concurring in part and dissenting in part:
In Atkins v. Virginia, 536 U.S. 304, 321, 122 S.Ct. 2242,153 L.Ed.2d 335 (2002), the Supreme Court held that the imposition of the death penalty on individuals with mental retardation violates the Eighth Amendment. While the Supreme Court “le[ft] to the States the task of developing appropriate ways to enforce the constitutional restriction upon their execution of sentences,” id. at 317, 122 S.Ct. 2242 (quoting Ford v. Wainwright, 477 U.S. 399, 405, 416-17, 106 S.Ct. 2595, 91 L.Ed.2d 335 (1986)) (internal quotations omitted), it surely set some floor for determining who is mentally retarded.
The Supreme Court found that “[b]e-cause of their disabilities in areas of reasoning, judgment, and control of their impulses, ... [mentally retarded persons] do not act with the level of moral culpability that characterizes the most serious adult criminal conduct. Moreover, their impairments can jeopardize the reliability and fairness of capital proceedings against mentally retarded defendants.” Id. at 306-07, 122 S.Ct. 2242. Thus, there is a substantive component of the Eighth Amendment that must not be eviscerated by judicial interpretation of state definitions of mental retardation. Yet, today’s result allows the Commonwealth to defeat any mental retardation claim, no matter how compelling in its entirety, by cherry picking evidence relevant to a single component of Virginia’s definition. This does satisfy Atkins’ mandate.
While I agree that both prongs of the Virginia definition must be satisfied for a defendant to prove mental retardation, I depart from the majority’s assertion that each component of the definition may always be read in isolation. The Virginia statute itself states that the defendant must “prov[e] that he is mentally retarded by a preponderance of the evidence” and not that the defendant must prove each component by a preponderance of the evidence. Va.Code Ann. § 19.2-264.3:1.1(0 [hereinafter Section 3:1.1]. Furthermore, though we are guided by state law, this Court has an obligation to fulfill the federal Supreme Court directive from Atkins that persons with mental retardation shall not be executed. The spirit of Atkins cannot be upheld by an interpretation of a mental retardation definition that ignores vast quantities of relevant evidence as well as blatant constitutional violations that prevent a petitioner from presenting his claim. For example, it must not be that a court tasked with following Atkins may find it immaterial that the defendant was diagnosed as mentally retarded by scoring more than two standard deviations below the mean on an I.Q. test, regardless of how low the score, simply because a defendant can have intimate relations with a woman or obtain a driver’s license.
The majority cites to our decision in Green v. Johnson, 515 F.3d 290 (4th Cir.2008) to support its conclusion that it is unnecessary to reach Walker’s intellectual functioning arguments. In Green, we found that under Virginia law, “Green must prove both prongs of Virginia’s statutory definition for mental retardation in order to establish that he is mentally retarded.” Id. at 301 (citation omitted). We then affirmed the district court’s dismissal of petitioner’s Atkins claim based on evidence relevant to the adaptive functioning *332prong of Virginia’s definition of mental retardation. Id. at 301-03.
I believe Green comes perilously close to the substantive floor set in Atkins by seemingly analyzing the two prongs of the Virginia definition independently.1 In Green, however, we affirmed the dismissal of the mental retardation claim only after finding that “the district court’s (and magistrate judge’s) consideration of the evidence [relevant to petitioner’s adaptive functioning] was thorough.” Id. (“The magistrate judge set forth in detail the expert testimony presented by both sides at the evidentiary hearing and explained its findings and conclusions with respect to [petitioner’s] adaptive skills.”). Thus, although close, the constitutional floor set in Atkins was protected by a thorough and complete review of petitioner’s evidence of mental retardation. The same cannot be said of the case before us now.
In this capital case, a case that has generated an appellate record of over 2,300 pages, the district court’s opinion spanned fewer than ten pages. In particular, the district court’s “consideration of the evidence” on the adaptive functioning prong in this case amounted to at most three sentences. J.A. 2370-71. Based on this record, we cannot have the degree of confidence required to affirm the district court’s dismissal of Walker’s Atkins claim by looking exclusively to the adaptive functioning prong.
