concurring in part and dissenting in part.
I join in the majority’s conclusions regarding the sufficiency of evidence for Richards’s conviction of obstruction of justice, the rejection of Richards’s coercion claim, the applicability of 18 U.S.C. § 1512(c) to Richards’s conduct, the district court’s loss calculation, and the defendants’ acceptance of responsibility. I disagree, however, with the majority’s conclusion that the defendants’ sentencing on securities and wire fraud charges on the basis of the “one book” of the Sentencing Guidelines in effect in 2005, long after those violations had been completed, does not violate the constitutional prohibition against ex post facto laws. The defendants’ commission of subsequent obstruction of justice offenses, though related to the underlying securities and wire fraud charges and committed at a time when the 2005 Guidelines would apply, does not, in my view, render those Guidelines applicable to the securities and wire fraud charges because at the time the defendants committed the securities and wire fraud offenses, they did not have “fair notice” of the severity of the penalties to which they might be subjected for them under the later, harsher Guidelines. To the extent that the majority conclude otherwise, I respectfully dissent.
Ex Post Facto Clause
The Ex Post Facto clause of Article I, Section 9, reads: “No Bill of Attainder or ex post facto Law shall be passed.”1 U.S. Const. Art. I, § 9, cl. 3. Some eleven years after it was adopted as part of the original United States Constitution, Justice Samuel Chase, a President Washington appointee, observed that this “prohibition ... necessarily requires some explanation; for, naked and without explanation, it is unintelligible, and means nothing.” Calder v. Bull, 3 U.S. 386, 390, 3 Dall. 386, 1 L.Ed. 648 (1798) (Chase, J.). He then established as black letter Constitutional law that, inter alia, “[ejvery [criminal] law that changes the punishment, and inflicts a greater punishment, than the law annexed to the crime, when committed,” violates the Ex Post Facto clause. Id.2 (emphasis
*639omitted).3
Nearly two hundred years later, Chief Justice Rehnquist, writing for the Court, explained:
Early opinions of the Court portrayed [the enumeration by Justice Chase in Calder ] as an exclusive definition of ex post facto laws [citing three Nineteenth Century Supreme Court decisions]. So well accepted were these principles that the Court in Beazell v. Ohio, 269 U.S. 167 [46 S.Ct. 68, 70 L.Ed. 216] (1925), was able to confidently summarize the meaning of the Clause as follows:
“It is settled, by decisions of this Court so well known that their citation may be dispensed with, that any statute which punishes as a crime an act previously committed, which was innocent when done; which makes more burdensome the punishment for a crime, after its commission, or which deprives one charged with crime of any defense available according to law at the time when the act was committed, is prohibited as ex post facto.” Id., at 169-170 [46 S.Ct. 68].
See also Dobbert v. Florida, 432 U.S. 282, 292 [97 S.Ct. 2290, 53 L.Ed.2d 344] (1977).
The Beazell formulation is faithful to our best knowledge of the original understanding of the Ex Post Facto Clause: Legislatures may not retroactively alter the definition of crimes or increase the punishment for criminal acts.
Collins v. Youngblood, 497 U.S. 37, 42-43, 110 S.Ct. 2715, 111 L.Ed.2d 30 (1990) (holding change in State law allowing reformation of improper criminal verdicts 35 not to violate Ex Post Facto clause) (footnotes omitted).
As we observed in somewhat different circumstances, “the ex post facto doctrine is concerned not just with notice, but with the inherent injustice associated with retroactivity itself.” Sash v. Zenk, 439 F.3d 61, 64 (2d Cir.2006) (denial of panel rehearing); see also id. (contrasting the ex post facto doctrine and the rule of lenity, which is “more narrowly focused” on the sole issue of notice).4 The ex post facto *640doctrine therefore requires not only “notice,” but notice that is “fair.” See, e.g., Weaver v. Graham, 450 U.S. 24, 30, 101 S.Ct. 960, 67 L.Ed.2d 17 (1981).
The “One-Book Rule”
Section 1B1.11 of the United States Sentencing Guidelines provides in pertinent part:
Use of Guidelines Manual in Effect on Date of Sentencing (Policy Statement)
(a) The court shall use the Guidelines Manual in effect on the date that the defendant is sentenced.
(b)(1) If the court determines that use of the Guidelines Manual in effect on the date that the defendant is sentenced would violate the ex post facto clause of the United States Constitution, the court shall use the Guidelines Manual in effect on the date that the offense of conviction was committed.
(2) The Guidelines Manual in effect on a particular date shall be applied in its entirety....
(3) If the defendant is convicted of two offenses, the first committed before, and the second after, a revised edition of the Guidelines Manual became effective, the revised edition of the Guidelines Manual is to be applied to both offenses.
The notion that only one set of Guidelines should be applied in imposing a single sentence, even where that sentence covers multiple crimes the commission of which straddles the effective dates of two sets of Guidelines, appears to derive from the principle that each set of Guidelines is meant to act as a “cohesive whole.” United States v. Bailey, 123 F.3d 1381, 1404 (11th Cir.1997). “ ‘A sentencing court has no authority to pick and choose, taking one provision from an earlier version of the guidelines and another from a later version.... [The one-book rule] avoids twisting the guidelines, depriving them of uniformity and consistency.’ ” Id. (quoting United States v. Keller, 58 F.3d 884, 890 (2d Cir.1995) (brackets and ellipsis in Bailey )).
