dissenting.
I agree with my colleagues that Matthew Wrinkles’s trial attorneys were deficient in failing to object to the trial court’s insistence on the use. of restraints absent judicial findings that Wrinkles presented a security threat or otherwise required physical restraints. I cannot agree, however, that Wrinkles was not prejudiced by counsels’ error. The natural reading of the opinion of the Indiana Supreme Court in Wrinkles II is that several jurors were aware of the stun belt during the trial. In light of that finding, Wrinkles has shown both the inherent prejudice that accompanies visible restraints and other detriments specific to his case. Only through a tortured interpretation of the Indiana Supreme Court’s opinion, with which not even the respondent agrees, does the majority conclude that Wrinkles was not prejudiced by his attorneys’ error. I would not wager a man’s life on the correctness of the majority’s grammatical parsing, and therefore I respectfully dissent.
During the state postconviction proceedings, Wrinkles submitted affidavits from two jurors who attested that, during the trial, they were “aware” that Wrinkles was wearing a shock belt. One juror “believe[d]” that a bailiff told the jurors of the belt and the other was unsure how he became aware but “believe[d]” the judge told the jurors. A third juror (juror Kraft) attested that she was aware of the belt and even saw it on Wrinkles during the trial; she described the belt as looking like a cummerbund. Kraft also stated that she may have become aware of the belt after the trial from a newspaper article. Despite ample time to investigate and prepare counter-affidavits, the state submitted only one, from one of the three bailiffs who worked at Wrinkles’s trial. The bailiff attested that he never communicated to the jurors that Wrinkles was wearing a stun belt. Nothing, however, contradicted the three jurors’ testimony that they knew about the belt; only how each learned of it remained open to question. Despite the testimony by all three jurors that they knew about the stun belt, and with nothing in the record to the contrary, the postcon-viction court found that the jurors were not aware of the stun belt. The court reasoned that the affidavits were not credible because of inconsistencies and because the jurors were not subject to cross-examination.1
*824Wrinkles vigorously contested the court’s finding. He moved the postconviction court to reopen the evidence and to correct error, and supplemented the record with four additional juror affidavits. Kraft, one of the three original affiants, clarified that she had “no doubt” that during the trial she knew the belt she saw on Wrinkles was a stun belt, although she was still unsure as to how she came by that knowledge. Three additional jurors attested that they were aware of the belt during the trial, although none could recall the source of that information. Two of them supplied the additional detail that they understood that the belt could be activated remotely by a deputy. Finally, in a supplemental filing, one additional juror attested that she saw the stun belt during the trial. The postconviction court declined to revisit the issue and denied Wrinkles’s motions.
Wrinkles argued on appeal to the Indiana Supreme Court that the finding that the jurors were unaware of the stun belt was erroneous. In its opinion, the Indiana Supreme Court implicitly accepted that argument by making statements inconsistent with the postconviction court’s finding. If the jurors knew about the stun belt, we must presume that Wrinkles was prejudiced, and so the interpretation of the supreme court’s statements on this subject is the linchpin of this case. ■
In its first reference to the issue, the Indiana Supreme Court stated: “Wrinkles asserts that utilization of the stun belt, which was conspicuous to at least seven jurors, undermined his presumption of innocence and made him appear dangerous and uncontrollable in front of the jurors who would help decide whether he would live or die.” Wrinkles v. State, 749 N.E.2d 1179, 1192 (Ind.2001) (Wrinkles II). The statement is ambiguous at best, and it could be interpreted as the court’s recapitulation of one of Wrinkles’s arguments. However, Wrinkles’s proffered interpretation, that the supreme court itself believed that the stun belt was conspicuous, is also plausible because the court did not specifically ascribe that assertion to Wrinkles.2 The sentence reads more naturally as an acknowledgment by the court that seven jurors were aware of the belt.
