GORSUCH, Circuit Judge.

Nearly a decade after Keith Prost pleaded guilty to engaging in a money laundering conspiracy, the Supreme Court issued United States v. Santos, 553 U.S. 507, 128 S.Ct. 2020, 170 L.Ed.2d 912 (2008), interpreting the statute under which Mr. Prost stood convicted. In light of that new interpretation, Mr. Prost says his guilty plea should be revisited and his conviction undone. The problem is that Mr. Prost never pursued a statutory interpretation argument in his own trial court proceedings, on appeal, or in his initial collateral challenge to his conviction. And he concedes that 28 U.S.C. § 2255(h) doesn't permit him a second collateral challenge to raise such an argument now. So, it would seem, Mr. Prost's long-final criminal conviction must remain just that: final.

Not so, says Mr. Prost. Even now, Mr. Prost insists, he should be allowed to proceed with his statutory interpretation argument under 28 U.S.C. § 2241, if not § 2255. The district court rejected this argument and so must we. Congress has told us that federal prisoners challenging the validity of their convictions or sentences may seek and win relief only under the pathways prescribed by § 2255. To this rule, Congress has provided only one exception: a federal prisoner may resort to § 2241 to contest his conviction if but only if the § 2255 remedial mechanism is "inadequate or ineffective to test the legality of his detention." 28 U.S.C. § 2255(e). And that exception doesn't apply here. Mr. Prost was free to bring a Santos-type statutory interpretation argument in his initial § 2255 motion, and an initial § 2255 motion offered him an adequate and effective means for testing such an argument. The fact that § 2255 bars Mr. Prost from bringing his statutory interpretation argument now, in a second § 2255 motion almost a decade after his conviction, doesn't mean the § 2255 remedial process was ineffective or inadequate to test his argument. It just means he waited too long to raise it.

I

In 1998, Mr. Prost found himself, along with a dozen colleagues, indicted in the Eastern District of Missouri for participating in a drug trafficking operation. Eventually, Mr. Prost pleaded guilty to one count of conspiring to possess and distribute methamphetamine, in violation of 21 U.S.C. § 846, and two counts of conspiring to launder proceeds derived from a drug dealing operation, in violation of 18 U.S.C. § 1956. See United States v. Prost, No. 98-CR-264-ERW (E.D.Mo. Jan. 22, 1999) (unpublished). After being convicted and sentenced in 1999, Mr. Prost filed a collateral challenge pursuant to 28 U.S.C. § 2255 in the Eastern District of Missouri seeking to vacate his sentence, though not his conviction. The district court rejected that effort in 2004, Prost v. United States, No. 00-CV-98-ERW (E.D.Mo. Jan. 12, 2004) (unpublished), and the Eighth Circuit agreed with the district court's disposition, Prost v. United States, No. 04-1394 (8th Cir. July 6, 2004) (unpublished).

That seemed to mark the end of the legal road for Mr. Prost. But then, nearly a decade after his conviction became final, the Supreme Court handed down Santos. There, at least in the context of an illegal lottery operation, the Court held that the term "proceeds" in the federal money laundering statute, 18 U.S.C. § 1956, means "profits," and not just "gross receipts." 553 U.S. at 514, 128 S.Ct. 2020. So, to establish a violation of § 1956, the Supreme Court instructed, the government had to show that the defendant in Santos laundered the lottery's profits, not merely its gross receipts.1

Appreciating Santos's potential significance for his case, Mr. Prost filed a petition for habeas corpus pursuant to 28 U.S.C. § 2241, asking to have his money laundering conspiracy convictions reopened and overturned. He claimed that the funds he laundered were only the gross receipts, not the profits, of the drug dealing operation, and so he was innocent of any money laundering crime. Because § 2241 petitions must be brought in the district of incarceration, Mr. Prost filed his petition in federal district court in Colorado as, by this time, he was serving his sentence at the federal prison camp in Florence, Colorado.

The district court immediately identified one problem with Mr. Prost's § 2241 petition. Congress long ago decided that a federal prisoner's attempt to attack the legality of his conviction or sentence generally must be brought under § 2255, and in the district court that convicted and sentenced him—here, the Eastern District of Missouri. Meanwhile, § 2241 petitions, brought in the district where the prisoner is confined, are generally reserved for complaints about the nature of a prisoner's confinement, not the fact of his confinement. Before the Colorado district court, Mr. Prost acknowledged all this, and conceded that his petition does not challenge the condition of his confinement but rather attempts a frontal assault on his conviction.

Under these circumstances, a district court would normally dismiss Mr. Prost's § 2241 petition without prejudice so that he might refile it as a § 2255 motion in the appropriate sentencing court. See Bradshaw v. Story, 86 F.3d 164, 167 (10th Cir. 1996). Alternatively, a district court might construe Mr. Prost's petition as a § 2255 motion and transfer it to the sentencing court for review. But Mr. Prost sought to avoid both these outcomes. Before the Colorado district court, he admitted that years earlier he had filed (and lost) one motion under § 2255—a motion in which he attacked only his sentence. Mr. Prost also admitted that Congress, in the Antiterrorism and Effective Death Penalty Act of 1996 ("AEDPA"), has restricted the availability of second or successive § 2255 motions to claims involving either newly discovered evidence strongly suggestive of innocence or new rules of constitutional law made retroactive by the Supreme Court. See 28 U.S.C. § 2255(h).2 And, Mr. Prost admitted, a new statutory interpretation, such as the one announced in Santos, is neither of those things. Mr. Prost thus acknowledged that he is prohibited by § 2255(h) from bringing a Santos-based challenge to his money laundering convictions in a second motion.

But, Mr. Prost noted, § 2255(e) includes a so-called "savings clause" which sometimes allows a federal prisoner to resort to § 2241 to challenge the legality of his detention, not just the conditions of his confinement. To fall within the ambit of savings clause and so proceed to § 2241, a prisoner must show that "the remedy by motion [under § 2255] is inadequate or ineffective to test the legality of his detention." 28 U.S.C. § 2255(e). Section 2255 is both of these things in his case, Mr. Prost submitted, because § 2255(h) bars him from pursuing an argument that, in light of the Supreme Court's statutory interpretation in Santos, he should've been found not guilty of the money laundering charges against him.

The district court rejected this contention, holding that "[t]he fact that Mr. Prost may be barred [by § 2255(h)] from raising his claims in a second or successive motion . . ., by itself, does not demonstrate that the remedy provided in § 2255 is inadequate or ineffective." See Prost v. Wiley, 08-CV-2246-BNB, 2008 WL 4925667, at *2 (D.Colo. Nov. 13, 2008) (unpublished). It is from this ruling that Mr. Prost now appeals.3

II

A criminal conviction is "a decisive and portentous event." Wainwright v. Sykes, 433 U.S. 72, 90, 97 S.Ct. 2497, 53 L.Ed.2d 594 (1977). Before any conviction, the accused enjoys a presumption of innocence, a right to trial by jury, and a range of evidentiary and procedural guarantees secured by the Constitution and multifold statutes. All of this stems from our society's aspiration to protect the innocent against the possibility of a wrongful conviction, an aspiration long given voice in the common law by Blackstone's maxim that it is a better thing for ten guilty persons to escape punishment than for one innocent to suffer wrongly. See 4 William Blackstone, Commentaries *358; see also Alexander Volokh, "n Guilty Men," 146 U. Penn. L.Rev. 173 (1997). While perhaps no criminal justice system in history can convincingly claim to have succeeded entirely in preventing the conviction of the innocent, ours concentrates its considerable resources "at [the] time and place [of trial or plea in an effort] to decide, within the limits of human fallibility, the question of guilt or innocence of one of [our fellow] citizens." Wainwright, 433 U.S. at 90, 97 S.Ct. 2497. We then double- and sometimes triple-check the result through our layered appellate system. Only after that appellate process is exhausted does the criminal proceeding yield what our legal system recognizes as a "final judgment."

The principle of finality, the idea that at some point a criminal conviction reaches an end, a conclusion, a termination, "is essential to the operation of our criminal justice system." Teague v. Lane, 489 U.S. 288, 309, 109 S.Ct. 1060, 103 L.Ed.2d 334 (1989); see also McCleskey v. Zant, 499 U.S. 467, 491, 111 S.Ct. 1454, 113 L.Ed.2d 517 (1991). In every case there comes a time for the litigation to stop, for a line to be drawn, and the parties encouraged to move forward rather than look back. "A procedural system which permits an endless repetition of inquiry into facts and law in a vain search for ultimate certitude," the Supreme Court has explained, "implies a lack of confidence about the possibilities of justice that cannot but war with the effectiveness of underlying substantive commands. . . . There comes a point where a procedural system which leaves matters perpetually open no longer reflects humane concern but merely anxiety and a desire for immobility." McCleskey, 499 U.S. at 492, 111 S.Ct. 1454 (quoting Paul M. Bator, Finality in Criminal Law and Federal Habeas Corpus for State Prisoners, 76 Harv. L.Rev. 441, 452-53 (1963)). Anxiety and immobility, of course, are accompanied by other social costs—to victims, their families, to future potential victims, to the government, and to the courts—that revisiting and retesting convictions five or ten years old—or (as here) even older—can involve.

