dissenting.
The majority has resurrected a new substantive right of their own invention, which made its first appearance in Stumpf v. Mitchell, 367 F.3d 594 (6th Cir.2004), vacated by Bradshaw v. Stumpf, 545 U.S. 175, 125 S.Ct. 2398, 162 L.Ed.2d 143 (2005), and apparently had all along been lurking somewhere within the Fourteenth Amendment. In its current iteration, the new right protects a convicted murderer from being sentenced to death where mitigating evidence (i.e., evidence that does not undermine the murder conviction itself but that might have counseled towards a more lenient sentence) discovered after sentencing is later used by the prosecution against a different defendant. Notably, the due process violation is not that mitigating evidence exists that is later discovered, which would not by itself offend the Constitution, Noel v. Norris, 322 F.3d 500, 504 (8th Cir.2003); see Herrera v. Collins, 506 U.S. 390, 400, 113 S.Ct. 853, 122 L.Ed.2d 203 (1993),1 but, curiously, that the newly discovered evidence is later used by the prosecution against a different defendant. Somehow, that purely later conduct retroactively renders the earlier sentence unconstitutional.
Consider a hypothetical case. B is killed in a horrifying fashion. A is tried for the murder of B, is convicted, and due to the terrible nature of the crime, is sentenced to death. A’s trial is not merely in compliance with constitutional standards, but is a model trial. The most scrutinizing criminal lawyers available comb through the trial record in search of some plausible legal claim to bring on appeal, but they find nothing, and A’s sentence is correctly affirmed on direct review. The majority’s new right has no import at this time. Many years later, a witness comes forward with new information — the witness explains that C and A killed B in concert, and that C’s conduct was more vile. Still, the majority’s new right has no import. But wait — the prosecutor acts on the witness’s testimony and tries C for B’s murder, arguing that, even though A was already convicted of the murder, it was in fact C that committed the most horrifying aspects of the crime. Sure, the jury acquits C, but that is apparently besides the point. Now, the majority’s new right finally jumps into action. The prosecutor in C’s trial has denied A — yes, A — the right to be sentenced fairly, in violation of the Due Process Clause of the Fourteenth Amendment.
I do not agree with that application of the Constitution in the slightest,2 and as I *441explained in my first dissent in this case, I believe that the out-of-circuit cases that the majority relies upon for its new rule are readily distinguishable. Stumpf v. Mitchell, 367 F.3d at 618-22 (Boggs, J., dissenting). I would therefore affirm the decision of the district court dismissing Stumpfs petition for a writ of habeas corpus.
If ever there were a case in which to adopt the majority’s new right, this is not it. Indeed, the majority’s opinion in many places obfuscates the timeline of what happened here, and argues as if the state withheld information from the Ohio courts, or argued factually inconsistent theories at the original sentencing or at the same time. That is simply not correct. Stumpfs trial and sentencing were based on all the evidence then available, and there was no error in those proceedings.3 Subsequent to that time, new evidence came to the attention of the state in the form of a statement by Eastman, a cellmate of co-defendant Wesley. Such statements are always viewed with some skepticism, and in any case, there is no claim that the state was responsible for bringing forth this evidence, or that the state was in any way derelict in its duties in not discovering it earlier.
Clearly, there is a timeline problem for the majority — how did the prosecution violate the Constitution at Stumpfs trial with respect to evidence that did not yet exist? The majority attempts to deal with this issue by focusing in part on the prosecutor’s actions in response to Stumpfs post-trial motion for resentencing, which occurred after Wesley’s trial. The majority faults the prosecutor for apparently making an argument that they disapprove of (by ostensibly trying to have it both ways) rather than making the argument that they apparently would approve of and simply confessing that he was wrong at Wesley’s trial. See maj. op. at 437-38. But this is not a situation where the prosecutor misled or bamboozled a jury. Rather, the prosecutor made an argument, referencing all available evidence, to a panel of judges. The state did not hide the ball, and the judges were not bamboozled. I know of no case or principle in which an otherwise unblemished prosecution can be held to have violated the Constitution because a federal appellate panel disapproves of the nature of the prosecutor’s oral arguments to a judge at a post-trial proceeding.