Consistent with that principle, Application Note 2 to Section IB 1.11 provides, in relevant part, that
the approach set forth in subsection (b)(3) should be followed regardless of whether the offenses of conviction are the type in which the conduct is grouped under § 3D1.2(d). The ex post facto clause does not distinguish between groupable and nongroupable offenses, and unless that clause would be violated, Congress’ directive to apply the sentencing guidelines in effect at the time of sentencing must be followed.
U.S.S.G. § 1B1.11 App. Note 2.
Relevant Facts
The facts relevant to the ex post facto issue before us are relatively simple and straightforward.
The defendants engaged in securities and wire fraud, and conspired to commit such fraud. The last overt act of the conspiracy was committed in May of 2000, and the fraud itself—the use of a so-called “35-day accounting month”—ended no later than October 2000. At the time the defendants committed these crimes, the November 1, 1998, edition of the Guidelines was in effect.
Between 2001 and 2003, the Guidelines were revised so as to increase the punishment for crimes of the sort that the defendants had previously committed. The revisions resulted in an increase in the offense levels applicable to those crimes *641based upon the amount of money lost by the victims of the crime, and in an expansion of the loss table through the addition of new categories for losses of $200 and $400 million, with the latter calling for a 30-level enhancement compared to the 20-level enhancement that would have applied to a $400 million loss under the 1998 version of the Guidelines. The revised Guidelines also created a new 6-level enhancement for fraud involving 250 or more victims, and a 4-level enhancement for violations of securities laws by defendants who were officers or directors of public companies. See U.S.S.G. § 2b1.1(b)(2) & (b)(13).
Under the Guidelines in effect at the time of the defendants’ commission of the fraud and conspiracy crimes, the applicable offense level was 30, which translated into a Guidelines range of 97 to 121 months’ imprisonment. As a result of the subsequent revisions, the offense level was raised to 50, resulting in' a recommended sentence of life imprisonment. See Maj. Op. at 625.
Beginning in September 2002 and continuing until April 2004, the defendants engaged in various acts designed to cover up their previously committed conspiracy and fraud. In a 2005 superseding indictment—the original indictment had been handed down in 2004—the defendants were charged with committing fraud and conspiracy to commit fraud over a period of time ending in 2000.5 But the superseding indictment also charged Richards with committing perjury thereafter, Kumar with making false statements to a Special Agent of the FBI thereafter, and both defendants with filing false SEC filings, obstructing justice and conspiring to obstruct justice thereafter. Both defendants pled guilty to all of these charges in April 2006, and were sentenced on all of them in November of that year. No relevant changes in the Guidelines were made between the time of the commission of the coverup crimes and the filing of the superseding indictment, the plea, and the sentencing.
Relying on the one-book rule, the district court sentenced the defendants on all of the charges.to which they pled guilty under the 2005 Guidelines. The court decided that no ex post facto issue was presented by the fact that this version of the Guidelines was not in effect at the time the defendants committed the fraud and conspiracy crimes; it therefore did not apply the exception set forth in subsection (b)(1) that instructs a court not to apply the Guidelines manual in effect on the date of sentencing if so doing would violate the Ex Post Facto clause. Using the 2005 Guidelines, the district court sentenced each defendant to what would appear to be (depending on the actual longevity of each defendant) well below his Guidelines-indicated life sentence: Kumar to 144 months’ incarceration, followed by three years of supervised release; Richards to 84 months’ incarceration, followed by three years of supervised release.
The Majority Opinion
There is an inherent tension between the one-book rule and the Ex Post Facto *642clause—between a rule prescribing the application of Guidelines that are current as of the day of sentencing with respect to crimes that were committed before those Guidelines went into effect, and a constitutional requirement that “[l]egislatures[6] not retroactively increase ... the punishment for criminal acts.” Collins, 497 U.S. at 43, 110 S.Ct. 2715. Indeed, that tension is explicitly acknowledged in subsection (b)(1) of the one-book rule, quoted above. Judge Easterbrook once remarked on the “gymnastics performed” by the Seventh Circuit in the case there on appeal in resolving that tension in order “to show that a[n] ... increase in [the defendant’s] offense level [was] compatible with the ex post facto clause.... ” United States v. Vivit, 214 F.3d 908, 924 (7th Cir.2000) (Easterbrook, J., concurring); see also id. (opining that such gymnastics were “unnecessary, because the sentencing guidelines are not ‘laws’ within the scope of [the Ex Post Facto ] clause”). That might be said of similar attempts by federal courts nationwide since the Guidelines regime began. I fear, though, that in the case at bar, the majority have turned from gymnastics to contortions.
