I concur with the majority opinion regarding the issues of alleged judicial bias and ineffective assistance of counsel. It is only the issue of prosecutorial misconduct that causes me to dissent from the opinion of my colleagues.
Initially, and as a matter of law, I must agree with the majority that the prosecu-torial misconduct in this case is subject to the doctrine of harmless constitutional error as set forth in Brecht v. Abrahamson, 507 U.S. 619, 113 S.Ct. 1710, 123 L.Ed.2d 353 (1993).1 If, however, a reviewing court *649is in “grave doubt” as to the constitutional error’s impact on the verdict, the court should deem the error as not being harmless and should reverse the conviction. O’Neal v. McAninch, 513 U.S. 432, 436, 115 S.Ct. 992, 130 L.Ed.2d 947 (1995).
In Kincade v. Sparkman, 175 F.3d 444 (6th Cir.1999), the Sixth Circuit set forth the criteria for determining whether habe-as relief should be granted for prosecutorial misconduct:
To grant relief in such cases as this, we must find that the prosecutor’s comments constituted more than simply trial error under state law. The misconduct must be “so fundamentally unfair as to deny him due process,” Donnelly v. DeChristoforo, 416 U.S. 637, 645, 94 S.Ct. 1868, 40 L.Ed.2d 431 (1974), based on “the totality of the circumstances” of the case, taking into account
the degree to which the remarks complained of have a tendency to mislead the jury and to prejudice the accused; whether they are isolated or extensive; whether they were deliberately or accidentally placed before the jury, and the strength of the competent proof to establish the guilt of the accused.
Angel v. Overberg, 682 F.2d 605, 608 (6th Cir.1982) (en banc) (quoting from United States v. Leon, 534 F.2d 667, 677 (6th Cir.1976)).
175 F.3d at 445-46.
The majority opinion appears to apply the test for determining whether the pros-ecutorial misconduct in this case violated the defendant’s right to due process, as set forth in Kincade and the earlier case of Angel v. Overberg, 682 F.2d 605, 608 (6th Cir.1982) and, with the exception of the prosecutorial misconduct being isolated, “every other prong of the [Angel] test” has been satisfied. Opinion, Section 11(D). The majority nevertheless then applies the Brecht standard for harmless error and, because a second degree murder conviction in Michigan does not require an intent to kill, and there was sufficient evidence to support a finding of intent to create a high risk of death or bodily harm, the conclusion is reached that the prosecutorial misconduct did not have a substantial injurious effect on the outcome of the trial. I have several problems with this.
Fust, the issue before us, under the Antiterrorism and Effective Death Penalty Act (“AEDPA”), is whether the decision of the Michigan Court of Appeals involved an unreasonable application of clearly established federal law, as determined by the Supreme Court of the United States. I think it is clear from the state court’s opinion that it did not apply the criteria set forth in Darden v. Wainwright, 477 U.S. 168, 181, 106 S.Ct. 2464, 91 L.Ed.2d 144 (1986) or, for that matter, the Angel criteria adopted in this Circuit, to determine whether the admitted prosecutorial misconduct violated defendant’s constitutional right to a fair trial. Furthermore, as discussed later, I think it is also clear *650that the state court did not apply the Chapman standard in its finding of harmless error. In my view, applying the Supreme Court’s Darden criteria or the Angel criteria of this Circuit, the result is that defendant’s constitutional right to a fair trial was violated.
Second, even if the violation of defendant’s constitutional right to a fair trial is subject to the Brecht standard of harmless error, I believe that the prosecutorial misconduct in this case was so outrageous that it falls within the exception noted in the following footnote in Brecht:
Our holding does not foreclose the possibility that in an unusual case, a deliberate and especially egregious error of the trial type, or one that is combined with a pattern of prosecutorial misconduct, might so infect the integrity of the proceeding as to warrant the grant of habe-as relief, even if it did not substantially influence the jury’s verdict. Cf. Greer v. Miller, 483 U.S. 756, 769,107 S.Ct. 3102, 3110, 97 L.Ed.2d 618 (1987) (STEVENS, J., concurring in judgment). We, of course, are not presented with such a situation here.