Furthermore, this Court in Green was not faced with serious errors by the district court that affected the petitioner’s ability to present his mental retardation claim in the first instance: a blatant due process violation affecting Walker’s ability to present evidence of significantly subaverage intellectual functioning and a failure by the district court to consider Walker’s evidence of generally accepted professional practices in interpreting the results of intellectual functioning tests. These errors cannot be harmless, as the majority finds, when the floor set by Atkins is not secure.
Nor did we face in Green a case in which the district court specifically failed to adhere to a previous order from this Court. In Walker v. True II, 399 F.3d 315 (4th Cir.2005), we described the nature of the evidentiary hearing that Walker should have been afforded on remand: “Walker is entitled under law both to an evidentiary hearing in which he is afforded an opportunity to fully develop the factual basis of his mental retardation claim and to consideration by the courts of all of the evidence that is relevant to that claim under Virginia’s statutory framework.” Id. at 327 (internal citations omitted) (emphasis added). If our decision in Walker v. True II is to have any significance, we must remand this case so that Walker may receive the consideration to which he is entitled.
Because I find that the district court committed clear error by not properly considering Walker’s evidence of significant limitations in adaptive behavior, I also consider Walker’s claims on the intellectual functioning prong. On that prong, I find both that the district court violated Walker’s procedural due process rights and erred in failing to consider Walker’s evidence of generally accepted professional practices in interpreting the results of intellectual functioning tests. Based on these errors, this Court should remand for the district court to consider all evidence relevant to Walker’s Atkins claim. Moreover, I do not read our precedent to mean that any evidence on one prong, no matter *333how slight, may always be cited by the district court to preclude our review of a petitioner’s full Atkins claim under Virginia law. Atkins requires a reasonable appellate review to uphold the constitutional proscription of the use of capital punishment on offenders who are mentally retarded. Thus, I respectfully dissent from section II of this opinion.
I.
A.
On this record, I would find that the district court committed clear error by not properly considering Walker’s evidence of significant limitations in adaptive behavior. Based on this error alone, I believe this Court has an obligation to consider Walker’s claim that his due process rights were violated and his claim that the district court failed to follow this Court’s previous decision and consider Walker’s evidence of generally accepted professional practices in interpreting intellectual functioning tests’ results.
The majority finds that Walker’s Atkins claim fails because Walker possesses some adaptive skills. However, the district court made no finding that Walker was unequivocally functioning at such a level that he could not be classified as a person with mental retardation. The district court simply parroted several findings of the Commonwealth’s expert that should have been considered along with all other evidence proffered on the issue of adaptive functioning. This three-sentence reasoning by the district court should not withstand any level of scrutiny.
While the district court acknowledged that Walker “suffers from below average mental intelligence, struggles to perform some basic activities, exhibits anti-social behavior, and obtains financial support from others,” the court dismissed Walker’s claim by citing evidence that he “has committed various crimes requiring the ability to relate to others, associated with women on a personal and intimate level, engaged in homemaking activities, seduced under-aged girls, used others to help him avoid authorities, independently invoked his Miranda rights, used his brother’s identity to obtain a driver’s license, and obtained goods for himself while in prison.” J.A. 2370-71.
Although the clear error standard under which we review the district court’s findings is quite deferential, this Court may reverse the district court where its determination was “made without properly taking into account substantial evidence to the contrary.” Miller v. Mercy Hosp., Inc., 720 F.2d 356, 361 (4th Cir.1983). Thus, the question here is whether the district court properly took into account Walker’s expert testimony regarding his limitations in adaptive behavior.
In its decision, the district court did not specifically discuss any of Walker’s evidence; instead, it merely summarized the testimony of the Warden’s expert, Dr. Hagan. On its face, this seems like a classic “battle of the experts,” and thus it would be extremely difficult to find that the district court erred in believing one expert’s testimony over another.2 However, the district court did not purport to lend more credence to Dr. Hagan’s testimony than to Walker’s experts. Rather, the court apparently found that Dr. Hagan’s testimony somehow negated Walker’s showing of significant limitations in adaptive behavior. But as Walker’s experts testified, the ac*334tivities cited by the district court are entirely consistent with mental retardation.3 This adheres to the logic that the American Association on Mental Retardation (“AAMR”) says one must follow when applying the definition of mental retardation:
“Within an individual, limitations often coexist with strengths.” This means that people with mental retardation are complex human beings who likely have certain gifts as well as limitations. Like all people, they often do some things better than other things. Individuals may have capabilities and strengths that are independent of their mental retardation.