The majority and I begin on common ground. We first assume that the ex post facto doctrine applies to the Sentencing Guidelines after the Supreme Court decided, in United States v. Booker, 543 U.S. 220, 244-268, 125 S.Ct. 738, 160 L.Ed.2d 621 (2005), that the Guidelines are advisory.7 We then agree that “[fjor a law to contravene the Ex Post Facto clause, ‘two critical elements must be present: First, the law must be retrospective, that is, it must apply to events occurring before its enactment; and second, it must disadvantage the offender affected by it.’ ” Maj. Op. at 624-25 (quoting Miller v. Florida, 482 U.S. 423, 430, 107 S.Ct. 2446, 96 L.Ed.2d 351 (1987)). We also agree that the second element is present here: “[T]here is no question that application of the 2005 Guidelines disadvantaged the defendants by subjecting them to the higher ranges of the 2005 Guidelines compared to the 1998 version of the Guidelines in effect when the 35-day month practice was discontinued in October 2000.” Maj. Op. at 625 (footnote omitted). The use of the 2005 Guidelines raised the recommended Guidelines range from 97 to 121 months’ incarceration to life imprisonment.
Where I depart from the majority’s view is on the question of retroactivity. The majority recognize that
“central to the ex post facto prohibition is a concern for ‘the lack of fair notice and governmental restraint when the legislature increases punishment beyond what was prescribed when the crime was consummated.’ ” Miller, 482 U.S. at 430, 107 S.Ct. 2446 (quoting Weaver v. Graham, 450 U.S. 24, 30, 101 S.Ct. 960, 67 L.Ed.2d 17 (1981)). The existence of *643an ex post facto violation [thus] turns on whether an individual was deprived of fair notice....
Id. at 628. They find sufficient notice to the defendants by concluding that
the adoption of the one-book rule prior to the commission of the defendants’ obstruction offense had placed them on notice of the consequences of committing that second offense. That the consequences of the second offense included the application of the post-amendment Guidelines to all offenses considered at the defendants sentencing was fully apparent prior to the commission of the crimes that triggered those consequences.
Id. (emphasis added).
But it seems to me that the notice that the defendants received here was notice as to punishment for the wrong crime: not as to the fraud and conspiracy crimes for which punishment was revised markedly upward, but the subsequent obstruction offenses for which the Guidelines have not changed. This notice was inconsequential because the defendants were not subjected to an increased sentence for obstruction; they were subjected to an increased sentence for already completed frauds. And I think that the majority give insufficient attention to the quality of notice that ex post facto jurisprudence requires. It is not notice simpliciter, but notice that is “fair.” As the Supreme Court said in Miller, “The constitutional prohibition against ex post facto laws cannot be avoided merely by adding to a law notice that it might be changed;” in that case the Court found an ex post facto violation because the defendant was not aware of the prescribed range of punishment for his offense at the time he committed it. Miller 482 U.S. at 431, 107 S.Ct. 2446. “Ex post facto is as much a doctrine of retroactivity as it is a doctrine of notice,” Sash, 439 F.3d at 64-65, and “the inherent injustice associated with retroactivity itself’ must guide our analysis, id. at 64.
What we said in the related context of the change of rules regarding supervised release bears repeating here:
We are unpersuaded by the government’s argument that the Ex Post Facto Clause is not implicated so long as the penalty for a supervised-release violation is enhanced before the defendant engages in his supervised-release-violative conduct because the violator then has notice and fair warning that that conduct will result in the enhanced penalty. While it is true that a defendant would have notice of that enhancement before he committed his violation of supervised release, it is equally true that he would have had no such notice before he committed the original offense. Indeed, the enhanced punishment of supervised-release revocation simply would not be applicable to him had he not committed his original offense and been sentenced, for that offense, to supervised release. Thus, the government’s notice argument is not helpful to resolving the issue presented by this appeal.
United States v. Meeks, 25 F.3d 1117, 1122 (2d Cir.1994), abrogated on other grounds by Johnson v. United States, 529 U.S. 694, 120 S.Ct. 1795, 146 L.Ed.2d 727 (2000).8 *644While it may be true that, in the case at hand, the one-book rule provided Richards and Kumar constructive notice of the enhancement in sentencing before they obstructed justice, “it is equally true that [they] would have had no such notice before [they] committed the original offensefs]” of fraud and conspiracy. Id.; see also, United States v. Zagari, 111 F.3d 307, 323-24 (2d Cir.1997) (finding, where the government sought to apply a revised version of the Guidelines to a fraud committed before the revision based upon charged but later dismissed conduct that the government alleged to amount to a continuation of the fraud after the revision, that use of the revised Guidelines violated the Ex Post Facto clause where the alleged frauds before and after the revision were “not properly subsumed into one huge scheme to defraud* ... but rather should be thought of as separate schemes.... ”).
