Todd Carta appeals the district court’s ruling that he is a “sexually dangerous person” subject to civil commitment under the Adam Walsh Child Protection and *3Safety Act of 2006 (“the Act”).1 After review of the record and the district court’s factual - findings and legal conclusions, we affirm.
I.
Carta pled guilty to federal child pornography charges in 2002 and was sentenced to five years in prison and three years of supervised release. Prior to his scheduled release in 2007, the Bureau of Prisons certified that Carta was a “sexually dangerous person” and commenced proceedings pursuant to the Act, which authorizes civil commitment of a person in federal custody. 18 U.S.C. § 4248(a). A “sexually dangerous person” is one “who has engaged or attempted to engage in sexually violent conduct or child molestation and is sexually dangerous to others.” 18 U.S.C. § 4247(a)(5). A determination that an individual is a “sexually dangerous person” requires the government to prove by clear and convincing evidence that he “suffers from a serious mental illness, abnormality, or disorder as a result of which he would have serious difficulty in refraining from sexually violent conduct or child molestation if released.” Id. § 4247(a)(6).2
After a district judge ruled that the government had failed to establish that Carta’s diagnosis of “paraphilia not otherwise specified characterized by hebephilia” was a “serious mental illness, abnormality or disorder” within the meaning of the Act, United States v. Carta, 620 F.Supp.2d 210 (D.Mass.2009) (“Carta I”), a panel of this court reversed, holding that the district court erred in ruling that the government had failed to establish the serious mental illness element. United States v. Carta, 592 F.3d 34, 44 (1st Cir.2010) (“Cartall”). The case was.remanded for consideration of “whether the requisite dangerousness exists,” Id. A different district court judge subsequently conducted a seven-day trial, ultimately ruling in the government’s favor. See United States v. Carta, No. 07-12064-PBS, 2011 WL 2680734 (D.Mass. July 7, 2011) (“Carta III”). This timely appeal followed.
II.
We begin by observing that Carta, who was 42 years old when he pled guilty to the child pornography charges for which he was incarcerated, does not deny that the Act’s first element — that he has engaged in child molestation in the past — is satisfied. His lengthy history of committing sexual abuse, beginning when he was 11 years old and including many such acts over the course of three decades, is detailed in both district’ court orders. See Carta III, 2011 WL 2680734 at *3-7; Carta I, 620 F.Supp.2d at 212-14. We eschew repeating the details of his history here. In addition to his prior abusive conduct, Carta also displayed troubling behavior while in a sex offender treatment program in federal prison. After being transferred at his request to a prison in North Carolina that offered the program, Carta dropped out in part because of his inappro*4priate interest in the program’s younger members.. Carta II, 592 F.3d at 37.
Given Carta’s concession as to the first element, the remaining issues before us are whether he suffers from a serious mental illness, abnormality or disorder and, if so, whether such condition would result in his having serious difficulty refraining from sexually violent conduct or child molestation.
A.
At the heart of the first issue is the diagnosis proffered by the government expert, Dr. Amy Phenix, who testified that Carta was afflicted with a mental disorder known as “paraphilia not otherwise specified (“NOS”) characterized by hebephilia.” The Diagnostic and Statistical’ Manual of Mental Disorders (“DSM IV” or “DSM”) describes the “essential features” of paraphilia as follows:
[R]ecurrent, intense sexually arousing fantasies, sexual urges, or behaviors generally involving 1) nonhuman objects, 2) the suffering or humiliation of oneself or one’s partner, or 3) children or other nonconsenting persons, that occur over a period of at least 6 months ... [and that] cause clinically significant distress, or impairment in social, occupational, or other important areas of functioning.
Carta II, 592 F.3d at 38 (quoting Am. Psychiatric Ass’n, DSM 522-23 (4th ed. 2000)). The “not otherwise specified” portion of the diagnosis stems from the fact that hebephilia is not one of the specific conditions listed in the DSM TV, either separately or as an example of paraphilia. Id. While the precise contours of hebephilia are the subject of debate, it suffices to say that the disorder consists of a sexual attraction to adolescents, as opposed to, for example, a specified paraphilia such as pedophilia, a sexual attraction to pre-pubesqent children. Id. (citing DSM at 527-28).
The parties dispute whether our determination in Carta II that the government had established the mental disorder element is binding as law of the case. It is.
