Michigan Supreme Court Lansing, Michigan Syllabus Chief Justice: Justices: Stephen J. Markman Brian K. Zahra Bridget M. McCormack David F. Viviano Richard H. Bernstein Joan L. Larsen Kurtis T. Wilder This syllabus constitutes no part of the opinion of the Court but has been Reporter of Decisions: prepared by the Reporter of Decisions for the convenience of the reader. Kathryn L. Loomis PEOPLE v LYLES Docket No. 153185. Argued on application for leave to appeal March 9, 2017. Decided August 1, 2017. William Lyles, Jr., was convicted following a jury trial in the Wayne Circuit Court before James Lacey, J., of first-degree murder, MCL 750.316, for the 1983 stabbing death of Andrew “Melvin” Weathers. At the time of his death, Weathers lived in a house with several individuals, including Louise Kountz, whom defendant had previously dated, and Kountz’s two teenage daughters. Defendant had previously lived in the home for about four years but had moved out the summer before the murder. In December 1983, a perpetrator broke into the home through a basement window, disabled the electricity, placed the family dog in the basement freezer, and then proceeded upstairs to stab Weathers in his bedroom, using a knife from the home’s kitchen. While leaving the house to call for help, Kountz’s daughters saw a figure whom they believed to be defendant; the build of the figure matched that of defendant, and one daughter smelled a distinctive stale cigarette odor that she associated with defendant. Police obtained a warrant for defendant’s arrest but were unable to locate him until 2012, at which point defendant was arrested and charged with first-degree murder. At trial, the prosecution introduced evidence of defendant’s prior acts of domestic violence toward Kountz during their relationship, including the testimony of Kountz’s daughters, Weathers’ sister, another individual who had lived in the Kountz home, and three neighbors, two of whom had been Kountz’s long-time next-door neighbors. In response, defendant introduced evidence of his character for peacefulness, including the testimony of a close family friend who had moved to California to attend college three years before the murder but returned to the neighborhood during school breaks. After the court instructed the jury, defendant objected to the adequacy of the instructions, requesting that the trial court provide the model instruction explaining to the jury that defendant’s evidence of good character alone could create a reasonable doubt. The court did not provide the requested instruction and instead provided an instruction regarding testimony about witnesses’ truthfulness and testimony about defendant’s lack of good character. Defendant again objected, and the trial court noted the objection but did not alter the instructions. Defendant was convicted and sentenced to life in prison without the possibility of parole. Defendant appealed. In an unpublished per curiam opinion, issued July 22, 2014 (Docket No. 315323), the Court of Appeals, BECKERING, P.J., and HOEKSTRA and GLEICHER, JJ., reversed defendant’s conviction and remanded for a new trial on the ground that the trial court had failed to instruct the jury regarding defendant’s good-character evidence. The prosecution sought leave to appeal in the Supreme Court, and the Supreme Court ordered and heard oral argument on whether to grant the application or take other action. 497 Mich 960 (2015). On October 30, 2015, the Supreme Court issued an order remanding the case to the Court of Appeals to properly apply the harmless-error standard set forth in People v Lukity, 460 Mich 484 (1999), because the Court of Appeals had improperly relied on cases that did not apply the current standard. 498 Mich 908 (2015). On remand, the Court of Appeals again reversed defendant’s conviction in an unpublished per curiam opinion, issued December 22, 2015 (Docket No. 315323), concluding that the court’s instructional error was not harmless under the Lukity standard. The prosecution sought leave to appeal in the Supreme Court, and the Supreme Court ordered and heard oral argument on whether to grant the application or take other action. 500 Mich 875 (2016). In an opinion by Justice LARSEN, joined by Chief Justice MARKMAN and Justices ZAHRA and WILDER, the Supreme Court, in lieu of granting leave to appeal, held: Preserved, nonconstitutional errors are subject to harmless-error review. MCL 769.26 provides that no judgment or verdict shall be set aside or reversed or a new trial be granted by any court of this state in any criminal case, on the ground of misdirection of the jury, or the improper admission or rejection of evidence, or for error as to any matter of pleading or procedure, unless in the opinion of the court, after an examination of the entire cause, it shall affirmatively appear that the error complained of has resulted in a miscarriage of justice. A defendant carries the burden of showing that it is more probable than not that the error complained of was outcome-determinative. In making this determination, the reviewing court should focus on the nature of the error presented to it in light of the weight and strength of the untainted evidence. In this case, despite the dissent’s allusion to other nonpreserved errors, the only error presented to this Court was the trial court’s denial of defendant’s request for an instruction informing the jury that defendant’s good-character evidence alone could create a reasonable doubt. The Court of Appeals concluded that the trial court’s instructional error was not harmless because it “eviscerated the significance of defendant’s character evidence,” which was defendant’s “best defense” to the prosecution’s evidence of his past violence. But when considering whether the failure to give an instruction was harmless error, the question is whether the instruction would have made a difference in the outcome, and answering this question requires a court to consider not only the relationship between the instruction and the defendant’s defense strategy, but also whether that strategy would have been sufficient to create a reasonable doubt given the proofs as a whole. In this case, the prosecution presented evidence that the perpetrator was familiar with the house in which the murder occurred. And the prosecution presented evidence that defendant was identified the night of the murder; an arrest warrant was issued for defendant a few months after the murder, but defendant had fled the state. Additionally, numerous witnesses testified to defendant’s abuse of Kountz: Weathers’ sister saw defendant hit Kountz nearly every day; Kountz’s daughters witnessed physical abuse, and one witnessed sexual assault; a former housemate testified to defendant hitting Kountz; and several neighbors recounted repeated physical abuse. In his defense, defendant called three witnesses: his sister, a former girlfriend, and his lone character witness—a family friend who had grown up on the same block as defendant but who had moved to California for school three years before the murder. She occasionally returned to Michigan, including during the summers, where she would visit defendant’s family home daily but would only see defendant if he happened to be visiting the family home. She testified that defendant was a peaceful person who did not have a reputation for being abusive. She also testified that she saw defendant with Kountz about three or four times and did not observe any abuse. In reviewing the weight and strength of the evidence, the prosecution’s evidence, including evidence of defendant’s bad character offered by witnesses who had many chances to observe defendant in