Six of our sister circuits that have addressed this issue have limited their conclusion of constitutionality either to cases that involve a continuing course of conduct or to cases in which the offenses are sufficiently similar to be subject to “grouping” under the Guidelines. See United States v. Duane, 533 F.3d 441, 449 (6th Cir.2008) (concluding the “better argument in favor of § 1B1.11(b)(3)’s constitutionality” to be “that a Guidelines revision, § 1B1.11(b)(3) itself, and the § 3D 1.2(d) grouping rules provide a criminal fair warning that committing future similar crimes may subject him to increased penalties for similar prior offenses”); United States v. Sullivan, 255 F.3d 1256, 1262-63 (10th Cir.2001) (“[T]he grouping rules and the relevant conduct provisions gave [the defendant] notice that his three consecutive failures to file would be considered part of the same course of conduct and would collectively determine his sentence.”); Vivit, 214 F.3d at 919 (deciding that the Guidelines, including in particular the grouping rules, succeed in “providing] notice to criminals that engaging in ongoing fraudulent behavior involving the same type of harm risks grouping of convictions, which because of the one-book rule, will all be sentenced according to the Guidelines in effect when the latest conduct occurred.”); United States v. Kimler, 167 F.3d 889, 895 (5th Cir.1999) (“[The defendant] had proper notice that, if he continued to commit related offenses that would be grouped under § 3D1.2(d), he would be sentenced under the guidelines in use when he committed the last offense in the grouped series.”); Bailey, 123 F.3d at 1404-05 (“[T]he one book rule, together with the Guidelines grouping rules and relevant conduct, provide that related offenses committed in a series will be sentenced together under the Sentenc*645ing Guidelines Manual in effect at the end of the series. Thus, a defendant knows, when he continues to commit related crimes, that he risks sentencing for all of his offenses under the latest, amended Sentencing Guidelines Manual.”); United States v. Regan, 989 F.2d 44, 48 (1st Cir.1993) (deciding, before the establishment of the one-book rule, that it was not an ex post facto violation to use revised Guidelines where offenses committed prior to revision had been grouped with offenses committed after the revision and “were manifestly part of the same ongoing scheme of embezzlements”).
Two other circuits have found the one-book rule to be constitutional even in cases not involving continuing courses of conduct or grouped offenses. These circuits only reached a conclusion of constitutionality, however, in cases that involved repeated commission of the same offense before and after the revision of the Guidelines. See United States v. Lewis, 235 F.3d 215, 218 (4th Cir.2000) (finding the application of the one-book rule to be constitutional in a case involving multiple acts of tax evasion before and after a revision of the Guidelines); United States v. Cooper, 35 F.3d 1248, 1251 (8th Cir.1995) (concluding that the Ex Post Facto clause was not violated where “a series of firearms offenses” straddled a revision of the Guidelines, and comparing series of offenses to a “conspiracy that straddles the Sentencing Guidelines’ effective date”).
The two remaining circuits that have addressed this issue9 have unequivocally concluded, to the contrary, that the one-book rule is unconstitutional in these circumstances, even where the convictions that straddle a revision of the Guidelines are grouped for sentencing purposes. See United States v. Ortland, 109 F.3d 539 (9th Cir.1997); United States v. Bertoli, 40 F.3d 1384 (3d Cir.1994).
So, it seems, we now stand alone: The panel concludes that any offense that has been committed by a defendant after a revision of the Guidelines may be used as a basis to apply the revised Guidelines to crimes committed before the revision so long as the pre- and post-revision crimes are prosecuted together but irrespective of the relationship, if any, between them.10 *646The majority take this position in a case in which the two sets of crimes, the substantive crimes and the coverup offenses, are in fact related—indeed, they have been grouped together for sentencing purposes—and no such broad holding is required for resolution of this appeal.11 As discussed in more detail below, I do not think the relationship of these offenses to be sufficiently close to conceive of them as groupable continuing offenses and therefore to overcome the ex post facto problem, but such a holding would at least acknowledge that there is a notice problem inherent in allowing a Guideline revised after a crime was committed to be used to sentence for that crime.
The majority seek to find in the one-book rule a form of constructive notice to defendants as to the consequences of their crimes because they knew at the time they committed any successive crime that the sentence for all prior completed crimes indicted together with the successor crime would be increased as a consequence of that successive crime. But the Ex Post Facto clause requires not only notice, but also that the notice be “fair.”12 In the case of Messrs. Kumar and Richards, it was not.
I agree with the sentiments of Judge Kelly, of the Tenth Circuit, dissenting under somewhat similar circumstances: “[T]he only notice ... providefd to the defendants] at the time of commission of the ... pre-amendment offenses is that the sentence could be determined in accordance with guideline provisions that may or may not be amended. Even if the notice is sufficient to inform a defendant that the last offense could determine the sentence, only a defendant with the prescience of a clairvoyant could anticipate an actual sentence based upon a yet-to-be amended guideline.” United States v. Sullivan, 255 F.3d 1256, 1266 (10th Cir.2001) (Kelly, J., dissenting).
Grouping
The majority and I agree that grouping is not determinative of whether the sentences here comport with the Due Process Clause. We reach this conclusion for dif*647ferent reasons, however. The issue of grouping of offenses affects the ex post facto inquiry, it seems to me, in at least two ways—each relevant to the case at bar, and only one of which is acknowledged in the majority opinion.
