Jonathan Blades v. United States

Beckwith, Associate Judge, dissenting:

Several clear principles have emerged from the case law on a criminal defendant's Sixth Amendment right to a public trial. We know that the purpose of this right is to benefit the accused, Gannett Co. v. DePasquale , 443 U.S. 368 , 380, 99 S.Ct. 2898 , 61 L.Ed.2d 608 (1979), and that the right is at least as protective as the press and the public's First Amendment right to public trial, Waller v. Georgia , 467 U.S. 39 , 46, 104 S.Ct. 2210 , 81 L.Ed.2d 31 (1984). We know that contemporaneous review of court proceedings is critical to the right's protections, 1 and that a transcript is therefore not generally a meaningful substitute. 2 Most pertinent to the present case, we know that the right extends to jury selection, Presley v. Georgia , 558 U.S. 209 , 213, 130 S.Ct. 721 , 175 L.Ed.2d 675 (2010) (describing this as "well settled"), and that closing the voir dire proceedings in a criminal trial will violate the right to public trial unless it satisfies the four-part test of Waller v. Georgia : "To close a proceeding: (1) the party seeking closure must advance an 'overriding interest that is likely to be prejudiced'; (2) the closure must be 'no broader than necessary to protect that interest'; (3) the court must consider 'reasonable alternatives' to closure; and (4) the court must 'make findings adequate to support the closure.' " Rodriguez v. Miller , 537 F.3d 102 , 108 (2d Cir. 2008) (citing *249 Waller , 467 U.S. at 48 , 104 S.Ct. 2210 ). 3 And finally, we know that a closure that does not satisfy the four-part Waller test is structural error that requires reversal. 4

In this case, the trial court decided that spectators of Mr. Blades's criminal trial would be able to see-but not hear-the questioning of prospective jurors regarding their responses on jury questionnaires based on a generalized view that jurors are more forthcoming and candid under the cover of the husher. My colleagues in the majority do not dispute that such a generalized justification would fail the Waller test. They conclude, however, that "the husher procedure in this case ... does not amount to a closure or partial closure of the courtroom, but is more appropriately viewed as an alternative to closure." Ante at 240. Such an "alternative to closure," they assume, is not subject to Waller 's prerequisites.

Although I am cognizant of the ramifications of recognizing a constitutional problem with an apparently widely used Superior Court practice, I dissent from my colleagues' decision to uphold that practice here because it is impossible to reconcile the majority's view of the husher procedure as somehow exempt from Waller 's requirements for proposed limitations on a public trial with the Supreme Court's public trial cases and with the decisions of this and other courts. 5 As an initial matter, my colleagues do not explain why limitations short of closing the courtroom door are not subject to the Waller test. And in fact, with the exception of Copeland v. United States , 111 A.3d 627 (D.C. 2015), which I address infra , the only case I have found that addresses a challenge on public-trial grounds to the particular practice used in this case never questioned that the use of a husher during voir dire was a type of closure subject to the requirements of Waller . See In re Memphis Pub. Co. , 887 F.2d 646 , 648-49 (6th Cir. 1989) (citing Waller 's precursor Press-Enterprise Co. v. Superior Court of California , 464 U.S. 501 , 510, 104 S.Ct. 819 , 78 L.Ed.2d 629 (1984) ). In Memphis Publishing , the Sixth Circuit reversed without any harm analysis because the trial court's justification for using the husher during jury selection was too general and "without any specific finding of fact to support [it]." Id. Similarly, although the majority relies heavily upon obstructed view (hear-but-not-see) cases in support of its view that the see-but-not-hear approach employed here is an "alternative to closure," those cases all make clear that a trial court's use of a screen to obstruct the public's view of a witness or a proceeding is deemed constitutional only where the court first satisfied the Waller criteria. See infra. Whatever we call the husher procedure-a closure, a partial closure, or an alternative to closure-and *250 however longstanding and accepted the Superior Court practice may be, its routine employment based on generalized concerns about juror candor is the antithesis of applying and satisfying the Waller factors, which require, among other things, individualized and adequately supported findings demonstrating the necessity of the closure in the specific circumstances of that case.