Even though a criminal conviction is generally said to be "final" after it is tested through trial and appeal, Congress is free to provide still further safeguards against wrongful convictions.4 And Congress has done just that in many statutes over the course of our history. Most relevant for purposes of this case, in 28 U.S.C. § 2255 Congress has chosen to afford every federal prisoner the opportunity to launch at least one collateral attack to any aspect of his conviction or sentence. This, of course, is exactly the opportunity Mr. Prost exercised unsuccessfully back in 2004.

But Congress didn't stop there. If a prisoner's initial § 2255 collateral attack fails, as Mr. Prost's did, Congress has indicated that it will sometimes allow a prisoner to bring a second or successive attack. Recognizing the enhanced finality interests attaching to a conviction already tested through trial, appeal, and one round of collateral review, however, Congress has specified that only certain claims it has deemed particularly important—those based on newly discovered evidence suggestive of innocence, or on retroactively applicable constitutional decisions—may be brought in a second or successive motion. See 28 U.S.C. § 2255(h); supra n. 2.

Yet, even here Congress has provided an out. A prisoner who can't satisfy § 2255(h)'s conditions for a second or successive motion may obviate § 2255 altogether if he can show that "the remedy by motion" provided by § 2255 is itself "inadequate or ineffective to test the legality of his detention." 28 U.S.C. § 2255(e). In these "extremely limited circumstances," Caravalho v. Pugh, 177 F.3d 1177, 1178 (10th Cir.1999) (internal quotation omitted), a prisoner may bring a second or successive attack on his conviction or sentence under 28 U.S.C. § 2241, without reference to § 2255(h)'s restrictions. It is, however, the prisoner's burden to show that these conditions, prescribed by § 2255(e)'s so-called "savings clause," apply to his case. See Miller v. Marr, 141 F.3d 976, 977 (10th Cir.1998).

And that's where the rubber meets the road in this case. Mr. Prost seeks to avoid § 2255(h) and proceed directly to § 2241. But to do so, to overcome the presumption of finality attaching to his conviction, he must carry the burden of showing that § 2255(e)'s savings clause applies to his case. It is that question that lies at the heart of this case and to which we now turn.

III

In asking whether § 2255 is "inadequate or ineffective to test the legality of [a prisoner's] detention," the question naturally arises: compared to what? When trying to ascertain whether something is "inadequate or ineffective," after all, we usually ask: inadequate or ineffective to what task? Dictionaries define "inadequate" to mean "not equal to requirement," and "ineffective" as "[o]f such a nature as not to produce any, or the intended, effect." See 7 Oxford English Dictionary 770, 902 (2d ed.1989). Both definitions presuppose some metric or measure—some "requirement" or "effect"—that should be but isn't met. In what follows, we first define the yardstick against which we must measure § 2255's adequacy and effectiveness before then assessing how Mr. Prost's arguments stack up.

A

The relevant metric or measure, we hold, is whether a petitioner's argument challenging the legality of his detention could have been tested in an initial § 2255 motion. If the answer is yes, then the petitioner may not resort to the savings clause and § 2241. That this is the appropriate metric for measuring savings clause claims is clearly indicated by the clause's plain language, its context and history, as well as our own precedent.

First, the saving clause's text. The clause begins by juxtaposing the terms "inadequate or ineffective" with the phrase "to test the legality of [a prisoner's] detention." From this, the clause's intended requirement, effect, or metric starts to emerge. If a petitioner's argument challenging the legality of his detention could've been tested in a § 2255 motion, the clause is satisfied. In this way, the clause is concerned with process—ensuring the petitioner an opportunity to bring his argument—not with substance—guaranteeing nothing about what the opportunity promised will ultimately yield in terms of relief. The clause's text emphasizes this point further by proceeding to focus on the question whether the overall § 2255 motions process offers an adequate and effective "remedy." Section 2255(e) expressly distinguishes between the terms remedy and relief, stating that § 2241 is not available to a petitioner simply because a "court has denied him relief"; to invoke the savings clause, it must "also appear[] that the remedy by motion is inadequate or ineffective." § 2255(e) (emphasis added); see also 28 U.S.C. § 2254(b)(1)(A) (using the word remedy—in the context of requiring state prisoners to exhaust state remedies—to mean avenue for relief, not relief itself). Here again, the clause emphasizes its concern with ensuring the prisoner an opportunity or chance to test his argument. Here again it underscores that with this opportunity comes no guarantee about outcome or relief. The ultimate result may be right or wrong as a matter of substantive law, but the savings clause is satisfied so long as the petitioner had an opportunity to bring and test his claim.

Recognizing these features of the savings clause's plain language, we have long and repeatedly said that a petitioner's "[f]ailure to obtain relief under § 2255 does not establish that the remedy so provided is either inadequate or ineffective," Bradshaw, 86 F.3d at 166 (quoting Williams v. United States, 323 F.2d 672, 673 (10th Cir.1963)) (emphasis added), and that an "erroneous decision on a § 2255 motion" doesn't suffice to render the § 2255 remedy itself inadequate or ineffective, Sines v. Wilner, 609 F.3d 1070, 1073 (10th Cir.2010). Many other circuits have reached similar conclusions. See, e.g., Cradle v. U.S. ex rel. Miner, 290 F.3d 536, 539 (3d Cir.2002) ("Section 2255 is not inadequate or ineffective merely because the sentencing court does not grant relief.. . ."); In re Vial, 115 F.3d 1192, 1194 n. 5 (4th Cir.1997) (en banc) ("[T]he remedy afforded by § 2255 is not rendered inadequate or ineffective merely because an individual has been unable to obtain relief under that provision.").

Second, the immediate context in which the savings clause appears reveals that it doesn't guarantee multiple opportunities to test a conviction or sentence. When seeking a statute's ordinary meaning we must of course take care to study not just the particular isolated clause at issue but also its surrounding context. See U.S. Nat'l Bank of Or. v. Indep. Ins. Agents of Am., 508 U.S. 439, 454, 113 S.Ct. 2173, 124 L.Ed.2d 402 (1993) ("[T]ext consists of words living a communal existence, . . . the meaning of each word informing the others and all in their aggregate taking their purport from the setting in which they are used." (internal quotation and alteration omitted)). Here, the savings clause's near neighbor, § 2255(h), added in 1996 as part of AEDPA, restricts second and successive motions to those raising newly discovered evidence or new constitutional rulings. When Congress adopted § 2255(h), it was undoubtedly aware that prisoners might wish to press other sorts of arguments in second or successive motions. And it was surely aware that prisoners might seek to pursue second or successive motions based on newly issued statutory interpretation decisions. See Davis v. United States, 417 U.S. 333, 347, 94 S.Ct. 2298, 41 L.Ed.2d 109 (1974) (holding long before AEDPA that claims of innocence based on new statutory interpretations could be brought in a § 2255 motion); see also Merck & Co. v. Reynolds, ___ U.S. ___, 130 S.Ct. 1784, 1795, 176 L.Ed.2d 582 (2010) ("We normally assume that, when Congress enacts statutes, it is aware of relevant judicial precedent."). Indeed, cognizant that prisoners would seek to pursue motions based on new statutory interpretations, Congress in § 2255(f)(3) aimed to ensure that they may assert such claims in initial § 2255 motions without risk of being time-barred, explaining that the one-year statute of limitations for bringing a first § 2255 motion begins to run only from "the date on which the right asserted was initially recognized by the Supreme Court, if that right has been newly recognized by the Supreme Court and made retroactively applicable to cases on collateral review." See Haugh v. Booker, 210 F.3d 1147, 1149-50 (10th Cir. 2000) (recognizing that the new statutory interpretation announced in Bailey establishes the kind of "right" needed to take advantage of § 2255(f)(3)).5 But, just as evidently, in subsection (h) Congress chose to preclude petitioners from raising such statutory innocence claims—among many other kinds of claims—in second or successive § 2255 motions.6

Reading the savings clause in this context, it is evident that a prisoner generally is entitled to only one adequate and effective opportunity to test the legality of his detention, in his initial § 2255 motion. If the rule were otherwise—if the § 2255 remedial mechanism could be deemed "inadequate or ineffective" any time a petitioner is barred from raising a meritorious second or successive challenge to his conviction—subsection (h) would become a nullity, "a meaningless gesture." United States v. Barrett, 178 F.3d 34, 50 (1st Cir.1999). If the rule were otherwise—if, say, courts were to read subsection (h) as barring only losing second or successive motions—the statute's limitations would be effectively pointless and, as the Second Circuit has recognized, Congress would have "accomplished nothing at all in its attempts—through statutes like the AEDPA—to place limits on federal collateral review." Triestman v. United States, 124 F.3d 361, 376 (2d Cir.1997); see also McGhee v. Hanberry, 604 F.2d 9, 10 (5th Cir.1979) ("It is well established that a prior unsuccessful § 2255 motion is insufficient, in and of itself, to show the inadequacy or ineffectiveness of the remedy."); cf. Gray-Bey v. United States, 201 F.3d 866, 876 (7th Cir.2000) (Easterbrook, J., dissenting) ("Judicial emphasis must be on `test': a § 2255 motion is not `inadequate or ineffective' merely because the petitioner loses. Nor do the changes made by the AEDPA, which limit the number of § 2255 motions (and the time to file them) render § 2255 inadequate or ineffective to test the lawfulness of detention. No one is entitled to more than one collateral attack."); United States v. Guerrero, Nos. 10-3177 & 10-3179, ___ Fed.Appx. ___, 2011 WL 286362 (2011) (unpublished) ("That [defendant] may be barred from bringing another § 2255 motion . . . does not establish that the remedy set out in § 2255 is inadequate or ineffective.").

