the statute constituted “a deprivation of the right of fair warning” guaranteed by due process. Id. at 352, 84 S.Ct. 1697. The Court reasoned that “an unforeseeable judicial enlargement of a criminal statute, applied retroactively, operates precisely like an ex post facto law.... If a state legislature is barred by the Ex Post Facto Clause from passing such a law, it must follow that a State Supreme Court is barred by the Due Process Clause from achieving precisely the same result by judicial construction_ If a judicial construction of a criminal statute is ‘unexpected and indefensible by reference to the law which had been expressed prior to the conduct in issue,’ it must not be given retroactive effect.” Id. at 358-54, 84 S.Ct. 1697 (citations omitted). See also Marks v. United States, 430 U.S. 188, 97 S.Ct. 990, 51 L.Ed.2d 260 (1977), where, following Bouie, the Court held that due process barred retroactive application of the Court’s own recent decision announcing new Constitutional standards for obscenity prosecutions which relaxed requirements that the Court had enunciated in a previous decision. Although Bouie and Marks involved the judicial enlargement of criminal liability, the reasoning of those cases applies equally to the judicial enlargement of punishment for existing offenses such as appellants allege this court accomplished in Noble. See Johnson v. Kindt, 158 F.3d at 1063; Helton v. Fauver, 930 F.2d 1040, 1045 (3d Cir.1991). But see United States v. Newman, 203 F.3d 700, 702 (9th Cir.2000) (holding that due process principles of Bouie are not applicable to after-the-fact increase in degree of punishment).
There is force to appellants’ argument. Prior to Noble the prevailing view in this jurisdiction — as expressed in legal opinions and briefs of the Corporation Counsel, a regulation promulgated by the Department of Corrections, decisions of the Superior Court, dicta in two opinions of this court, and the first district court decision on Matthew Noble’s habeas petition — was that the GTCA impliedly repealed the street time forfeiture provision of § 24-206(a). The Department of Corrections officially informed prisoners who were recommitted to its custody after the revocation of their parole that they would retain credit for their street time. Over the course of a decade the Department calculated the sentences of thousands of prisoners in accordance with that rule. Although in Noble the D.C. Circuit deemed the issue a clouded one on which we had sent “mixed signals,” 317 U.S.App. D.C. at 308, 82 F.3d at 1112, we must give appellants their due: our decision in Noble, though ultimately 8-1 against implied repeal, contradicted expectations in the District that were encouraged by authoritative pronouncements and that were reasonably held.
Under Bouie, however, the due process issue turns not solely on whether Noble was unexpected, but on whether it was “unexpected and indefensible by reference to the law which had been expressed prior to the conduct in issue.” Bouie, 378 U.S. at 354, 84 S.Ct. 1697. The question, in short, is whether our decision was so unforeseeable that appellants had no fair warning that it might come out the way it did. This is a stringent test; for courts frequently and inevitably issue legal rulings that are more or less unanticipated— particularly to those on the losing side— but that are routinely and properly applied to the parties in whose cases the rulings *218come.8
We agree with the Tenth Circuit’s conclusion in Johnson v. Kindt that our decision in Noble was not unforeseeable in the Bouie sense. Noble did not overrule a previous decision of this court. Cf. Marks, 430 U.S. at 195-96, 97 S.Ct. 990. Nor did Noble employ indefensible or even novel legal reasoning; our decision was grounded on the well established principle of statutory construction that repeals by implication are strongly disfavored. And both the holding and the rationale of Noble were forecasted explicitly by the 1991 opinion of the Ninth Circuit in Tyler, “the only appellate decision that directly addressed the issue” prior to our decision. Johnson v. Kindt, 158 F.3d at 1063. Cf. Rose v. Locke, 423 U.S. 48, 53, 96 S.Ct. 243, 46 L.Ed.2d 185 (1975) (distinguishing Bouie because “[o]ther jurisdictions had already reasonably construed identical statutory language to apply to such acts”); United States v. Newman, 203 F.2d at 703 (“even if Bouie applies here, no due process violation occurred because the decision ... was reasonably foreseeable given the circuit split on the meaning of’ the statute in question). Imagine if, at any time prior to Noble, appellants had asked their lawyers to examine the state of the law and advise them as to whether the GTCA repealed the street time forfeiture provision of § 24-206(a). On the ultimate merits of that question, reasonable lawyers could have differed, as in actuality they did; but a competent lawyer would have warned appellants that the issue was controverted and unresolved, that implied repeals are disfavored, and that the GTCA might eventually be construed not to limit § 24-206(a). Appellants would have been warned, in other words, that despite the enactment of the GTCA, the issuance of implementing regulations by the Department of Corrections, and the seemingly uniform views of officials in the District, revocation of parole might turn out to entail loss of street time after all.
The preceding discussion might suggest that appellants and other D.C.Code offenders were only theoretically on notice that § 24-206(a) might require forfeiture of street time. We think that is not so. D.C.Code offenders were given actual warning of our eventual holding in Noble by the actions taken by the United States Parole Commission after the GTCA took effect. Pursuant to D.C.Code § 24-425, *219the Attorney General could and frequently did place offenders under federal rather than District supervision. See District of Columbia v. Cooper, 483 A.2d at 322 (D.C. prisoners have “no legitimate expectation” that they will remain in D.C. prison facility); Curry-Bey v. Jackson, 422 F.Supp. at 933 (no “justifiable expectation” to be considered for parole by D.C. Board rather than Commission). It was no secret that D.C.Code offenders who found themselves under federal supervision were divested of their street time if their parole was revoked because the Commission did not agree with the District that the GTCA repealed the street time forfeiture provision of § 24-206(a). D.C.Code offenders were, therefore, on actual notice that under § 24-206(a) they might lose street time upon revocation of parole, notwithstanding the GTCA, at least in the event that the Attorney General designated them to serve their prison sentences in federal facilities. It follows that D.C. offenders were on notice that the law in this area was unsettled, and that the disagreement between the Commission and the District might one day be settled in favor of the Commission’s view, as in Noble it was.
It is true that prisoners placed in the custody of the Department of Corrections after revocation of their parole were officially told that they would receive credit for their street time. Appellants argue that those prisoners understandably relied on what they were told as they planned for their release from prison and the end of their sentences. Their reasonable expectations were raised and then frustrated when the Department of Corrections corrected their sentences by subtracting street time credit in compliance with Noble. This was indeed regrettable. Cf. Breest v. Helgemoe, 579 F.2d 95 (1st Cir.1978) (prospect of release on date certain may “assume a real and psychologically critical importance” to prisoner). That fact does not, however, mean that Noble was unforeseeable within the meaning of Bouie, or — absent “other, more tangible prejudice,” Lerner v. Gill, 751 F.2d 450, 459 (1st Cir.1985) — that the belated correction of sentences otherwise violated due process.
