dissenting:
The Supreme Court has remanded this case to us “for further consideration in light of Teague v. Lane, 489 U.S. -, [109 S.Ct. 1060, 103 L.Ed.2d 334] (1989).” Yet the Cox plurality opinion1 effectively ignores Teague. I believe that it is our duty to follow the Supreme Court’s remand order and consider Teague, that it is improper for us to revisit issues that we previously resolved en banc, and that our prior en banc determination in this case was correct. Accordingly, I dissent.
Because Judge Johnson has written a dissent detailing the effect of Teague on Moore’s petition, I will not repeat what he has already said. I join in parts I, II-B, II-C-1, II-C-2, III-A, III-C, and IV of Judge Johnson’s dissent. I write separately because I disagree with his analysis of certain issues.
I.
OUR ROLE ON REMAND
Two years ago this court, sitting en banc, concluded that Moore’s failure to assert his Estelle and Proffitt claims in his first federal habeas petition was not an abuse of the writ. The en banc court also directed the district court to determine whether the ends of justice required the court to consider Moore’s Gardner claim. Now, Judge Cox, apparently believing that he is writing *1519on a clean slate, simply concludes that our prior decision was wrong. Yet the facts have not changed in the interim, nor has the relevant law.
The plurality clutches at the fact that the Supreme Court vacated our prior opinion and remanded the case for our reconsideration in light of Teague. The only way for the Court to have us consider the unique way Teague interacts with the abuse of the writ doctrine, however, was by vacating our prior opinion. We should draw no inference from a remand order, one way or the other, as to the Court’s view of the correctness of our prior en banc opinion.2 We should, however, interpret the remand order as meaning what it says: our task on remand is to reconsider our prior decision in light of Teague, that is, whether Teague affected our earlier opinion. We show no greater fealty to the Supreme Court than when we construe the Court to mean what it says, but the plurality has chosen a different path.
Of course, we have the power to revisit any issue determined by the prior en banc court. In my view, however, we should not exercise that power, particularly in this case, where doing so flouts the plain meaning of the Supreme Court’s remand order.
Courts have long recognized that principles of finality, fairness, and efficiency counsel against redetermining issues that have already been decided by the same court. Under the rubric “law of the case,” these principles give rise to the rule that once a court has decided an issue in a case, that issue remains settled unless or until it is reversed or modified by a higher court. Unaware of the irony, the Cox opinion ignores such principles of finality, while purporting to vindicate those same principles. See, e.g., ante, at 1504.
II.
APPLYING TEAGUE TO MOORE’S PETITION
A.
The Georgia Supreme Court affirmed Moore’s conviction in 1975. When the Supreme Court denied Moore’s petition for certiorari on October 4, 1976 his conviction became “final” for the purposes of our analysis under Teague.3 Cf. Griffith v. Kentucky, 479 U.S. 314, 107 S.Ct. 708, 712 n. 6, 93 L.Ed.2d 649 (1987) (“By ‘final,’ we mean a case in which judgment of conviction has been rendered, the availability of appeal exhausted, and the time for a petition for certiorari elapsed or a petition for certiorari finally denied.”).
Having ascertained the date Moore’s conviction became final, the next step is to determine whether a claim he is raising is a “new rule” claim, i.e., a claim in which he seeks the benefit of a rule announced after his conviction became final.4 If a claim is one seeking the benefit of a “new rule,” then under Teague the federal habeas court may not entertain the claim unless the rule fits into one of two exceptions.5 See Penry v. Lynaugh, — U.S. -, 109 S.Ct. 2934, 2944, 106 L.Ed.2d 256 (1989).
*1520I agree with Judge Johnson that Moore’s Proffitt and Gardner claims are “new” for retroactivity purposes. I also agree that under Battie v. Estelle, 655 F.2d 692 (5th Cir.1981), Moore’s Estelle v. Smith claim is not “new” for retroactivity purposes. Therefore, we may entertain the Proffitt and Gardner claims only if they fall into one of the two exceptions outlined in Teag-ue. On the other hand, Teague does not affect the Estelle v. Smith claim at all.
I disagree with Judge Johnson’s suggestion that because a claim is not “new law” for retroactivity purposes it may not be “new law” for purposes of the abuse of the writ, or vice versa.6 Thus, I do not perceive any tension between our decision in Battie holding that Estelle v. Smith is not new law for retroactivity purposes and our prior en banc ruling that the Estelle v. Smith claim was a new law claim for abuse of the purposes.