Third, our reading comports with the even larger statutory context in which subsections (e), (f), and (h) appear. The emphasis on providing a single opportunity to test arguments, rather than any guarantee of relief or results, can be found throughout §§ 2255 and 2254. Federal prisoners seeking to take advantage of new rulings of constitutional magnitude that would render their convictions null and void are not always allowed to do so in second or successive motions. See, e.g., 28 U.S.C. § 2255(h) (permitting federal prisoners to take advantage only of new constitutional rules that the Supreme Court has expressly declared to have retroactive application); see also Dodd v. United States, 545 U.S. 353, 125 S.Ct. 2478, 162 L.Ed.2d 343 (2005). State prisoners seeking to show that the state court's adjudication on the merits was contrary to clearly established federal law may not rely on Supreme Court precedent decided after that adjudication, even in their initial federal habeas petitions. See 28 U.S.C. § 2254(d)(1); see also Brown v. Greiner, 409 F.3d 523, 533-34 (2d Cir.2005). And the ineffectiveness or incompetence of counsel during collateral post-conviction proceedings "shall not be a ground for relief in a proceeding arising under section 2254." § 2254(i). When viewed in the context of AEDPA as a whole, it is unsurprising that the plain language of § 2255 means what it says and says what it means: a prisoner can proceed to § 2241 only if his initial § 2255 motion was itself inadequate or ineffective to the task of providing the petitioner with a chance to test his sentence or conviction.7

Fourth, an examination of the history of the savings clause confirms our reading. Before § 2255 was enacted in 1948, federal prisoners who wanted to challenge any aspect of their detention—its legality or the conditions of their confinement—had to do so in a § 2241 petition filed in the district where they were incarcerated. This meant that all federal collateral attacks were concentrated in those few districts housing federal penitentiaries. United States v. Hayman, 342 U.S. 205, 212-14, 72 S.Ct. 263, 96 L.Ed. 232 (1952). That situation proved problematic when Congress, in 1867, statutorily expanded the availability of collateral relief, flooding the federal district courts whose jurisdictions encompassed federal prisons with a tide of post-conviction motions. Id. To alleviate the burden, Congress passed § 2255, requiring federal prisoners to bring challenges to the lawfulness of their detentions in the districts where they were originally convicted and sentenced. See 28 U.S.C. § 2255(a). From this history, the Supreme Court has concluded that § 2255 wasn't adopted to expand or "impinge upon prisoners' rights of collateral attack upon their convictions," but only to address the "difficulties that had arisen in administering" habeas corpus. Hayman, 342 U.S. at 219, 72 S.Ct. 263.8 Simply put, Congress included the savings clause to ensure that those who couldn't comply with § 2255's new venue mandate were still provided with at least one opportunity to challenge their detentions. This isn't to say that the language Congress employed in the savings clause is necessarily limited only to remedying venue problems. But it is to say that the history of the clause illustrates that Congress's purpose in enacting it surely wasn't to ensure that a prisoner will win relief on a meritorious successive motion, or receive multiple bites at the apple.

Fifth, our reading of the savings clause is in harmony with the decisions of this court that have permitted resort to the clause. So, for example, in Spaulding v. Taylor, the defendant's sentencing court had been abolished by the time the prisoner sought to bring his initial collateral attack. 336 F.2d 192 (10th Cir.1964). Because the defendant's § 2255 motion had to be brought in the (now nonexistent) sentencing court, that remedial mechanism was necessarily inadequate and ineffective to test the legality of his detention and so we permitted him to bring his challenge in the district court where he was confined. Id. at 193-94. Likewise, we've held that resort to § 2241 is the norm rather than the exception when a military prisoner seeks to challenge the results of his court martial. This is due to the evanescent nature of court martial proceedings: the sentencing court literally dissolves after sentencing and is no longer available to test a prisoner's collateral attack. See Ackerman v. Novak, 483 F.3d 647, 649 (10th Cir.2007); see also Witham v. United States, 355 F.3d 501, 505 (6th Cir.2004). In every case in this circuit dealing with the savings clause, we have recognized the narrowness of the language and allowed resort to § 2241 sparingly, only when an adequate or effective means for testing a § 2255 petition was genuinely absent.

Having held that the savings clause is satisfied so long as a petitioner could've raised his argument in an initial § 2255 motion, it is obvious Mr. Prost's effort to invoke the clause must fail. He does not contend that he faced any difficulty in bringing an initial § 2255 motion in the court where he was sentenced. Neither does he dispute that an initial § 2255 motion would've been up to the task of testing the Santos argument he now presses. After all, Mr. Santos included in his initial § 2255 motion the argument that the money laundering statute's use of the term "proceeds" encompassed only profits, not gross revenues. Despite repeated challenges by the government, taken all the way to the Supreme Court, Mr. Santos ultimately prevailed. Plainly, § 2255 is up to the job of testing the question whether the money laundering statute requires proof of profits. Of course, § 2255(h)'s restrictions on second and successive motions bar Mr. Prost from trying a Santos argument now, nearly a decade after his conviction and long after pursuing his initial § 2255 motion. But that fact doesn't mean that the § 2255 remedial regime is inadequate or ineffective to test such an argument. It only means that, in Congress's considered view, finality concerns now predominate and preclude relitigation of Mr. Prost's criminal judgment.

B

Seeking to avoid this result, Mr. Prost complains that the notion endorsed in Santos —that § 1956 requires proof of profits, not just proceeds—wasn't in circulation at the time of his first § 2255 motion. In his view, a federal prisoner should have recourse to § 2241 through the savings clause any time he can demonstrate that his initial § 2255 proceeding finished before the Supreme Court announced a new statutory interpretation that would likely undo his conviction. Put differently, Mr. Prost says he should be excused for failing to bring a "novel" argument for relief that the Supreme Court hadn't yet approved in Santos. In aid of his cause, he says the Ninth Circuit has adopted his view, see Harrison v. Ollison, 519 F.3d 952 (9th Cir.2008), and Stephens v. Herrera, 464 F.3d 895 (9th Cir.2006), though (as we shall see) other circuits have not followed suit, see infra Sections III.C. and III.D.

Respectfully, we cannot agree that the absence of Santos from the U.S. Reports at the time of a prisoner's first § 2255 motion has anything to do with the question whether § 2255 was an inadequate or ineffective remedial mechanism for challenging the legality of his detention. As we've explained, it is the infirmity of the § 2255 remedy itself, not the failure to use it or to prevail under it, that is determinative. To invoke the savings clause, there must be something about the initial § 2255 procedure that itself is inadequate or ineffective for testing a challenge to detention. Problems of that magnitude surely existed in Spaulding and Ackerman, but just as clearly they do not exist here.

We readily acknowledge that, at the time of his first § 2255 motion, it is likely that neither Mr. Prost nor his counsel imagined the particular statutory interpretation argument Santos ultimately vindicated. But in much the same way that a student's failure to imagine a novel or creative answer to an exam question doesn't make the exam an inadequate or ineffective procedure for testing his knowledge, the fact that Mr. Prost or his counsel may not have thought of a Santos-type argument earlier doesn't speak to the relevant question whether § 2255 itself provided him with an adequate and effective remedial mechanism for testing such an argument. The § 2255 remedial vehicle was fully available and amply sufficient to test the argument, whether or not Mr. Prost thought to raise it. And that is all the savings clause requires.

At bottom, Mr. Prost's novelty test represents no more than a frank policy disagreement with § 2255(h). His real position seems to be less that an initial § 2255 motion would have been inadequate or ineffective to test his argument and more that he has a good excuse for having failed to pursue that argument earlier. The problem, of course, is that in subsection (h) Congress identified the excuses it finds acceptable for having neglected to raise an argument in an initial § 2255 motion. Failing to pursue novel statutory interpretations is not on that list, though Congress was aware situations like this one might arise and fully intended § 2255(h) to bar otherwise meritorious successive petitions. The simple fact is that Congress decided that, unless subsection (h)'s requirements are met, finality concerns trump and the litigation must stop after a first collateral attack. Neither is this court free to reopen and replace Congress's judgment with our own.9

C

Perhaps recognizing the problems with his proposed novelty test, Mr. Prost suggests that, at the very least, he should be excused from having failed to pursue a Santos-type argument in his initial § 2255 motion because Santos's reading of the money laundering statute was erroneously foreclosed under Eighth Circuit law at the time he was convicted and sentenced. But when it comes to abiding the plain language of § 2255(e) and the larger statutory context and purpose we outlined in Section III.A, Mr. Prost's alternative "erroneous circuit foreclosure" test fares no better than his novelty test.