In case after case, the Supreme Court and other appellate courts have held that limitations upon the public's access to jury selection procedures were unconstitutional, not based on the particular form of closure, but because the trial court imposed it without satisfying the criteria of Waller . In Cable News Network, Inc. v. United States , 824 F.2d 1046 (D.C. Cir. 1987), for example, the D.C. Circuit summarily reversed the district court for failing to provide individualized, case-specific justification for allowing each prospective juror in a criminal case to choose whether to be "questioned in open court or in camera " regarding their answers to the questionnaire they were given. Id. at 1048-49 . And more recently, in Jury Questionnaires , 37 A.3d 879 , where the Washington Post challenged a D.C. Superior Court judge's ruling denying the newspaper access to the completed jury questionnaires of prospective jurors in a high profile murder case, we considered the questionnaires to be part of the jury selection process and thus subject to the presumption of access (and to the Waller test) under the Supreme Court public trial cases. Id. at 885-87 . Much like in the present case, the trial court in Jury Questionnaires expressed generalized concerns about the candor of prospective jurors in justifying the decision to deny access to the questionnaires. Because the trial court did not "articulate specific protectible privacy interests" or "consider alternatives to complete closure to protect those interests," we held that this "blanket closure" ran afoul of the constitutional right of access to the trial-a right the press enjoys as a surrogate for the public. Id. at 882, 887-89 .

As mentioned above, the obstructed-view cases the majority cites as examples of "alternatives to closing" likewise expose the flaw in the notion that these purported "alternatives"-and by analogy, the see-but-not-hear procedure in this case-are not impermissible closures if they fall short of the requirements the Supreme Court has repeatedly stated must be met before a trial court restricts the public's access to a criminal trial. Every case the majority cites is a case in which the approved-of alternative was tailored to the case based on a finding of compelling need grounded in the specific circumstances, as opposed to the husher procedure in this case, which was employed as a matter of longstanding practice based on generalized concerns and without individualized findings about a case-specific privacy interest at stake.

In Pearson v. James , 105 F.3d 828 (2d Cir. 1997), for example, the Second Circuit granted habeas relief where the trial court closed the courtroom during an undercover officer's testimony because, even though the trial court had met three of the four factors of Waller , 6 it failed to consider reasonable alternatives to the proceeding, the third requirement of Waller . 7

*251 Id. at 830-31 . When the court stated that placing a screen between the public and the testifying officer would have been an alternative to closure, id. at 830 , that meant it would have been a reasonable alternative after the court had satisfactorily established, with individualized findings supported by the record, that there was an overriding interest warranting some type of closure. Each of the other obstructed-view cases on which the majority relies holds exactly that: that the use of a screen to prevent the public from seeing a particular witness was not a violation of the right to a public trial because the court employing this method satisfactorily jumped through the Waller hoops. 8 These courts' exacting adherence to Waller 's stringent test underscores the distinction between their particularized scrutiny and the routine employment of the husher in Superior Court voir dire proceedings and shows why the jury selection procedure in this case violated Mr. Blades's public trial right. Whether these sorts of restrictions on jurors' ability to see or hear certain court proceedings are called closures or alternatives, they only become reasonable-and constitutional-by surviving the test designed to confine them to the most exceptional circumstances.