The court’s second mention of the critical factual issue is not ambiguous. It first states that Wrinkles’s counsel, lacking the guidance of the later decision to ban stun belts outright, reasonably chose between shackles and stun belt based on the effect each might have on the jurors. Id. at 1195. The court continues: “Counsel believed that the chance of the jury seeing the shackles was fairly high. On the other hand, counsel opted for the stun belt because they thought that jurors would not be able to see it. Obviously, they were later proven wrong. However, at the time the decision was made, it was a prudent one.” Id. (emphasis added). Given the preceding sentence, the. italicized sentence must be read as an acknowledgment by the supreme court that, despite counsels’ pretrial predictions that the stun belt would go undetected, some jurors indeed were aware of it.
The majority, however, concludes that the attorneys were “later proven wrong” in their decision to evaluate “the choice of restraint through the lens of juror-prejudice alone.” Ante at 820.’ In order to arrive at this tortured result, my colleagues devote no less than ten pages to what they describe, fittingly,. as “Our *825Reading of ‘Obviously, they were later proven wrong.’ ” Ante at 818. The subheading is particularly apt because it is the majority’s interpretation and the majority alone. At oral argument, even the State of Indiana did not deny that the statement means that the supreme court believed that jurors were aware of the belt. Cf. McCaskill v. SCI Mgmt. Corp., 298 F.3d 677, 680 (7th Cir.2002) (“The verbal admission by SCI’s counsel at oral argument is a binding judicial admission, the same as any other formal concession made during the course of proceedings.”). Instead, the State characterized this finding as “an aside” the supreme court inserted while making the point that “the fact that later the jurors may have seen it ... doesn’t matter for counsel’s decision at the time he made it.” Indeed, nowhere in its brief does the State mention the passage of the Indiana Supreme Court’s opinion that divides the panel, a fact which makes the majority’s ten-page analysis of the disputed passage all the more extraordinary. See, e.g., Kochert v. Adagen Med. Int'l., Inc., 491 F.3d 674, 679 (7th Cir.2007) (Plaintiff “did not raise, much less develop this argument, and undeveloped arguments are waived.”).
Then my colleagues, who accuse Wrinkles of taking the statement out of context, proceed to rearrange the entire paragraph to reach the conclusion that the Indiana Supreme Court was explaining why “counsel’s decision to choose the stun belt was a ‘prudent one’ even though the attorney’s were ‘later proven wrong’ to examine their choice solely based on ‘the effect of the jurors.’ ” Ante at 819. Not only does this interpretation strain common sense, it is inconsistent with the supreme court’s analysis of counsels’ decision. As the majority concedes, the Indiana Supreme Court erroneously concluded that counsel were not deficient for failing to object to the stun belt because in light of the trial court’s supposed policy of requiring restraints, “an objection to wearing restraints would not have been sustained by the trial judge even if made.” Wrinkles II, 749 N.E.2d at 1195. Thus, the Indiana Supreme Court excused counsels’ failure to know the law with the illogical reasoning that the trial judge’s illegal policy obviated the need for an objection.
Given this view, it would make no sense for the Indiana Supreme Court to in the same breath conclude that counsel were “later proven wrong” to have evaluated the situation solely through the lens of Wrinkles’s appearance before the jury. Quoting a piece of the disputed sentence and then adding on its own “clarification” the majority comes up with the following conclusion: “Thus, ‘[ojbviously, they were later proven wrong’ to have evaluated the choice of restraint through the lens of juror-prejudice alone. Nonetheless, because the attorneys could not be faulted for failing to predict the form of prejudice announced in Wrinkles II,’ at the time the decision was made, it was a prudent one.’ ” Ante at 820. But being “proven wrong” about the possibility that the defendant would be prejudiced by the jury seeing him restrained is a nearly nonsensical concept. After all, the fact that the Indiana Supreme Court emphasized the effect of the restraint on the defendant (a form of prejudice long-recognized, not, as both the Indiana Supreme Court and my colleagues seem to believe, a novel proposition) as opposed to the prejudice occasioned by the jury seeing the belt does not mean that prejudice from a visible restraint is no longer a legitimate concern of counsel— i.e., that counsel would be “wrong,” as the majority suggests, for considering that form of prejudice. That borders on the absurd.