the early 1980s, far outweighed defendant’s good-character evidence, which was both substantively weaker and was offered by a witness who had little contact with defendant during the relevant period and who no longer lived in the same city. Defendant was not denied the opportunity to present his good-character evidence; he was only denied a proper instruction that good-character evidence, if believed by the jury, could alone produce a reasonable doubt. The Court of Appeals’ critical error was focusing on the importance of the good-character instruction to defendant’s defense strategy instead of evaluating the likelihood of defendant’s prevailing on that strategy. Given the evidence at trial and considering the only claim of error presented by defendant to the Supreme Court—the trial court’s failure to give one instruction on how the jury could consider the testimony of defendant’s lone good-character witness—defendant failed to show that it was more probable than not that the outcome would have been different if the jury had been properly instructed. Therefore, the Court of Appeals erred by holding that defendant had met the Lukity harmless-error standard. Court of Appeals’ judgment reversed; defendant’s conviction reinstated. Justice MCCORMACK, joined by Justices VIVIANO and BERNSTEIN, dissenting, would have upheld the Court of Appeals’ conclusion that the failure to instruct the jury regarding defendant’s good-character evidence was not harmless to the jurors’ verdict and therefore necessitated a new trial. Contrary to the majority’s characterization, the Court of Appeals panel conducted a thorough and complete harmless-error analysis under Lukity’s standard, and an examination of the entire cause confirmed that the panel’s conclusion was correct. The defendant’s trial was a pure contest of credibility and character. By the time of the trial in 2013, the original case file, including the reports, witness statements, and any other evidence it might have contained, had been completely lost, and neither Kountz nor many of the officers involved in the original investigation were alive or available to testify. Therefore, to reconstruct the events that occurred on the night of the murder, the prosecution heavily relied on the 30-year- old, unsubstantiated memories of Kountz’s then-teenaged daughters—with a majority of the testimony about that night (and defendant’s place within it) coming solely from the one daughter who admittedly hated defendant at the time of the murder and was responsible, 30 years later, for setting in motion the events that led to his trial. The prosecution also focused intently on proving, through testimony from these daughters and other family friends, that defendant had persistently and violently abused and harassed Kountz over the years leading up to the murder. Like the murder charge, there was no physical evidence or documentation to substantiate these decades-old, uncharged acts of domestic violence; furthermore, none of the prosecution’s witnesses could explain how or why defendant’s alleged treatment of Kountz might have led defendant to kill Weathers that night. Nonetheless, the prosecution turned defendant’s alleged abuse of Kountz into the overriding theme of his trial for murdering Weathers. To counter this character attack, defendant presented three witnesses, including a long-time school principal who had known defendant her entire life and who testified regarding defendant’s character for peacefulness. The jurors were tasked with determining defendant’s guilt on the basis of these proofs; the court, however, only provided the jurors with instructions regarding the prosecution’s proofs. The court failed to instruct the jurors that defendant’s proofs—his good-character evidence, the only evidence he presented—could, alone and as a matter of law, create a reasonable doubt and foreclose a finding of guilt. Accordingly, the jurors were given no legal guidance regarding whether and how they could consider defendant’s good-character evidence in reaching a verdict, let alone that this evidence could in itself determine that verdict. Instead, the jurors heard nothing from the court regarding defendant’s evidence and were instructed only on the ways in which the prosecution’s evidence could lead them to a conclusion of guilt beyond a reasonable doubt. Considering the nature of this error and the weight and strength of the evidence untainted by the error, it affirmatively appears that the court’s failure to instruct the jury regarding the potentially dispositive legal effect of defendant’s good-character evidence was not harmless to the jurors’ verdict. Namely, it affirmatively appears more probable than not that, had the jurors been properly instructed, at least one of those jurors would have recognized the doubt cast by defendant’s evidence as the “reasonable” sort countenanced by the law to foreclose a finding of guilt and would not have voted to convict defendant. Accordingly, and as the Court of Appeals panel thoroughly and correctly explained, a new trial is required under Lukity to ensure that any verdict on this murder charge is reliable under the law. Justice MCCORMACK would have denied leave and allowed a new trial to go forward. ©2017 State of Michigan Michigan Supreme Court Lansing, Michigan OPINION Chief Justice: Justices: Stephen J. Markman Brian K. Zahra Bridget M. McCormack David F. Viviano Richard H. Bernstein Joan L. Larsen Kurtis T. Wilder FILED August 1, 2017 STATE OF MICHIGAN SUPREME COURT PEOPLE OF THE STATE OF MICHIGAN, Plaintiff-Appellant, v No. 153185 WILLIAM LYLES, JR., Defendant-Appellee. BEFORE THE ENTIRE BENCH LARSEN, J. In his trial for first-degree murder, the trial court improperly denied defendant’s request for an instruction informing the jury that his evidence of good character could create a reasonable doubt. We now consider whether defendant has shown that it is more likely than not that this error was outcome-determinative. See People v Lukity, 460 Mich 484, 495-496; 596 NW2d 607 (1999). Defendant was permitted to introduce his good- character evidence; it was, however, minimal and strongly contradicted by the prosecution’s witnesses. Given this and the other evidence implicating defendant in the murder, we cannot conclude that the absence of the instruction, the only error alleged here, made the difference. Defendant has not shown that it is more likely than not that the outcome would have been different if the jury had been given this instruction. Accordingly, we reverse the judgment of the Court of Appeals and reinstate defendant’s conviction for first-degree murder. I. FACTS In 2013, defendant was convicted by a jury of first-degree murder, MCL 750.316, for the 1983 stabbing death of Andrew “Melvin” Weathers. At the time of his death, Weathers lived in a house in Highland Park with a relative, Louise Kountz, her teenage daughters, Melissa Kountz and Kimberly Stokes, as well as a family friend, Jimmy Godwin. Defendant had previously dated Louise Kountz and had lived in the Highland Park home for about four years, before moving out the summer before the murder. According to many witnesses, defendant had been violent and abusive toward Kountz during their relationship. After defendant had moved out of the home, he told Kountz’s daughters that he blamed Weathers for the break-up of his relationship with Kountz, saying that “[t]his is Melvin’s fault.” He threatened that “if it’s the last thing I do I’m going to get Melvin.” In the early morning of December 28, 1983, while the household was asleep, a perpetrator broke into the home through a basement window, disabled the electricity, placed the family’s dog in the basement freezer, and then proceeded upstairs to stab Weathers in his bedroom, using a knife from the home’s kitchen. The household awoke 2 to the sounds of an intruder and Weathers groaning. Kountz told her daughters to go next door and call the police. As they walked down the dark stairway from the upstairs bedrooms to the main floor, the daughters saw someone ahead of them whom they believed to be defendant. The build of the figure matched that of defendant, and one daughter smelled a distinctive stale cigarette odor that she associated with defendant. The daughters ran to a neighbor’s home and told their neighbor that defendant had killed Weathers. The police arrived and, in the dining room, discovered defendant’s brown loafers with a sponge taped to one of their soles. 