Given the strong presumption of openness, the majority's effort to distinguish cases that could more naturally be cited in support of Mr. Blades's argument in this case is perplexing. And yet my colleagues say that Cable News does not support reversal here because it held the public trial right was violated when the questioning of individual jurors was conducted " in camera ," a phrase that was left unexplained in the decision but that my colleagues are comfortable concluding does not encompass private or inaudible proceedings at the bench. But see, e.g. , Bennett v. United States , 797 A.2d 1251 , 1254 (D.C. 2002) (stating that "the prosecutor provided the requested material to the trial judge in camera at an ex parte bench conference");

*252 United States v. Minsky , 963 F.2d 870 , 874 (6th Cir. 1992) ("The court stated that the ex parte bench conference was part of its in camera review of the FBI 302s for the express purpose of ruling on the defense motions under both the Jencks Act and Brady. ")

The majority similarly distinguishes Jury Questionnaires on remarkably narrow grounds. According to my colleagues, for example, Jury Questionnaires does not support reversal in this case (1) because it did not address the specific issue in the present case-namely, "whether the press had a right to listen contemporaneously to oral voir-dire questioning of individual prospective jurors," ante at 241 n.10, and (2) because it involved access only to the questionnaires completed by empaneled jurors. These are both true, but are not good reasons to treat the use of the husher differently. My colleagues in the majority also recognize the potential relevance of the court's statement in Jury Questionnaires that the trial court "had no problem keeping oral voir dire open to the public," which seems to suggest that voir dire was conducted openly in that case and that the court recognized the importance of that openness. But the majority downplays the significance of this statement because the court did not specifically say whether a husher was used in Jury Questionnaires . Aside from the unlikelihood that the Washington Post would have failed to challenge any procedure that prevented its reporters from hearing the voir dire in that case, if the trial court had utilized a husher at voir dire, the judge probably would not have said at a hearing that "the press had heard all the individual questioning of prospective jurors," Jury Questionnaires , 37 A.3d at 884 (quotation in the appellate opinion and emphasis added). See also id. at 888 (referring, in the course of a discussion about jury candor, to "keeping oral voir dire open to the public").

In the end, the idea that the right to public trial is not violated by a practice that keeps the public from hearing what is going on during jury selection cannot be squared with the values the Supreme Court has said the public trial right serves. Ensuring "that judge and prosecutor carry out their duties responsibly," "encouraging witnesses to come forward," and "discouraging perjury," Waller , 467 U.S. at 46 , 104 S.Ct. 2210 ; see also Jury Questionnaire , are best accomplished by people who can hear the proceedings. We should follow the Sixth Circuit's lead in holding that making voir dire inaudible to spectators can violate that right-and that it will violate that right if it is not supported by case-specific findings that there is a compelling interest justifying some form of closure and that the chosen method of protecting that interest is no broader than necessary to do so. See Memphis Publishing , 887 F.2d at 648 -49 . "[T]he sensible course is for the trial judge to recognize that open trials are strongly favored, to require persuasive evidence of serious risk to an important interest in ordering any closure, and to realize that the more extensive is the closure requested, the greater must be the gravity of the required interest and the likelihood of risk to that interest." Ayala v. Speckard , 131 F.3d 62 , 70 (2d Cir. 1997) (en banc) ( Ayala II ).

Finally, although the majority does not ultimately decide the matter, the parties in this case have thoroughly briefed and argued the question whether our decision in Copeland v. United States , 111 A.3d 627 (D.C. 2015), precludes reversal in this case because we are bound by its statement that it was "not persuaded by the argument that [the defendant's] right to public trial was violated by the procedures used during the selection of his jury"-procedures the court described as "conducting a limited amount of individual voir dire at the bench with a 'husher' on." Id. at 633 .

*253 In the government's view, this constitutes a holding by this court that the practice at issue in this case does not violate the right to a public trial. Mr. Blades counters that the court's statement in Copeland was unnecessary to its decision that trial counsel was ineffective by failing to inform Mr. Copeland of his right to be present at the bench during the voir dire of jurors. 9 Mr. Blades also argues that to the extent Copeland did resolve the question in this case, that holding was inconsistent with Jury Questionnaires and we are required to follow the earlier decision. See Thomas v. United States , 731 A.2d 415 , 420 n.6 (D.C. 1999).