The majority claims its strained interpretation is the more plausible one in light *826of first, Indiana procedural law, and second, the Indiana Supreme Court’s subsequent decision in Stephenson. As for state procedural law, it seems odd indeed to be assessing whether the Indiana Supreme Court breached its own procedural rules in the context of a federal habeas claim, see, e.g., McCloud v. Deppisch, 409 F.3d 869, 875-76 (7th Cir.2005) (construction of state law irrelevant in habeas proceeding), where it goes without saying that state law cannot be the basis for relief. Conversely, if our decision turns on the answer to a question of state law, something is amiss. On the issue of the additional affidavits Wrinkles sought to have admitted in the postconviction court, the state conceded at oral argument that the Indiana Supreme Court reviewed those affidavits “in some fashion.” It seems both unusual and unhelpful to analyze whether we believe the Indiana Supreme Court is allowed under Indiana law to do precisely what the state concedes that it did. Moreover, I do not think the additional affidavits are determinative — the Indiana Supreme Court could easily have rejected the postconviction court’s conclusion based on the three affidavits indisputably in the record, making the majority’s preoccupation with whether the supreme court had the authority under state law to review the additional affidavits largely irrelevant.
The majority’s unlikely interpretation also ignores the procedural context. First, the factual issue was squarely presented to the Indiana Supreme Court. Second, the court suggested that it reviewed all seven juror affidavits despite the postconviction court’s refusal to do so; it referred to the belt being conspicuous to “at least seven” jurors. See Wrinkles II, 749 N.E.2d at 1192. Third, because no jurors testified at the postconviction hearing, the supreme court was at no disadvantage compared to the trial court in evaluating the credibility of the jurors’ statements, and so there is no reason to defer to the trial court’s interpretation. Finally, the supreme court made statements inconsistent with the factual finding of the postconviction court. The only logical conclusion is that the Indiana Supreme Court supplemented the record with its own finding that a number of jurors were aware that Wrinkles was restrained by a stun belt.
The majority protests that such a conclusion cannot be drawn because, “[h]ere, Wrinkles did not appeal the post-conviction court’s refusal to admit the additional affidavits into evidence,” ante at 817 (emphasis in original). Not only is this line of argument a red herring, it is simply untrue. In fact, Wrinkles did appeal the post-conviction court’s refusal to admit the additional affidavits. Specifically, Wrinkles’s brief on appeal to the Indiana Supreme Court states that “Wrinkles attempted to admit affidavits from four additional jurors who knew Wrinkles was restrained. The post-conviction court erroneously denied Wrinkles’ motion to supplement the record with these affidavits.” (Brief for Petitioner-Appellant at 19 n. 6, Wrinkles v. Indiana, No. 82C01-9407-CF-447.) The fact that, contrary to the majority’s repeated insistence otherwise, see ante at 817, Wrinkles placed the issue before the Indiana Supreme Court makes the court’s reference to the jurors’ awareness of the belt all the more straightforward.
Moreover, whether the Indiana Supreme Court did or did not formally admit the additional affidavits is in no way as determinative as my colleagues suggest. The three affidavits that were originally admitted all establish the jurors’ knowledge of the stun belt; and Wrinkles vigorously argued to the supreme court in a properly preserved appeal that the postconviction court’s contrary finding was clearly erroneous. Those three affidavits alone estab*827lished the jurors’ knowledge; it is only the source of that knowledge that was unclear. Neither the testimony of Wrinkles’s attorneys (regarding the belt’s visibility) nor the affidavit from one of the bailiffs3 (regarding his own communication with jurors and not addressing any other possible source) contradicts the jurors’ testimony that they knew about the belt. And the state did not produce a single counteraffi-davit from a juror who was not aware of the belt. As I have stated, to the extent there was no oral testimony by the jurors, and the issue was decided on the basis of the affidavits alone, there is no reason to defer to the postconviction court’s interpretation of the written testimony over the Indiana Supreme Court’s.