First, there is considerable support for the argument that when offenses are grouped for sentencing purposes because “the behavior is ongoing or continuous in nature and the offense guideline is written to cover such behavior,” U.S.S.G. § 3D1.2(d), ex post facto concerns are met. According to the Eighth Circuit, for example, “it has been held that applying the Sentencing Guidelines to a conspiracy that straddles the Sentencing Guidelines’ effective date is not violative of the ex post facto clause. [Courts have] noted that with conspiracy and other• continuing offenses it is the completion date of the offense that controls the version of the Sentencing Guidelines to be applied.” Cooper, 35 F.3d at 1251 (emphasis added; citations omitted). This argument would support a finding of constitutionality despite the fact that the penalty prescribed for the crime is made harsher during the course of the criminal behavior because the defendants at least theoretically made a choice to continue their unlawful activity while the penalties for it were being increased.13
This is precisely the argument the government makes here, contending that the obstruction offenses are part of a “continuing course of conduct” with the fraud and conspiracy offenses. Appellee’s Br. at 49. While I do not think this argument persuasive in this particular case,14 I recognize *648that the argument at least attempts to address the ex post facto problems inherent in applying a revised version of the Guidelines to the sentence of an act completed before the revision. Had this been the basis of the majority’s, decision today, our disagreement on the law would seem to me to be a narrow one.
But the majority appear to view the impact of grouping in a second fashion instead, one that has indeed been endorsed by several of our sister circuits. It was expressed explicitly by the Seventh Circuit in Vivit: “[T]he adoption of the one-book rule and the grouping rules put[s] criminals on notice that ‘the version of the sentencing guidelines in effect at the time he committed the last series of grouped offences will apply to the entire group.’ ” Vivit, 214 F.3d at 918 (quoting Kimler, 167 F.3d at 895). Before they violated the securities fraud and conspiracy laws, the defendants “knew” the Guidelines sentence at the time, but they also “knew” that it could be increased if they later committed an offense—in this case obstruction of justice—that would be grouped with it.
The majority base their unique holding on the one-book rule rather than the grouping theory described here, but the two analyses bear similarities: The one-book rale as applied here is constitutional because the Guidelines provide notice that the law is subject to change. Even if they do not provide notice of what that change will be at the time an act is committed, they provide such notice before a subsequent act. Whether this analysis is conducted under the grouping rules or the one-book rale seems largely beside the point to me. In either event it permits notice that is insufficient under the observation in Miller that “[t]he constitutional prohibition against ex post facto laws cannot be avoided merely by adding to a law notice that it might be changed.” Miller, 482 U.S. at 431, 107 S.Ct. 2446.
This sort of compound, abstract notice hardly seems to me to be “fair” notice at the time the fraud and conspiracy crimes were committed of what punishment the defendants might receive if they misbehaved in that manner—giving notice only that if they did what they did, and later committed another (potentially “groupable”) crime, their sentence could be incx-eased by some unknown amount. I am reminded again of Judge Kelly’s reference to the “defendant with the prescience of a clairvoyant [who alone] could anticipate an actual sentence based upon a yet-to-be amended guideline.” Sullivan, 255 F.3d at 1266 (Kelly, J., dissenting). It is, in short, speculative and incomplete notice and therefore not “fair.”
Recidivism Cases
The majority reason by analogy to decisions, including those of the Supreme Coux’t, upholding recidivism statutes— those that punish crimes committed by a person with a specified level of criminal record more harshly than those committed by a person without such a record. “The one-book rule,” the majority say, “when it leads to a higher sentencing range than would be applied to a single offense, operates in a manner similar to that of the recidivist statutes and ‘three strikes’ laws upheld by the Supreme Court and our sister circuits in the past.” Maj. Op. at 629. They cite Gryger v. Burke, 334 U.S. 728, 68 S.Ct. 1256, 92 L.Ed. 1683 (1948), as having “x-ejected the defendant’s argument that the consideration of his past offenses in determining his sentence for a later offense was foreclosed by the Ex Post Facto clause. Id. at 732, 68 S.Ct. 1256.” Maj. Op. at 629.
The majority then assert that “[t]he fact that the impetus for enacting the recidivist statutes was to reflect the greater culpability associated with the latter offenses, whereas the impetus for the enactment of *649the one-book rule is to avoid ‘piecemeal’ sentencing is of no relevance for the purposes of determining the retroactivity of the legal consequences of the defendant’s actions.” Maj. Op. at 629 (citation omitted). The majority therefore conclude that “the distinction between the recidivist statutes and the one-book rule makes neither a practical nor a logical difference for purposes of an analysis under the Ex Post Facto clause,” because “the defendants’ obstruction offense is the ‘actual crime’ triggering the application of the one-book rule, the defendants had prior notice of the consequences of that crime, and therefore the application of the one-book rule is proper.” Maj. Op. at 630.
I think, to the contrary, that there is a crucial difference between the legislative branch deciding that a particular crime is more serious when committed by—and that the public is in need of more protection from-—a person who has a specified level of past criminal behavior than someone who does not, and increasing a penalty for a completed crime “triggered” by the commission of a subsequent one—indeed, irrespective, in the majority’s view, of whether there is a connection between those crimes committed before the change and those committed afterward. Justice Jackson, writing for the Court in Gryger, put it thus:
Nor do we think the fact that one of the convictions that entered into the calculations by which petitioner became a fourth offender occurred before the Act was passed, makes the Act invalidly retroactive or subjects the petitioner to double jeopardy. The sentence as a fourth offender or habitual criminal is not to be viewed as either a new jeopardy or additional penalty for the earlier crimes. It is a stiffened penalty for the latest crime, which is considered to be an aggravated offense because a repetitive one.