AAMR, User’s Guide: Mental Retardation: Definition, Classification, and Systems of Support 8 (10th ed. 2002) (2007). In particular, individuals with mental retardation can have intimate relationships, perform homemaking activities, craft simple lies, and invoke4 their Miranda rights. With regard to the crime cited by Dr. Hagan which purportedly demonstrated that Walker had the “ability to relate to others,” Walker’s expert testified that his execution of the scheme, including giving the victim his phone number, did not demonstrate a high level of cognitive or conceptual ability.
Because the district court did not properly consider Walker’s evidence of significant limitations in adaptive behavior, the district court committed clear error. This basis is sufficient for this Court to proceed to Walker’s arguments on the intellectual functioning prong. In addition, I believe that Atkins requires more in this case than simply relying on the Commonwealth’s evidence of Walker’s adaptive functioning, to the exclusion of all else, in order to dismiss Walker’s claim of mental retardation.
B.
Because the majority relies only on the Commonwealth’s evidence of Walker’s adaptive functioning, it fails to recognize a blatant due process violation by the district court. This constitutional violation resulted in the exclusion of two empirical tests that satisfied the intellectual functioning component of Virginia’s definition of mental retardation, thus preventing Walker from fully developing the basis for his claim, in contradiction of our previous decision in Walker v. True II, 399 F.3d at 327.
In order to establish “significantly sub-average intellectual functioning,” Virginia requires the “administration of at least one standardized measure generally accepted by the field of psychological testing.” Section 3:1.1(B)(1). The statute further directs the Commissioner of Mental Health, Mental Retardation and Substance Abuse Services (the “Department”) to “maintain an exclusive list of standardized measures of intellectual functioning generally accepted by the field of psychological testing.” Id.
On January 4, 2005, the Department published the exclusive list of standardized measures as required by Section 3:1.1 (the “January list”). It is undisputed that the January list was in effect at the time that this Court reversed the district court’s dismissal of Walker’s petition. Included in the January list were three standardized tests administered to Walker: the Comprehensive Test of Nonverbal Intelligence (“CTONI”), the General Ability Measure for Adults (“GAMA”), and the Wechslerseries tests.
*335Walker’s scores on two tests, the GAMA and CTONI, were indisputably two standard deviations below the mean, and both of those tests were included in the January list. Walker relied on the January list in developing the record to support his Atkins claim, and Walker was required to submit all expert disclosures by September 7, 2005. The Department only publishes its list on its website, and according to Walker, the January list remained posted until after the commencement of Walker’s evidentiary hearing, which began on November 1, 2005. On the first day of the hearing, Dr. James Morris, the Director of the Office of Forensic Services for the Department and the person with “final responsibility for what’s included or not included on the list,” J.A. 1598, testified that the GAMA qualified as an approved measure on Virginia’s published list. Dr. Morris also testified that the public could rely on the list that was published on the Department’s website, which contained both the GAMA and the CTONI.
Then, in January 2006, after the close of the evidence, the Warden filed a post-hearing brief. This brief referenced a new list of I.Q. tests approved by the Department for use in capital proceedings. That new list (“the October list”) bore a date of October 31, 2005, which was the day prior to Walker’s hearing. According to Walker, neither the Warden nor Dr. Morris disclosed the existence of the new list to the district court or himself before this post-hearing brief. The October list did not include either the GAMA or the CTO-NI as approved measures.
Following the evidentiary hearing, the district court found that Walker could not prove that he scored two standard deviations below the mean on an approved measure. In reaching this conclusion, the district court relied on the October list and excluded Walker’s scores on the GAMA, and CTONI.