The Indiana Supreme Court’s discussion of Wrinkles II in Stephenson likewise does nothing to undercut the plain language of the disputed passage. In a confusing passage devoted to “explaining” why “obviously, they were later proven wrong” means wrong about some other issue than the one identified in the preceding sentence, the majority resorts to yet another Indiana Supreme Court case on stun belts. But it is unclear how Stephenson, which does indeed discuss Wrinkles II, sheds any light on whether the court believed the jurors knew about the stun belt in Wrinkles’s case. The majority first explains its reliance on Stephenson by analogizing it to a situation where “an ensuing state supreme court decision affects a disputed finding in a previous decision.” Ante at 821. But the examples cited provide no precedent for resorting to a later opinion to clarify a state court’s finding of fact in an earlier, unrelated proceeding. The supposedly “comparable” case relied on by the majority — Tibbs—demonstrates the point. There the United States Supreme Court referred to a later pronouncement by the Florida Supreme Court to resolve ambiguity in the earlier opinion in the same case; to be sure, subsequent pronouncements in the same case may illuminate the basis of a state court decision — as in Tibbs when a case is reversed and then retried and the court in the second appeal comments on its rationale in the first appeal. But using an unrelated subsequent state court opinion to interpret the meaning of case-specific language in a previous case strikes me as, if not unprecedented (certainly the majority points to no truly analogous scenario), highly unusual.
At all events, Stephenson is hardly so illuminating as the majority suggests. My colleagues point out that Stephenson “tracks” the reasoning from Wrinkles II by recreating the decision facing Wrinkles’s counsel at the time and their concern with whether the jury would see the belt as opposed to what effect the device would have on Wrinkles. But the fact that the Indiana Supreme Court repeated in Stephenson its mistake in Wrinkles — excusing counsels’ objectively deficient performance — sheds no light on the meaning of “obviously, they were later proven wrong.” Notably, the court in Stephenson concluded that the jurors in that case did see the defendant’s stun belt. But despite the discussion that “tracks” its reasoning in Wrinkles II, the court nowhere distinguishes Wrinkles II on the basis that the jurors in that case did not know about the *828stun belt. Stephenson’s explicit finding that the jurors in that case were aware of the stun belt still did not lead the court to conclude that counsel were deficient. See Stephenson, 864 N.E.2d at 1034-40. If anything, the supreme court’s repetition in Stephenson of its mistake regarding counsels’ effectiveness confirms that the court failed, in both Wrinkles II and Stephenson, to see that failure to object to restraints imposed without particularized justification amounts to objectively deficient representation. The court’s inability to appreciate this in Stephenson makes it all the more obvious that it recognized in Wrinkles II that the jurors were aware of the belt, but erroneously deemed that fact irrelevant in light of its misplaced focus on the fact that the trial court would have (incorrectly) overruled an objection to the stun belt. Compare Wrinkles II, 749 N.E.2d at 1195 and Stephenson, 864 N.E.2d at 1040-41.
Finally, the majority places great weight on the fact that the Indiana Supreme Court was considering the “choice of restraint facing Wrinkles’s attorneys at trial in light of the only theory of prejudice then available — the ‘effect on the jurors.’ ” Ante at 819. But neither the majority nor the Indiana Supreme Court is correct that the jury’s diminished impartiality was the only legally recognized form of prejudice at the time of Wrinkles’s trial. Both the United States Supreme Court and lower courts have long recognized that the harm flowing from visible restraints is threefold. In addition to the potential effect on the jury’s impartiality, the Supreme Court in 1970 recognized that restraints may interfere with the accused’s right to assist in his defense. See Illinois v. Allen, 397 U.S. 337, 344, 90 S.Ct. 1057, 25 L.Ed.2d 353 (1970) (“[Ojne of the defendant’s primary advantages of being present at the trial, his ability to communicate with his counsel, is greatly reduced when the defendant is in a condition of total physical restraint.”). In Deck v. Mo., 544 U.S. 622, 631, 