Critically, Mr. Prost doesn't—and can't—dispute that he was entirely free to raise and test a Santos-type argument in his initial § 2255 motion. Instead, he argues only that a Santos-type argument likely would have been rejected on the merits at the district court and circuit panel levels because of adverse circuit precedent, leaving him with only en banc and certiorari petitions to try to undo that precedent. But, as we have explained, the plain language of the savings clause does not authorize resort to § 2241 simply because a court errs in rejecting a good argument. The savings clause doesn't guarantee results, only process. Neither does this fact change merely because the court's error on the merits happens to be induced by preexisting circuit precedent. We readily acknowledge that circuit precedent sometimes requires judges to reject a claim on its merits, and sometimes that precedent is quite wrong in doing so. But as we've already detailed at length, the possibility of an erroneous result—the denial of relief that should have been granted—does not render the procedural mechanism Congress provided for bringing that claim (whether it be 28 U.S.C. §§ 1331, 1332, 2201, 2255, or otherwise) an inadequate or ineffective remedial vehicle for testing its merits within the plain meaning of the savings clause. Whether a statutory interpretation argument is rejected on the merits by a circuit court on the basis of a newly crafted but deficient test, or by application of an old but equally bad test found in circuit precedent makes no difference. Legal error has occurred. And whenever legal error occurs it may very well mean circuit law is inadequate or deficient. But that does not mean the § 2255 remedial vehicle is inadequate or ineffective to the task of testing the argument.

Santos helps illustrate the point. As it happens, Mr. Santos didn't face any adverse precedent in his circuit when he brought an initial § 2255 motion arguing that proceeds must mean profits under § 1956. But Mr. Santos's motion succeeded in overturning plenty of adverse § 1956 precedent in other circuits, and there's no reason to think an initial § 2255 motion would've been any less of an adequate or effective means for testing the § 1956 question in the Supreme Court had Mr. Santos's motion originated from a circuit where the precedent happened to be against him. It is no more than a quirk of fate that the case the Supreme Court used to decide the § 1956 question happened to come from a circuit where the argument was open rather than foreclosed by erroneous circuit precedent. The U.S. Reports are, after all, replete with instances where the Supreme Court has rewarded litigants who took the trouble to challenge adverse circuit precedent. While there is of course no guarantee that any en banc or certiorari petition will be granted, Mr. Prost can't dispute that § 2255, as a procedural vehicle, was (and has proven to be) an adequate and effective means for testing the question he now seeks to pose.10

Just like his novelty test, Mr. Prost's erroneous circuit foreclosure test also asks us to disregard § 2255(h). Mr. Prost's proffered test rests on the view that asking petitioners to pursue arguments precluded by existing circuit law is an imprudent waste of time. On this basis, he effectively (and once again) says that he should be excused for having failed to raise his Santos-argument in his first § 2255 motion. But it remains an implacable fact that when enacting subsection (h) Congress didn't consider Mr. Prost's excuse strong enough to overcome the finality interests attaching to a conviction already tested through trial, appeal, and one round of collateral review. Neither can we see how we might permit Mr. Prost to proceed with his claim through the § 2255(e)'s savings clause without nullifying (or at least doing much violence to) the restrictions on second and successive motions Congress has imposed in § 2255(h). And this we decline to do.

Mr. Prost's erroneous circuit foreclosure test overlooks as well § 2255(f). As we have noted, Congress enacted the savings clause against the backdrop of the Supreme Court's holding in Davis, 417 U.S. at 346-47, 94 S.Ct. 2298, permitting federal prisoners to press statutory claims of innocence in § 2255 motions. Cognizant of such statutory claims, Congress added § 2255(f)(3) to AEDPA to ensure that petitioners may assert new statutory innocence claims in a first § 2255 motion without being time-barred. This suggests that Congress was aware of claims like Mr. Prost's when it passed AEDPA but chose not to authorize them in successive collateral attacks. To adopt Mr. Prost's proffered test would require us to ignore this judgment as well.

Although Mr. Prost suggests there is something unusual about barring a claim that rests on a correct and previously foreclosed statutory interpretation, the fact is that many other provisions of AEDPA limit the ability of prisoners to reap the benefit of unforeseeable but helpful new legal developments. The result in his case is thus hardly as anomalous as he suggests. For example, it is well-settled that not every claim of innocence—even one based on new constitutional rules— is authorized in a second or successive collateral challenge. As a result of the interaction between § 2255(f)(3) and § 2255(h)(2) (two provisions included in AEDPA), federal prisoners who file a second or successive § 2255 motion seeking to take advantage of a new rule of constitutional law will usually be time-barred, except in those rare cases where the Supreme Court "announces a new rule of constitutional law and makes it retroactive within one year." See Dodd, 545 U.S. at 359, 125 S.Ct. 2478 (emphasis added). State prisoners face this same result. See § 2244(d)(1)(C); § 2244(b)(2)(A). Even in a first federal collateral attack, a state prisoner may not rely on a novel Supreme Court decision issued after the state's adjudication of his claim to demonstrate that adjudication on the merits was "contrary to, or involved an unreasonable application of, clearly established Federal law," as required by § 2254(d)(1). And though the writ of habeas corpus in its earliest form was largely a remedy against confinement imposed by a court lacking jurisdiction, see McCleskey, 499 U.S. at 478, 111 S.Ct. 1454, this court has barred a state prisoner convicted of murder and sentenced to death by the wrong sovereign from bringing a successive collateral attack to contest his conviction on this basis. See In re Wackerly, No. 10-7062, at 5 (10th Cir. Sept. 3, 2010). This is because, like a statutory claim of innocence, lack of jurisdiction is not one of the two authorized grounds upon which a successive § 2254 motion may be filed. Id.

Untethered from the plain language of the savings clause and in disregard of that clause's neighboring provisions, Mr. Prost's erroneous circuit foreclosure test would also undermine Congress's obvious purposes by upsetting the balance it sought to achieve. Concerned with protecting the innocent, Congress chose in § 2255 to afford federal prisoners the opportunity to attack their convictions or sentences even after the extensive procedures provided for in the plea process, at trial, and on direct appeal. But seeking to temper that interest with the need to provide a degree of finality to criminal convictions, Congress also placed specific restraints on this collateral review. And it is hardly surprising that these restraints become increasingly restrictive as a prisoner's conviction or sentence is tested through more rounds of review—a fact the Supreme Court has told us falls "well within the compass" of the "evolutionary process" of statutory habeas developments. Felker, 518 U.S. at 664, 116 S.Ct. 2333 ("[J]udgments about the proper scope of the writ are normally for Congress to make."). We can well imagine an alternative statutory regime that might strike the balance differently than Congress has done. But it is not our place to adopt a test that replaces the balance Congress reached with one of our own liking.11

Mr. Prost emphasizes that the Seventh Circuit has used the erroneous circuit foreclosure test, see In re Davenport, 147 F.3d 605, 610 (7th Cir.1998), and that the test has been employed by a couple other courts, see, e.g., Wofford v. Scott, 177 F.3d 1236, 1244 (11th Cir.1999); Reyes-Requena v. United States, 243 F.3d 893, 904 (5th Cir.2001). But Mr. Prost (and the concurrence) overstate the test's acceptance and offer no persuasive reason to adopt it. As we've already noted, the Ninth Circuit has offered a very different test. See supra Section III.B. And, as we will discuss in a moment, the Second and Third Circuits have vigorously pursued another test still. See infra Section III.D. (Indeed, in creating the erroneous circuit foreclosure test, the Seventh Circuit expressly rebuked the Second Circuit's test. See In re Davenport, 147 F.3d at 611.) Even among those cases that have employed the erroneous circuit foreclosure test, moreover, none has addressed the textual and structural clues we've discussed. For example, the Seventh Circuit in In re Davenport seemed to rely primarily on "practical" considerations, see infra n. 12, and did not consider that the language of the savings clause asks whether § 2255 is adequate to test the legality of a prisoner's detention, suggesting a concern with assuring some mechanism for assessing the prisoner's claim. The court likewise did not examine how the savings clause's meaning is affected by Congress's inclusion in § 2255(h) of two specific kinds of claims that may be brought in a second or successive collateral attack. It did not discuss how § 2255(f) fits into this picture, or how the saving clause coexists with the larger statutory structure limiting many otherwise successful successive petitions. And, notably too in light of more recent events, it is unclear whether even the Seventh Circuit itself continues to ascribe to the erroneous circuit foreclosure test. See Morales v. Bezy, 499 F.3d 668, 672-73 (7th Cir.2007); id. at 674 (Rovner, J., dissenting) (questioning whether, in light of the majority's analysis, a petitioner in the Seventh Circuit would still have to establish circuit foreclosure to access § 2241 via the savings clause).

D

While we reject Mr. Prost's novelty and erroneous circuit foreclosure tests, we pause to acknowledge one more way in which he might have sought to invoke the savings clause and proceed to § 2241. According to the Second and Third Circuits, a petitioner may proceed to § 2241 not only when a court is unavailable to entertain a § 2255 petition—as we have long held in Spaulding and Ackerman —but also when the application of § 2255(h)'s bar against a second or successive motion for collateral review would seriously threaten to render the § 2255 remedial process unconstitutional. See Triestman, 124 F.3d at 377 (savings clause may be triggered where "the failure to allow for collateral review would raise serious constitutional questions"); In re Dorsainvil, 119 F.3d 245, 248 (3d Cir.1997) ("Were no other avenue of judicial review available for a party who claims that s/he is factually or legally innocent as a result of a previously unavailable statutory interpretation, we would be faced with a thorny constitutional issue.").