Under the Due Process Clause of the Fifth Amendment to the United States Constitution, no person shall be “deprived of life, liberty, or property, without due process of law.” At a minimum, procedural due process requires both fair notice and an opportunity to be heard. Mathews v. Eldridge, 424 U.S. 319, 333, 96 S.Ct. 893, 47 L.Ed.2d 18 (1976); Mullane v. Cent. Hanover Bank & Trust Co., 339 U.S. 306, 314-15, 70 S.Ct. 652, 94 L.Ed. 865 (1950). In order to determine whether an individual has received fair notice, this Court “must examine the relevant facts of each case.” United States v. Hoechst Celanese Corp., 128 F.3d 216, 224 (4th Cir.1997). While “[generally, ignorance of the law or a mistake of the law is no defense,” courts will determine “fair notice” by examining whether particular persons would reasonably know the consequences of their conduct. See id. (internal citations and quotations omitted); see also Goodman v. United States, 151 F.3d 1335, 1337 (11th Cir.1998) (per curiam).
This concept of fair procedure is of “special importance ... in the capital sentencing context.” Lankford v. Idaho, 500 U.S. 110, 125, 111 S.Ct. 1723, 114 L.Ed.2d 173 (1991). This Circuit has impliedly recognized this concept as well. In United States v. Barnette, 211 F.3d 803, 824 (4th Cir.2000), this Court vacated a death sentence where the defendant was unable to present surrebuttal evidence to counter rebuttal evidence of an aggravating factor because such an action violated the concept of “simple fairness.”
In the present case, the district court’s application of the October list did not give Walker “fair notice” of the requirements for proving his Atkins claim, and thus violated the Due Process Clause. First, the October list was not discovered by the *336Warden until after the evidentiary hearing, and Walker had relied on the January list in developing and presenting evidence of mental retardation. Second, there was no evidence in the record regarding whether the October list was in fact published on the Department website prior to the hearing. Without this evidence, it is unknown whether Walker would have learned of the existence of the October list had he examined the website on that date. Indeed, Walker’s own counsel avers that it checked the website on October 31st, and the October list was not posted. Third, and perhaps most importantly here, Dr. Morris, who is the person with “final responsibility for what’s included or not included on the list,” J.A. 1598, testified during the hearing that the GAMA qualified as an approved measure on Virginia’s published list, J.A. 1603-04. It would be patently unfair to require Walker to conform his evidence to a list when a representative of the Department represented that a different test was on that list. Given this, the district court should have at least afforded Walker the opportunity to develop a claim of mental retardation based upon the approved measures in the October list.
The district court excluded two of Walker’s I.Q. scores that indisputably were two standard deviations below the mean. In applying the October list instead of the January list without affording Walker the opportunity to develop evidence of mental retardation in accordance with the October list, the district court thus violated Walker’s procedural due process rights.
C.
The district court also determined that Walker failed to show that he was mentally retarded before age eighteen. In particular, the district court found that
[e]ach WISC-R [“Weschler Intelligence Scale for Children-Revised”] administered to Petitioner before the age of eighteen supports a finding that Petitioner, while suffering from below average mental intelligence, is not mentally retarded. The evidence presented by Petitioner of his below average mental intelligence and limitations in conceptual, practical, and social skills do not constitute mental retardation or outweigh the results of the three WISC-R administrations in 1982,1984, and 1987.
J.A. 2368. Apparently, the district court did not consider the impact of the Flynn effect and Standard Error of Measurement (“SEM”) on the 1984 administration of the WISC-R.5 See id.
I find that the district court erred in determining that the 1984 WISC-R full scale score of seventy-six could not satisfy the intellectual functioning prong of the Virginia mental retardation definition. For its part, the district court did not express any opinion regarding whether the *337Flynn effect could be taken to account,6 but rather found that this Court’s decisions in Walton v. Johnson, 440 F.3d 160 (4th Cir.2006), and Hedrick v. True, 443 F.3d 342 (4th Cir.2006), foreclosed consideration of the SEM.
The district court rejected Walker’s contention that his 1984 WISC-R full scale score of seventy-six could support his claim of mental retardation because it interpreted this Court’s decisions in Walton and Hedrick as standing for the proposition that courts should “refus[e] to use the standard error of measurement to lower IQ scores in Atkins cases due to the inherent speculation of using the standard error of measurement to lower an IQ score when it could just as likely be used to raise an IQ score.” J.A. 2370. However, the district court misconstrued the import of these two decisions. In Walton, this Court found that the petitioner’s claim regarding the SEM should be dismissed because “Walton does not explain what this ‘standard error of measurement’ is or why it should reduce his particular score to 70 or less. Walton can only speculate that this standard measure error ... actually lowered his given score of 77 enough to meet Virginia’s mental retardation standard.” 440 F.3d at 178. Similarly, in Hedrick, the Court declined to consider the SEM because the petitioner did not provide expert evidence regarding the application of the SEM, and thus “only speculation on [the Court’s] part would lower Hedrick’s IQ score of 76.” 443 F.3d at 368. Fairly read, Walton and Hedrick stand for the proposition that bare allegations, without more, are insufficient to state a claim for Atkins relief, and the decisions do not imply that the SEM can never be considered on account of “inherent speculation.”