1 In February 1984, the police obtained a warrant for defendant’s arrest. The police immediately placed defendant’s warrant into the Law Enforcement Information Network (LEIN) so that his warrant would show up if defendant encountered law enforcement anywhere in the country. The police searched for defendant locally, including at the homes of his relatives. Defendant’s family did not provide any information as to defendant’s location, but the police learned from other sources in the community that defendant had left the state shortly after Weathers’ death. The police reached out to their out-of-state counterparts, but they could not locate defendant. At one point, due to budgetary issues, the Highland Park Police Department’s participation in the LEIN program ended. And for several years, the Highland Park Police Department was closed. When it reopened, a new detective was assigned to Weathers’ murder case. In 2012, the police finally located defendant at his mother’s house in Oak Park. 1 The prosecution theorized that defendant had used the sponge to quiet his movements as he moved through the occupied house. 3 When the police first approached defendant and informed him that they were investigating a cold murder case, defendant lied and stated that his name was “Mark Jackson.” He could not provide any identification but consented to having the officers take his photograph. The officers sent the photo to one of Kountz’s daughters, who confirmed that it was defendant, the man she believed killed Weathers. The officers then arrested defendant for Weathers’ murder. After waiving his Miranda rights, 2 defendant told police that he had left town shortly after the murder because he feared Kountz and her family. He did not return to Michigan for more than two decades. He denied stabbing anyone in 1983 and denied ever hurting or threatening Kountz. Defendant was tried in 2013. At trial, the prosecution introduced evidence of defendant’s prior acts of domestic violence. 3 In response, defendant introduced evidence of his character for peacefulness. After the trial court instructed the jury, defendant objected to the adequacy of the instructions, including the omission of his requested character-evidence instruction, M Crim JI 5.8a(1), which reads: You have heard evidence about the defendant’s character for [peacefulness / honesty / good sexual morals / being law-abiding / (describe 2 Miranda v Arizona, 384 US 436, 444; 86 S Ct 1602; 16 L Ed 2d 694 (1966). 3 Before trial, the Court of Appeals reversed the trial court’s pretrial determination that evidence of the prior acts of domestic violence was inadmissible. See People v Lyles, unpublished order of the Court of Appeals, entered January 7, 2013 (Docket No. 313665); see also MCL 768.27b(1) (“[I]n a criminal action in which the defendant is accused of an offense involving domestic violence, evidence of the defendant’s commission of other acts of domestic violence is admissible for any purpose for which it is relevant, if it is not otherwise excluded under Michigan rule of evidence 403.”). Defendant did not appeal that decision, nor has he argued posttrial that the prosecution’s evidence was improperly admitted. 4 other trait)]. You may consider this evidence, together with all the other evidence in the case, in deciding whether the defendant committed the crime with which (he / she) is charged. Evidence of good character alone may sometimes create a reasonable doubt in your minds and lead you to find the defendant not guilty. The trial court then gave several more instructions to the jury, including the following: [Y]ou’ve heard the testimony of --- about witness’ [sic] truthfulness. You may consider this evidence together with all other evidence in the case in deciding whether you believe the testimony of the witness, in[] deciding how much weight to give to that witness. The prosecutor has examined some of the defendant’s character witnesses as to whether or not they heard anything bad about the defendant. You should consider such cross- examination only in deciding whether or not you believe the character witness and whether they described the [defendant] fairly. The prosecutor also has called witnesses who have testified that the defendant did not have good character of the other acts. Defendant again objected because the requested instruction related to the evidence of defendant’s character for nonviolence and peacefulness, not to his witnesses’ truthfulness. He also argued that the prosecution did not call any witnesses to rebut his good-character evidence. 4 The trial court noted the objection but did not alter the instructions. The jury convicted defendant of first-degree murder. The trial court sentenced him to life in prison without the possibility of parole. The Court of Appeals reversed defendant’s conviction and remanded for a new trial on the ground that the trial court had failed to instruct the jury regarding defendant’s evidence of his good character. People v Lyles, unpublished per curiam opinion of the Court of Appeals, issued July 22, 2014 (Docket No. 315323), p 5. This Court heard oral 4 Although the prosecution did not call witnesses to rebut the testimony of defendant’s good-character witnesses, the prosecution had already introduced evidence that defendant had committed acts of domestic violence in its case-in-chief. 5 argument on the prosecution’s application for leave to appeal and then issued an order remanding the case to the Court of Appeals to properly apply the governing standard, set forth in Lukity, for evaluating whether the error was harmless. See People v Lyles, 498 Mich 908, 908 (2015) (criticizing the Court of Appeals for citing older cases from this Court for the proposition that the failure to give a character-evidence instruction automatically requires reversal). On remand, the Court of Appeals purported to apply the Lukity harmless-error standard and again reversed defendant’s conviction. The court reasoned that the trial court’s error in failing to give the proper instruction had “eviscerated the significance of defendant’s character evidence, which was crucial to his defense, and wholly undermined his effort to establish that he was a peaceful person who could not have committed such a vicious murder.” People v Lyles (On Remand), unpublished per curiam opinion of the Court of Appeals, issued December 22, 2015 (Docket No. 315323), p 3. The prosecution again seeks this Court’s leave to appeal, arguing that the Court of Appeals erred by holding that the instructional error was harmful. We ordered oral argument on the application. People v Lyles, 500 Mich 875 (2016). II. ANALYSIS Preserved, nonconstitutional errors are subject to harmless-error review, governed by MCL 769.26: No judgment or verdict shall be set aside or reversed or a new trial be granted by any court of this state in any criminal case, on the ground of misdirection of the jury, or the improper admission or rejection of evidence, or for error as to any matter of pleading or procedure, unless in the opinion of the court, after an examination of the entire cause, it shall affirmatively appear that the error complained of has resulted in a miscarriage of justice. 