While Copeland is not easily reconciled with Jury Questionnaires or, for that matter, with the whole body of Supreme Court case law on the right to public trial, what qualifies as a holding in a case and whether one of our decisions fails to adhere to prior authority are difficult questions. The court's decision today exacerbates and prolongs the uncertainty caused by this tension between Jury Questionnaire 's holding (that the preclusion of media access to jury questionnaires in a criminal trial based on generalized concerns alone violates the public's right to be present at jury selection) and the language in Copeland , now reinforced by this court (that a practice that precludes all members of the public from hearing voir dire in a criminal case does not violate a defendant's right to public trial, even when the trial court imposes the practice without making individualized findings justifying the closure and without satisfying the other requirements of Waller ). The majority does not give us a unifying principle for understanding what types of closures are subject to the Waller test and what types are not. The significance of the public trial right and the importance of maintaining uniformity among our cases and consistency with Supreme Court precedent may therefore make this case a good candidate for en banc consideration.

Although I would reverse the judgment of the Superior Court in this case based on the violation of the public trial right, I also take issue with some aspects of my colleagues' resolution of the remaining issues in this case-particularly their harmless error analysis.

In my concurring colleague's view, "[t]he hardest issue in this appeal" is the trial court's error in overruling Mr. Blades's objection to the prosecutor's statements in closing argument. Ante at 247. Judge Farrell and I are in agreement that "the prosecutor's repeated ... suggestion in rebuttal argument that appellant had forfeited self-defense by the act of bringing 'the gun to the fistfight' or 'to the neighborhood' " was "erroneous." Ante at 247-48. Unlike my colleagues, however, I am not confident the misstatements were harmless, particularly given that, as Judge Farrell also notes in his concurrence, the trial court's provocation instruction "gave arguable support" to the prosecutor's misleading arguments by asserting that a defendant "cannot claim self-defense" if he puts himself in a position to provoke trouble. Ante at 247-48.

My colleagues nevertheless think the prosecutor's misstatements could not have caused the jury to convict on improper grounds because the government's evidence, particularly the physical evidence, all but compelled the jury to reject Mr. Blades's self-defense claim by demonstrating his "excessive use of force" when he *254 fired off nine shots while walking toward Mr. Campbell. Judge Farrell quotes the prosecutor's dramatic reenactment, in closing argument, of Mr. Blades "walking forward, shooting as he goes." Ante at 248. The majority likewise highlights the government's evidence "that appellant was shooting at and pursuing targets who were running away from him[.]" 10 Ante at 244. And to be sure, this is what the government's firearm examination expert surmised when he looked at a photograph showing where the expended shell casings landed. The expert's testimony is good evidence as far as it goes. The problem with making it the centerpiece of the majority's determination of harmless error, however, is that the overall evidence of the walking-shooter scenario is weak, there is evidence of a self-defense counter-narrative that the jury could have credited, and the record does not feature the kind of evidentiary lopsidedness that might assure us that the prosecutor's misstatements could not have played a role in the jury's verdict because no juror could have maintained any doubt that the government had disproved self-defense beyond a reasonable doubt.

There are several reasons the ballistics evidence is inadequate footing for my colleagues' conclusion that the jurors could not possibly have been distracted by the prosecutor's misstatements and followed their false lead to conviction. As an initial matter, the expert's pattern-recognition testimony was the only evidence of a walking-shooter scenario, and none of the actual witnesses to the shooting described anything like this. Contrary to Judge Farrell's view that "the combined ballistics evidence and testimony of Campbell and Paige strongly favored this scenario," ante at 248, Mr. Campbell testified that he saw Mr. Blades shooting "with the car door open," while Mr. Paige said Mr. Blades shot in a "standing up-straight position" "in the middle of the street." Neither eyewitness described Mr. Blades as walking while shooting, and their accounts contradicted each other. Some of these discrepancies were minor and some were more consequential, 11 but it is fair to say the walking-shooter narrative that was so critical to the government's effort to disprove self-defense was not so airtight as to warrant my colleagues' faith in the harmlessness of the prosecutor's misstatements.