I will not rush in where the Supreme Court has hesitated to tread and try to define what is “new law” for either re-troactivity or abuse of the writ purposes.7 For this dissent it will suffice simply to note that the equitable principles underlying the abuse of the writ properly focus on the petitioner’s conduct, or that of his counsel if he was not travelling pro se. See generally Gunn v. Newsome, 881 F.2d at 957-96 (11th Cir.1989) (en banc). By contrast, the conduct of the habeas petitioner or his attorney is irrelevant to determining whether a rule of law is new for retroactivity purposes. The retroactivity analysis focuses solely on the relationship of the new rule to prior law:
In general, however, a case announces a new rule when it breaks new ground or imposes a new obligation on the States or the Federal Government. To put it differently, a case announces a new rule if the result was not dictated by precedent existing at the time the defendant’s conviction became final.
Teague, 109 S.Ct. at 1070 (citations omitted) (plurality opinion). We should not conflate the questions of whether a rule is new for nonretroactivity purposes as opposed to abuse of the writ purposes.
B.
Because both the Gardner and Proffitt claims are new for retroactivity purposes, the next step is to determine whether the claims fit into either of the two exceptions presented in Teague.8
The first exception for new rules that will be applied retroactively is where the new rule places “certain kinds of primary, private individual conduct beyond the power of the criminal law-making authority to proscribe.” Teague, 109 S.Ct. at 1075 (citation omitted) (plurality opinion). This exception is plainly not applicable to either the Proffitt or Gardner claim.
The second exception to Teague’s general rule that a new rule will not be applied retroactively on collateral review encompasses fundamental rules “without which the likelihood of an accurate conviction is seriously diminished.” Teague, 109 S.Ct. at 1076-77 (plurality opinion).9
*1521I agree with Judge Johnson that both the Gardner and Proffitt claims fit within the terms of this second exception. Both are based on the right of confrontation, and our adversarial system — unlike the inquisitorial method — depends above all else upon the right of confrontation to arrive at an accurate result.
III.
ABUSE OF THE WRIT
Because the Gardner and Proffitt claims may be applied retroactively on petitions for collateral relief, as a threshold matter these claims are available to Moore. Because the Estelle v. Smith claim is not “new” for retroactivity purposes, Teague does not come into play, and that claim is also, as a threshold matter, available to Moore. The next step is to determine whether Moore has abused the writ, and thus disentitled himself from presenting these claims through the equitable remedy of the writ of habeas corpus. See, e.g., Sanders v. United States, 373 U.S. 1, 17, 83 S.Ct. 1068, 1078, 10 L.Ed.2d 148 (1963); Gunn v. Newsome, 881 F.2d at 954-56.
I agree with Judge Johnson that Moore did not abuse the writ in failing to raise the Proffitt claim. While Judge Johnson suggests that the Supreme Court’s discussion of new law for retroactivity purposes implicates our prior en banc determination that Moore had not abused the writ with respect to his Estelle v. Smith claim, I do not. In my view the prior en banc court was correct and Teague does not affect that determination.10
I also agree with Judge Johnson that the district court should consider whether the ends of justice call for the court to entertain Moore’s Gardner claim. As the prior en banc court made clear, Moore’s Gardner claim does indeed have merit. Moore has raised a genuine factual dispute as to whether and when his counsel received the presentence investigation report, a report which indisputably contained many inaccuracies.
In finding the Gardner claim without merit, the majority reasons that because a defendant may be sentenced to death with only one aggravating circumstance, e.g., committing the murder during the commission of a felony, Moore must challenge that very aggravating circumstance for the ends of justice to apply. I emphatically disagree with this suggestion. The court that imposed the death penalty on Moore had an inaccurate presentence investigation report. The very purpose of this report is to enable the sentencing court to make a reasoned and informed decision on whether to impose that most final of penalties. The choice to impose the death penalty based on all the available and accurate information is the state’s. Does the majority really believe that the state court would not be *1522justified in imposing a penalty of less than death once the inaccuracies of the presen-tence investigation report were brought to the state’s attention? I find it an intrusion into the proper role of the state courts for a federal court to step in and declare that the state must impose the death penalty unless Moore challenges all applicable aggravating circumstances, and I know of no authority to support such an unprecedented rule. The ends of justice should require the district court to reach the merits of Moore’s Gardner claim.
IV.
CONCLUSION
The abuse of the writ issue decided by the court today was decided before by this court en banc; it was, however, decided the other way. This reversal of our previous decision is, at best, unseemly as there has been no intervening factual or legal development to explain or excuse reconsideration of the abuse of the writ issue. The Supreme Court vacated our prior en banc decision for reconsideration in light of Teague. It did not give us carte blanche to reexamine the entire case.
Five years ago I dissented from the original panel that affirmed the district court's determination that Moore had abused the writ. Two years ago I joined the majority opinion of the en banc court explaining why Moore had not abused the writ. Even if I did not dissent from the result reached by the majority today, I would like to think that I would dissent from the decision to revisit our prior en banc opinion and to ignore the Supreme Court’s remand order. Accordingly, I once again dissent.