In this way, the Second and Third Circuits acknowledge Congress's authority, in enacting subsection (h), to prescribe which excuses are sufficiently weighty to overlook a petitioner's failure to bring an argument in a first motion and allow him a second or successive collateral attack. And in doing so, these circuits reject Mr. Prost's apparent view that courts may use the savings clause as a means to adorn subsection (h) with new excuses they think sensible, the essential foundation on which he seeks to construct the novelty and circuit foreclosure tests. At the same time, if and when the narrowness of subsection (h) poses a difficulty of constitutional dimension, the Second and Third Circuits say, a court may step in to permit the petition to proceed. In support of their view, the Second and Third Circuits rest on the traditional canon of statutory construction that courts should seek to interpret Congress's statutory handiwork in light of and consistent with the Constitution's commands.12

Whether the savings clause may be used in the fashion the Second and Third Circuits have suggested to avoid serious constitutional questions arising from application of § 2255(h) is an important question. So are the questions whether, when, and how the application of § 2255(h)'s limits on second or successive motions might (ever) raise a serious constitutional question. These are, however, matters we decline to pursue in this particular case. We decline to pursue them in part because Mr. Prost hasn't. Although Mr. Prost briefly intimates that denying him access to § 2241 would present "constitutional issues," Opening Br. at 13-14, at no point does he develop any argument why that is so, or even identify what provision of the Constitution he thinks would be offended by the imposition of § 2255(h)'s bar in his case. We generally avoid entertaining arguments for reversing a district court's judgment that were not adequately developed by a petitioner in his opening brief. See Hill v. Kemp, 478 F.3d 1236, 1250-51 (10th Cir.2007); Headrick v. Rockwell Int'l Corp., 24 F.3d 1272, 1277-78 (10th Cir. 1994) (White, J., sitting by designation). Neither do we see any reason to depart from our general practice in this case, given that significant and largely uncharted questions of the Constitution's meaning, questions whose proper outcome is far from certain, hang in the balance. See generally In re Alvarado, No. 10-4205 (10th Cir. Dec. 2, 2010) (unpublished) (rejecting a suspension clause challenge to § 2255(h) restrictions); Felker, 518 U.S. at 664, 116 S.Ct. 2333 (rejecting a suspension clause challenge to AEDPA's analogous limitations on second and successive § 2254 petitions); Herrera v. Collins, 506 U.S. 390, 401, 113 S.Ct. 853, 122 L.Ed.2d 203 (1993) (holding that "[d]ue process does not require that every conceivable step be taken, at whatever cost, to eliminate the possibility of convicting an innocent person [as this] would all but paralyze our system for enforcement of the criminal law" (internal quotation omitted)). So it is that we leave these constitutional questions for another day and another case.

IV

The concurrence reaches exactly the same result we do but by a different course. It takes a separate route in part because, it says, this court's decision "creat[es]" a circuit split. See Concurrence at 599, 605. But of course our decision does nothing of the sort. Long before we arrived on the scene the circuits were already divided three different ways on how best to read the savings clause. We hardly "creat[ed]" any of this. Id. Rather, this case has called on us to enter an already messy field where we have sought to tread carefully, studying the competing points of view expressed by our sister circuits before offering our own considered independent judgment, as we are duty-bound to do.

The concurrence likewise errs when it charges (repeatedly) that we have "bar[red]" or "preclude[d]" any means for federal prisoners to take advantage of new statutory decisions in second or successive habeas petitions, leaving this court somehow isolated from "every other" circuit and leaving the parties without the chance to "brief[ ] . . . the constitutional [avoidance] issue[ ]." See Concurrence at 599, 603, 605, 607. Again, we have done no such thing. As we have explained, it remains possible that this court might permit a petitioner in Mr. Prost's position to invoke the savings clause and to do so for the constitutional avoidance reasons stated by the Second and Third Circuits. Today, we close only two roads through the savings clause (novelty and erroneous circuit foreclosure), leaving at least a third open for future litigants to brief and argue (constitutional avoidance). Should this court ultimately join the Second and Third Circuits, we would reach the same substantive outcome as the other circuits that have interpreted the savings clause. It is unclear why the concurrence chooses to disregard these facts.

In places, the concurrence suggests that it would've been better—that is, more judicially restrained—for this court first to assume without deciding that the erroneous circuit foreclosure test comports with the savings clause's language and structure and, then, proceed to hold that Mr. Prost fails even under its terms. See, e.g., Concurrence at 600. But the concurrence fails to explain how this would have been the more restrained approach, and in the end the concurrence tellingly abandons the project.

To be sure, assuming without deciding the validity of a particular test is often the narrower and easier approach to resolving a case, and we recognize that it very well may have been easier and narrower in our earlier savings clause cases to assume without deciding the propriety of the erroneous circuit foreclosure test. See Concurrence at 7-8 (citing unpublished cases). But the same just doesn't hold true in this case. Far from easily and narrowly resolving Mr. Prost's claim, applying the "erroneous circuit foreclosure" test to this case would require us to decide many novel questions of law, a fact the concurrence fails to acknowledge.

To begin, we would have to ask and answer what it means to be "circuit foreclosed." In Mr. Prost's view, of course, any argument on the meaning of "proceeds" in the money laundering statute was barred by an earlier Eighth Circuit decision holding that the term "proceeds" in the Racketeer Influenced and Corrupt Organizations Act ("RICO") was not limited to the profits of illegal activity. But can a decision regarding one statute sufficiently "foreclose" a similar argument based on another? The concurrence disagrees with Mr. Prost and says no. But such an answer itself only raises another question: how much deference, if any, do we accord the Eighth Circuit's view of its own precedent? After all, the Eighth Circuit appears to agree with Mr. Prost that its RICO precedent did preclude the reading of the money laundering statute later adopted in Santos. See United States v. Williams, 605 F.3d 556, 567 (8th Cir.2010). So to reach the result it advocates, the concurrence has to disagree with the Eighth Circuit's reading of its own precedent. (The concurrence casts aside the problematic portion of the Eighth Circuit's decision in Williams as "careless." See Concurrence at 603 n. 2.)

Even if we could get past all that, still other novel questions remain to be decided under the concurrence's approach. A prisoner seeking to satisfy the erroneous circuit foreclosure test must of course demonstrate that he is "actually innocent" under a new statutory interpretation issued by the Supreme Court. But what kind of showing of innocence is sufficient to invoke the savings clause? And on what record? The original criminal record? Or may the parties introduce new evidence now in the course of a habeas proceeding brought nearly a decade after conviction? The proper answers to these questions are far from clear.

Neither is even this the end of it. Confusingly, the concurrence doesn't dispute that Mr. Prost's favored novelty test, endorsed by the Ninth Circuit, is inconsistent with the statute and must be rejected by this court. And it can hardly be the case that judicial restraint allows the concurrence to disagree freely with two circuits (the Eighth and Ninth) while simultaneously requiring this court to suppress any note of respectful disagreement with another (the Seventh).

In the end, the concurrence's claim to judicial restraint is simply unconvincing. Under its approach, the concurrence would have us bypass the question what Congress actually intended and apply the circuit foreclosure test without worrying about its provenance in the statutory text; handle plenty of knotty and novel legal questions about the test's application, in the process creating a significant and entangling body of advisory law about a test Congress never authorized; disagree with the Eighth Circuit's understanding of its own precedent; and still invite a separate schism with the Ninth Circuit. All this, despite the fact that the meaning of the savings clause and its application to this case are clear. Respectfully, when faced with the two possible paths presented to us by this case, we do not question for a moment that the easier, narrower, and far more judicially restrained course is for this court to apply faithfully the law Congress has written, not one it hasn't. After all, it is the "duty of the judicial department to say what the law is," Marbury v. Madison, 5 U.S. 137, 1 Cranch 137, 177, 2 L.Ed. 60 (1803) (emphasis added), not to develop a rococo jurisprudence about the application of a hypothetical law that Congress might've enacted—but clearly didn't—enact.13

Worse still, after chiding this court for failing to assume-without-deciding the propriety of the erroneous circuit foreclosure test, the concurrence abandons the very approach it advocates. Indeed, toward the end of its opinion, the concurrence reveals that its primary purpose is to defend the merits of the erroneous circuit foreclosure test, doing so in depth and at length. See Concurrence at 603-06. Here the concurrence finally, if implicitly, acknowledges that its real difficulty is just one of substance, a good faith disagreement about the validity of the erroneous circuit foreclosure test. After all, it surely can't be the case that the principles of judicial restraint the concurrence seeks to invoke in its opinion work only one way, requiring this court but not the concurrence to assume-without-deciding the propriety of the erroneous circuit foreclosure test.

What defense the concurrence does muster, however, is unpersuasive and serves only to confirm the correctness and comparative restraint of our approach. The concurrence doesn't attempt to reconcile the erroneous circuit foreclosure test with the plain language of the savings clause or surrounding provisions. Or, for that matter, with the congressional purposes we've outlined above. Neither does the concurrence explain why we should privilege a petitioner who was foreclosed by circuit precedent over a petitioner whose challenge wasn't yet recognized by the Supreme Court at the time of his initial § 2255 motion, the effect of adopting circuit foreclosure to the exclusion of the standard Mr. Prost and the Ninth Circuit advocate. Instead, the concurrence defends the erroneous circuit foreclosure test only on the basis that, even if "Congress failed to amend § 2255 expressly to permit" claims like Mr. Prost's, it is essential to find a way to allow them to avoid "a miscarriage of justice." Concurrence at 606.