6 We interpreted this statute in Lukity and held that a defendant carries the burden of showing that “it is more probable than not that the error was outcome determinative.” Lukity, 460 Mich at 495-496. “In making this determination, the reviewing court should focus on the nature of the error in light of the weight and strength of the untainted evidence.” People v Elston, 462 Mich 751, 766; 614 NW2d 595 (2000). Here, the Court of Appeals’ critical error was focusing on the importance of the good-character instruction to defendant’s defense strategy instead of evaluating the likelihood of defendant’s prevailing on that strategy. When considering whether the error was harmless, the question is whether the instruction would have made a difference in the outcome. See Lukity, 460 Mich at 495-496. Answering this question requires a court to consider not only the relationship between the instruction and defendant’s defense strategy, but also the strength of that strategy relative to the proofs as a whole. Defendant’s character evidence was extremely weak. But he was not denied the opportunity to present it; he was only denied a proper instruction that the jury could consider this evidence with all the other evidence in the case and that good-character evidence, if believed by the jury, could alone produce a reasonable doubt. Even if the jury had been properly instructed, it is not more likely than not that defendant’s character evidence would have been sufficient to create a reasonable doubt, given the prosecution’s evidence; and so defendant failed to show that the instructional error was harmful. 7 A. THE PROSECUTION’S EVIDENCE We state at the outset two hurdles in the prosecution’s case: the trial took place almost 30 years after the murder and any physical evidence 5 had been inadvertently destroyed in the intervening years. 6 Yet defendant had been identified as a suspect immediately after the murder. Kountz’s daughters accused defendant on the night of the murder, and an arrest warrant for defendant was issued just a few months later in February 1984. Defendant, however, had fled the state. His sudden absence was palpable. For months before the murder, beginning when he had moved out of the house, defendant had not left Kountz or her family alone. On one occasion, he returned to the house and got into an argument with Kountz that resulted in her having a bloody nose. He repeatedly called the house and asked about Kountz. Several times, bricks were thrown at the house, and although no one saw defendant throwing them, the bricks, as well as the harassing visits and phone calls, stopped after the murder and defendant’s leaving the state. The family did not see or hear from defendant again until he was arrested in 2012. 5 A laboratory report from the Detroit Police Department indicates that two pieces of evidence were sent to the forensic laboratory for testing during the original investigation. The murder weapon was tested for fingerprints, but the prints were not usable. No description of the second piece of tested evidence has survived, though the original detective on the case believed that some bloody clothing and bed sheets were collected as evidence. 6 Any physical evidence was inadvertently destroyed due to a series of unfortunate events: the city of Highland Park went into receivership, there was flooding at the police department building, the police department building was torn down, and the courthouse was closed and abandoned. 8 In addition to the suspicions raised by defendant’s flight, there was evidence that the perpetrator of the crime had inside knowledge of the house and the habits of its occupants. The residents’ dog stayed in the basement at night; the perpetrator broke in through a basement window and placed the dog in the basement freezer. The perpetrator unscrewed two fuses, disabling the house’s electricity. Then, the perpetrator went upstairs to the kitchen, grabbed a knife from the knife block, and, without waking the other occupants of the home, entered Weathers’ bedroom, where he stabbed Weathers before fleeing. Kountz’s daughters identified defendant as the perpetrator on the night of the murder. After waking to the sound of someone in the house, Kountz told her daughters to go next door and call the police. As the girls walked down the stairs, they could see another person heading downstairs ahead of them. One of them was able to see and smell the man, 7 and she was certain it was defendant. The other girl saw the size and build of the man and thought it was defendant. The girls went straight to their neighbor’s home to call the police and told their neighbor and her daughter that defendant had killed Weathers. After returning to the house, one of the daughters found a pair of men’s 7 She had lived in the same house as defendant for four years prior to defendant moving out the summer before the murder occurred and so was presumably quite familiar with defendant’s scent. She described defendant as having a distinctive smell of stale cigarettes. Another witness testified that around the 1980s defendant had a distinctive “weird odor,” such as “how some people don’t wash a lot and it’s like coming through their pores, he always stank.” A detective who encountered defendant in 2012 described defendant as “very unkept,” with “I’m not sure how to describe the smell, an unwashed smell, possibly with smoker’s smell.” 9 loafers; one of the loafers had a sponge taped to the bottom. The shoes had not been there the night before. The daughter identified the shoes as belonging to defendant. B. DEFENDANT’S CHARACTER EVIDENCE Defendant called only three witnesses during the trial. The first was his sister, Geraldine Johnson. She testified that defendant graduated from high school in 1971 and then worked at General Motors before enlisting in the United States Army from 1974 to 1977, after which he returned to Highland Park and worked at Chrysler. Defendant’s second witness was Jo Ann Davenport, a former girlfriend who had dated defendant for about a year around 1972, 11 years before the murder. She testified that defendant had not abused her during their relationship. A few years after their relationship ended, she had moved away from Highland Park. She did not stay in touch with defendant but would occasionally see him in passing when she visited her family back home. She could not remember the last time she had seen him. She testified that when she had visited her old neighborhood, she had heard nothing about defendant’s abusing any girlfriends in the early 1980s. Defendant’s third witness, Kim Harden, described defendant as being “like a cousin” to her because their parents were very close. They grew up on the same block, and although Harden had moved to California for school three years before the murder, she returned to Michigan on occasion, including during the summers. She testified that during these return trips, she would visit defendant’s family home daily, but she would only see defendant if he happened to be visiting his family home. She saw defendant with Kountz about three or four times during the 1980s, and not once did she observe 10 verbal or physical abuse. She described defendant as a peaceful person who did not believe in violence. She also testified that during the 1980s, defendant did not have a reputation for being an