125 S.Ct. 2007, 161 L.Ed.2d 953 (2005), the Supreme Court refers to the “ ‘ancient’ English rule” forbidding shackles and bonds absent a compelling justification — a rule formed in part out of concern that the restraints not interfere with a defendant’s presentation of his defense: “ ‘If felons come in judgment to answer, ... they shall be out of irons, and all manner of bonds, so that their pain shall not take away any manner of reason, nor them constrain to answer, but at their free will.’” See Deck, 544 U.S. at 626, 125 S.Ct. 2007 (quoting 3 E. Coke, Institutes of the Law of England). Lastly, Allen points out that the use of visible restraints is an affront to the dignity of judicial proceedings. Allen, 397 U.S. at 344, 90 S.Ct. 1057. Thus, at least from Allen onward, courts have recognized three distinct harms flowing from the use of visible restraints: (1) prejudice to the jury’s impartiality, (2) prejudice to the defendant’s ability to participate in his defense, and (3) damage to the dignity of the proceedings. See, e.g., Deck, 544 U.S. at 630-32, 125 S.Ct. 2007 (recognizing “three fundamental legal principles ” animating the “judicial hostility” towards visible restraints) (emphasis added); Harrell v. Israel, 672 F.2d 632, 635 & n. 3 (7th Cir.1982) (citing Allen to support three reasons given for the rule against physical restraints); Coates v. State, 487 N.E.2d 167, 169 (Ind.App.1985) (recognizing that restraints distract defendant’s “thought process”); People v. Brown, 45 Ill.App.3d 24, 3 Ill.Dec. 677, 358 N.E.2d 1362, 1363 (Ill.App.1977) (recognizing the prejudicial effect of restraints on jury’s feelings about defendant, the possibility that shackles would impair defendant’s ability to communicate with counsel, and the fact that shackles detract *829from “dignity and decorum of judicial process”).
Thus, there is no basis for the majority’s attempt to explain away the clear import of the phrase “[ojbviously, they were later proven wrong” by reasoning that the Indiana Supreme Court must have been excusing counsels’ failure to predict that the effect on a defendant would one day become a legal rationale forbidding the use of restraints at trial. That rationale was available to counsel at the time of Wrinkles’s trial. And the fact that counsel failed to “predict” what was in fact a long-settled rule of law is not remotely surprising: lead trial counsel testified at the post-conviction hearing that, “I did not know that there was a law about shackling.” If anything, the Indiana Supreme Court’s failure to acknowledge the longstanding recognition that restraints also prejudice the accused’s ability to participate in his defense simply reaffirms that the Indiana Supreme Court unreasonably applied clearly established law. Cf. Williams v. Taylor, 529 U.S. 362, 407, 120 S.Ct. 1495, 146 L.Ed.2d 389 (2000) (state court unreasonably applies Supreuso Court precedent when it “unreasonably refuses to extend” a legal principle “to a new context where it should apply”). As this court recently recognized, “law” refers not just to Supreme Court holdings, but “legal principles derived from the holdings in Supreme Court opinions.” Samuel v. Frank, 525 F.3d 566, 569 (7th Cir.2008). Thus, the fact that earlier law may not have addressed stun belts in particular as opposed to restraints generally is of no consequence.
It is well established that our obligation to defer to the factual findings of state courts extends to appellate courts. See Sumner v. Mata, 449 U.S. 539, 546-47, 101 S.Ct. 764, 66 L.Ed.2d 722 (1981); Miranda v. Leibach, 394 F.3d 984, 999 (7th Cir.2005); Mendiola v. Schomig, 224 F.3d 589, 592-93 (7th Cir.2000); Sprosty v. Buchler, 79 F.3d 635, 643 (7th Cir.1996); Holland v. McGinnis, 963 F.2d 1044, 1048 (7th Cir.1992). The finding that some jurors knew about the stun belt is amply supported by the record, and we must defer to it. See Sprosty, 79 F.3d at 643 (presumption of correctness applies to “implicit resolution of a factual dispute that can be fairly inferred from the state court record”). In contrast, the majority, like the state post-conviction court, can point to no record evidence supporting the conclusion that the jurors were not aware of the belt. Instead, it shores up its unlikely interpretation with a foray into state law positing that under Indiana law, the Indiana Supreme Court probably would not do what common sense suggests that it did when it commented that Wrinkles’s attorneys did not think jurors would see the stun belt but “they were later proven wrong.”