The general rule is otherwise. An expectation of early release from prison (or from service of a sentence) that is induced not by a valid statute or regulation but by the mistaken representations of officials does not without more give rise to a liberty interest entitled to protection under the Due Process Clause. See Jago v. Van Curen, 454 U.S. 14, 17-19, 102 S.Ct. 31, 70 L.Ed.2d 13 (1981) (rescission without hearing of prisoner’s promised parole prior to his release held not violative of due process). “Only in rare circumstances have courts allowed the misconstructions of officials to estop the proper execution of state or federal law, and such cases have involved prejudice and harm beyond frustrated expectations.” Lerner, 751 F.2d at 459; see also Office of Pers. Mgmt. v. Richmond, 496 U.S. 414, 423, 110 S.Ct. 2465, 110 L.Ed.2d 387 (1990); Heckler v. Community Health Servs. of Crawford County, Inc., 467 U.S. 51, 60, 104 S.Ct. 2218, 81 L.Ed.2d 42 (1984). Thus, it is a well established rule that a prisoner has no constitutional right to object to the correction of a miscalculation of his sentence. See, e.g., United States v. Merritt, 478 F.Supp. 804, 807 (D.D.C.1979) (“A convicted person will not be excused from serving his sentence merely because someone in a ministerial capacity makes a mistake with respect to its execution.”)9 Rather, “[t]he *220public have rights in such matters which it is beyond the power of the public officer to barter away,” Leonard v. Rodda, 5 App. D.C. 256, 270 (1895). The offender’s expectation and reliance interests in sentence mistake cases are ordinarily trumped by the strong public interest in crime prevention, see United States v. Salerno, 481 U.S. 739, 750, 107 S.Ct. 2095, 95 L.Ed.2d 697 (1987), public safety, see Pell v. Procunier, 417 U.S. 817, 822-23, 94 S.Ct. 2800, 41 L.Ed.2d 495 (1974), and punishing criminals, see Bearden v. Georgia, 461 U.S. 660, 669, 103 S.Ct. 2064, 76 L.Ed.2d 221 (1983).
We conceive of the possibility that, under “extreme” circumstances, a belated correction of a sentence might be “so unfair that it must be deemed inconsistent with fundamental notions of fairness embodied in the Due Process Clause.” De-Witt v. Ventetoulo, 6 F.3d 32, 35 (1st Cir.1993). In Noble this court allowed for the possibility that this substantive due process concern might be implicated if the District were to reincarcerate ex-offenders whose sentences had been deemed satisfied and who had readjusted to society. See 693 A.2d at 1105 (citing Johnson v. Williford, 682 F.2d 868, 871-73 (9th Cir.1982)). That particular situation is not before us, however, since the District has chosen not to re-arrest persons who were previously granted unconditional release. Appellants do not contend that across-the-board retroactive application of Noble to persons still in the custody of the Department of Corrections would violate substantive due process, nor could they plausibly do so. As much as there are substantial questions about how the requirements of substantive due process apply to the correction of sentences,10 one thing is certain. Only the most egregious case, involving for example governmental culpability and unusual prejudice to the affected prisoner, would support a substantive due process claim. No such claim has been asserted in this case, nor have we been told that any D.C. prisoner has asserted such a claim in any other post-Noble case. The mere fact that we can conceive of the bare possibility that a prisoner in unusual circumstances might have a substantive due process claim is no justification for prohibiting the general recomputation of sentences in accordance with Noble.
*221Finally, our conclusion that retroactive application of Noble comports with due process is not undermined by the possibility that some prisoners may have sustained actual prejudice as a result of rebanee on the pr e-Noble misunderstanding of the law in the District. Appebants’ principal claim in this regard is that there may be some cases in which the Board of Parole would not have revoked parole at ab if it had known that forfeiture of street time would result. Appebants rely on the concession of the District (made in its statement supporting rehearing en banc in Noble) that “in cases where the parolee had spent years successfuby on parole before incurring technical or minor violations of parole, such as failure to keep an appointment with a parole officer (as opposed to new criminal violations), the Board might have revoked parole based on the understanding that the parolee was entitled to street time, but might not have revoked parole if the parolee faced years more incarceration without the street time credit.” Such parolees would have been prejudiced by the Board’s misapprehension of the law if our post-revocation decision in Noble resulted in their actual incarceration for longer than the Board intended.11
This possibihty of actual prejudice in a few cases does not require us to conclude that due process bars retroactive application of Noble in all cases. There is another remedy. If the Board would not have revoked parole in a particular case but for its misapprehension of the law, its decision would be vulnerable to a chabenge in court for abuse of discretion. “ ‘A district court by definition abuses its discretion when it makes an error of law,’ Koon v. United States, 518 U.S. 81, 100, 116 S.Ct. 2035, 135 L.Ed.2d 392 (1996), and the same is true of the [Board of Parole].” Teachey v. Carver, 736 A.2d 998, 1004 (D.C.1999); see also Hall v. Henderson, 672 A.2d 1047, 1055 (D.C.1996) (applying “abuse of discretion” standard to decision of Parole Board). If the Board did abuse its discretion by revoking parole because of its misunderstanding of the appbcable law, a remand to the Board (or its successor, since the Board of Parole was recently abolished12) for a new revocation determination in bght of correct legal principles would be appropriate. Cf. Teachey, 736 A.2d at 1007.13
*222B. Non-Constitutional Challenge to Retroactivity
We have concluded that retroactive application of Noble comports with due process. The remaining question is whether equitable considerations may yet require us to deny retroactive effect to our holding in Noble even where the Constitution does not. In other words, though lacking constitutional justification, are we nonetheless required — are we even permitted — to disregard the governing law when we happen to think it “fair” to do so, and thus to excuse prisoners from serving their statutorily mandated sentences?
1. Non-Retroactivity Under Mendes v. Johnson
In asking us to exempt Noble from “the fundamental rule of ‘retrospective operation’ that has governed ‘judicial decisions ... for nearly a thousand years,’ ”14 appellants invoke the flexible approach to retro-activity that the Supreme Court first articulated in Linkletter v. Walker, 381 U.S. 618, 85 S.Ct. 1731, 14 L.Ed.2d 601 (1965),15 and that this court adopted in Mendes v. Johnson, 389 A.2d 781 (D.C.1978) (en banc). In Mendes we held that whether and to what extent a decision of this court announcing a new rule of law will be retroactive is a matter of “judicial policy.” Id. at 788. That policy, we stated, requires “an individualized, case-by-case basis” analysis of “ ‘the prior history of the rule in question, its purpose and effect, and whether retrospective operation will further or retard its operation.’ ” Id. at 788-89 (quoting Linkletter, 381 U.S. at 629, 85 S.Ct. 1731). We identified four “specific criteria” that the court would consider in conducting this analysis: “(1) the extent of the reliance of the parties on the old rule (including the degree of justifiable reliance and the hardship which might result to the litigants as a result of retrospective appli*223cation); (2) avoidance of altering vested contract or property rights; (3) the desire to reward plaintiffs who seek to initiate just changes in the law; and (4) the fear of burdening the administration of justice by disturbing decisions reached under the overruled precedent.” Mendes, 389 A.2d at 789. These criteria, we stated, were “central” in determining whether a new rule of law should have “total retroactive as well as prospective application, partial retroactive application (i.e., retroactive application of the new rule to the parties to the case in addition to prospective application), [or] purely prospective application.” Id.16
Although Mendes itself addressed the retroactive effect of a decision that overruled prior precedent, we have deemed the principles of that case applicable to non-overruling decisions which otherwise announce a “new rule of law.” The test, we have said, “is whether the rule is a ‘clear break’ from the past, a ‘newly minted principle,’ ... or a rule that an attorney ‘should have known’ was ‘about to be changed, because of either judicial or legislative intimations to that effect.’ ” Nimetz v. Cappadona, 596 A.2d 603, 608 (D.C.1991) (citations omitted). Echoing Chevron Oil, 404 U.S. at 106, 92 S.Ct. 349, we have stated that “[a] court may establish a new rule of law by (1) overruling clear past precedent, or (2) deciding a matter of first impression in a manner not clearly foreshadowed.” Sanders v. Sanders, 602 A.2d 663, 669 n. 5 (D.C.1992); accord, French v. District of Columbia Bd. of Zoning Adjustment, 658 A.2d 1023, 1031 (D.C.1995).