abusive person. The jury heard all of this evidence; the error was in failing to properly instruct the jurors on the use of the good-character evidence: You may consider this evidence, together with all the other evidence in the case, in deciding whether the defendant committed the crime with which (he / she) is charged. Evidence of good character alone may sometimes create a reasonable doubt in your minds and lead you to find the defendant not guilty. [M Crim JI 5.8a(1).][8] We note that the jury was instructed that “[a] reasonable doubt is a fair, honest doubt growing out of the evidence or the lack of evidence,” that “[y]ou must think about all of the evidence and decide each piece of evidence, what it means and how important you think it is. That includes whether you believe what each of the witnesses said,” and that “it is your job to decide what the facts of the case are. You must decide which witnesses you believe and how important you think their testimony is.” We cannot conclude that the absence of the good-character instruction made the difference in the outcome. Or to put it differently, we cannot say that if only the jury had been properly instructed, the verdict would have been altered because of this good- character evidence. In terms of character evidence alone, the prosecution’s evidence of defendant’s past violence toward Kountz was far stronger. The victim’s sister, Carrie 8 Under MCR 2.512(D)(2), “[p]ertinent portions of the instructions approved by the . . . Committee on Model Criminal Jury Instructions . . . must be given in each action in which jury instructions are given” if they are applicable, they accurately state the applicable law, and they are requested by a party. 11 Weathers, who had lived in the house for a time, described defendant’s relationship with Kountz as “brutal,” testifying that she saw defendant hit Kountz “just about every other day.” One of Kountz’s daughters testified that once, when Kountz asked defendant if Weathers’ sister could move back into the house, defendant refused, demonstrating his anger by stabbing a butcher knife in the middle of a bed. Kountz’s daughters also testified to defendant’s abuse toward their mom, including rape and repeated instances of other violence. George Arnold, who had lived in the house for less than a year around 1981 or 1982, testified that he saw defendant hit Kountz and that once, when he had tried to intervene, defendant had struck him. Jeffrey Trent, a neighbor, testified that defendant and Kountz had a violent relationship and that on one occasion defendant asked Kountz to look under the hood of his car; when she bent over, he kicked her and slammed the hood on the upper half of her body. Kountz’s long-time next-door neighbors, Carolyn and Camille Rhodman, similarly testified to observing defendant abuse Kountz on multiple occasions. At face value, the prosecution’s evidence of bad character, offered by those who had many chances to observe defendant in the early 1980s, far outweighed defendant’s good-character evidence, which was both substantively weaker and was offered by witnesses who had little contact with defendant during the relevant time period and who no longer lived in the same city. And, of course, the jury did hear and could weigh defendant’s proffered good-character evidence; the only error was that the jurors were not instructed on character evidence in particular, including that, if they believed this evidence, it was permissible for them to conclude, based on this evidence alone, that there was reasonable doubt. 12 The Court of Appeals erred by focusing on the instruction’s importance to defendant’s defense rather than focusing on its importance to the verdict. Compare Lyles (On Remand), unpub op at 3 (“In effect, this error by the trial court eviscerated the significance of defendant’s character evidence, which was crucial to his defense, and wholly undermined his effort to establish that he was a peaceful person who could not have committed such a vicious murder.”), with Lukity, 460 Mich at 495-496 (“[A] preserved, nonconstitutional error is not a ground for reversal unless after an examination of the entire cause, it shall affirmatively appear that it is more probable than not that the error was outcome determinative.”) (quotation marks omitted). But even the Court of Appeals’ conclusion that “[d]efendant’s best defense . . . was his evidence regarding his peaceful character,” 9 Lyles (On Remand), unpub op at 5, is belied by the fact that defense counsel did not think it important to make a good-character argument in either the defendant’s opening or closing arguments. 10 The prosecution argues, here and in the Court of Appeals, that the failure to include the good-character argument in opening or 9 Despite its conclusion that the instructional error “eviscerated the significance of defendant’s character evidence, which was crucial to his defense,” Lyles (On Remand), unpub op at 3, the Court of Appeals did not recount this crucial evidence or describe it at all other than in this one sentence: “In particular, in contrast to the violence described by several of the witnesses, defendant presented evidence of his peaceful character in the form of reputation and opinion testimony from a woman who had known defendant all her life and lived on defendant’s street for many years.” Lyles (On Remand), unpub op at 2. 10 Instead, defendant argued that the prosecution’s evidence of domestic violence should not be believed because there were no police reports, the victim (Louise Kountz) did not testify (she had passed away), and Kountz’s daughters had incentives to lie. The rest of defendant’s opening and closing statements consisted of pointing out the lack of physical evidence and the age of the case. 13 closing arguments is relevant in considering the importance of the omitted instruction to the defense theory. The Court of Appeals dismissed the prosecution’s argument, reasoning that “[r]egardless of whether counsel emphasized this evidence during closing, a properly instructed jury should have been told that the character evidence could, on its own, create a reasonable doubt.” Lyles (On Remand), unpub op at 4 n 1. We agree that a properly instructed jury would have received the instruction. But that goes to whether there was error, not to whether the error was harmless. Even if the Court of Appeals were right, however, that defendant’s good-character evidence was his best defense, defendant was permitted to introduce this evidence. The instructional error is not reversible error unless the defense likely would have succeeded at trial. 