Under the law of the case doctrine, “when a court decides upon a rule of law, that' decision should continue to govern the same issues in subsequent stages of the same case.” United States v. Matthews, 643 F.3d 9, 12 (1st Cir.2011) (citing Arizona v. California, 460 U.S. 605, 618, 103 S.Ct. 1382, 75 L.Ed.2d 318 (1983)). Two branches of the doctrine are pertinent here. The “mandate rule” proscribes relitigation in the trial court of matters that were decided by an earlier appellate decision in the same case. Id. at 13. The other branch, related to the “law of the circuit” principle, binds a successor appellate panel in the same case to “honor fully the original decision.” Id. Here, both varieties of the rule apply.
The only issue before us in Carta II was whether the district court in Carta I had correctly ruled that the diagnosis of paraphilia NOS characterized by hebephilia was not a serious mental illness, abnormality or disorder. We determined that the statutory definition of “serious mental illness” is not limited to either the consensus of the medical community or to maladies identified in the DSM. Carta II, 592 F.3d at 39-40. Additionally, we noted that even if hebephilia lacks sufficient indicia to fit it within the statutory praxis, paraphilia itself is listed in the DSM, as is the category “paraphilia not otherwise specified.” Id. at 40. Thus, we discounted the testimony of defense expert Dr. Leonard Bard that hebephilia is not a generally accepted diagnosis and does not fit within the DSM definition of paraphilia. In the end, we concluded that “Dr. Phenix’s report, Carta’s past history’ of sexually abusing mi*5nors, his in-prison behavior and his expressed attitudes justify classifying him as suffering from paraphilia: he has a decades-long sexual fixation on minors that plainly has ‘caused significant distress or impairment’ in his life.” Id. at 40.3
On remand to determine whether Carta would have serious difficulty in refraining from sexually violent conduct or child molestation, the district court nevertheless acceded to Carta’s request to accept additional evidence on the mental condition issue. Carta III, 2011 WL 2680734 at *2. The court need not have done so.
In Carta II, we explicitly decided the serious mental illness question. Under the mandate rule, relitigation is ordinarily foreclosed. See United States v. Rivera-Martinez, 931 F.2d 148, 150 (1st Cir.1991) (“When a case is appealed and remanded, the decision of the appellate court establishes the law of the case and it must be followed by the trial court on remand.”) (emphasis in original). The trial court possesses “some limited discretion” to reopen an issue, but only in “very special situations.” United States v. Wallace, 573 F.3d 82, 88-89 (1st Cir.2009) (quoting United States v. Bell, 988 F.2d 247, 250-51 (1st Cir.1993)). We have described the “exceptional circumstances” that must be present to overcome the mandate rule as requiring a showing of a change in controlling legal authority, significant new evidence not previously obtainable, or the prospect of serious injustice. Id.
Similarly, the law of the circuit principle requires that we follow Carta II “unless and until the decision is modified or overturned by a higher court” or “if the initial ruling was made on an inadequate record or was designed to be preliminary or tentative, ... if newly discovered evidence bears on the question, or if reconsideration would avoid manifest injustice.” Wallace, 573 F.3d at 89 (quoting Ellis v. United States, 313 F.3d 636, 647-48 (1st Cir.2002)) (internal quotation marks omitted). “A finding of manifest injustice requires a definite and firm conviction that a prior ruling ... is unreasonable or obviously wrong [and] a finding of prejudice.” Id. (citation and internal quotation marks omitted).
None of the factors that would obviate application of either branch of the law of the case doctrine to the holding in Carta II is present here. There has been no intervening authority that would undermine the earlier panel decision, Carta II was not decided on an inadequate record, and there was nothing about the opinion that could be regarded as preliminary or tentative. Carta, however, argues that new evidence compels re-examination of the issue. Indeed, the district court allowed him to introduce an article written after Carta II, in which two drafters of the DSM IV express their doubts about the propriety of a forensic diagnosis of paraphilia NOS with a hebephilia descriptor. The court also appointed, at Carta’s request, Dr. Robert Prentky, who interviewed Carta for six hours and reviewed his records. Although Dr. Prentky testified that Carta did not suffer from a serious mental illness, he agreed with Dr. Phenix that paraphilia NOS with a descriptor of hebephilia can be an appropriate diagnosis. But this additional disagreement among experts does nothing to alter the scenario that we analyzed in Carta II, and therefore our determination that Carta fits -within the “serious mental illness” prong of the Act was (and remains) binding.