Respectfully, this is less an argument for interpreting § 2255 than it is one for amending § 2255. Congress has expressed its views and purposes in § 2255 clearly and unequivocally. In § 2255, Congress has guaranteed every federal prisoner, after a trial and appeal, one additional adequate and effective opportunity to pursue any argument he wishes against his conviction or sentence, so long as it is brought within the applicable limitations period. In setting this rule, Congress has sought to balance the competing interests of vindicating the potentially innocent and providing a degree of finality to criminal convictions without pursuing either interest blind to the other. No doubt Congress could have struck a different balance than the one it did between these important ends. No doubt it might strike a different balance in the future. But unless and until Congress's currently expressed balance can be said to violate the Constitution, a question (again) not presented in this appeal, it is the job of this court, respecting the principles of judicial restraint, to enforce Congress's expressed purposes, not to replace them with our own.14 See Harrington v. Richter, ___ U.S. ___, 131 S.Ct. 770, 178 L.Ed.2d 624 (2011) ("Judges must be vigilant and independent in reviewing petitions for the writ, a commitment that entails substantial judicial resources. Those resources are diminished and misspent, however, and confidence in the writ and the law it vindicates undermined, if there is judicial disregard for the sound and established principles that inform its proper issuance."); Lamie v. U.S. Trustee, 540 U.S. 526, 542, 124 S.Ct. 1023, 157 L.Ed.2d 1024 (2004) ("[I]t is beyond our province to rescue Congress from its drafting errors, and to provide for what we might think . . . is the preferred result." (internal quotation omitted)); John F. Manning, Textualism and the Equity of the Statute, 101 Colum. L.Rev. 1, 18 (2001) ("Because statutory details may reflect only what competing groups could agree upon, legislation cannot be expected to pursue its purposes to their logical ends; accordingly, departing from a precise statutory text may do no more than disturb a carefully wrought legislative compromise.").15

The judgment of the district court is

Affirmed.

SEYMOUR, Circuit Judge, concurring in part and dissenting in part.

I agree with the majority that:

We agree on little else. In my view, the answer to the jurisdictional question we face is a straightforward "No." Mr. Prost had an adequate and effective opportunity to test the legality of his conviction because he was not foreclosed by adverse circuit precedent from bringing his actual innocence claim when he filed his initial § 2255 petition. Accordingly, we should dismiss the instant habeas petition for lack of jurisdiction. It is on that ground alone that I concur in the judgment and would affirm dismissal of Mr. Prost's petition.

The majority opinion, in contrast, sweeps far beyond this jurisdictional imperative. The majority takes the unnecessary position that the § 2255 remedy still would have been "adequate and effective" even if Mr. Prost faced adverse circuit precedent at the time he filed his initial § 2255 petition, which would have prevented success on his claim of actual innocence. The majority's rejection of what it terms the "erroneous circuit foreclosure test," see maj. op. at 589-93; see also id. at 593-94, flies in the face of judicial restraint—creating an unnecessary circuit split on an issue that was neither raised by the parties nor implicated by the facts of this case.

Respectfully, any implication by the majority that it is not creating a circuit split, see maj. op. at 594-96, is flatly wrong. Every other circuit deciding the issue has held, contrary to the majority, that § 2255 is "inadequate or ineffective" as a remedy in the extremely narrow situation in which it would procedurally bar a claim of actual factual innocence, like the one raised here by Mr. Prost, and where success on the actual innocence claim was previously barred by circuit precedent. This simple fact is undeniable. Moreover, the majority's conclusion that there is no effective remedy for an actually innocent person squarely raises serious constitutional questions, which the majority declines to address. To the extent the majority reaches the merits of the safety valve issue, I dissent from its interpretation of § 2255(e).

I.

"[T]he `cardinal principle of judicial restraint' is that `if it is not necessary to decide more, it is necessary not to decide more.'" Morse v. Frederick, 551 U.S. 393, 431, 127 S.Ct. 2618, 168 L.Ed.2d 290 (2007) (Breyer, J., concurring in the judgment in part and dissenting in part) (quoting PDK Labs., Inc. v. Drug Enforcement Admin., 362 F.3d 786, 799 (D.C.Cir.2004) (Roberts, J., concurring in part and concurring in the judgment)); see also Valley Forge Ins. Co. v. Health Care Mgmt. Partners, 616 F.3d 1086, 1094 (10th Cir.2010) (quoting favorably same language). Accordingly, we generally "answer[ ] only the questions we must, not those we can." Valley Forge Ins. Co., 616 F.3d at 1094. To protect against "improvident or ill-advised opinion[s]", Hill v. Kemp, 478 F.3d 1236, 1251 (10th Cir.2007), we typically require that the questions we answer be not only necessary to our disposition, but also adequately developed by the parties during the adversarial process. See id. at 1250-51; cf. Fed. R.App. P. 28. We should be especially cautious when we risk creating or exacerbating a circuit split. Cf. Am. Atheists, Inc. v. Davenport, ___ F.3d ___, 2010 WL 5151630, at *10 (10th Cir.2010) (Gorsuch, J., dissenting from denial of rehearing en banc). The majority fails to heed these rules, and it is not the least restrained in so doing.

A.

That Mr. Prost's claim is one of actual factual innocence is of utmost importance. He contends the erroneous interpretation prior to United States v. Santos, 553 U.S. 507, 128 S.Ct. 2020, 170 L.Ed.2d 912 (2008), of the term "proceeds" in the federal money laundering statute, see 18 U.S.C. § 1956(a)(1), resulted in his conviction for a nonexistent crime. Mr. Prost claims actual "factual innocence," not mere "legal innocence." See Bousley v. United States, 523 U.S. 614, 623, 118 S.Ct. 1604, 140 L.Ed.2d 828 (1998) ("`[A]ctual innocence' means factual innocence, not mere legal insufficiency."); see also Ellis v. Hargett, 302 F.3d 1182, 1186 n. 1 (10th Cir.2002) (discussing "legal innocence" and "factual innocence"); Beavers v. Saffle, 216 F.3d 918, 923 (10th Cir.2000) (same); cf. Herrera v. Collins, 506 U.S. 390, 419-21, 113 S.Ct. 853, 122 L.Ed.2d 203 (1993) (O'Connor, J., concurring) (recognizing a person may be both "legally and factually innocent"). The fact that Mr. Prost relies on a Supreme Court decision announcing a new substantive rule of statutory interpretation does not transform his claim into anything other than what it is: a claim of "actual innocence." See Bousley, 523 U.S. at 623, 118 S.Ct. 1604 (remanding analogous petition, despite procedural default, to permit petitioner to prove "actual innocence"); cf. Davis v. United States, 417 U.S. 333, 341-46, 94 S.Ct. 2298, 41 L.Ed.2d 109 (1974).

Our task is to decide whether the savings clause of § 2255(e) permits Mr. Prost to bring his claim of actual factual innocence in the instant § 2241 petition. That is the only question before us.

B.

Mr. Prost's claim, though one of first impression in this court, is easily resolved on narrow grounds, without questioning or rejecting the circuit foreclosure tests adopted by other circuits in cases of actual innocence.

Claims of actual factual innocence have been recognized in constitutional and habeas jurisprudence as among "the most compelling case[s] for habeas review." Murray v. Carrier, 477 U.S. 478, 501 n. 8, 106 S.Ct. 2639, 91 L.Ed.2d 397 (1986) (Stevens, J., concurring); see also id. at 495-96, 106 S.Ct. 2639 (majority opinion) (describing actual innocence as "an extraordinary case" in which "principles of comity and finality . . . must yield to the imperative of correcting a fundamentally unjust incarceration") (internal quotation marks omitted); Bousley, 523 U.S. at 623, 118 S.Ct. 1604 (recognizing actual innocence exception to post-AEDPA jurisprudential bar on procedurally defaulted claim); Schlup v. Delo, 513 U.S. 298, 317-23, 115 S.Ct. 851, 130 L.Ed.2d 808 (1995) (recognizing actual innocence exception to pre-AEDPA jurisprudential bar on second or successive petitions); id. at 324-25, 115 S.Ct. 851 ("[T]he individual interest in avoiding injustice is most compelling in the context of actual innocence . . . merit[ing] . . . a somewhat less exacting standard of proof. . . ."); Lopez v. Trani, 628 F.3d 1228, 1230-31 (10th Cir.2010) (recognizing actual innocence as basis for tolling AEDPA's one-year statute of limitations). In a non-binding unpublished decision, we recently observed that an actual innocence claim based on "erroneous specification of [petitioner's] offense in the indictment, plea agreement, and judgment of conviction" presented "unusual and compelling circumstances for federal post-conviction relief." Robinson v. Ledezma, No. 10-6123, 399 Fed.Appx. 329, 329, 2010 WL 4159461, at *1 (10th Cir. Oct. 14, 2010) (unpublished).