The majority’s opinion here emphasizes the dilemma in which the state was placed and the nature of today’s ruling. If the state had simply ignored the evidence, it apparently would have now been in no difficulty. See Dist. Attorney’s Office for Third Judicial Dist. v. Osborne, 557 U.S. 52, 129 S.Ct. 2308, 2319-20, 174 L.Ed.2d 38 (2009) (no Brady obligation to disclose exculpatory evidence discovered post-trial). Indeed, the majority is apparently clear that, if the prosecution had simply decided not to present the cellmate evidence against Wesley and achieved the same life sentence that it did here, there would be no inconsistency, and thus no due-process violation. See maj. op. at 437-38, 439-40. That concession emphasizes the bizarro nature of the majority’s argument that you can somehow violate the rights of A, who *442has had a meticulously fair proceeding, by the decision that you make as to how to proceed in the trial of C.
The inconsistent-theory cases raised by the majority virtually all relate to a state’s contemporaneous presentation of inconsistent theories that could have an effect on the trial at issue, and that could result in factually impossible inconsistencies impacting both defendants. Here, the alleged violation is how the prosecutor chose to deal with the later-discovered evidence, in a wholly separate proceeding. It is much closer to the alleged inconsistency in our case of Getsy v. Mitchell, 495 F.3d 295 (6th Cir.2007), where we rejected, en banc, a similar argument. At bottom, what happened here is that some additional evidence, the statement of the cellmate, came to light after the proceedings were concluded. That new material was presented in a proper way to the courts of Ohio, was litigated by Stumpf in the proper order, and Stumpf lost.
Further, even if I were to agree that the prosecutor4 somehow retroactively violated Stumpfs constitutional rights, I fail to see how Stumpf was prejudiced by that violation. Once the Eastman testimony became available to him, Stumpf went back to the same Ohio panel that sentenced him.5 The particulars of Eastman’s testimony, and of its handling at Wesley’s trial, were all fully explored on the record in Wesley’s trial, and were then placed before the Ohio panel hearing Stumpfs motion for resentencing based on that new evidence. That panel had the power to undo Stumpfs death sentence, but declined to do so, even after being fully informed of all of the facts, including the ones that the majority finds dispositive as to the prosecutor’s action.
The majority emphasizes, at page 437, a statement by one judge, made during oral argument on Stumpfs motion for resentencing, that “if [the panel] had not been satisfied that Stumpf was, in fact, the trigger man, the principal offender, and we were satisfied that he was, in fact, an aider and abettor, that may very well have had an effect upon this Court’s determination of whether the death penalty should follow. I’m not saying it would, but it’s possible.” It should first be emphasized just how qualified that statement is — the judge merely expressed orally, in the course of a discussion with the prosecutor, the possibility that the panel might have declined to impose the death penalty if it were satisfied that Stumpf was not the principal offender. Indeed, the judge simply emphasized the prosecutor’s framing of the issue (i.e., whether Stumpf was the principal offender and, if not, whether the panel would still have imposed the death penalty), and in no way indicated a leaning as to how the panel felt about the matter. That statement does not appear in any way in the court’s written decision and order, indicating that the panel either concluded, after deliberation, that it was satisfied that *443Stumpf was the principal offender or, alternatively, that a contrary finding would not have made a difference in its decision to impose the death penalty. The panel had full opportunity and ability to state any skepticism as to its earlier action, or to indicate that it had some belief in the statement by the cellmate, and it did not do so. Thus, the majority’s speculation at pages 436 through 438 as to its view of the evidence is simply unsupported.
Certainly, the panel had the same options that it did at the original hearing. The panel could have concluded either that (1) Stumpf was the killer, (2) Wesley was the killer, or (3) either way, Stumpf deserved the death penalty for his involvement in this heinous crime, for which he was criminally responsible even if Wesley pulled the trigger. Notably, the panel also had inculpatory evidence that it did not have before, in the form of Wesley’s trial statements that supported the prosecution’s view of events — that Stumpf was the person who killed Mary Jane Stout. And, of course, it is worth mentioning the fact that the Eastman testimony was not credited by the Wesley jury, which apparently did not believe the cellmate’s version of the facts, or, in any event, did not find it so compelling as to sentence Wesley to death. In light of the Ohio panel’s review of that evidence, and its decision to leave Stumpfs death sentence intact, I do not see how Stumpf has demonstrated any probability that he was prejudiced by the prosecutor’s decision to present the new evidence and argue a new theory at Wesley’s trial.