Apart from its lack of support from the other evidence, the government's expert's testimony also had its own vulnerabilities. The expert acknowledged that his moving-shooter *255 conclusion was "based on generalities" and that he formed his opinion just that morning, shortly before taking the stand, after being shown for the very first time the photograph depicting the shell casings' location. He acknowledged that he had not actually visited the scene and was not aware of the slope of the street where the casings were found. The expert admitted that "[m]any, many variables" can affect where casings land and that they can bounce up to 21 feet and they can roll. Though he had not consulted any studies or reference books before reaching his conclusion that the shooter in this case was moving while shooting, during his testimony he recalled one comprehensive study that concluded there could be as much as 75 percent error in shell casing pattern recognition. The question is not whether the expert's walking-shooter opinion was relevant, or whether it supported the government's theory of the case. It was and it did. The question is whether it was so strong as to overwhelm the potential appeal, for any juror, of Mr. Blades's self-defense claim, and thus renders harmless the prosecutor's erroneous suggestion that Mr. Blades had relinquished his right to claim self-defense by bringing a gun to the neighborhood. It was not.

My colleagues' certainty about the power of the government's physical evidence is even less justified when one considers that the defense had a firearm expert too-one who, unlike the government's expert, had visited the scene, taken measurements at the scene, and reviewed the police reports before forming his opinion that the pattern of shell casings supported a scenario involving a stationary shooter. Mr. Blades followed his expert's testimony with his own account of the incident and his take on the multiple gunshots and on Mr. Campbell's proximity: he testified that he committed the shooting in self-defense and that he fired his gun blindly after being struck in the head and upon threat of being stabbed, just trying to get the men away from him. Mr. Blades was undoubtedly an interested witness with an incentive to avoid being convicted, and the prosecutor's cross-examination gave the jury additional reasons to question his testimony as well. For its part, the government had its own problems with the credibility of its main witnesses, and acknowledged in its brief that the jury may have had reasons to doubt some of the testimony of Mr. Campbell and Mr. Paige. Most significantly, the jurors did not believe that Mr. Blades had tried to run Mr. Campbell over with his car, and their acquittal of Mr. Blades on the vehicular assault counts suggests they would have viewed the entirety of Mr. Campbell's testimony with some degree of skepticism. All of this is to say that the record in this case is not conducive to a finding of harmless error that is dependent upon the relative strength of the government's case-particularly where the error in question went to the central question whether Mr. Blades shot Mr. Campbell in self-defense and where it was the government's burden to disprove self-defense beyond a reasonable doubt. 12

*256 As to the admission of the mugshots in this case, although the majority states that it is not "definitively deciding" that the government had "no reason to show identification" at trial, it also concludes-and in his concurrence, Judge Farrell agrees-that Mr. Blades "has the better of the argument about whether there was a demonstrable need to introduce the photo arrays." Ante at 241. We are in apparent agreement, then, that the government had no reason to introduce Mr. Blades's actual mugshot in a case where defense counsel admitted in opening statement that Mr. Blades was the shooter (in self-defense) and where the government presented "more than sufficient testimony regarding positive photographic identification" from its two eyewitnesses, Mr. Campbell and Mr. Paige. See ( Kirk ) Williams v. United States , 382 A.2d 1 , 5 (D.C. 1978). As demonstrable need is one of three requirements the government must satisfy to justify admission of a mugshot-like photo of a defendant in a criminal trial, Bishop v. United States , 983 A.2d 1029 , 1034 (D.C. 2009), the admission of the array in this case was error.