The Supreme Court recently reiterated the importance of actual innocence claims in In re Davis, ___ U.S. ___, 130 S.Ct. 1, 174 L.Ed.2d 614 (2009), confirming that such claims require careful scrutiny even when they are brought in a successive collateral attack. The petitioner in In re Davis filed an original habeas petition in the Supreme Court after being denied permission to file a second or successive petition in the Court of Appeals for the Eleventh Circuit. Id. The Supreme Court transferred the petition to the district court for an evidentiary hearing to "receive testimony and make findings of fact as to whether evidence that could not have been obtained at the time of trial clearly establishes petitioner's innocence." Id. at 1. The Court did so over the vigorous objection of Justice Scalia, joined by Justice Thomas, who emphasized that petitioner's actual innocence claim already had been rejected multiple times, by the State Supreme Court, a State Board of Pardons and Paroles, and a Federal Court of Appeals, and that the Court's action was in direct violation of AEDPA. See id. at 2-4 (Scalia, J., dissenting). Justice Stevens, joined by Justices Breyer and Ginsburg, concurred in the Court's action and, writing separately, pointed to the potential constitutional problems that might arise if AEDPA were interpreted narrowly to bar judicial review of such actual innocence claims. See id. at 1-2 (Stevens, J., concurring) (citing with approval Triestman v. United States, 124 F.3d 361, 377-80 (2d Cir.1997)).

Nevertheless, neither the Supreme Court nor this court has had occasion to decide whether, and under what conditions, the savings clause of § 2255(e) permits claims of actual innocence to be brought for habeas review when they are otherwise procedurally barred. See Robinson, 2010 WL 4159461, at *3 ("[T]his circuit . . . has yet to definitively adopt, let alone fully work out the content of, an actual-innocence exception to the exclusivity of § 2255."). So far, in unpublished decisions, we have deferred deciding those issues until they are adequately presented and we are actually required to reach them. See, e.g., id. at *3 (refusing to define actual-innocence exception where petitioner failed to establish threshold eligibility for relief, "[w]hatever the finer contours of such an exception might be"); see also, e.g., Saleh v. Davis, 398 Fed. Appx. 331, 332-33 (10th Cir.2010) (unpublished) (similar); Davis v. Ledezma, 393 Fed.Appx. 564, 565-66 (10th Cir.2010) (unpublished) (similar). So, too, in the present case we can, and should, refrain from delving into the thicket of defining the contours of an actual-innocence exception to the exclusivity of § 2255, much less critiquing extant formulations.

Every other circuit reaching the issue has concluded that the savings clause of § 2255(e) does, in fact, permit claims of actual innocence to be brought pursuant to § 2241 where a defendant was "foreclosed" by circuit precedent from successfully bringing his claim earlier. See infra Part II.A. (discussing cases). We favorably recognized this position in a prior nonbinding unpublished decision. See United States v. Apodaca, 90 Fed.Appx. 300, 304 & n. 10 (10th Cir.2004) (unpublished). In that case, we construed a pro se petitioner's submissions as a request to proceed with a successive § 2255 petition and denied relief on the ground that the petition contained neither newly discovered evidence nor a new, retroactively-applicable rule of constitutional law. Id. at 303-04. In so doing, we explained that "a petition for writ of habeas corpus under § 2241 demonstrating actual innocence may be an available remedy." Id. at 304. We cited with approval the Fifth Circuit's decision in Reyes-Requena v. United States, 243 F.3d 893 (5th Cir.2001), a case which presents an analogous question to the one we face in the present case. See Apodaca, 90 Fed.Appx. at 304 n. 10.

The Fifth Circuit in Reyes-Requena considered a § 2241 petition asserting actual innocence based on a new rule of statutory interpretation announced by the Supreme Court after the petitioner's conviction, direct appeal, and initial § 2255 motion. See Reyes-Requena, 243 F.3d at 895, 900-06. The petitioner there relied on Bailey v. United States, 516 U.S. 137, 116 S.Ct. 501, 133 L.Ed.2d 472 (1995), in which the Supreme Court held that a conviction for "use" of a firearm in violation of 18 U.S.C. § 924(c)(1) requires the government to show "active employment of a firearm." Reyes-Requena, 243 F.3d at 900 (quoting Bailey, 516 U.S. at 144, 116 S.Ct. 501).1 The petitioner claimed that, in of light Bailey, he was not guilty of violating 18 U.S.C. § 924(c)(1), the crime of which he was convicted. Id. at 895, 904.

Consistent with other circuits that have considered similar Bailey claims, the Fifth Circuit permitted the petitioner to raise his claim, despite having failed to satisfy the statutory gatekeeping requirements for bringing a second or successive § 2255 petition. See id. at 900-06. Following other circuits, the court enunciated a two-part test for assessing whether claims of actual innocence could proceed under § 2241:

Id. at 904; see also id. (observing that "most circuits have included an actual innocence component in their savings clause tests") (collecting cases). Concluding that the petitioner satisfied this circuit foreclosure test, the Fifth Circuit permitted him to proceed under § 2241. See id. at 904-06; see also Garland v. Roy, 615 F.3d 391, 396-402 (5th Cir.2010) (applying the Reyes-Requena test and holding that a Santos claim qualified under the § 2255(e) savings clause and could therefore be brought pursuant to § 2241).

Significantly, in the present case, not even the government asks us to reject the circuit foreclosure test, which I discuss more fully below. The government asserts that Reyes-Requena's two-part circuit foreclosure test governs and that Mr. Prost does not meet its requirements. See, e.g., Aple. Br. at 17, 19 (citing Reyes-Requena, 243 F.3d at 903-04 and Apodaca, 90 Fed.Appx. at 304 n. 10). Mr. Prost in his Reply Brief asks us to expand the test to permit not only those claims that were "foreclosed by circuit law," but also those which lacked affirmative Supreme Court precedent in support. See, e.g., Aplt. Reply Br. at 12-13. Neither party asks us to reconsider or reject the circuit foreclosure tests used in other circuits as a baseline for assessing whether claims of actual innocence are the exception that may proceed under § 2241 when they are otherwise procedurally barred from proceeding under § 2255.

Nor do we need to address that issue in this case. Mr. Prost clearly was not foreclosed by circuit precedent from raising his claim of actual innocence at the time of his initial petition. While Mr. Prost tries to suggest that United States v. Simmons, 154 F.3d 765 (8th Cir.1998), "erroneously foreclosed" any argument on the meaning of "proceeds" under the federal money laundering statute at the time of his initial § 2255 motion, a careful study of that decision reveals it does no such thing. In Simmons, the Eighth Circuit interpreted the Racketeer Influenced and Corrupt Organizations Act ("RICO"), 18 U.S.C. § 1962, to determine whether the term "proceeds" in that statute's forfeiture provisions meant the gross receipts obtained from, or the net profits of, illegal activity. The court's decision on this point relied on (1) the legislative history of RICO, which indicated that Congress intended the word "proceeds" to be read more broadly than "profits," and (2) Congress's explicit directive that RICO be liberally construed. Id. at 770-71. Simmons thus did nothing to address the money laundering statute, let alone foreclose an interpretation of it that equated proceeds with profits. If anything, one would think that Simmons, decided in 1998, more than four months before Mr. Prost pleaded guilty, would have alerted a defendant in his shoes to the possible argument that "proceeds" could conversely mean "profits," not "gross receipts," for purposes of the federal money laundering statute. Moreover, while Mr. Prost's § 2255 petition was pending, the Seventh Circuit adopted the "profits" definition of "proceeds," see United States v. Scialabba, 282 F.3d 475, 478 (7th Cir. 2002), yet Mr. Prost never sought to amend his motion to add a claim based on that statutory argument. And it was not until 2005, after Mr. Prost's first § 2255 motion was finally adjudicated, that the Eighth Circuit held that a federal money laundering conviction may stand even if the defendant laundered only gross receipts, not profits. See United States v. Huber, 404 F.3d 1047, 1058 (8th Cir.2005).2

The fact that Mr. Prost's claim fails any circuit's foreclosure test is sufficient grounding to dismiss the petition. I would go no further.

C.

The majority is not so restrained. Even as it extols the virtues of judicial restraint as a rationale for avoiding "constitutional issues," see maj. op. at 593-94, it simultaneously disregards this principle throughout its opinion. The majority reaches into uncharted territory to reject any circuit foreclosure test, see id. at 589-94, reaching a conclusion contrary to every other circuit that has decided this question, see infra Part II.A, without acknowledging that it is doing so. Worse, it does so on an issue which was neither adequately presented by the parties nor necessary to our disposition of this case. I do not join in the majority's frolic.

II.

The majority indisputably stands alone in its view of § 2255(e)'s savings clause. The majority cites no case, and I have found none, in which a court has posited that the § 2255 remedy is "adequate and effective" even when circuit precedent squarely foreclosed a petitioner from succeeding on a claim of actual innocence based on a new rule of substantive law announced by the Supreme Court after the petitioner had appealed and collaterally attacked his conviction on other grounds. Instead, courts of appeals that have considered this issue overwhelmingly have reasoned to the contrary. The majority's rejection of those decisions, see, e.g., maj. op. at 589-94, is unpersuasive on the merits, and I dissent from its analysis.

A.