In addition, even if I were to share the majority’s vision of the Fourteenth Amendment as applied to this case, Teague v. Lane clearly bars the retroactive application of the majority’s new constitutional right to Stumpf. See 489 U.S. 288, 306, 109 S.Ct. 1060, 103 L.Ed.2d 334 (1989). Notably, all of the cases that the majority relies upon for its rule were decided long after Stumpfs conviction became final in 1988, when the Supreme Court denied certiorari. See Stumpf v. Ohio, 484 U.S. 1079, 108 S.Ct. 1060, 98 L.Ed.2d 1022. Accordingly, “the result was not dictated by precedent existing at the time the defendant’s conviction became final,” and the rule is new for purposes of Teague. Teague, 489 U.S. at 301, 109 S.Ct. 1060. Further, neither of Teague’s two exceptions — new limitations on criminal lawmaking authority and new “watershed” rules of criminal procedure — apply here. DeCastro v. Branker, 642 F.3d 442, 458 & n. 4 (4th Cir.2011).
Although their conclusion conflicts with the views of the only two Justices to consider the issue, pursuant to Circuit precedent,6 I agree with the majority that the state has waived its Teague defense. Bradshaw v. Stumpf, 545 U.S. at 191, 125 S.Ct. 2398 (Thomas, J., concurring) (“The Court’s opinion does not preclude the State from advancing [a Teague ] defense[ ] on remand in support of Stumpfs death sentence.”). However, whether we must apply Teague is only the threshold inquiry. As the Supreme Court has held, “a federal court may, but need not, decline to apply Teague if the State does not argue it.” Horn v. Banks, 536 U.S. 266, 271, 122 S.Ct. 2147, 153 L.Ed.2d 301 (2002) (quoting Bohlen, 510 U.S. at 389, 114 S.Ct. 948). Accordingly, after determining that the state waived Teague and we are not required to apply it, step two is to consider whether we should apply Teague.
On that point, the majority’s analysis is unconvincing. The majority suggests that *444the “stakes are simply too high” to allow the state’s Teague defense “to forestall review of the merits” of Stumpfs due-process claim. Maj. op. at 434-35 n. 3. That, however, is flatly inconsistent with Teague itself. The very holding of Teague is that courts must avoid reaching the merits where, such as here, the petitioner seeks “[a]pplieation of constitutional rules not in existence at the time [his] conviction became final.” Teague, 489 U.S. at 309, 316, 109 S.Ct. 1060. Accordingly, “forestalling review” would not undermine the application of Teague not just in this case, but in all cases. Further, one of the Court’s primary considerations in adopting Teague’s nonretroactivity rule was that similarly situated defendants should be treated similarly. Id. at 315, 109 S.Ct. 1060. Accordingly, a determination in a particular case that, for one reason or another, “the stakes are too high,” is no reason to ignore Teague. That conclusion is reinforced here, where the majority’s rule will have only the most sporadic impact. Significantly, the majority’s rule is not clearly established Supreme Court precedent, see 28 U.S.C. § 2254(d), and post-AEDPA habeas petitioners are therefore unable to avail themselves of it, regardless of whether the state properly raised Teague.