Gryger, 334 U.S. at 732, 68 S.Ct. 1256 (emphasis added).
The later crime may, as the majority say, “trigger” the change in the sentence for the earlier crimes, but what it triggers is what the Gryger court said was improper: an “additional penalty for the[ir] earlier crimes.” Whatever the trigger, the revisions increased the offense levels applicable to the earlier fraud and conspiracy crimes, not the later obstruction of justice offenses. The revisions added an increase based upon the amount of money lost by the victims of the fraud and conspiracy crimes; they expanded the loss table for the fraud and conspiracy offenses, not the obstruction of justice offenses, by adding new categories for losses of $200 and $400 million; and they created a new 6-level enhancement for fraud, not obstruction of justice, involving 250 or more victims; and added a 4-level enhancement for violations of securities laws, not the obstruction of justice laws, by defendants who were officers or directors of public companies. See U.S.S.G. § 2b1.1(b)(2) & (b)(13); cf. Cooper, 35 F.3d 1248, 1251 (8th Cir.1994) (“It is well settled that habitual offender statutes do not offend the ex post facto clause, even though such statutes impact imposition of an instant sentence in consideration of past criminal conduct. [Citing Gryger, 334 U.S. at 732, 68 S.Ct. 1256.] The enhanced sentence is considered to impose a stiffer penalty for the latest crime, which is considered to be an aggravated offense due to its repetitive nature.” (emphasis added)). This is, of course, the critical difference between recidivist statutes and the circumstances we are faced with here: that while, as the majority point out and as Gryger makes clear, “the consideration of [] past offenses in determining [the] sentence for a later offense” may be permissible for ex post facto purposes, Maj. Op. at 629, it is *650the consideration of a later offense in determining the sentence for past offenses, as we do here, that runs afoul of ex post facto considerations.
As a “practical” matter, to be sure, Congress, or perhaps the Sentencing Commission, might have—and may still—adopt a permissible recidivism statute to cover a circumstance very much like the present one: A person who commits an obstruction of justice in order to cover up a fraud in which the losses inflicted by the fraud are $ X will receive a Y level increase in offense level. It does not follow, “logically” or otherwise, that reaching a similar result by the present method—by retroactively increasing the punishment for fraud—is constitutional. The hypothetical law would provide punishment for behavior of which the potential violator would be fully and fairly warned before engaging in that behavior. It would reflect the perceived seriousness of future obstruction offenses, publically disseminated before any such offense is committed, and not an attempt to re-punish completed past acts. To increase the punishment for fraud and conspiracy after they are completed provides no “fair notice” and evokes “the inherent injustice associated with retroactivity itself.” Sash, 439 F.3d at 64.
CONCLUSION
We have been instructed for more than 200 years that “a law that changes the punishment, and inflicts a greater punishment, than the law annexed to the crime, when committed” violates the Ex Post Facto clause. Calder, 3 U.S. at 390. That seems to me to describe the change in the Guidelines that was applied to the defendants’ fraud and conspiracy crimes upon sentencing long after those crimes had ended.
Again, I have no reason to doubt that the Guidelines Commission may promulgate obstruction of justice guidelines that take into account the gravity, as it then views the gravity to be in the context of the crime for which punishment is being evaded, so as prospectively to prescribe harsher sanctions for the obstruction. But then it must prescribe those sanctions to the obstruction itself and not to crimes that are over. I think that to be a violation of the Ex Post Facto clause.
For the foregoing reasons and to the extent indicated, I respectfully dissent.
. Section 10, which applies to the States as opposed to Congress, contains a similar prohibition against ''pass[ing] any Bill of Attainder [or] ex post facto Law....” U.S. Const. Art. I, § 10, cl. 1.
. Justice Chase's complete enumeration of prohibited ex post facto laws reads:
1st. Every law that makes an action, done before the passing of the law, and which was innocent when done, criminal; and punishes such action. 2nd. Every law that aggravates a cnme, or makes it greater than it was, when committed. 3rd. Every law that changes the punishment, and inflicts a greater punishment, than the law annexed to the crime, when committed. 4th. Every law that alters the legal rules of evidence, and receives less, or different, testimony, than the law required at the time of the commission of the offence, in order to convict the offender.
Id. (emphasis in original). The continued viability of the fourth category, at least with the breadth suggested by Justice Chase, is in doubt, see Collins v. Youngblood, 497 U.S. 37, *63943 n. 3, 110 S.Ct. 2715, 111 L.Ed.2d 30 (1990) ("As cases subsequent to Calder make clear, this language was not intended to prohibit the application of new evidentiary rules in trials for crimes committed before the changes.”), but that aspect of ex post facto law is not implicated in this appeal.