11 See Lukity, 460 Mich at 495-496. Given the evidence at trial, it is not “more 11 The Court of Appeals relied on two of our cases for its conclusion that there was reversible error here because “[b]y failing to give the key part of the instruction requested by defendant, the trial court prevented the jury from having the opportunity to properly weigh defendant’s character evidence.” Lyles (On Remand), unpub op at 4. The Court of Appeals failed to note that the first, People v Silver, 466 Mich 386; 646 NW2d 150 (2002) (opinion by TAYLOR, J.), is a fractured decision of our Court and that the opinion from which it quoted represents the views of only two Justices. As for the second, People v Rodriguez, 463 Mich 466, 474; 620 NW2d 13 (2000), the Court of Appeals relied on our statement that the instruction was “crucial to the defendant’s defense and was clearly supported by the evidence.” That was not the extent of our analysis of the instructional error in Rodriguez, however. The defendant was convicted of evading the use tax on a van that he claimed he planned to sell at auction after repairing; the requested instruction would have informed the jury that property purchased for resale is exempt from the use tax. We concluded that “[u]nder [MCL 205.94(c)], the defendant was—if a properly instructed jury were to believe his version of the facts—exempt from the tax.” Id. at 472. Thus, “[t]here is no question that the error undermined the reliability of the verdict . . . .” Id. at 474. Accordingly, we examined “the entire cause,” MCL 769.26, and determined that the specific circumstances of that case rendered the instructional error outcome-determinative. Rodriguez does not stand for the proposition that any instructional error that affects an issue that is “crucial to the defendant’s 14 probable than not” that the outcome would have been different if the jury had been properly instructed on how it could consider the evidence defendant offered regarding his good character. See id. at 496. Therefore, the Court of Appeals erred by holding that defendant had met the Lukity harmless-error standard. 12 III. CONCLUSION Defendant has failed to show that the instructional error more likely than not affected the outcome of his trial for first-degree murder. Accordingly, we reverse the judgment of the Court of Appeals and reinstate defendant’s conviction. Joan L. Larsen Stephen J. Markman Brian K. Zahra Kurtis T. Wilder defense”—regardless of the strength of the evidence in support of that defense—requires reversal. 12 Much of the dissent seems to us flavored by issues not raised in this Court, such as whether the jurors should have heard the evidence, held admissible by the Court of Appeals, regarding defendant’s persistent violence toward Kountz, whether the trial court made other instructional errors, or whether defense counsel should have objected at various points in the proceeding. Whether or not those errors might be cognizable in another forum, they are not before us here. We wholeheartedly agree, however, with the first paragraph of the dissent: determining whether an error was harmless can be a difficult task. In considering the facts of this case, in light of the claim of error presented to us, we simply see this case differently than the dissent. 15 STATE OF MICHIGAN SUPREME COURT PEOPLE OF THE STATE OF MICHIGAN, Plaintiff-Appellant, v No. 153185 WILLIAM LYLES, JR., Defendant-Appellee. MCCORMACK, J. (dissenting). Of the many judgments I must make in this job, harmlessness is among the hardest and most humbling. Identifying errors is simple by comparison; dig deeply enough into the record and law, and you’ll see if one is there. Determining that error’s harm to the outcome, however, requires much more—particularly when that outcome is a jury verdict. As we well know, appellate courts are not juries, and we do not become them when reviewing their verdicts for harm. Rather, in performing that review, we must respect the role and work of the jurors in reaching the verdict they did, while still scrutinizing the reliability of that verdict now that a defect in its making has been exposed. We must try to discern, from a cold and silent record, what those jurors may have been thinking, but without thinking for them. We must measure their verdict based not on whether we agree with it ourselves, but on whether we can still trust it enough, as a matter of law, to reflect how they would have decided the case had it not been wrongly presented to them. And of course, the stakes, particularly in a criminal matter, couldn’t be higher: whether a criminal conviction, with all it entails, remains intact despite a failing in the system of justice through which it was secured; or whether that failing instead requires another trial—and all the expenditures that attend it, emotional and otherwise—to ensure any verdict on the charge is reliable under the law. This task is as daunting as it is daily. And so, as courts tend to do, we (with the Legislature’s help) have developed various standards to guide us in handling it. Those standards have changed over time (sometimes rather abruptly), and they are tailored to match the general nature and circumstances of the error at issue. Occasionally, a case makes their application relatively straightforward: an error is so clearly important to a verdict, or so clearly irrelevant to it, that its harm can be measured in fairly short order. That is not this case. We are no strangers to the defendant’s jury trial for the 1983 murder of Andrew “Melvin” Weathers. This is the second time we’ve considered it. It’s also the second time we have been asked by the prosecutor, and have agreed, to consider the specific question before us: whether the trial court’s misinstruction of the jury regarding the defendant’s character evidence was harmless, contrary to what the Court of Appeals panel (now twice) concluded. 1 Indeed, this is the only question about this case we have ever taken up. The standard governing that question is provided by People v Lukity, 460 Mich 484, 495-496; 596 NW2d 607 (1999), and didn’t change from the first time we considered the question to the second. After the first time, we concluded that, while the Court of Appeals panel had recited this settled standard in its unpublished opinion, it may 1 See People v Lyles, unpublished per curiam opinion of the Court of Appeals, issued July 22, 2014 (Docket No. 315323); People v Lyles (On Remand), unpublished per curiam opinion of the Court of Appeals, issued December 22, 2015 (Docket No. 315323). 2 not have truly applied the standard in its analysis. We chose to remand for that Court to perform that fact-intensive task, as intermediate appellate courts so routinely do. People v Lyles, 498 Mich 908 (2015). In another unpublished opinion, the panel did as we directed, articulating the harmlessness standard set forth in Lukity and detailing its application to the facts of this case. The result remained unchanged: the court’s error was not harmless to the jurors’ verdict, and thus a new trial was required. And so, here we are again. Only this time, a majority of this Court has concluded that the panel’s award of relief should be reversed. I cannot see why. I. THE PANEL’S OPINION ON REMAND As the panel summarized, the following standard governs the harmlessness question before us: MCL 769.26 “presumes that a preserved, nonconstitutional error is not a ground for reversal unless after an examination of the entire cause, it shall affirmatively appear that it is more probable than not that the error was outcome determinative.