Nor can the district court’s ruling be sustained on the ground that Walker provided no evidence on the accepted professional practice regarding the SEM. The definitions of mental retardation promulgated by the AAMR and the American Psychiatric Association state that the SEM should be considered in diagnosing mental retardation. J.A. 879, 888-89. Moreover, Walker’s experts testified that the SEM creates a band of approximately five points around a particular I.Q. score, which means that an individual who scores within five points of two standard deviations below the mean should be diagnosed as having mental retardation.
Had the district court taken into account the Flynn effect and SEM, Walker’s score of seventy-six on the 1984 WISC-R may have been two standard deviations below the mean. Given this, the district court erred in not considering this score for purposes of subaverage intellectual functioning or for onset before the age of eighteen. In particular, even if a score such as this is not conclusive proof of subaverage intellectual functioning, it is certainly consistent with onset before the age of eighteen, and thus it should have been considered in that regard. Thus, the district court erred in interpreting our precedent as foreclosing consideration of the SEM in assessing scores on I.Q. tests and in not considering the Flynn effect. Like the due process violation, these errors thwarted Walker’s efforts to develop the basis for his mental retardation claim, in contradiction of our previous decision in Walker v. True II, 399 F.3d at 327.
II.
The Supreme Court has unambiguously held that a person with mental retardation *338can not be executed. I fear that the substance of this constitutional prohibition is put in jeopardy by affirming a district court’s dismissal of an Atkins claim based on scant evidence exclusively relevant to the adaptive component of a state’s mental retardation definition where, as here, the empirical tests forming the basis of the intellectual functioning component were excluded. Accordingly, I would find that in this case, the district court erred by finding that Walker had not shown significant limitations in adaptive behavior without meaningfully considering Walker’s expert testimony in that regard, applying the October list without giving Walker the opportunity to develop and present evidence in accordance with the new list, and refusing to consider Walker’s evidence of generally accepted professional practices in interpreting the results of intellectual functioning tests, in particular his evidence regarding the SEM. On this basis, I must dissent.
. Green at least analyzes both prongs. This Court in Green, in contrast to the majority here, undertook an analysis of the intellectual functioning prong as an alternative basis for its holding. 515 F.3d at 300.
. Indeed, this is the majority’s interpretation of the district court’s opinion. Supra at 328. However, even the majority seems to assume that the court credited Dr. Hagan’s testimony over Walker's experts’ testimony.
. Thus, the district court was not necessarily "presented with conflicting evidence,” as the majority states. Supra at 328.
. In fact, the record indicates that Walker did not affirmatively invoke his Miranda rights but rather merely remained silent.
. Dr. Daniel Reschly testified that the Flynn effect and SEM should be considered in interpreting I.Q. scores, and thus Walker’s score of seventy-six on the 1984 administration of the WISC-R would be within the range of mental retardation.
I.Q. tests are almost universally scaled so that the mean for the population is 100, with a standard deviation of fifteen. For such tests, an individual must score seventy or lower for the score to be at least two standard deviations below the mean. However, several phenomena have been identified that may be considered for purposes of adjusting or interpreting I.Q. scores.
The Flynn effect is the theory that I.Q. tests rise by .3 points per year after they are normed, and thus one must consider when an individual is administered an I.Q. test compared to when that test was last normed.
The SEM is an interpretative tool that reflects the idea that I.Q. scores have a ninety-five percent confidence interval, and thus an individual’s true score is somewhere within a five-point range of the reported score.
. The failure to consider the Flynn effect itself was error because this Court in Walker v. True II, 399 F.3d at 323, held that "on remand the district court should consider the persuasiveness of Walker's Flynn Effect evidence.”