The scope of § 2255's savings clause has been considered by other circuits in the analogous context of Bailey v. United States, 516 U.S. 137, 116 S.Ct. 501, 133 L.Ed.2d 472 (1995). In those cases, every court of appeals to reach the issue has held—albeit using varying language—that the § 2255 remedy is "inadequate or ineffective" when, at the time of the initial appeal and § 2255 petition, circuit precedent foreclosed success in overturning a conviction for conduct the Supreme Court later held is not criminal (i.e., the petitioner has a colorable claim of actual factual innocence). See Triestman v. United States, 124 F.3d 361, 363 (2d Cir.1997) (permitting resort to § 2241 where there was pre-Bailey circuit precedent precluding relief such that petitioner "could not have effectively raised his claim of innocence at an earlier time");3 In re Dorsainvil, 119 F.3d 245, 251 (3d Cir.1997) (permitting resort to § 2241 where petitioner "had no earlier opportunity to challenge his conviction for a crime that an intervening change in substantive law may negate"); In re Jones, 226 F.3d 328, 332-34 (4th Cir.2000) (permitting resort to § 2241 where it would have been "futile" to bring Bailey claim earlier in light of "settled law of [the] circuit or the Supreme Court"); Reyes-Requena v. United States, 243 F.3d 893, 903-06 (5th Cir.2001) (permitting resort to § 2241 where innocence claim was "foreclosed by circuit law" due to adverse precedent); Martin v. Perez, 319 F.3d 799, 804-05 (6th Cir.2003) (permitting petitioner to proceed under § 2241, in non-Bailey context, where post-conviction Supreme Court decision arguably made petitioner "factually innocent" of federal crime); In re Davenport, 147 F.3d 605, 610-11 (7th Cir.1998) (permitting resort to § 2241 where petitioner "had no reasonable opportunity" to raise claim earlier because "[t]he law of the circuit was so firmly against him");4 Wofford v. Scott, 177 F.3d 1236, 1244 (11th Cir.1999) (holding, in non-Bailey context, that § 2241 remedy is available where "circuit law squarely foreclosed" petitioner from bringing claim earlier).5

Some courts of appeals, like this one, have refrained from specifying the circumstances in which the § 2241 remedy remains available to bring such a claim of actual innocence. See supra Part I.B. (discussing Tenth Circuit cases); see also United States v. Barrett, 178 F.3d 34, 52 (1st Cir.1999) (leaving for another day the task of articulating when § 2241 remedy remains available); Abdullah v. Hedrick, 392 F.3d 957, 958-59, 963 (8th Cir.2004) (rejecting § 2241 petition where petitioner "did have an unobstructed procedural opportunity" to cite Bailey in prior petition, without considering whether such remedy would be available on different facts) (emphasis added); In re Smith, 285 F.3d 6, 6-9 (D.C.Cir.2002) (directing petitioner to file § 2241 habeas petition in the Seventh Circuit, where he was incarcerated, because § 2255 was "inadequate" in that circuit to resolve his invalid conviction claim, without expressly defining a savings clause test for the District of Columbia). But no circuit has interpreted the § 2255(e) savings clause to preclude a claim of actual innocence based on a new rule of substantive law subsequently announced by the Supreme Court.

The majority obfuscates the fact that, regardless of the phraseology employed by the other circuit courts, the results are the same. Every court of appeals to reach the issue has held that § 2255(e)'s savings clause does permit habeas review of actual innocence claims brought in § 2241 petitions where a prisoner was foreclosed by circuit precedent from succeeding on his claim in an initial § 2255 petition, and I would do so as well. Any implication by the majority that it is not creating a circuit split by interpreting § 2255(e) contrary to all the other circuits, see maj. op. at 594-96, is simply incorrect.

B.

The majority posits that the "circuit foreclosure" approach taken by other circuits is "untethered from the plain language the savings clause," maj. op. at 591-92, and "disregard[s]" other provisions of AEDPA specifying situations in which a prisoner may bring a collateral attack without resorting to § 2255(e)'s savings clause, see, e.g., id. at 591-92. Neither criticism is persuasive.

The majority's interpretation of the "plain language" of § 2255(e)'s savings clause stands alone. Our sister circuits recognize what the majority here does not: the fundamental purpose of habeas corpus and collateral review—even post-AEDPA—is to afford a prisoner a "reasonable opportunity to obtain a reliable judicial determination of the fundamental legality of his conviction and sentence," In re Davenport, 147 F.3d at 609 (emphasis added), and the text of the savings clause must be interpreted in view of this purpose, id. Against that measure, the § 2255 remedy necessarily is "inadequate and ineffective" when a petitioner is foreclosed by circuit precedent from succeeding on a claim that he has been convicted and imprisoned for a non-existent offense. See, e.g., id. at 609-11; Triestman, 124 F.3d at 373-77; In re Dorsainvil, 119 F.3d at 248-52; Reyes-Requena, 243 F.3d at 903-04 (adopting reasoning of earlier cases); In re Jones, 226 F.3d at 333-34 (similar). The defect in the § 2255 remedy is not that it fails to guarantee any particular "outcome or relief," maj. op. at 585, or that it fails to afford "multiple opportunities to test a conviction or sentence," id. at 585. The defect is that, as a result of erroneous circuit precedent directly on point, the § 2255 remedy failed to provide a prisoner with even one meaningful opportunity to raise his claim of actual innocence.6 The notion that an actually innocent prisoner can adequately and effectively "test" the legality of his conviction when he has no legal basis in his circuit for doing so cannot be squared with this central purpose of habeas review or the plain language of the savings clause.

Despite the majority's claims to the contrary, the fact that AEDPA does not expressly permit claims of actual innocence in a second or successive § 2255 motion, see maj. op. 585-86, 591-92, does not mean that such claims should be precluded from § 2241 habeas review via § 2255(e)'s savings clause. Long before AEDPA, the Supreme Court expressly recognized that a change in law which establishes a prisoner has been convicted and sentenced "for an act that the law does not make criminal. . . inherently results in a complete miscarriage of justice and present(s) exceptional circumstances that justify collateral relief under § 2255." Davis, 417 U.S. at 346-47, 94 S.Ct. 2298 (alteration in original) (internal quotation marks omitted). If "it is a `complete miscarriage of justice' to punish a defendant for an act that the law does not make criminal, thereby warranting resort to . . . § 2255, it must follow that it is the same `complete miscarriage of justice' when the AEDPA amendment to § 2255 makes that collateral remedy unavailable." In re Dorsainvil, 119 F.3d at 251; cf. Holland v. Florida, ___ U.S. ___, 130 S.Ct. 2549, 2560-62, 177 L.Ed.2d 130 (2010) (recognizing equitable exception to AEDPA's statutory bar on filing petitions outside one-year limitation period); Bousley, 523 U.S. at 620-21, 118 S.Ct. 1604 (relying on logic of Davis in refusing to preclude habeas review of Bailey claim). That Congress failed to amend § 2255 expressly to permit such claims of actual innocence does not obviate the potential for a miscarriage of justice if such claims are precluded from habeas review.7

If anything, the Court's rationale in Davis for permitting collateral review is even greater now than it was before AEDPA. As other circuits have recognized, interpreting AEDPA to bar judicial review of actual innocence claims raises serious constitutional concerns—under the Eighth Amendment and the Due Process Clause, if not also the Suspension Clause. See, e.g., Triestman v. United States, 124 F.3d 361, 377-80 (2d Cir.1997) (discussing the "serious constitutional questions" that would arise if an actual innocence claim was barred from habeas review); In re Dorsainvil, 119 F.3d 245, 248 (3d Cir.1997) (similar); Wofford v. Scott, 177 F.3d 1236, 1243-44 (11th Cir.1999) (discussing with approval other circuits' analyses of possible constitutional problems with a narrow reading of the savings clause); see also In re Davis, ___ U.S. ___, 130 S.Ct. 1, 1, 174 L.Ed.2d 614 (2009) (Supreme Court transferred original habeas petition to district court for evidentiary hearing on actual innocence claim); id. at 1 (Stevens, J., concurring) (pointing to discussion of "`serious' constitutional concerns that would arise if AEDPA were interpreted to bar judicial review of certain actual innocence claims") (quoting Triestman, 124 F.3d at 377-80); Herrera v. Collins, 506 U.S. 390, 432 n. 2, 113 S.Ct. 853, 122 L.Ed.2d 203 (1993) (Blackmun, J., dissenting) ("It may also violate the Eighth Amendment to imprison someone who is actually innocent."); Robinson v. California, 370 U.S. 660, 667, 82 S.Ct. 1417, 8 L.Ed.2d 758 (1962) ("Even one day in prison would be a cruel and unusual punishment for the `crime' of having a common cold.").

Such constitutional concerns are not triggered in the present case if, as explained above, the case is decided on the ground that Mr. Prost had a fair opportunity to bring his Santos-based claim at the time of his initial § 2255 petition. The majority does not do that, however. Instead, it interprets § 2255(e) as inapplicable to Mr. Prost's actual innocence claim in a manner not raised by the parties, and refuses to decide the constitutional issues squarely raised by its conclusion that there is no relief for an innocent person. The parties here agreed at a minimum that some form of circuit foreclosure test applied to this case, and hence did not address the constitutional concerns other than in cursory fashion. In fairness to the parties, and to avoid an improvident decision, we should ask them for supplemental briefing on the constitutional issues before interpreting § 2255 to completely preclude a claim of actual innocence.

For the reasons set forth above, I dissent from the reasoning of the majority decision.