. Written by Judge Cox and joined by Judges Tjoflat, Fay, Vance and Edmondson.
. The Cox opinion suggests that simply applying Teague to our prior opinion must implicitly rest on the belief that the Supreme Court has approved of our prior determination. The remand order does not constitute an implicit approval of our prior opinion. Nor does it constitute implicit disapproval or constitute a direction to "go back and do whatever you want.” The remand order means what it says, no more, no less: "reconsider your prior opinion in light of Teague."
. In Penry v. Lynaugh, — U.S. -, 109 S.Ct. 2934, 106 L.Ed.2d 256 (1989), the Supreme Court applied Teague to a capital case, settling a question that Teague itself had not resolved: the nonretroactivity principles of Teague apply to capital cases.
. By contrast, a “new law” claim in the abuse of the writ context means a claim seeking the benefit of a rule announced after the petitioner filed a prior habeas petition.
.Judge Johnson suggests that Teague and the principles of nonretroactivity are an affirmative defense that may be waived. I disagree. Permitting a state to waive Teague in some cases and not in others would create the very unfairness and disparate treatment of similarly situated petitioners that Teague sought to prevent. I believe that a federal habeas court must conduct an analysis of whether Teague applies as a threshold matter, before reaching, for example, the abuse of the writ. See Penry v. Lynaugh, — U.S. -, 109 S.Ct. 2934, 2944, 106 L.Ed.2d 256 (1989) ("Penry is currently before the Court on his petition in federal court for a writ of habeas corpus. Because Penry is before us on collateral review, we must determine, as a threshold matter, whether granting him the relief he seeks *1520would create a ‘new rule.’ Teague v. Lane, 489 U.S. -, -, 109 S.Ct. 1060, -, 103 L.Ed.2d 334 (1989). Under Teague, new rules will not be applied or announced in cases on collateral review unless they fall into one of two exceptions. Id. at -, 109 S.Ct. at -.”).
. I also disagree with Judge Johnson’s reading of the Supreme Court’s remand as a suggestion "that only decisions which are new law for retroactivity purposes may excuse a successive petition.”
. See, e.g., Teague, 109 S.Ct. at 1070 ("It is admittedly often difficult to determine when a case announces a new rule, and we do not attempt to define the spectrum of what may or may not constitute a new rule for retroactivity purposes.") (plurality opinion).
. The Estelle v. Smith claim drops out of my Teague analysis because it is not new law for retroactivity purposes, therefore Teague is not implicated.
. Justice Harlan explained why a fundamental rule should be applied retroactively in Mackey v. United States:
[I]n some situations it might be that time and growth in social capacity, as well as judicial perceptions of what we can rightly demand of the adjudicatory process, will properly alter our understanding of the bedrock procedural elements that must be found to vitiate the fairness of a particular conviction. For example, such, in my view is the case with the right to counsel at trial now held a necessary condition precedent to any conviction for a serious crime.
*1521Mackey, 401 U.S. 667, 693-94, 91 S.Ct. 1160, 1180, 28 L.Ed.2d 404 (1971) (separate opinion) (quoted in Teague, 109 S.Ct. at 1075-76 (plurality opinion)).
A rule must implicate the accuracy of a conviction in order to meet this second exception to Teague’s rule of nonretroactivity because one of the main purposes of the writ of habeas corpus is, as Justice Harlan noted in Desist, "to assure that no man has been incarcerated under a procedure which creates an impermissibly large risk that the innocent will be convicted.” Thus "all ‘new’ constitutional rules which significantly improve the pre-existing factfinding procedures are to be retroactively applied on habeas.” Desist v. United States, 394 U.S. 244, 262, 89 S.Ct. 1030, 1041, 22 L.Ed.2d 248 (1969) (Harlan, J., dissenting) (quoted in Teague, 109 S.Ct. at 1076 (plurality opinion)).
The Teague plurality fashioned the second exception by combining these two concerns, and "limiting the scope of the second exception to those new [fundamental] procedures without which the likelihood of an accurate conviction is seriously diminished.” Teague, 109 S.Ct. at 1076 (plurality opinion). Justice Stevens, however, while concurring in the use of Harlan’s analytic framework, disagreed with the plurality’s modification of the second exception, and adhered to Harlan’s own view rejecting the linkage of fundamental fairness to factual innocence. In addition, Stevens noted that "a touchstone of factual innocence would provide little guidance in certain important types of cases, such as those challenging the constitutionality of capital sentencing hearings." Teague, 109 S.Ct. at 1080-81 (Stevens, J., joined by Black-mun, J., concurring in part) (footnote omitted).
. Although Chief Judge Roney and Judge Hill agree that the court first should have applied Teague to the issues presented before considering abuse of the writ, they do not agree with the five dissenting judges’ conclusion that Teague does not bar Moore’s claims.