*6That said, however, nothing about our law of the case analysis should be read to suggest that we have cause to doubt the district court’s conclusion that the proof provided during the court’s supplemental inquiry satisfied the second element of the Act. The judge’s analysis of the serious mental condition issue, taking into account the additional evidence, does not come close to qualifying as clear error. See Fed.R.CivJP. 52(a)(6).
B.
Turning to the third element, we note that the Act does not define “serious difficulty in refraining” from sexually violent conduct or child molestation. Nevertheless, sufficient guidance comes from the Supreme Court’s pre-Act review of a Kansas civil commitment statute in Kansas v. Crane, 534 U.S. 407, 122 S.Ct. 867, 151 L.Ed.2d 856 (2002). In Crane, the Court held that a showing of “serious difficulty in controlling behavior” was a prerequisite to civil commitment of a sexually dangerous person. Id. at 413, 122 S.Ct. 867. Such lack of control need not be “demonstrable with mathematical precision,” but the defendant’s lack of control “must be sufficient to distinguish the dangerous sexual offender whose serious mental illness, abnormality, or disorder subjects him to civil commitment from the dangerous but typical recidivist convicted in an ordinary criminal ease.” Id.
In ruling in the government’s favor, the district court relied on transcripts from Carta I, testimony from Drs. Phenix, Prentky and Bard, and Carta’s own testimony and history. As for the experts, Dr. Phenix concluded that Carta would have serious difficulty refraining from sexually molesting children if released; Dr. Prentky and Dr. Bard disagreed.
Although Carta refused to allow her to interview him, Dr. Phenix based her opinion on her review of Carta’s records, as well as her use of three different risk-assessment actuarial tools to yield a re-offense rate. Carta III, 2011 WL 2680734 at *14-16, *23. The district court did not entirely accept Dr. Phenix’s actuarial conclusions, however, because her analysis included a subjective step that contributed to the finding of a higher re-offense rate, but which has not been empirically validated. Id. at *14 n. 1.
In addition to applying the actuarial tools, Dr. Phenix observed other risk factors present in Carta’s life, including the absence of social support, the presence of significant intimacy deficits and poor sexual and non-sexual self-regulation, and a demonstrated lack of cooperation with supervision. In sum, Dr. Phenix testified that Carta was the same person that he was when he entered treatment — still believing that sexual relationships with thirteen year-old boys were permissible.
Dr. Prentky, who did meet with Carta, used a diagnostic tool (known as the SVR-20) that excluded from consideration sex crimes that had not been legally adjudicated. Id. at *20. Dr. Prentky combined the score from this model with other factors, including, e.g., the lack of documented sexual offenses and significant sanctions, as well as Carta’s infraction-free presence in the community pending sentencing and the fact that he neither had re-offended after a criminal sanction nor had a sex-related disciplinary problem in prison. Id. Ultimately, although Dr. Prentky believed that Carta would have serious difficulty refraining from general anti-social behavior, he did not think that such difficulty would carry over into sexually violent conduct or child molestation. Id.
Dr. Bard relied on an actuarial test that assessed Carta’s likelihood of recidivision at seven to fifteen percent. He acknowl*7edged, however, that the tool did not account for Carta’s then-current situation, but instead combined factors from his past and compared them to others with similar data points. Id. at *18. To overcome this limitation, Dr. Bard created a “dynamic risk assessment” that combined the test with other factors. He placed considerable weight on the fact that Carta had not engaged in any sexual misconduct while in prison. He placed little significance, however, on Carta’s actions during his time in the federal prison treatment program. Essentially, Dr. Bard took the position that since there is nothing illegal or deviant about having a relationship with a twenty-five year-old, Carta’s actions did not point to a likelihood of re-offense. To the contrary, Dr. Bard suggested that these actions demonstrated that Carta was “mov[ing] his level of attraction from teenagers to twenty-somethings.” Id. Moreover, Dr. Bard accepted Carta’s explanation for having dropped out of the treatment program — it resulted from his stubbornness about admitting that he had made a mistake, rather than from inappropriate contact with other program members. Id.