[”] Lukity, 460 Mich at 495-496 (quotation marks omitted). Consequently, the burden is on defendant to “demonstrate that after an examination of the entire cause, it shall affirmatively appear that the error asserted has resulted in a miscarriage of justice.” Id. at 495 (quotation omitted). “In making this determination, the reviewing court should focus on the nature of the error in light of the weight and strength of the untainted evidence.” People v Rodriguez, 463 Mich 466, 474; 620 NW2d 13 (2000), quoting People v Elston, 462 Mich 751, 766; 614 NW2d 595 (2000). See also People v Mitchell, 301 Mich App 282, 286; 835 NW2d 615 (2013). “The object of this inquiry is to determine if it affirmatively appears that the error asserted undermines the reliability of the verdict.” Lukity, 460 Mich at 495 (quotation marks and citations omitted). “In other words, the effect of the error is evaluated by assessing it in the context of the untainted evidence to determine whether it is more probable than not that a different outcome would have resulted without the error.” Id. [People v Lyles (On Remand), unpublished per curiam opinion 3 of the Court of Appeals, issued December 22, 2015 (Docket No. 315323), p 3.] The majority does not take issue with this articulation of the Lukity standard. Nor is there any reason to do so. Instead, the majority seems concerned that the panel went astray in its application of this standard, “purport[ing]” to apply it but, once again, not actually doing so. More specifically, the majority posits that the panel’s “critical error was focusing on the importance of the good-character instruction to defendant’s defense strategy instead of evaluating the likelihood of defendant’s prevailing on that strategy”— that is, “focusing on the instruction’s importance to defendant’s defense rather than focusing on its importance to the verdict.” In support, the majority repeatedly points to the panel’s conclusion that the trial court’s instructional error “eviscerated the significance of defendant’s character evidence, which was crucial to his defense, and wholly undermined his effort to establish that he was a peaceful person who could not have committed such a vicious murder”—at one point holding it right up against Lukity’s fully stated harmlessness standard, as if to better expose its inadequacy. It is true: the panel did evaluate how the court’s misinstruction to the jury interacted with the defense presented by the defendant at his trial. To the extent the majority suggests this was itself a legal error, I can’t agree. A (properly instructed) jury arrives at its verdict only after assessing both the prosecution’s proofs and the defendant’s defense to those proofs; the importance of an error to that defense is thus quite relevant to its importance to the ensuing verdict. I do not read Lukity or any other 4 case to suggest otherwise, nor has the majority cited any authority to that effect. 2 To the contrary, assessing just how big an impact an error has had on a defendant’s defense falls comfortably within Lukity’s mandate that the reviewing court “focus[] on the nature of the error . . . .” Lukity, 460 Mich at 495 (citation and quotation marks omitted). That is precisely what the panel did here. To read the majority’s opinion, you’d think that’s all the panel did—that it simply looked at “the nature of the error,” found it very significant to the defendant’s defense, and called it a day. Were that the case, then the panel’s analysis might very well have fallen short of the requirement that it “assess[] [the error] in the context of the untainted evidence to determine whether it is more probable than not that a different outcome would have resulted without the error.” Id. But that’s not what the panel did. Rather, immediately after summarizing the “eviscerat[ing]” effect of the court’s error on the defendant’s character evidence and defense, the panel went on to explain that, 2 The majority criticizes the panel for citing, in support of this portion of its analysis, two cases from this Court: People v Silver, 466 Mich 386; 646 NW2d 150 (2002), and People v Rodriguez, 463 Mich 466, 474; 620 NW2d 13 (2000). The panel, however, merely cited these cases as general support for its conclusion that “the nature of the error” in this case was significant because, “[b]y failing to give the key part of the instruction requested by defendant, the trial court prevented the jury from having the opportunity to properly weigh defendant’s character evidence.” Lyles (On Remand), unpub op at 4. Both Silver and Rodriguez strike me as fairly read to provide such general support (and although, as the majority notes, Silver was a fractured decision, I do not see, with respect to the point for which the panel cited it here, disagreement among the two Justices who joined the opinion and the two who concurred). While both cases may be distinguishable in other respects from the instant case, the panel—by signaling its citation to them with a “Cf.”—did not purport to suggest otherwise. Simply put, I fail to see any error in the panel’s limited reliance on these cases, or in the more general point it cited them to support. 5 “considering the nature of this error in the context of the evidence presented by the prosecution, we are persuaded that defendant has met his burden of demonstrating that it is more probable than not that a different outcome would have resulted without the error.” Lyles (On Remand), unpub op at 3-4. The panel then detailed at length why that was the case, evaluating over multiple paragraphs the very same proofs the majority summarizes here and concluding that, “[u]ltimately, considering the entire cause, we are persuaded that the trial court’s failure to give the requested character evidence instruction eviscerated the effect of defendant’s proffered character evidence and that such an error cannot be considered harmless in light of the evidence relied upon by the prosecution,” and thus “defendant has met his burden of establishing that it is more probable than not that the error was outcome determinative.” Id. at 4-6. The majority does not engage with, or even acknowledge, this substantial portion of the panel’s harmlessness analysis, and so it is difficult to discern what exactly they might think is so wrong with it: what critical evidence the panel clearly neglected or misstated, or what fatal mistake of law it made in carrying out its fact-intensive application of Lukity’s settled standard. From my own review, I can’t find any such thing, or any reason why this Court should reverse the panel’s analysis and conclusion. We gave the panel a job to do. It did that job, and showed its work. That work was deliberate and thorough, well beyond what we ourselves are often asked to review. And as “an examination of the entire cause” confirms, it was correct. 6 II. THE ENTIRE CAUSE As acknowledged by both the panel and the majority, the defendant’s trial for the 30-year-old murder of Andrew “Melvin” Weathers was exceptional for what it lacked. Though a warrant for the defendant’s arrest was issued within a few months of the crime in 1983, the case went cold after initial efforts to locate the defendant failed—and only got colder as the years progressed. 3 By the time of the defendant’s trial in 2013, the original case file had been completely lost, along with any reports, witness statements, or other evidence it might have contained. 4 Gone too were most officers who had been involved in the original investigation 5 as well as Louise Kountz, who had died a few years before the trial and was the only conceivable link between the defendant and the crime. 6 As the majority says, these circumstances certainly posed “hurdles” in the case. 3 As the majority notes, a number of circumstances seemingly contributed to this state of affairs, including the defendant’s move to another state, the shuttering of the police department originally in charge of the investigation, and the absence of the arrest warrant from the Law Enforcement Information Network. 4 A copy of the original arrest warrant, it seems, survived; it was mentioned at trial, but not entered into evidence and thus not submitted to the jury. 5 One such officer testified at trial, but as the majority notes, he could recall little about the evidence collected beyond that the murder weapon had been tested for fingerprints and the prints proved unusable. Nonetheless, and without further substantiation or explanation, the officer averred that he did not “ever receive any lab results, additional evidence, or any information that caused [him] to change [his] mind about this defendant being the right person” to “charge[] with this murder”—and even suggested in passing that the evidence “just strengthened” that impression. Defense counsel raised no objection to this testimony, and the defendant did not challenge it on appeal. 