Of course, even finding that some jurors were aware of the stun belt did not lead the Indiana Supreme Court to conclude that Wrinkles’s attorneys were deficient for failing to object. That conclusion, rightly rejected by my colleagues, is largely based on the court’s determination that any objection would have been futile due to the trial court’s practice of routinely requiring restraints, as well as the fact that Wrinkles’s guilt was not in question. I will not dwell on the erroneousness of that analysis, but it is worth emphasizing that counsel’s obligation to object for the record was more, not less, urgent where the judge imposed an extralegal burden on Wrinkles without even attempting to justify it. Moreover, counsels’ failure to object— whatever the probable ruling — contributed in large part to the procedural hurdles Wrinkles now must clear in order to get relief, compounding their error. The majority appropriately concludes, therefore, that Wrinkles’s counsel were deficient for *830failing to object to the use of restraints without justification.
After parting ways with the Indiana Supreme Court on the first prong of Strickland, however, the majority then relies on its strange interpretation of that Court’s factual finding to conclude that Wrinkles was not prejudiced by his attorneys’ failure to object. “Without evidence that the jurors saw the stun belt, or that he was otherwise affected by the stun belt throughout trial, Wrinkles cannot demonstrate prejudice.”4 Ante at 823. Constrained by the Indiana Supreme Court’s finding that a number of jurors knew about the stun belt as well as the record evidence that the stun belt did indeed affect Wrinkles throughout trial, I would reach a different result.
It has long been established that visible restraints are so prejudicial that they are permissible only where a “special need” is present. Deck, 544 U.S. at 626, 125 S.Ct. 2007; see Holbrook v. Flynn, 475 U.S. 560, 568-69, 106 S.Ct. 1340, 89 L.Ed.2d 525 (1986); Allen, 397 U.S. at 344, 90 S.Ct. 1057. Routine use of restraints is prohibited; their employment must be preceded by a judicial finding that an essential state interest such as physical security, escape prevention, or courtroom decorum requires the use of restraints on a particular defendant. See Deck, 544 U.S. at 628, 125 S.Ct. 2007. So “inherently prejudicial” are visible restraints, Holbrook, 475 U.S. at 568, 106 S.Ct. 1340, that no “actual prejudice” need be demonstrated by a defendant asserting a deprivation of due process based on their unjustified use, Deck, 544 U.S. at 635, 125 S.Ct. 2007.
Wrinkles’s situation cannot be distinguished from the line of cases addressing visible restraints because the stun belt was “visible” in the only meaningful sense to any juror who was aware that he was restrained. See Roche v. Davis, 291 F.3d 473, 483 (7th Cir.2002). In this sense, the majority has seized upon yet another red herring, repeatedly referencing Wrinkles’s inability to establish that the belt was “visible” or “seen.” See, e.g., ante at 806-07, 810, 811, 812 n. 3, 815-16, 819. But of course the prejudice caused by restraints stems from the jurors’ knowledge of them, not their “visibility,” and the false distinction the majority weaves throughout its opinion is an unfortunate distraction.
Given the jurors’ awareness of the belt, every type of prejudice that the Supreme Court has associated with the use of restraints is implicated in this case. First, the message sent by restraints — that the judicial system itself already believes the defendant to be uncontrollably dangerous — undermines the presumption of innocence. Deck, 544 U.S. at 630, 125 S.Ct. 2007. Although Wrinkles admitted to killing the victims, the jury had to decide what level of homicide he had committed, and the burden was on the state to prove that he committed knowing murder. The jury’s decision on the level of homicide, as much as any other determination of guilt, could be tainted by their knowledge of the stun belt.
Second, physical restraints can interfere with the defendant’s ability to participate in his own defense. Id. at 631, 125 S.Ct. 2007. The most obvious example of that in this case is when Wrinkles’s attorney *831warned him not to “make any sudden moves” in response to Wrinkles asking where he should put his hands while testifying. This exchange came on the heels of an incident midway through trial when the belt began “buzzing.” The trial had to be halted as a result and the same belt was put back on Wrinkles after investigation revealed that the buzzing was caused by a low battery. No doubt this incident and the constant fear of an unannounced, unstoppable 50,000-volt shock impaired Wrinkles’s ability to participate in his defense. Not surprisingly, Wrinkles’s attorney described his client as “petrified” by the vibrations that accompanied the “buzzing” belt. Third, the dignity of judicial proceedings suffers when a participant is in restraints. Id. at 631-32, 125 S.Ct. 2007. That the trial had to be halted due to the belt’s “buzzing” provides a stark example of this last concern.