My colleagues nonetheless go on to conclude that the introduction of the mugshot was harmless beyond a reasonable doubt. Judge Farrell's focus, in concurrence, is again upon "the strong inconsistency" between Mr. Blades's defense theory and the physical evidence. Ante at 247. The government likewise argues in its brief that it was "the obvious inconsistency of appellant's testimony with the physical evidence, and not any vague inference of prior arrest supposedly implied by the arrays, that produced the verdicts." But Mr. Blades's testimony was not inconsistent with the physical evidence-it was inconsistent with the government's expert's opinion of the physical evidence. The government concedes that Mr. Blades's testimony was consistent with the defense expert's opinion of the physical evidence, but dismisses the expert's testimony because "his analysis suffered from two flaws"-namely, that he used a different gun at a different location when conducting a test for comparison purposes and that he made a "self-contradictory claim" about the randomness of how expended casings land and the ability to see a pattern. These "flaws" were grounds for impeachment, not the absolute repudiation of Mr. Blades's theory of self-defense. As in ( Kirk ) Williams , Mr. Blades's account of the shooting "was not a wholly unlikely possibility," see 382 A.2d at 7 , and the jury could well have believed it. 13 I see no grounds in this record for rejecting the possibility that the combination *257 of Mr. Blades's testimony and the defense expert's analysis could have created a doubt in the mind of at least one juror, particularly where it was the government's burden to disprove self-defense beyond a reasonable doubt.

As for the harmlessness analysis in the majority opinion, while I agree with my colleagues that the "unnecessary publication to District of Columbia juries of photo arrays of African-American men also poses a danger of fostering unconscious bias," ante at 242 n.11, the majority's grounds for nonetheless finding the admission of the mugshot of the African-American defendant in this case harmless beyond a reasonable doubt do not assure me that the jury in this case did not "consider [Mr. Blades's] bad character in deciding whether to convict [him] of the charged crime." Bishop v. United States , 983 A.2d at 1034, 1038 (citation omitted). At the outset, that the men photographed were not in prison garb, that someone removed the mugshot serial numbers, and that the detective who created the arrays did not call them "mugshots" mean little when, as the majority acknowledges, the "unsmiling expressions" on the faces of the men pictured in the photo array still make the photos look like mugshots. See ante at 242. And the array still bore indications that it had been altered, in the rough edge that was left when "mugshot.com" was removed and in the mark at the top hiding the reference to serial numbers. In any event, as this court noted in Williams , the exceptional circumstances in which cleaned-up mugshots might be deemed admissible are limited to those circumstances in which the government proves a demonstrable need. See 382 A.2d at 5 .

I also cannot agree that the introduction of the mugshot was harmless beyond a reasonable doubt on the ground that the jury had already learned from the evidence at trial that Mr. Blades actually broke the law by bringing an unregistered firearm and ammunition into the district. According to Judge Thompson, the jury's acquittal on the car-related assault showed that the jury could weigh the evidence unaffected by knowledge that Mr. Blades was a lawbreaker. Ante at 242-43. On this point I align with Judge Farrell, who states in his concurrence that we should not "draw speculative inferences" from these acquittals. See ante at 247 n.1. But in any event, given how many states have far less restrictive gun laws than the District's and how some have no registration requirements at all, I am not as confident as Judge Thompson that Mr. Blades's possession of a gun contrary to the laws of the District "went beyond what an arrest mugshot implies." Ante at 243.

In Williams , this court held that the use of mug shots was not harmless where the complainant's account of a violent attack was not corroborated by physical evidence. 382 A.2d at 7 . Here, by contrast, my colleagues are confident, albeit for different reasons, that the inference of prior arrest from the admission of the mugshot did not cause the jury to reject Mr. Blades's self-defense claim. But this is a case where the parties presented competing expert testimony supporting their respective accounts of how things happened, where there were strengths and weaknesses to both versions but no glaring lopsidedness or foregone conclusions, where the defendant had a motive to lie but presented a viable story of self-defense consistent with the defense expert's testimony, where the jury discredited the complainant's testimony about a separate incident after the shooting, and where the government's witnesses were neither united nor concretely helpful in corroborating the government's ballistics expert. Under these circumstances, the jury "might well have been influenced because of the improper, indirect proof of *258 [Mr. Blades's] criminal past." Bishop , 983 A.2d at 1039 .