Appellants argue that our holding in Noble announced a new rule of law subject to the retroactivity principles of Mendes because — even if Noble was foreseeable and did not overrule prior precedent — it decided “a matter of first impression in a manner not clearly foreshadowed” in the District of Columbia. This court’s decision in French lends support to appellants’ claim.17 In French, much as in Noble, we overturned a “standard practice” that was supported by an opinion of the Corporation Counsel, after finding that the opinion was erroneous because it contravened the “plain language” of the applicable statute. French, 658 A.2d at 1030-31. Our decision in French was unquestionably foreseeable, given that the statute in that ease was unambiguous and, we said, “could not be clearer.” Id. Moreover, as we acknowledged, “opinions of the Corporation Counsel are not valid legal authority” and our decision in French did not overrule prior law. Id. at 1031. Nevertheless, because, we said, “no decision of this court has ever squarely addressed the precise issue at *224hand .... [and the] Corporation Counsel’s opinion was issued in 1977 and, until now, has never been challenged in this court,” id., we readily concluded in French that our holding “certainly decides ‘a matter of first impression in a manner not clearly foreshadowed.’ ” Id. (quoting Sanders, 602 A.2d at 667 n. 5).
The same points could be made about Noble, which also rejected a Corporation Counsel opinion of long standing. Indeed, appellants have a stronger argument that our holding in Noble resolved a matter of first impression in this jurisdiction and was not clearly foreshadowed. The Corporation Counsel’s opinion that the GTCA impliedly repealed the street time forfeiture provision of D.C.Code § 24-206(a) was embodied in a formal regulation of the Department of Corrections. That opinion had prevailed in Superior Court. And it was supported by dicta in opinions of this court, including dicta that specifically cast doubt on the Ninth Circuit decision in Tyler which rejected the implied repeal argument. Moreover, unlike in French, the language of the statute in question in Noble (the GTCA) was inconclusive, requiring us to examine the legislative history to interpret it. See Noble, 693 A.2d at 1088; and see id. at 1106 et seq. (Schwelb, J., dissenting).
Appellants further argue that, as in French, the factors enumerated in Mendes weigh in favor of limiting Noble’s new rule of law to purely prospective application.18 Appellants chiefly emphasize that for years D.C.Code offenders and District officials alike reasonably relied on the pre-Noble understanding that revocation of parole would not result in loss of street time, and that prisoners in the District sustained palpable hardship when the Department of Corrections recomputed and lengthened their sentences by months or years in response to Noble. In Mendes this court agreed that “[w]here retroactive application of a new rule would result in substantial disruption of settled transactions and/or injustice to a party because of reliance on the continued validity of the prior legal rule especially one of long standing courts are extremely reluctant to accord retroactive effect to overruling decisions.” Mendes, 389 A.2d at 789. We also stated that the significance of this factor depends in part on “the degree of hardship that the parties before the court, and others in general, may sustain as a result of the retroactive application.” Id. at 790.
Appellants may well overstate the strength of the reliance factor in this case. Disappointment of expectations is not the same thing as detrimental reliance. The record does not establish that appellants (or any other D.C.Code offenders who were adversely affected by Noble) actually did rely to their detriment on the pre-Noble “rule,” in the sense that they irrevocably changed their position, or would have acted differently if they had known that revocation of parole would entail loss of street time credit. The record likewise *225does not affirmatively establish that any D.C.Code offenders were actually prejudiced because the Board of Parole would not have revoked their parole if it had correctly apprehended the consequences. In contrast, French did present a case in which the element of detrimental reliance was clearly established. That said, the reliance factor in this case is not de min-imis. We think that appellants make a plausible prima facie case, if not necessarily an overwhelming or unanswerable one, for non-retroactive application of Noble under the principles of Mendes.
The soundness of those retroactivity principles is, however, in doubt, because “[i]n the intervening years, the Supreme Court has discarded its previous approach to retroactivity — the very approach on which this court based its decision in Mendes.” Washington v. Guest Services, Inc., 718 A.2d 1071, 1074 (D.C.1998). We granted rehearing en banc in this case in part to reexamine whether the retroactivity doctrine of Mendes deserves our continued allegiance. To that reexamination we now turn.
2. Rethinking Retroactivity
The Mendes court derived its approach to retroactivity from Linkletter, the case in which the Supreme Court abandoned its historic adherence to the rule that judicial decisions necessarily apply retrospectively and adopted instead of that rule a tripartite balancing test. Pursuant to that test, the retroactive application of a new rule of law turned on the results of a judicial inquiry into the purpose of the new rule, the extent of reliance on prior understanding of the law, and the effect of retroactive application on the administration of justice.19 See Linkletter, 381 U.S. at 629, 636, 85 S.Ct. 1731; Stovall, 388 U.S. at 297, 87 S.Ct. 1967; Chevron Oil, 404 U.S. at 106-07, 92 S.Ct. 349. The underlying premise of Linkletter and its progeny was that the courts “are neither required to apply, nor prohibited from applying, a decision retrospectively” and must therefore “weigh the merits and demerits [of retroactive application of a new rule] in each case,” Linkletter, 381 U.S. at 629, 85 S.Ct. 1731.
We need not trace the evolution of retro-activity doctrine in the Supreme Court in detail. As Judge Mack, the author of the majority opinion in Mendes, observed for the court in Kirk v. United States, 510 A.2d 499 (D.C.1986), “[t]he outcome of the [.Linkletter] balancing test varied considerably,” and there was widespread criticism that it “rested on no principled basis.” Id. at 505. Justice Harlan spoke for many both inside and outside the Supreme Court when he concluded a few years after Link-letter that “ '[r]etroactivity’ must be rethought.” Desist v. United States, 394 U.S. 244, 258, 89 S.Ct. 1030 (1969) (Harlan, J., dissenting).