The prejudice inflicted by restraints is particularly dangerous in a case such as this, where any one factor could have been decisive for the jury in both the guilt and penalty phases given the nature and strength of Wrinkles’s defense. As appellate counsel explained, the use of the belt negated the entire theory of the defense: that Wrinkles was not a dangerous or violent person by nature but had “snapped” under extreme circumstances such as the bitter separation from his family, a recent involuntary commitment to a psychiatric facility, and a severe drug addiction. Indeed, sufficient evidence was adduced at trial to instruct the jury on voluntary manslaughter (which entails “sudden heat”) and reckless homicide in addition to knowing murder. Surely a presumption that Wrinkles was so dangerous as to require restraints would make a conviction for knowing murder more likely than it might have been based on the evidence alone. See Harrell, 672 F.2d at 637 (visible restraints “could instill in the jury a belief that the defendant is a dangerous individual who cannot be controlled, an idea that could be devastating to his defense.”) The prejudice was renewed during the sentencing phase, when the jurors who were aware of the stun belt had to decide whether to recommend death or imprisonment while believing Wrinkles was still violent and dangerous. Particularly where the mitigating factors far outnumbered the one aggravating factor allowing for the death penalty, see Roche, 291 F.3d at 484, the potential influence of the stun belt cannot be overstated. Accordingly, I must conclude that Wrinkles was prejudiced by the failure of his attorneys to object to the use of a stun belt.
Finally, even if I could accept my colleagues’ strained characterization of the Indiana Supreme Court’s statements on the visibility of the stun belt, I would find it difficult to accept their reflexive conclusion that Wrinkles was not prejudiced. The jurors’ awareness of the restraint aside, Wrinkles also argues that he could not fully and meaningfully participate in his trial while strapped to a torture device. This argument was a logical application of existing Supreme Court precedent on restraints, and competent counsel would surely have raised it in response to the trial court’s illegal “policy.” The majority dismissively concludes that Wrinkles did not present “evidence” that the stun belt “affected his abilities to participate in his own defense,” ante at 823. Not only did he present such evidence (see discussion ante at 830-31), in my view a court need not abandon its common sense when considering whether being forced to wear, with no justification, a device that delivers an unstoppable, 8-second, 50,000-volt shock might affect a defendant’s participation and demeanor, and, relatedly, the jury’s impressions of him. For this reason, and more importantly because of the *832prejudice stemming from the jurors’ awareness that Wrinkles was restrained with a stun belt, I respectfully dissent.
. The court also did not explain why it credited the bailiff's affidavit, which was also not subject to cross-examination. The affidavits are duly sworn under the penalty of perjury. They were prepared four years after the trial.
. As the majority points out, ante at 815 n. 4, the supreme court otherwise consistently used such language as “Wrinkles contends” or "He claims” immediately preceding his arguments.
. Notably, although the majority emphasizes the bailiff's affidavit, his sworn testimony that he was “sequestered with the jury for the entire duration of the trial’’ was later shown to be inaccurate. That bailiff eventually submitted a supplemental affidavit clarifying that he was temporarily absent from the trial and another bailiff took over his duties during that time. Nor was this bailiff the only one assigned to the trial; thus there is little to the postconviction court's suggestion that the bailiff’s affidavit contradicted Kenneth Ranes’s affidavit that he believed "the bailiff” told jurors about the stun belt.
. I have already commented that the majority’s interpretation of the Indiana Supreme Court's statements on this matter strains credulity. But, in the face of the juror affidavits admitted into the record, the majority's statement that Wrinkles is “without evidence” that jurors knew about (the more appropriate term than "saw”) the stun belt is wholly inaccurate.