. The Supreme Court has consistently and repeatedly held that a law that increases the punishment of a crime after the completion of the offending act violates the Ex Post Facto clause. See, e.g., Garner v. Jones, 529 U.S. 244, 249, 120 S.Ct. 1362, 146 L.Ed.2d 236 (2000) ("One function of the Ex Post Facto Clause is to bar enactments which, by retroactive operation, increase the punishment for a crime after its commission.”); Collins, 497 U.S. at 43, 110 S.Ct. 2715 ("Legislatures may not retroactively ... increase the punishment for criminal acts.”); Weaver v. Graham, 450 U.S. 24, 30, 101 S.Ct. 960, 67 L.Ed.2d 17 (1981) (finding that the Ex Post Facto clause "forbids the imposition of punishment more severe than the punishment assigned by law when the act to be punished occurred”); Calder, 3 U.S. at 397 (Paterson, J.) ("The enhancement of a crime, or penalty, seems to come within the same mischief as the creation of a crime or penalty; and therefore they may be classed together.”). These decisions recognize no distinction between the scope of protection under the Ex Post Facto clause for laws that increase a sentence and laws that criminalize conduct after the commission of the relevant act.
. We continued:
For this reason, the Supreme Court has associated the ex post facto doctrine with the Fifth Amendment's Takings Clause, which "prevents the Legislature ... from depriving private persons of vested property rights except for a 'public use’ and upon payment of 'just compensation,’ " and with the “prohibitions on 'Bills of Attainder' in Art. I, §§ 9-10, [which] prohibit legislatures from singling out disfavored persons and meting out summary punishment for past conduct.” Landgraf v. USI Film Prods., 511 U.S. 244, 266, 114 S.Ct. 1483, 128 L.Ed.2d *640229 (1994). Ex post facto is as much a doctrine of retroactivity as it is a doctrine of notice.
Id. at 64-65.
. The indictment purports to charge the frauds as occurring "[o]n or about and between April 1, 1998 and April 6, 2004, both dates being approximate and inclusive,” but alleges no overt act after May of 2000 and itself describes “The Scheme to Defraud” as occurring "[p]rior to and during CA's fiscal year 2000, which ended March 31, 2000.” On appeal, the government does not dispute that the frauds were completed before the revision of the Guidelines in 2001 and 2002 and the subsequent increases in the penalties for securities and wire fraud. See, e.g., Appellees’ Br. at 46 (“This case involves the application of a Guidelines manual to two sets of crimes, one occurring before, and one after, a revision of the Guidelines.”).
. We all assume for present purposes that the Guidelines were promulgated by a "legislature.” Cf. footnote 7, infra.
. We have not yet explicitly decided whether the now-advisory Guidelines can implicate the Ex Post Facto clause, although we implicitly assumed this to be the case in United States v. Kilkenny, 493 F.3d 122 (2d Cir.2007) where, post-Booker, we analyzed the application of revised Sentencing Guidelines in light of the Ex Post Facto. It is the subject of some disagreement among the various circuit courts. Compare, e.g., United Stales v. Demaree, 459 F.3d 791 (7th Cir.2006) (Clause not directly applicable after Booker) with, e.g., United States v. Turner, 548 F.3d 1094, 1100 (D.C.Cir.2008) ("It is enough ... [for purposes of the Ex Post Facto clause] that using the [post-crime] Guidelines created a substantial risk that [the defendant's] sentence was more severe” than it would have been under the Guidelines in effect at the time of the crime.); and cf. United States v. Thompson, 518 F.3d 832, 870 (10th Cir.2008) (assuming that the Guidelines may still implicate ex post facto concerns after Booker).
. Although Meeks was subsequently abrogated by Johnson, it was not because the Supreme Court thought our logic on the present score unpersuasive. To the contrary, there are repeated suggestions in the dicta in Johnson that had the issue been properly presented to the Court, it would have ruled the same way this Court did in Meeks. But the government in Johnson specifically "disavow[ed]’’ the argument that it had made to, and had been endorsed by, the district court in Meeks, and that is made by the majority here: that there is no ex post facto violation because “revocation of supervised release imposes punishment for defendants’ new of*644fenses for violating the conditions of their supervised release.” Johnson, 529 U.S. at 699-700, 120 S.Ct. 1795 (internal quotation marks omitted). The Johnson Court called such a disavowment ”wise[] in view of the serious constitutional question that would be raised by construing revocation and reimprisonment as punishment for the violation of the conditions of supervised release.” Id. at 700, 120 S.Ct. 1795. The Court then expressly noted that ‘‘[s]ince postrevocation penalties relate to the original offense, to sentence Johnson to a further term of supervised release under § 3583(h) would be to apply this section retroactively.” Id. at 701, 120 S.Ct. 1795. It would therefore appear that so long as the punishment for an act taken subsequently to a new law that increases the penalty for an offense relates to the punishment for an earlier offense, and not to the new act, the law is being applied retroactively and creates ex post facto concerns. See also id. at 699, 120 S.Ct. 1795 (recognizing that an ex post facto violation arises where a law “raises the penalty from whatever the law provided when [the defendant] acted ”) (emphasis added). These statements were dicta because the Court concluded that section 3583(h) was not retroactive, and was therefore inapplicable to cases where the original offenses was committed before passage of the law. Id. at 702-03, 120 S.Ct. 1795.
. The D.C. Circuit does not appear yet to have dealt with the situation with which we are faced here.