507 U.S. at 638 n. 9, 113 S.Ct. 1710. In my opinion, the deliberate and especially egregious prosecutorial misconduct in the present ease so infected “the integrity of the proceeding as to warrant the grant of ha-beas relief even if it did not substantially influence the jury’s verdict.”2
Third, even if the Brecht standard of harmless error is to be applied and this case does not fall within the exception noted in Brecht, I still would find that the defendant should be given habeas relief. Contrary to the majority view, I do not believe that simply because the state did not have to prove an intent to kill, the prosecutorial misconduct did not have a “substantial and injurious effect or influence in determining the jury’s verdict.” The defense was based entirely on defendant’s contention that the fatal shot was fired accidentally without any intent to kill or injure his girlfriend or to create a high risk of death or injury to her. The defendant’s intent was a critical element of the charged offense, and the defendant’s own testimony regarding his intent was the foundation of his defense. The prosecutor’s effort to destroy the defendant’s credibility before the jury and to put before the jury a matter that was not in evidence — the alleged testimony of a disinterested person that would prove an intent by the defendant to kill his girlfriend — in violation of defendant’s constitutional right to cross-examine such a witness, could not, in my opinion, fail to have an influence on the jury. As the majority opinion implicitly finds, the standard cautionary instruction was inadequate to “unring the bell,” and defendant’s reaction to the prosecutor’s statements, if anything, only served to pit his credibility against the credibility of a disinterested person who was not produced as a witness and subjected to cross-examination.
The majority finds that, apart from the attempt by the prosecutor to influence the jury in a totally improper way, there was sufficient admissible evidence in the record, in the form of petitioner’s statements to the police, to support the jury’s verdict. This, I believe, is an improper approach to determine whether the prosecutorial misconduct was harmless error. As the Supreme Court said in Kotteakos v. United States:
The inquiry cannot be merely whether there was enough to support the result, apart from the phase affected by the error. It is rather, even so, whether the error itself had substantial influence. If so, or if one is left in grave doubt, the conviction cannot stand.
328 U.S. 750, 765, 66 S.Ct. 1239, 90 L.Ed. 1557 (1946).
The majority states that if petitioner had been convicted of first degree murder *651they would agree that the prosecutorial misconduct had a substantial and injurious effect on the outcome of the trial, but since petitioner was convicted of second degree murder, the prosecutorial misconduct did not have a substantial and injurious effect on the outcome of the trial- I cannot, however, see that the distinction between first degree murder and second degree murder is a sufficient basis for denial of relief in this case. A defendant’s intent is a critical element of both crimes under Michigan law, and the prosecutorial misconduct involved proving that intent in an egregious manner and, at the same time, impugning the credibility of petitioner’s own testimony concerning his intent. In my view, basic fairness, as well as the constitutional right to due process, prohibits such conduct and, regardless of the degree of the offense in question, such conduct cannot avoid having a substantial and injurious effect on the outcome of the trial.
The Michigan Court of Appeals, in finding harmless error, was required to apply the test set forth in Chapman v. California, 386 U.S. 18, 87 S.Ct. 824, 17 L.Ed.2d 705 (1967), i.e., whether the prosecutorial misconduct was “harmless beyond a reasonable doubt.” There is absolutely no indication in its opinion that it followed this clearly established federal law as determined by the Supreme Court of the United States. To the contrary, it merely found that “it is unlikely that he (the defendant) suffered any prejudice as a result” (of the prosecutorial misconduct). J.A., Part 1, p. 51. Even under the high requirement of Nevers v. Killinger, 169 F.3d 352 (6th Cir.1999), in its interpretation of what must be shown to be an “unreasonable application of federal law,” if a habeas petitioner can demonstrate that the trial error meets the Brecht standard, “he will surely have demonstrated that the state court’s finding that the error was harmless beyond a reasonable doubt — the Chapman standard — was outside the realm of plausible credible outcomes, and therefore resulted from an unreasonable application of Chapman. Nevers at 371-372. In my view, this is such a case. The petitioner has shown that: (1) there is no indication whatsoever that the Michigan court applied the harmless error Chapman standard; (2) the prosecutorial misconduct meets the test of Brecht, if Brecht is to be applied, and was not harmless error; and (3) the Michigan court’s decision therefore involved an unreasonable application of clearly established federal law as determined by the Supreme Court of the United States.
It is my belief that two very recent decisions by this Court support the conclusion I have reached in this case. In Barker v. Yukins, 199 F.3d 867 (6th Cir.1999), the petitioner for a writ of habeas corpus had been convicted of first degree murder in Michigan. The defense was that petitioner killed the decedent while resisting a rape. The error of the trial court was in refusing to instruct the jury that petitioner was entitled to use deadly force to resist an imminent rape. The Michigan Supreme Court denied the petition on the ground that the error was harmless because, under the evidence presented at trial, no reasonable juror could have believed deadly force was necessary to prevent a rape.
This Court reversed the district court’s denial of habeas relief and ordered that a conditional writ of habeas corpus be granted unless Michigan commenced a new trial within 180 days of the opinion. This Court recognized the Supreme Court’s decision in O’Neal v. McAninch, 513 U.S. 432, 115 S.Ct. 992, 130 L.Ed.2d 947 (1995), in which the Supreme Court emphatically stated:
When a federal judge in a habeas proceeding is in grave doubt about whether a trial error of federal law had “substantial and injurious effect or influence in determining the jury’s verdict,” that error is not harmless. And, the petitioner must win.
Id. at 436, 115 S.Ct. 992.