It was not long before a majority of the Supreme Court agreed with Justice Harlan, and by 1993, Linkletter was overruled and its retroactivity doctrine swept away. The Supreme Court has summarized the change that took place by saying that “[w]hile it was accurate in 1974 to say that a new rule announced in a judicial decision was only presumptively applicable to pending cases, we have since established a firm rule of retroactivity.” Landgraf v. USI Film Products, 511 U.S. 244, 278 n. 32, 114 S.Ct. 1483 (1994) (citing Harper v. Virginia Dept. of Taxation, 509 U.S. 86, 113 S.Ct. 2510, 125 L.Ed.2d 74 (1993) and *226Griffith v. Kentucky, 479 U.S. 314, 107 S.Ct. 708, 93 L.Ed.2d 649 (1987)). In Griffith the Court explicitly overruled Linklet-ter and held that in criminal cases, all newly declared rules of law must be applied retroactively to all criminal cases pending on direct review or not yet final20 —“with no exception for cases in which the new rule constitutes a ‘clear break’ with the past.” Griffith, 479 U.S. at 328, 107 S.Ct. 708. In Harper the Court “heed[ed] the admonition that ‘the Court has no more constitutional authority in civil cases than in criminal cases to disregard current law or to treat similarly situated litigants differently.’ ” Harper, 509 U.S. at 97, 113 S.Ct. 2510 (citation omitted). Extending the rejection of the Linkletter doctrine to civil cases, the Harper Court stated that “we can scarcely permit ‘the substantive law [to] shift and spring’ according to ‘the particular equities of [individual parties’] claims’ of actual reliance on an old rule and of harm from a retroactive application of the new rule.” Id. (citation omitted; brackets in original).21 Harper squarely held that “[w]hen this Court applies a rule of federal law to the parties before it, that rule is the controlling interpretation of federal law and must be given full retroactive effect in all cases still open on direct review and as to all events, regardless of whether such events predate or postdate our announcement of the rule.” Id.
When this court took note of these Supreme Court developments in 1998, it recognized that they “substantially undermined” the retroactivity jurisprudence of Mendes. Washington, 718 A.2d at 1075. “Indeed,” the court added, “we think it probable that if ... Harper, and not Link-letter, had reflected the Supreme Court’s jurisprudence in 1978, this court would have given its decision in Mendes full retroactive effect.” Id. at 1076. The Washington court acknowledged that it was nonetheless bound to follow Mendes, because overruling a precedent of this court “can be effected only by this court en banc.” Id. (citing M.A.P. v. Ryan, 285 A.2d 310, 312 (D.C.1971)). In the present case, this court is en banc; and Mendes is ripe for reconsideration.
Preliminarily, our use of the term “reconsideration” suggests that we have some choice in the matter. In their arguments before us, the parties to this case have assumed that we are free to choose either to embrace Griffith and Harper or to adhere to Mendes, at least when we articulate new rules of local, i.e., District of Columbia, law.22 Cf. Great Northern R.R. Co. v. Sunburst Oil & Refining Co., 287 U.S. 358, 364-65, 53 S.Ct. 145, 77 *227L.Ed. 360 (1932) (state courts are free to adopt different rules regarding the retroactive effect of their own interpretations of state law). As a matter of “judicial policy,” the Supreme Court ordinarily defers to the decisions of this court on matters of District law. See Whalen v. United States, 445 U.S. 684, 687-88, 100 S.Ct. 1432, 63 L.Ed.2d 715 (1980); Pernell v. Southall Realty, 416 U.S. 363, 368-69, 94 S.Ct. 1723, 40 L.Ed.2d 198 (1974). That said, we need not dwell further on the question whether this court has the legal prerogative to retain Mendes; for on the assumption that we do have discretion to choose, we nonetheless opt to jettison Mendes in favor of Griffith and Harper. We proceed, therefore, to the merits. As we do so, we are mindful that the “commonality of sovereignty” of the District of Columbia and the United States, District of Columbia v. Ray, 305 A.2d 531, 534 (D.C.1973), lends support to, even if it does not settle, the argument that we should follow current rather than former Supreme Court teaching. Cf. Hornstein v. Barry, 560 A.2d 530, 536-37 n. 15 (D.C.1989) (en banc) (applying federal rule serves “the interest of harmony between court systems and uniformity of result in the same geographical area”).
On its merits, the Mendes retroactivity doctrine is subject to the same criticisms that felled Linkletter in the Supreme Court. First, on its own terms, the doctrine is difficult to apply in a principled and predictable fashion. There is, for instance, considerable ambiguity in the threshold requirement that this court must announce a “new” legal rule in order to trigger the Mendes inquiry. When is a legal rule “new” enough to meet this requirement? To say that the test is whether the new rule is a “clear break,” or “newly minted,” or “not clearly foreshadowed,” offers little real guidance; not only do such standards invite subjective adjudication, but they also are not consistent with each other. “Not clearly foreshadowed” is a more lenient standard than “clear break” while “newly minted” sounds like it falls somewhere else on the continuum of novelty — though exactly where is not easy to say. Our holding in Noble, for example, may not have been “clearly foreshadowed,” but it was hardly a “newly minted” principle given that it was based on a hoary old principle of statutory construction and that it ratified an interpretation of the GTCA that had held sway in the federal system for a decade. Whether our holding was a “clear break” from the past is debatable, because that depends on what is included in “the past.”
Passing the threshold inquiry of whether the court has announced a new legal rule, the criteria that Mendes requires the court to investigate and balance are also problematic — none more so than the rebanee factor on which appellants place so much weight in the present case. Under Mendes, the court must consider “[t]he degree of rebanee by the btigants before the court, and in some cases by the pubhc at large, on the legitimacy of the prior rule,” whether the rebanee was “reasonable” in view of any “intimations” that the prior rule might change, and “the degree of hardship that the parties before the court, and others in general, may sustain” if the rebanee interests are overridden by retrospective appbeation of the new rule. 389 A.2d at 789-90. This can fairly be characterized as an open-ended inquiry that is bereft of standards to guide it; exactly whose rebanee should be considered, how that reliance is to be established and measured, whether there were signals that the prior rule should not be rebed upon, whether it was reasonable to expect the btigants and others to pick up on those signals and conduct themselves accordingly, what kind and degree of hardship should be considered, how to determine *228and assess the hardship, are some of the questions that the appellate court is expected to resolve. Moreover, although these questions are partly factual, the court is expected to answer them without the benefit of a true evidentiary record detailing the extent to which the old rule of law was relied upon. The court is compelled to speculate, and to base its answers on hypotheses that are untested and probably untestable.
Appellants’ reliance claims in the present case exemplify these problems. Appellants contend that D.C.Code offenders may have relied on the understanding that revocation of parole would not result in loss of street time credit when they tendered guilty pleas, decided to violate conditions of their parole, defended themselves against revocation of parole, or planned for their anticipated reparole. Appellants further contend that Superior Court judges may have relied on the prior understanding when they sentenced some offenders, and that the Parole Board may have relied on that erroneous understanding when it revoked the parole of some offenders. These are all hypothetical possibilities, but what are we to make of them when the record does not tell us, for example, whether such reliance on the pre-Noble rule actually occurred, how many people were adversely affected, how those people were affected, or who those people were? Again, appellants contend that it was reasonable for D.C.Code offenders to rely on the information about their sentences which they received from the Department of Corrections prior to Noble. Yet there was information available to these offenders which could have alerted them that the Department’s sentence computations were subject to correction. This information included the regulation issued by the United States Parole Commission, the Ninth Circuit’s decision in Tyler, and the fact that for a decade after the GTCA was enacted, D.C.Code offenders in federal custody continued to forfeit street time credit when their parole was revoked. The record is silent as to whether and to what extent D.C.Code offenders were aware of these facts.