. The majority suggest that the Eighth Circuit has a similar rule based on that court's recent decision in United States v. Anderson, 570 F.3d 1025 (8th Cir.2009), which the majority understands to “base[] its holding entirely on the application of the one-book rule.” Maj. Op. at 627 n.13. I am not persuaded that Anderson stands for the proposition that the one-book rule alone is sufficient to overcome ex post facto problems inherent in sentencing a defendant under one version of the Guidelines for crimes committed both before and after that version came into effect. The majority properly note that “the crimes at issue [in Anderson ] were potentially subject to grouping under Guidelines section 2J1.6 and 3d1.2(c).” Id. But the crimes at issue in Anderson were not only "potentially" subject to grouping under the named sections; they were in fact grouped under those sections. See Appellee’s Br., United States v. Anderson, Nos. 08-3402, 08-3436, 2009 WL 2819251 (8th Cir. Feb.5, 2009) ("Per the grouping rules, the district court properly grouped the underlying wire fraud convictions with the failure-to-appear conviction and then applied the 'one-book rule' to the group.”). And while the language in that decision may be broad enough to suggest that the holding is based solely on the one-book rule, see Anderson, 570 F.3d at 1034 ("Pursuant to the one-book rule, because Anderson had been convicted of two offenses, the first—wire fraud—committed before, and the second— failure to appear—committed after, a revised edition of the guidelines manual became effective, the revised edition of the guidelines should be applied.”), the decision was based entirely on case law pertaining to grouped offenses, see id. at 1033 ("We have held that the one-book rule does not violate the Ex Post Facto Clause, even when it results in a higher *646guidelines range, because defendants have 'fair warning' that the revised guidelines manual will apply to grouped offenses.” (emphasis added)). Because the decision in Anderson was rendered in a case involving grouped crimes and relied entirely upon precedent that applied only to grouped offenses, I remain doubtful that such language can be read as an affirmative holding on the ex post facto implications of the one-book rule.
. It is worth noting that the government repeatedly makes clear that it only argues that there is no ex post facto problem in this case because the crimes straddling the revision of the Guidelines are related. See, e.g., Appellee’s Br. at 45 (''[Bjecause the securities fraud and the obstruction constitute 'relevant conduct' vis-a-vis one another, application of the later Guidelines did not violate the Constitution.” (citation omitted)); id. at 47 (“The grouping rules, together with the one-book rule, put the defendant on notice ....”); id. at 48 ("[T]his Court recognizes obstruction after-the-facl to be 'relevant conduct' to prior fraud.”); id. at 50 (“Because these counts were properly grouped pursuant to § 3D1.2, the application of the 2005 Guidelines pursuant to § 1B1.11 did not violate the Ex Post Facto Clause.”).
.
Blackstone illustrates the second purpose of the Ex Post Facto Clause, providing fair warning, by looking to the policies of the Roman despot Caligula. See 1 William Blackstone, Commentaries on the Laws of England 46 (1765). Caligula had laws written in fine print and hung them high up on pillars so that they were not available to nor readable by the Roman citizens affected by such laws. Id. They provided no fair warning and so, like laws made ex post facto, they would not have provided citizens fair notice to refrain from the criminalized conduct.
United States v. Kilkenny, 493 F.3d 122, 126 (2d Cir.2007).
. A similar argument, albeit a less convincing one, might be made to justify the use of a revised version of the Guidelines to sentence the repeated commission of the same crime committed both before and after the revision, as the Fourth Circuit did in Lewis and the Eighth Circuit did in Cooper. Such is not the case here. See Lewis, 235 F.3d at 218; Cooper, 35 F.3d at 1251.
. The argument that continuing offenses do not implicate ex post facto concerns might carry some day, but not this one. I do not think it proper to cast the abortive attempt to cover up the fraud and conspiracy several years later as a continuation of those crimes. They were complete when the defendants' fraudulent and conspiratorial activity ended. When the defendants engaged in fraud they likely hoped, even expected, to escape detection altogether. That they did not, and later engaged in obstruction of justice to try to avoid further exposure and thereby criminal prosecution, does not seem to me to be a part of the original criminal activity.
I think it significant in this respect that the obstruction and fraud/conspiracy offenses were grouped together under Section 3D 1.2(c), and not 3D 1.2(d). (The fraud and conspiracy offenses were in fact grouped with each other under Section 3D1.2(d), but these offenses were then grouped with the obstruction offenses under Section 3D 1.2(c)). These two sections govern different kinds of conduct and have distinct implications for ex post facto analysis. Grouping a continuing offense under Section 3D 1.2(d) may not cause ex post facto concern where crimes are grouped because "the behavior is ongoing or continuous in nature,” and therefore the crimes are arguably not completed until after the Guidelines are revised.
But there is no such implication of continuation under Section 3D 1.2(c), invoked here, which requires grouping "[w]hen one of the counts embodies conduct that is treated as a specific offense characteristic in, or other adjustment to, the guideline applicable to another of the counts.” U.S.S.G. § 3D1.2(c). That fraud and obstruction offenses share an offense characteristic or may lead to an adjustment to the sentence for one another has no bearing I can see on whether the fraud offenses can be understood as continuing through the time that the Guidelines were revised. In this case, as I have tried to explain, the fraud cannot be so understood. The fact that the two sets of crimes were grouped together under this section of the Guidelines has no implications with respect to the notice the defendants received of the consequences of their conduct committed before the Guidelines were revised.