Ultimately, Dr. Bard testified that Carta is “a very different person” and that his likelihood of re-offense would be tempered by the presence of a support system in the form of therapists, family members and a probation officer. Id. at *19. He testified that Carta’s attraction, impulsivity and anger “are not the same now as [they were] in the past” and that “[he] is able to control his behavior as evidenced by his nine plus years in prison without any serious violence.” Id.
The district court concluded that none of the experts’ testimony could be given full weight, noting that Dr. Bard believed that paraphilia NOS (hebephilia) is not a valid diagnosis under any circumstances, that Dr. Phenix’s methodology included a subjective step, and that Dr. Prentky’s analysis did not include Carta’s self-reported crimes. Id. at *23. Additionally, the court was less convinced than Dr. Prentky that Carta had undergone “considerable self-transformation” while in custody awaiting the hearing in this case.
The court also considered numerous other factors. In Carta’s favor were the lack of sexual misconduct either during the interval between his guilty plea and his incarceration or while he was in prison, the fact that he was seeking treatment during the pendency of the hearing, and his advancing age. Id. at *24. At the same time, however, the court observed that Carta’s history of sexual crimes and antisocial behavior was deeply troubling, and his impulsive actions and volatility when faced with adversity were highly concerning. Id. Additionally, the court found that Carta continued to engage in distorted thinking, exemplified by his attempt to rationalize his behavior by stating that both a thirteen year old boy and his daughter’s boyfriend had initiated sexual contact with him. M4 Finally, the court noted that Carta’s inability to refrain from engaging with younger participants while in the federal sex offender treatment program, and his departure from the program, demonstrated a disturbing degree of impulsiveness. Id. at *25.
Carta targets the weight that the district court assigned to the many strands of evidence in the mix. Two specific points *8stand out. First, Carta places great significance on the fact that he committed no sexual offenses or violations while incarcerated. The relevance of that behavior, however, is limited by the fact that the main targets of his attraction — adolescent boys — are inaccessible in prison. Next, he says that the district court placed insufficient weight on his lack of sexual misconduct while he was in the community pending sentencing, as well as on his lack of re-offense after being sanctioned. Compared to these relatively short time frames on which Carta focuses, however, he has a lengthy pre-sanction history of sexual misbehavior involving children. We have no difficulty accepting the premise that multiple instances of post-sanction recidivism can buttress a commitment decision. See, e.g. United States v. Shields, 649 F.3d 78, 89 (1st Cir.2011); United States v. Hunt, 643 F.Supp.2d 161, 181 (D.Mass.2009). But we do not read the Act as requiring previous post-sanction misconduct.
Aside from those two assertions, the essence of Carta’s argument is that more weight should have been given to the testimony of the experts who testified in his favor, with correspondingly less placed on Dr. Phenix’s testimony, his own past conduct, and his departure from the federal prison treatment program. After reviewing the district court’s factfinding for clear error, we have no warrant to upset that court’s assessment of the evidence. Caria II, 592 F.3d at 39.5 As the district court acknowledged, the “serious difficulty refraining” assessment presented the most vexing issue. Id. at *22. But it is abundantly clear that the court carefully weighed the evidence and the competing views espoused by each of the experts.
Our task is not to re-weigh the evidence or to make credibility assessments. See Adelson v. Hananel, 652 F.3d 75, 86 (1st Cir.2011) (“[Wjhen a case has been decided on the facts by a judge ... an appellate court must refrain from any temptation to retry the factual issues anew.”) (quoting Johnson v. Watts Regulator Co., 63 F.3d 1129, 1138 (1st Cir.1995)); Fed.R.Civ.P. 52(a)(6) (“reviewing court must give due regard to trial court’s opportunity to judge the witnesses’ credibility.”). The district court, in its role as factfinder, was free to “decide among reasonable interpretations of the evidence,” Shields, 649 F.3d at 89, and the factfinder’s choice between two permissible views of the evidence cannot be clearly erroneous. Adelson, 652 F.3d at 86 (quoting Johnson, 63 F.3d at 1138.).
As we have observed before, “there is no crystal ball that an examining expert or court might consult to predict conclusively whether a past offender will recidivate.” Shields, 649 F.3d at 87. Ultimately, we can find clear error only if “we are left with the definite and firm conviction that a mistake has been committed.” United States v. D'Andrea, 648 F.3d 1, 14 (1st Cir.2011) (citation and quotation marks omitted). No such error was committed here.
III.
The judgment of the district court is affirmed.