The equitable balancing that Mendes requires is subject to the further criticism that it is ad hoc and standardless. In the present case for example, Mendes would require us to weigh the extent of reliance on the pre-Noble rule against such considerations as the interest in treating D.C.Code offenders the same whether they are under federal or local supervision; public safety concerns over the premature release of felons who have demonstrated their inability to comply with parole conditions; and the longstanding principle that a convicted defendant must serve the sentence mandated by law, and cannot rely on an administrative mistake to insist on a more lenient sentence. Once these and other factors are identified, however, the court is on its own as far as balancing them is concerned.
In addition to the difficulty of applying Mendes in a principled way, there is a more fundamental objection to be made. By allowing courts to exercise their discretion to make their legal rulings prospective, the Mendes doctrine is antithetical to basic principles governing the role of courts. Unlike legislation, which is presumptively prospective in operation, judicial decisions are presumptively retrospective. In every case, the court’s ruling is a statement to the parties before it telling them what the law was that governed the acts and events leading up to the lawsuit— not what the law should have been or what it ought to be. Saying what the law was rather than what it should be is the judicial mandate in a democratic polity, distinct from the legislative mandate to make *229rules of law for the future. In the words of the Supreme Court, “ ‘the nature of judicial review1 strips us of the quintessentially ‘legislative]’ prerogative to make rules of law retroactive or prospective as we see fit.” Harper, 509 U.S. at 95, 113 S.Ct. 2510 (quoting Griffith, 479 U.S. at 322, 107 S.Ct. 708 (brackets in original)). For as Justice Harlan observed:
If we do not resolve all cases before us on direct review in light of our best understanding of governing ... principles, it is difficult to see why we should so adjudicate any case at all.... In truth, the Court’s assertion of power to disregard current law in adjudicating cases before us that have not already run the full course of appellate review, is quite simply an assertion that our constitutional function is not one of adjudication but in effect of legislation.
Griffith, 479 U.S. at 323, 107 S.Ct. 708 (quoting Mackey v. United States, 401 U.S. at 679, 91 S.Ct. 1160 (concurring opinion)).
As Justice Harlan’s statement implies, a prospective-only ruling is indeed a statement by the court that it is at liberty to “disregard” what it has determined to be the governing law in the case before it. The same may be said of a ruling that is only “partially” or “selectively” retrospective, i.e., that is retroactively applied only to the parties before the court in the case in which it is announced — it is a statement that the court may “disregard” the governing law when similarly situated parties come before it.23 The premise of Mendes is that this court has the authority to disregard governing law — including, as appellants assert in this case, a Congressional enactment — based on the notions of fairness that the judges on the court may be persuaded to entertain in the particular case. This premise begs the question of where we derive such authority. It is no answer to say that we may sit as a court of equity, because “it is well established that ‘courts of equity can no more disregard statutory and constitutional requirements and provisions than can courts of law.’ ” Immigration and Naturalization Serv. v. Pangilinan, 486 U.S. 875, 883, 108 S.Ct. 2210, 100 L.Ed.2d 882 (1988) (quoting Hedges v. Dixon County, 150 U.S. 182, 192, 14 S.Ct. 71, 37 L.Ed. 1044 (1893)). An overriding legal principle, such as the constitutional ex post facto prohibition in Bouie, may preclude retroactive application of a rule of law. Absent the intervention of such a countervailing principle, it is our judicial duty to apply the law in the cases before us, not to disregard it.
As Bouie illustrates, the Constitution does impose specific limits on the retroac-tivity of judicial decisions. The existence of these Constitutional limits, some of which overlap with the criteria identified in Mendes, responds to the fairness concerns which underlay our opinion in that case. In Landgraf the Supreme Court identified several provisions of the Constitution that embody antiretroactivity principles:
The Ex Post Facto Clause flatly prohibits retroactive application of penal legislation. Article I, § 10, cl. 1, prohibits States from passing another type of retroactive legislation, laws ‘impairing the Obligation of Contracts.’ The Fifth Amendment’s Takings Clause prevents the Legislature (and other government actors) from depriving private persons of vested property rights except for a ‘public use’ and upon payment of ‘just compensation.’ The prohibitions on ‘Bills of Attainder’ in Art. I, §§ 9-10, prohibit legislatures from singling out *230disfavored persons and meting out summary punishment for past conduct.... The Due Process Clause also protects the interests in fair notice and repose that may be compromised by retroactive legislation; a justification sufficient to validate a statute’s prospective application under the Clause ‘may not suffice’ to warrant its retroactive application.
Landgraf, 511 U.S. at 266, 114 S.Ct. 1483 (footnote and citations omitted). Accord, Lynce v. Mathis, 519 U.S. 433, 439-40, 117 S.Ct. 891, 137 L.Ed.2d 63 (1997) (“The specific prohibition on ex post facto laws is only one aspect of the broader constitutional protection against arbitrary changes in the law. In both the civil and the criminal context, the Constitution places limits on the sovereign’s ability to use its lawmaking power to modify bargains it has made with its subjects.”) Even though the cited Constitutional provisions are primarily a check on the legislature, Bowie shows that they may apply to judicial acts as well by virtue of the Due Process Clause.
The Constitution is not the only source of legal principles that may operate to limit the retroactivity of new rules of law announced in judicial decisions. The Supreme Court has referred to “the unsurprising fact that, as courts apply ‘retroactively’ a new rule of law to pending cases, they will find instances where that new rule, for well-established legal reasons, does not determine the outcome of the case.” Reynoldsville Casket Co. v. Hyde, 514 U.S. 749, 758-59, 115 S.Ct. 1745, 131 L.Ed.2d 820 (1995). For example, “a limitation inherent in the principle [of retroac-tivity] itself’ is that “[n]ew legal principles, even when applied retroactively, do not apply to cases already closed.” Id. at 758, 115 S.Ct. 1745.24 Along with this principle of finality, the Court in Reynoldsville Casket offered other examples of non-Constitutional limitations on retroactivity:
Thus, a court may find (1) an alternative way of curing the ... violation, or (2) a previously existing, independent legal basis (having nothing to do with retroac-tivity) for denying relief, or (3) as in the law of qualified immunity, a well-established general legal rule that trumps the new rule of law, or (4) a principle of law, such as that of “finality” present in the Teague context, that limits the principle of retroactivity itself.
Id., 514 U.S. at 759, 115 S.Ct. 1745. The Court has, however, ruled out “simple reliance (of the sort at issue in Chevron Oil) as a basis for creating an exception to Harper’s rule of retroactivity.” Id. at 759, 115 S.Ct. 1745.