This case raises an important question about the constitutionality of a practice frequently used in our trial court. Ultimately, I believe my colleagues' decision to uphold a practice that allowed the public to see, but not hear, the jury selection in Mr. Blades's criminal trial runs afoul of the applicable case law on the right to public trial. I also cannot agree that the prosecutor's repeated erroneous statements in closing argument and the unnecessary admission of the mugshots were harmless errors in the circumstances of this case. For these reasons, I respectfully dissent.

Unlike Judge Thompson, however, I do not believe we should draw speculative inferences, ante at [243], from the jury's acquittal of appellant of the assault charge based on his use of a car.

Other than the risk it provided of buttressing the prosecutor's mistaken argument, the provocation instruction itself was unobjectionable in a case where the jury could find, as the most reasonable hypothesis, that rather than withdraw from the scene appellant unnecessarily increased the danger of harm by going to his car and retrieving the gun or ammunition.

Similar to my earlier caveat, however, I cannot join Judge Thompson's speculation, ante at 244-45, that the jury's conviction of AWIKWA (firearm) rather than ADW was further indication that it was not prejudiced by the improper closing argument.

United States v. Gonzalez-Lopez , 548 U.S. 140 , 148-49, 126 S.Ct. 2557 , 165 L.Ed.2d 409 (2006) (describing structural errors as those that "defy analysis by 'harmless-error' standards because they affect the framework within which the trial proceeds," and including within the category of structural errors "the denial of the right to public trial" (internal brackets and quotation marks omitted) ); see also Littlejohn v. United States , 73 A.3d 1034 , 1042 (D.C. 2013) (stating that "if a defendant's right to a public trial has been violated-i.e., if the four Waller criteria were not met-he need not show specific prejudice resulting from that violation." (citing Waller v. Georgia , 467 U.S. 39 , 49, 104 S.Ct. 2210 , 81 L.Ed.2d 31 (1984) ).

The trial court itself acknowledged that there was "a lot of strong language out there" in case law addressing the right to public trial and that it was not "by any stretch obvious" or "abundantly clear" that the approach was proper.

The court did this by identifying an overriding interest in preventing exposure of the officer's identity and making specific and adequately supported findings that the closure was no more extensive than required to protect the interest asserted. Waller , 467 U.S. at 48 , 104 S.Ct. 2210 .

On rehearing en banc, the Second Circuit-granting review in Pearson and two other cases raising public trial issues on habeas-changed course and affirmed the district court's denial of Mr. Pearson's habeas petition after reconsidering and backing away from the holding of a prior Second Circuit decision, Ayala v. Speckard , 89 F.3d 91 (2d Cir. 1996) ( Ayala I ), that a trial court was required to sua sponte consider alternatives to closing the court during the testimony of one witness. Ayala v. Speckard , 131 F.3d 62 (2d Cir. 1997) (en banc) ( Ayala II ). More than a decade later, however, the Supreme Court clarified that its own precedent-specifically Waller and Press-Enterprise Co. v. Superior Court of California , 464 U.S. 501 , 510, 104 S.Ct. 819 , 78 L.Ed.2d 629 (1984) -had all along established, contrary to the view of the en banc Second Circuit, "that trial courts are required to consider alternatives to closure even when they are not offered by the parties." Presley v. Georgia , 558 U.S. 209 , 214, 130 S.Ct. 721 , 175 L.Ed.2d 675 (2010).