Because the concerns motivating Teague do not depend on whether the state promptly raises the issue, I agree with Judge Gilman’s concurrence in Lyons, in which he argued that “a federal court should presumptively apply the Teague analysis sua sponte whenever a defendant tries to raise a new constitutional rule for the first time on collateral review.” 188 F.3d at 346; accord id. at 339 n. 8 (Clay, J., concurring) (“[I]rrespective of the state’s preservation of its Teague defense, based upon the new rule proposed by Petitioner, if ever there was a time for this Court to raise and decide a Teague claim sua sponte, surely it is here.”). In this case, where there is no indication that the state strategically sought to “sandbag” this court, see id. at 346, where both parties fully briefed and argued the issue, see Albrecht v. Horn, 485 F.3d 103, 120 (3d Cir.2007), and where we are more than capable of making a reasoned judgment on the issue, see Goeke v. Branch, 514 U.S. 115, 118, 115 S.Ct. 1275, 131 L.Ed.2d 152 (1995), I would hold that Teague should apply, even though I must agree with the majority that the state waived the argument.
For these reasons, I respectfully dissent, both on the merits and pursuant to Teague’s nonretroactivity principle.7
. Although new evidence may be used to substantiate other constitutional claims, such as prosecutorial misconduct in the form of knowingly using perjured testimony, the existence of new evidence of innocence, or of “innocence of the death penalty” as in this case, is not in and of itself a constitutional violation. In any case, Stumpf's new evidence does not speak to any problematic conduct that occurred at his trial, as in the ordinary case, but rather at Wesley's trial, when the prosecutor adopted a theory based on the new evidence.
. I note that the majority’s repeated references to a federal court's obligation to ensure accuracy, maj. op. at 435-36, 436, 437-38, in the context of death-penalty habeas appeals simply has no support in Supreme Court jurisprudence. We do not sit to indulge our own sense of justice as to how state-court proceedings should come out. We do not have a freestanding license to decide how the strength of a case appeals to us without reference to existing laws and procedures. It is true that Justice Marshall on occasion, dissenting alone, used such a phrase, see, e.g. Barefoot v. Estelle, 463 U.S. 880, 938, 103 *441S.Ct. 3383, 77 L.Ed.2d 1090 (1983), but that vague statement, pregnant with the outcome preferences of the judges charged with implementing it, is appropriately not the law.
. See Bradshaw v. Stumpf, 545 U.S. at 186—87, 125 S.Ct. 2398 (unanimously reversing, without argument, the majority's holding invalidating Stumpfs guilty plea and its holding that inconsistent prosecutorial theories invalidated his conviction, and remanding to consider whether the inconsistent theories could have affected Stumpf's death sentence).
. Because Wesley's trial was wholly subsequent to and unconnected to Stumpfs, it is simply fortuitous that the same prosecutor was involved. The second trial could have just as easily been prosecuted by a different person or even, under some circumstances, in a different jurisdiction. In any case, I fail to see how the prosecutor's judgment at Wesley’s trial, which was based upon previously unavailable evidence, can somehow impeach his earlier motives or actions and render the original, previously constitutional, proceeding unconstitutional.
. The panel was comprised of two of the three original judges, the third having died in the interim. The two judges proceeded to hear Stumpfs motion for resentencing, premised on the new evidence, and they agreed that, so long as they were unanimous, there was no reason to bring in a third judge. Stumpf does not challenge that decision here.
. In a three-opinion split decision, a panel of this court held that a state waives its Teague defense by failing to raise it before the district court. Lyons v. Stovall, 188 F.3d 327, 345-46 (6th Cir.1999) (Gilman, J., concurring); id. at 347-48 & n. 1 (Moore, J., dissenting).
. This case was originally heard on December 11, 2002. After a decision on April 28, 2004, and a subsequent trip to the United States Supreme Court, it was heard again on July 26, 2007 — four years ago. During that interval, the state has filed two motions to expedite a decision, one on August 6, 2010, and a second on June 2, 2011. As is obvious, in a death penalty appeal, a failure to decide has the same short-run effect as a decision in favor of the condemned — it prevents his execution.
In the absence of a decision by the panel, there is no process allowing for the court as a whole to take the decision en banc. Thus, the only remedy available (from the Sixth Circuit) to a litigant in the position of the State of Ohio here is the purely hortatory motion that the state filed here — a motion that, like the very decision that it seeks to expedite, is also solely within the power of the panel and cannot be taken en banc (if it ever could be) in the absence of a decision on that motion. The only remedy available under current rules is a petition to the Supreme Court for *445certiorari before judgment. See Sup.Ct. R. 11.