We are persuaded by the foregoing considerations that the time has come to abandon the retroactivity doctrine of Mendes and conform our jurisprudence to that of the Supreme Court. We adopt the “firm rule of retroactivity,” Landgraf, 511 U.S. at 278 n. 32, 114 S.Ct. 1483, that the Court articulated in Griffith and Harper.25
*2313. Retroactive Application of Noble
It remains to apply that rule to our holding in Noble.26 We have found no due process or other legal prohibition against the retroactive application of Noble to appellants. Under Griffith and Harper, the equitable considerations that appellants advance do not suffice to preclude retroactivity. It is therefore our conclusion that the Department of Corrections acted properly when it recomputed appellants’ sentences to subtract their street time credit in compliance with our ruling in Noble that the GTCA did not repeal the street time forfeiture provision of D.C.Code § 24-206(a).
Appellants argue that Griffith and Harper do not require this result because the Supreme Court did not rule out the possibility of purely prospective rulings in exceptional cases. Rather, appellants emphasize, the Court ultimately held only that “[wjhen this Court does not ‘reserve the question whether its holding should be applied to the parties before it,’ ” then the new rule must be applied retroactively. Harper, 509 U.S. at 97, 113 S.Ct. 2510 (quoting James B. Beam Distilling Co. v. Georgia, 501 U.S. 529, 539, 111 S.Ct 2439, 115 L.Ed.2d 481 (1991) (opinion of Souter, J.)). Appellants argue that the present case presents the exceptional situation that the Supreme Court contemplated, inasmuch as this court did explicitly reserve the question of retroactivity when it rendered its decision in Noble. Since we did not apply the rule in Noble to Matthew Noble himself, say appellants, the rule of retroactivity enunciated in Griffith and Harper is inoperative, and instead the Linkletter-based equitable balancing test of Mendes continues to govern. See, e.g., Shah v. Pan American World Servs., Inc., 148 F.3d 84, 91 (2nd Cir.1998).
We are not persuaded by this argument, because its factual premise is flawed. Our opinion in Noble “reserved” the question of whether our ruling should be applied retroactively because we were only answering a question of law that was certified to us from the D.C. Circuit. It was not our place to do more. But when Matthew Noble’s case returned to the federal courts, those courts did apply our holding to him; and had the issue been before us, we would have been hard pressed to disagree with that outcome. Moreover, the rule that we announced in Noble has not been applied just to the prisoner in that case. By the time we decided Noble, that rule had already been applied, properly, to an untold number of D.C.Code offenders who were in federal custody. And following Noble, the Tenth Circuit held our ruling retroactive in Johnson v. Kindt. If the principal thrust of Griffith and Harper was to end selective retroactivity, those cases mandate retroactive application here.
In addition, we think that appellants’ argument is also flawed on the conceptual level. The Supreme Court did not “rethink retroactivity” and reject the Linklet-*232ter-Stovall-Chevron doctrine in favor of a “firm rule of retroactivity” only to revert back to that doctrine every time it has to decide whether to apply a new rule of law to the parties before it. Rather, the fundamental principles of Griffith and Harper must govern that decision. Appellants are correct that the Supreme Court has left the door open to the possibility that it might declare a new rule of law to be purely prospective in effect even if it is not required by the Constitution to do so. But that door is open by just a crack, and only for truly “exceptional cases.” Reynoldsville Casket, 514 U.S. at 761, 115 S.Ct. 1745 (Kennedy, J., concurring). It is not easy to envision the kind of case that would justify a purely prospective ruling that is not required to be prospective by due process or another Constitutional command.27 See id. at 761-62, 115 S.Ct. 1745 (“We cannot foresee the myriad circumstances in which the question might arise.... When a hard case presents the question of our authority to deny relief in a retroactivity case, that will be soon enough to resolve it....”). It is enough that we can echo here what the Court said when it rejected an argument for pure prospectivity in Ryder v. United States, 515 U.S. 177, 184-85, 115 S.Ct. 2031, 132 L.Ed.2d 136 (1995): “whatever the continuing validity of Chevron Oil after Harper ... and Reynoldsville Casket ..., there is not the sort of grave disruption or inequity involved in [retroactive application of Noble to appellants here] that would bring that doctrine into play.”
CONCLUSION
The Department of Corrections acted properly when it corrected appellants’ sentences in accordance with our holding in Noble that the GTCA did not repeal the street time forfeiture provision of D.C.Code § 24-206(a). The Superior Court therefore acted properly when it denied appellants’ petitions for writs of habeas corpus which challenged the correction of their sentences on Constitutional and equitable grounds. Accordingly, the orders on appeal are
Affirmed.
. In the National Capital Revitalization and Self-Government Improvement Act of 1997, Pub.L. No. 105-33 ("Revitalization Act”), § 11201, 111 Stat. 712, 734 (1997), Congress provided, among other things, for the transfer of the District’s prison system to federal authority. By December 31, 2001, Lorton Correctional Complex is to be closed and all Lorton inmates — D.C.Code felony offenders— are to be transferred to facilities operated or contracted for by the Bureau of Prisons. See D.C.Code § 24-1201 (2000 Supp.). For purposes of this opinion it is not necessary to address the potential impact of the changes wrought by the Revitalization Act on the claims of the parties, and we refrain from doing so.
. Id. at 454. See Morton v. Mancan, 417 U.S. 535, 550, 94 S.Ct. 2474, 41 L.Ed.2d 290 (1974) ("In the absence of some affirmative showing of an intention to repeal, the only permissible justification for a repeal by implication is when the earlier and later statutes are irreconcilable.”).
. The subtraction of improper street time credit extended what the affected prisoners previously had been told was their full term, and also resulted in later dates for mandatory release from prison. The recalculation did not necessarily mean longer periods of incarceration, however. Prisoners who, with street time credit, were about to be released did remain incarcerated longer than they had expected. Otherwise, where parole rehearing dates had already been scheduled (prior to the former mandatory release date), those rehearing dates were not changed. Where no parole rehearing had been scheduled prior to the former mandatory release date, the prisoner's file was reviewed (by the Parole Board or, following a transfer of jurisdiction, by the Commission) to determine if a re-parole hearing was warranted. Thus, the District contends, "there [has been] in place a process by which, on a case-by-case basis, the equities regarding particular prisoners can be weighed against the threat they pose to public safety in determining whether release is proper."
We nonetheless fully appreciate, as appellants argue, that even if the effect of the sentence recalculation required by Noble was mitigated by release from prison, the extension of the parole term (and the concomitantly increased risk of revocation and reimpris-onment) in itself represented a substantial restriction on liberty.
. Appellant Randall Patrick Martin was sentenced in 1992 to two to six years for attempted distribution of cocaine. He was released on parole in 1995. In February 1998, the D.C. Board of Parole issued a parole violator warrant for Martin, because Martin had been charged with assault. The Board revoked Martin's parole in April 1998 after a hearing. Crediting him with his accrued street time, the Department of Corrections advised Martin that his sentence would terminate the following month, in May 1998. In the wake of our en banc decision in Noble, however, the Department rescinded Martin’s street time credit shortly before he was scheduled to be released. Martin remained in prison for eight months before he was re-paroled, and the loss of street time credit effectively added more than two years to the time he spent serving his sentence.
The facts of appellant Mark Childs's case are similar. Childs was sentenced in 1993 to 16 to 48 months for attempted distribution of cocaine, and he was paroled (apparently for a second time) in June 1997. The Parole Board revoked Childs’s parole after a hearing in April 1998 based on his failure to meet with his parole officer. The Department of Corree-*214tions calculated that with credit for street time, Childs would be released from prison no later than June 1998. Following our en banc decision in Noble, however, the Department recomputed Childs's sentence. Childs remained in prison until he was re-paroled in July 1999. His total sentence was extended two and a half years to January 2001.