See United States v. Lucas , 932 F.2d 1210 , 1217-18 (8th Cir. 1991) (upholding the use of a screen, which was the government's initial request, where the trial court made specific individualized findings that an officer's life would be in danger if she testified publicly, concluded that the government had established two overriding interests likely to be prejudiced if the officer's identity was not concealed, and settled upon the use of a screen after considering complete closure of the courtroom and the use of a disguise); State v. Schultzen , 522 N.W.2d 833 , 836 (Iowa 1994) (rejecting criminal defendant's contention that ordering his family members to sit behind a screen during parts of the child victim's testimony violated his right to public trial where the trial court "met the requirements of Waller ," including making adequate findings to support the screening); Rodriguez v. Miller , 537 F.3d 102 , 110 (2d Cir. 2008) (holding that the proposed use of a screen during an undercover officer's testimony did not violate clearly established federal law where the state courts properly applied the Waller test and satisfied its factors with findings regarding two overriding interests that were firmly anchored in the record).

In support of this argument, Mr. Blades highlights the Copeland court's statement that "we base our decision solely on appellant's failure to satisfy the prejudice prong in Strickland [.]" Id. at 631 n.5.

The majority's harm analysis focuses on what "the government's evidence suggested" and what the jury had "ample basis" for concluding. Ante at 244-45. The test for harm is not whether the government presented sufficient evidence of the offense. At the very least, if we apply a Kotteakos standard, we must be able to conclude "with fair assurance" that the result of the trial was not substantially swayed by the error. Kotteakos v. United States , 328 U.S. 750 , 765, 66 S.Ct. 1239 , 90 L.Ed. 1557 (1946).

Mr. Paige testified that Mr. Campbell threw the first punch, but Mr. Campbell testified that he only hit Mr. Blades after Mr. Blades had already punched him in the face. As to the fight on the street, Mr. Campbell testified that he fought with Mr. Blades only on the sidewalk area and only until Ms. Mitchell chased him across the street. By contrast, Mr. Paige testified that Ms. Mitchell hit Mr. Campbell with a shoe until Mr. Paige pulled her off Mr. Campbell. Mr. Paige then explained that Mr. Blades and Mr. Campbell "kept wrestling and fighting," "steady punching each other" over the back of a trunk until they both "fell over to the middle of the street." And as to events earlier in the evening, Mr. Paige testified that the group consisted of "about five or six of us" while Mr. Campbell testified that there were "[a]bout nine of us." Mr. Paige testified that he and Mr. Campbell first saw Ms. Mitchell inside the Look Lounge, when she greeted and hugged them both, but Mr. Campbell testified that he did not see Ms. Mitchell in the club at all.

In a portion of the opinion Judge Farrell explicitly does not join, Judge Thompson contends that the fact that the jury convicted Mr. Blades of assault with intent to kill while armed instead of the lesser assault with a deadly weapon means the jury must have rejected the theory underlying Mr. Blades's defense of self-defense because it found that he intended to kill Mr. Campbell. Ante at 248 n.3. I agree with Judge Farrell. "[E]ven an intentional killing, if it comports with legally accepted notions of self-defense, is not malicious; it is excused and accordingly no crime at all." Comber v. United States , 584 A.2d 26 , 41 (D.C. 1990) (en banc) (citation omitted). The jury in this case could have believed that when Mr. Blades fired the gun, he was both defending himself against the threat of stabbing and intending to kill.

The government's own confidence that the admission of the mugshots was harmless error was substantially based upon an inadvertent but significant factual error that the government conceded at oral argument. Specifically, the government's assertions in its brief about the "obvious inconsistency" between the shell-casing evidence and Mr. Blades's account of the incident relied in part upon the force of the evidence that "[t]he expended casings were found lying almost in a straight line, over a distance of nearly 60 feet on 20th Street." The government subsequently repeated that evidence and emphasized its strength, asserting that Mr. Blades's expert "simply failed to give the jury any reason to accept the self-evident improbability that nine shell casings, fired from the same gun at the same time, would conveniently land in a nearly 60-foot line down the hill." But the government's brief was mistaken, and as the prosecutor noted at oral argument, the expended casings were arranged within a space that was closer to 20 feet in distance than 60 feet. The government's indication that "the expended casings were found lying almost in a straight line" is also different from the photograph of the casings' locations, which shows five casings in a line, two several feet off to the left of the line and two several feet off to the right.