Citing O’Neal, this Court in the Barker case said:
*652Only if a federal habeas court can say with certainty that a trial error had little to no impact on the judgment, should the judgment stand. See id. at 435-38, 115 S.Ct. at 994-995. In this matter, the only thing of which this court is certain is that the erroneous jury instruction left the door wide open. A reasonable juror could have very well walked through the door and rejected Barker’s claim for self defense because that juror believed that Madsen’s assault would not have led to death or serious bodily injury, thereby resulting in a substantial and injurious influence on the verdict. Since there is grave doubt as to whether the erroneous jury instruction created a substantial and injurious influence on the verdict, the error was not harmless. Accordingly, then, the Michigan Supreme Court engaged in an unreasonable application of Chapman’s harmless error test, and under § 2254(d), a writ of habeas corpus should issue.
199 F.3d at 874.
With only slight paraphrasing, this same language applies equally to this case:
Only if a federal habeas court can say with certainty that a trial error had little to no impact on the judgment, should the judgment stand. See O’Neal v. McAninch, 513 U.S. 432 at 435-38, 115 S.Ct. at 994-995, 130 L.Ed.2d 947. In this matter, the only thing of which this court is certain is that the prosecutorial misconduct left the door wide open. A reasonable juror could have very well walked through the door and rejected Maurino’s claim that he did not intend to create a high risk of death or great bodily harm to his girlfriend because that juror believed the prosecutor’s representation that there was an impartial person who said the petitioner had given her a bullet with the statement that “this one’s for Vicki,” thereby resulting in a substantial and injurious influence on the verdict. Since there is grave doubt as to whether the prosecutorial misconduct created a substantial and injurious influence on the verdict, the error was not harmless. Accordingly, then, the Michigan Supreme Court engaged in an unreasonable application of Chapman’s harmless error test, and under § 2254(d), a writ of habeas corpus should issue.
Also, even more recently, this Court held that an inflammatory cross-examination of the defendant and an improper closing argument (less egregious in my view than the prosecutorial misconduct in the present case), required the granting of a conditional writ of habeas corpus. In Boyle v. Million, 201 F.3d 711 (6th Cir.2000), this Court said:
It is true that the case against Boyle was relatively straightforward and strong. Given the egregious and inflammatory nature of the behavior and arguments of the prosecutor throughout trial, however, we are left with “grave doubt” as to whether the prosecutorial errors “had a substantial and injurious effect or influence in determining the jury’s verdict.” Brecht v. Abrahamson, 507 U.S. 619, 637, 113 S.Ct. 1710, 123 L.Ed.2d 353 (1993)(quoting Kotteakos v. United States, 328 U.S. 750, 776, 66 S.Ct. 1239, 90 L.Ed. 1557 (1946)); see also O’Neal v. McAninch, 513 U.S. 432, 436, 115 S.Ct. 992, 130 L.Ed.2d 947 (1995).
Id. at 717-18.
For the reasons stated above, I would reverse the judgment of the District Court and order that a conditional writ of habeas corpus be granted unless the State of Michigan commences trial proceedings against Maurino within 180 days of this opinion.
. Prosecutorial misconduct has not been considered a "structural defect” in the trial proceedings that would require an automatic reversal. See Arizona v. Fulminante, 499 U.S. 279, 111 S.Ct. 1246, 113 L.Ed.2d 302 (1991). I, nevertheless, have felt that if, in applying *649ihe criteria for determining whether prosecu-torial misconduct rises to the level of a violation of defendant's constitutional right to due process, see Darden v. Wainwright, 477 U.S. 168, 181, 106 S.Ct. 2464, 91 L.Ed.2d 144 (1986); Swofford v. DeTella, 101 F.3d 1218 (7th Cir.1996), it is found that this fundamental right to a fair trial has been violated, there should be no place for a harmless error analysis. If, in applying that criteria, the prosecu-torial misconduct does not rise to the level of a constitutional violation, there is no basis for habeas relief. Darden, 477 U.S. at 181 n. 15, 106 S.Ct. 2464. The misconduct in such a case can be deemed to be harmless error, but this is not the result of a Brecht harmless error analysis. By applying a harmless error analysis when there has been a constitutional violation, egregious prosecutorial misconduct is permitted if the state's case is deemed to be sufficiently strong, a result that not only reflects adversely on the judicial process in criminal cases, but also does not serve as a deterrent to prosecutorial abuse. The Sixth Circuit, however, has specifically held that, in cases involving prosecutorial misconduct, the reviewing court must apply a harmless error analysis. Eberhardt v. Bordenkircher, 605 F.2d 275 (6th Cir.1979); Pritchett v. Pitcher, 117 F.3d 959 (6th Cir.1997). I respect and, of course, am bound by those decisions.
. In Section 11(D) of the majority opinion, Judges Kennedy and Norris agree that "the prosecutor’s statements were egregious prose-cutorial misconduct.”