Appellant Maurice Delane Davis was sentenced in 1991 to 30 to 90 months for attempted distribution of cocaine. In April 1997, the Parole Board revoked Davis's parole for reasons not disclosed in the record before us and ordered that he be released on his mandatory release date in December 1998. Subsequently, in compliance with Noble, the Department of Corrections voided Davis's 110 days of street time credit and added the time to his mandatory release date. Davis was denied re-parole because of his "extensive record of repetitive criminal behavior, including robbery, other assaultive conduct, and multiple escapes;" a history of parole violations including criminal conduct; and the ultimate evaluation that he was "too serious a risk to merit a further release on parole.”
. To be precise, the Department of Corrections credited appellants and other parole violators with street time accrued up until the issuance of their parole violation warrants, but not for time on parole between the issuance of the warrants and the ultimate revocation of parole.
. U.S. Const., Art. I, § 9, cl. 3; Art. I, § 10, cl. 1. The Ex Post Facto clauses are directed at legislative enactments, not judicial decisions. See Frank v. Mangum, 237 U.S. 309, 344, 35 S.Ct. 582, 59 L.Ed. 969 (1915).
. Accord, Metheny v. Hammonds, 216 F.3d 1307, 1310-11 (11th Cir.2000); Stephens v. Thomas, 19 F.3d 498, 500 (10th Cir.1994); Cortinas v. United States Parole Comm’n, 938 F.2d 43, 46 (5th Cir.1991); Glenn v. Johnson, 761 F.2d 192, 194-95 (4th Cir.1985); Lerner v. Gill, 751 F.2d 450, 457 (1st Cir.1985). See also Crowley v. London, 780 F.2d 440, 444 (4th Cir.1985) (no ex post facto violation where court construes existing law which had been erroneously applied, since no new law is being applied retroactively).
. As the Supreme Court has explained:
The essence of judicial decisionmaking— applying general rules to particular situations — necessarily involves some peril to individual expectations because it is often difficult to predict the precise application of a general rule until it has been distilled in the crucible of litigation. See L. Fuller, Morality of Law 56 (1964) ("No system of law— whether it be judge-made or legislatively enacted — can be so perfectly drafted as to leave no room for dispute”).
Rivers v. Roadway Express, Inc., 511 U.S. at 312, 114 S.Ct. 1510.
The recent decision of the Supreme Court in Johnson v. United States illustrated the proposition that unexpected — but not indefensible or unforeseeable — -rulings will be applied retroactively even where the effect is to sanction increased criminal punishment. In that case the Court construed the Sentencing Reform Act of 1984 to authorize a district court to impose a second term of supervised release (a form of supervision analogous to parole) after revocation of the first term. The Court upheld the imposition of a second term of supervised release in the case before it even though the Court agreed that "the power to impose another term of supervised release following imprisonment [is] a power not readily apparent from the text” of the statute; the Court conceded that its construction of the statutory language was "unconventional;" and the Court acknowledged that courts of appeals in nine federal circuits had ruled, contrarily, that the statute did not permit imposition of post-revocation supervised release. See Johnson, 529 U.S. at 698 n. 2, and 706, 120 S.Ct. 1795.
. Cf. Bozza v. United States, 330 U.S. 160, 67 S.Ct. 645, 91 L.Ed. 818 (1947). In that case, involving judicial rather than administrative *220correction of an invalid sentence, the Supreme Court stated:
The Constitution does not require that sentencing should be a game in which a wrong move by the judge means immunity for the prisoner. In this case the court only set aside what it had no authority to do, and substituted directions required by the law to be done upon the conviction of the offender.
Id. at 166-67, 67 S.Ct. 645 (citations omitted). See also United States v. Campbell, 985 F.Supp. 158, 160 (D.D.C.1997) ("a court not only can, but must, increase a previously imposed sentence if that sentence is later found to be statutorily invalid”).
. In Johnson v. Williford, supra, the Ninth Circuit held that equitable estoppel and due process precluded the government from revoking the parole of a felon who was convicted under a statute requiring a minimum ten year prison term without the possibility of parole, but who was released on parole for fifteen months before the error was discovered. But see Hawkins v. Freeman, 195 F.3d 732 (4th Cir.1999) (en banc), where the Fourth Circuit held that reincarceration of a mistakenly released petitioner who spent two exemplary years on parole did not violate substantive due process, because the petitioner had no fundamental liberty interest in retaining freedom granted in error and his re-imprisonment was not so outrageous as to shock the conscience. See generally Timothy P. Lyden, Note, If the Parole Board Blunders, Does the Fourteenth Amendment Set the Prisoner Free? Balancing the Liberty Interests of Erroneously Released Prisoners, 88 Geo. L.J. 565 (2000) (criticizing Hawkins methodology and outcome, and recommending a balancing test to resolve substantive due process claims of erroneously released prisoners).
.We have no hard data, but we suspect that there were not many cases in which the retroactive application of Noble actually did thwart the Board’s intentions. In many cases, the Board would doubtless have revoked parole anyway, even if it had understood that street time might be lost, because the violations of parole were serious ones. If there were cases in which the Board would not have revoked parole had it understood that street time might be lost, those cases were presumably decided during the ten-year period before the issuance in April 1997 of the panel decision in Noble; for even though that decision was challenged, it alerted the Board to the potential consequence of revocation. But many if not most of the prisoners in those cases should have been able to complete their sentences without loss of street time credit; especially since the Department of Corrections did not begin to forfeit street time until a year later, after this court issued its en banc order in April 1998 adopting the original majority opinion in Noble.
. See D.C.Code § 24-1231 (2000 Supp.), enacted as part of the Revitalization Act, Pub.L. No. 105-33, § 11231, 111 Stat. 712, 745 (1997).
. Appellants postulate other ways in which reliance on the pr e-Noble understanding might have caused actual prejudice. Appellants suggest that some defendants might not have pled guilty, and that some parolees might not have violated their parole or might have defended more vigorously and successfully against revocation, if they had known that the consequences of revocation would include loss of street time. For purposes of due process analysis we discount these examples. In the first place, they are purely hypothetical and, in our view, highly improbable. *222in the second place, as explained above, it was no secret to defendants and parolees that forfeiture of street time might occur upon revocation of parole, if only because federal authorities disagreed with the District over the issue.
Appellants also suggest that it is possible that some Superior Court judges might have imposed lighter sentences had they known that parole revocation would entail loss of street time credit. However, we have been provided no reason to think either that Superior Court judges were unaware of this contingency or, if they were, that it made a difference given the inability to predict at sentencing whether a particular defendant will ever face revocation of parole.
. Harper v. Virginia Dep’t of Taxation, 509 U.S. 86, 94, 113 S.Ct. 2510, 125 L.Ed.2d 74 (1993) (quoting Kuhn v. Fairmont Coal Co., 215 U.S. 349, 372, 30 S.Ct. 140, 54 L.Ed. 228 (1910) (Holmes, J., dissenting)). See, e.g., United. States v. Schooner Peggy, 5 U.S. (1 Cranch) 103, 110, 2 L.Ed. 49 (1801). The courts of the District of Columbia historically adhered to the traditional view of retroactivity. See Jawish v. Morlet, 86 A.2d 96, 97 (D.C.1952); Ruppert v. Ruppert, 77 U.S.App. D.C. 65, 68, 134 F.2d 497, 500 (1942).
. In Linkletter the Supreme Court developed a balancing test for determining whether new rules of criminal law — in that case, the exclusionary rule of Mapp v. Ohio, 367 U.S. 643, 81 S.Ct. 1684, 6 L.Ed.2d 1081 (1961) — would be applied retroactively in collateral attacks on convictions. In Stovall v. Denno, 388 U.S. 293, 87 S.Ct. 1967, 18 L.Ed.2d 1199 (1967), the Court held the Linkletter balancing test applicable in direct appeals of criminal convictions as well. The Court thereafter extended Linkletter to civil cases, holding in Chevron Oil Company v. Huson, 404 U.S. 97, 106-07, 92 S.Ct. 349, 30 L.Ed.2d 296 (1971), that new legal rules could be denied retroactive effect if doing so would avoid injustice or hardship without unduly undermining the purpose and effect of the new rule.
As we discuss below, Linkletter, Stovall and Chevron have now all been overruled. The Supreme Court has renounced the flexible balancing test of those cases in favor of a return to the traditional rule that judicial decisions are retroactive.
. The four specific criteria identified in Mendes tilt the inquiry in favor of non-retro-activity. The overall balancing test must, however, take into account other pertinent considerations, including the benefits of retrospective application of the new rule. See Mendes, 389 A.2d at 788 (quoting Linkletter, 381 U.S. at 629, 85 S.Ct. 1731 (court must "weigh the merits and demerits in each case”)).
. The issue in French concerned the effect of a zoning regulation which provided that no order of the Board of Zoning Adjustment authorizing the erection or alteration of a structure would be valid for longer than six months unless a building permit was applied for within that period. The property owner in French did not apply for a permit because she relied on an advisory opinion of the Corporation Counsel which concluded that the running of the six-month period was tolled by the filing and pendency of a petition for review of the BZA order. We held that the advisory opinion was invalid because it was contrary to a statute which specifically provided that filing of a petition for review did not stay enforcement of the order. We refrained from applying that holding to the parties before us, however, in view of their reliance on the Corporation Counsel’s opinion. See French, 658 A.2d at 1030-32.
. Of the four factors listed in Mendes, we think that only the reliance factor weighs in favor of non-retroactivity here. The second and third factors — avoidance of altering property or contract rights, and rewards to plaintiffs who seek to change the law — are not relevant to this case. The fourth factor — the burdensome impact of retroactive changes in the law on the administration of justice — is relevant; however, we cannot conclude that it cuts against retroactivity in this case. The District does not complain that retroactive application of Noble has been administratively burdensome, and the record does not demonstrate that it has been. In point of fact, it could well be argued that equalizing the treatment of D.C.Code offenders wherever they are incarcerated and requiring them to serve the sentences which Congress intended them to serve are beneficial to the administration of justice.
. As Justice Harlan was later to observe, the Linkletter "doctrine was the product of the Court’s disquietude with the impacts of its fast-moving pace of constitutional innovation in the criminal field.” Mackey v. United States, 401 U.S. 667, 676, 91 S.Ct. 1160 (1971) (Harlan, J., concurring).
. "By 'final,' we mean a case in which a judgment of conviction has been rendered, the availability of appeal exhausted, and the time for a petition for certiorari elapsed or a petition for certiorari finally denied.” Griffith, 479 U.S. at 321 n. 6, 107 S.Ct. 708.
. See also Harper, 509 U.S. at 95 n. 9, 113 S.Ct. 2510 ("our decision today makes it clear that 'the Chevron Oil test cannot determine the choice of law by relying on the equities of the particular case’ and that the federal law applicable to a particular case does not turn on 'whether Ditigants] actually relied on [an] old rule [or] how they would suffer from retroactive application’ of a new one.")
.We are bound to follow Griffith and Harper in any case involving a new rule of federal law. See Harper, 509 U.S. at 100, 113 S.Ct. 2510 ("The Supremacy Clause, U.S. Const., Art. VI, cl. 2, does not allow federal retroac-tivity doctrine to be supplanted by the invocation of a contrary approach to retroactivity under state law. Whatever freedom state courts may enjoy to limit the retroactive operation of their own interpretations of state law ... cannot extend to their interpretations of federal law.”)
. Such “selective application of new rules [also] violates the principle of treating similarly situated [parties] the same.” Griffith, 479 U.S. at 323, 107 S.Ct. 708.
. In the context of collateral attacks on criminal convictions, this statement needs to be qualified. See Teague v. Lane, 489 U.S. 288, 109 S.Ct. 1060, 103 L.Ed.2d 334 (1989); United States v. McKie, 315 U.S.App.D.C. 367, 370-72, 73 F.3d 1149, 1152-54 (1996). Cf. Kirk v. United States, 510 A.2d 499, 504 (D.C.1986); Fields v. United States, 466 A.2d 822, 828 (D.C.1983).
. In her separate opinion, our colleague Judge Ruiz expresses the “fear” that a firm rule of retroactivity may "stunt the growth” of the common law in this jurisdiction. Post at 238. With respect, we are confident that this fear is unwarranted. For centuries, courts (including this court and its predecessors) "developed” the common law successfully, exactly as our colleague hopes we will continue to do, while adhering to the principle of retroactivity that we reinstate in this case. Moreover, in the decades following Mendes, this court did not once find it necessary to invoke that decision to hold that a new *231rule of common law would be non-retroactive. As exemplars of this court’s occasional announcement of new common law principles, which might be “chilled” by a rule of retroactivity, our colleague cites Carl v. Children’s Hospital, 702 A.2d 159 (D.C.1997) (en banc), and Williams v. Baker, 572 A.2d 1062 (D.C.1990) (en banc). Post at 234. Yet this court expressly held Carl retroactive in Washington v. Guest Servs., Inc., 718 A.2d 1071 (D.C.1998), and Williams retroactive in Jones v. Howard University, 589 A.2d 419 (D.C.1991). Retroactivity did not deter us in those instances from fulfilling our duty to "declare the common law of the District of Columbia.” Post at 234.
. Appellants argue that our adoption of the retroactivity principles of Griffith and Harper should not be applied in this case, but should be prospective-only. We perceive no basis for granting appellants' request.
. Perhaps a case where retroactive application of a new rule would involve grave disruption to the administration of justice would present a compelling argument for pure pros-pectivity. Cf. Northern Pipeline Constr. Co. v. Marathon Pipe Line Co., 458 U.S. 50, 88, 102 S.Ct. 2858, 73 L.Ed.2d 598 (1982), where the Supreme Court declared the broad grant of jurisdiction to Article I bankruptcy courts unconstitutional and applied its ruling prospectively only (except that the party that raised the constitutional challenge in that case did receive the relief it requested).