6 As the majority notes, Melvin was related to the Kountz family and was living in Louise’s house (along with her daughters and another man) at the time of the murder. The defendant and Louise had begun dating roughly four to five years before the murder; the defendant had lived in the house with Louise and her daughters during that time, 7 Whether moreso for the prosecution or for the defendant, however, I don’t think it’s possible to say. Whatever inculpatory or exculpatory information these now-gone sources may have held, neither the parties nor the jurors would have the benefit of knowing. They all would need to look elsewhere for proof of the defendant’s guilt. A. THE PROOFS AT TRIAL For the prosecution, this meant leaning heavily on Louise’s daughters, Melissa Kountz and Kimberly Stokes, 7 to reconstruct the events of that December night in 1983 when Melvin was murdered (and to place the defendant within them). At the time of the murder, Melissa and Kimberly were 13 and 16 years old, respectively. Per their testimony, they were both in the house when the murder occurred, and they both remembered seeing, through the darkness, a shadow cast by the perpetrator that struck them as roughly the same shape and size as the defendant. 8 Both also recalled running which overlapped to some extent with Melvin’s stay in the house. The defendant, however, had moved out of the house a number of months before the murder, after Louise had ended their relationship. 7 Melissa and Kimberly were Louise’s children from a prior marriage, which ended shortly before she began dating the defendant—though it seems Louise continued to go by her married name after the divorce. Louise’s prior husband had been Melissa’s biological father, and Kimberly’s adoptive one for all but the first year of her life. According to the defendant, Louise had left the girls’ father for him. The defendant then lived with Louise and the girls for roughly four or five years before moving out at Louise’s behest sometime in the summer of 1983. 8 In her 2012 statement to police, however, Kimberly was seemingly less certain about the source of this shadow, saying that she did not remember seeing the defendant in the house on the night in question. Also missing from both Melissa’s and Kimberly’s pre- trial accounts was any mention of the apparent source of light that permitted the shadow to be cast in the otherwise dark house—a detail both of them offered for the first time in their trial testimony. In any event, according to their trial testimony, Melissa and 8 out of their house right afterward and to their neighbors’ to call the police; according to the neighbors’ ensuing testimony, one of the girls blurted out that the defendant had killed Melvin (though one neighbor thought Melissa had said as much, whereas the other thought it had been Kimberly). 9 Lastly, both recounted a phone call with the defendant roughly a month or two before the murder, in which he blamed Melvin for the demise of his relationship with Louise and vowed to “get” Melvin. Why the defendant might have blamed Melvin—let alone why he might have blamed him so intensely as to plot and carry out his murder—neither they, nor anyone else at trial, could say. Kimberly had nothing more to offer regarding who the perpetrator of this crime might be. The remainder of what the jury heard on that score—all the details that were offered to lend some definition to the shadow in the house—came solely from Melissa: from the manner in which the perpetrator broke into the house and obtained the murder weapon; 10 to the odor of stale, Kool-brand cigarettes she smelled on him (which she Kimberly noticed this shadow when they were hurrying down the stairs right behind the murderer, who himself was rushing and stumbling. They were in this highly precarious position, they testified, because they had been told to go downstairs and across the street to tell the neighbors to call the police—despite the fact that the murderer had just headed downstairs and there apparently were still-functional rotary phones safely located in the house upstairs. 9 According to one of the neighbors, however, and in apparent tension with the daughters’ account of their immediate flight from the scene, police had already arrived at Louise’s house by the time the two girls reached her doorstep. Further confusing matters, Kimberly had originally told police during their 2012 investigation that it was Louise who had called the police after they “all” had gone over to the neighbors’ house. 10 While this testimony was offered to suggest, as the majority notes, that the perpetrator was familiar with the house, the testimony did not indicate any familiarity unique to the defendant or beyond that which could have been gained by the various other individuals 9 associated with the defendant); 11 to the pair of the defendant’s shoes, with a sponge taped to one their soles, which apparently materialized in the house after the killing. 12 And the jury also learned two other things from Melissa: that she hated the defendant at the time of the murder for his treatment of her and her mother; 13 and that she set in motion the events that led to his trial some 30 years later, contacting the police in 2012, seemingly out of the blue, to inquire about the status of the (by then, very cold) case. 14 who visited or stayed in the house over the years (or through speaking with any one of them). 11 Melissa said she smelled this odor at the same time as she glimpsed the shadow of the perpetrator. Kimberly, who was standing right by Melissa at that time (and who had been living with the defendant for just as long as Melissa), did not remember any such smell on the night of the murder. Nor did she, or anyone else, testify to the defendant having, at the relevant time, the distinctive smell Melissa attributed to both him and the perpetrator. And while another witness who was not present that night testified that he generally remembered the defendant having a “weird odor . . . [l]ike, you know how some people don’t wash a lot and it’s like coming through their pores,” he suggested no association between that odor and cigarettes and couldn’t remember if the defendant smoked. 12 According to Melissa, she found these shoes in the downstairs dining room, neatly arranged with their heels against the wall; she had not noticed them there the night before. She did not speculate as to why, if the defendant wore these shoes to commit the crime, he would have stopped mid-flight, after rushing and stumbling down the stairs, to remove and arrange them before proceeding out of the house into the winter’s night, seemingly unshod. 13 Kimberly, for her part, denied that she hated the defendant at the time of the murder, but said that she “disliked his ways” and that his treatment of her mother “made [her] angry.” 14 Melissa testified that she had contacted police about the case on one prior occasion, sometime between 2003 and 2006, after seeing one of the defendant’s relatives—but the officer she spoke with then was not “helpful.” 10 From these accounts, the jurors learned that Melissa and Kimberly believed the defendant killed Melvin—now and, per their testimony, then. Even if the jurors were to credit this belief, it left ample room for doubt. The prosecution thus focused on proving something else: that the defendant persistently and violently abused and harassed Louise. As the majority notes, the prosecution very nearly wasn’t permitted to do so. In a pretrial ruling, the trial court had deemed such evidence inadmissible under MRE 403 because it was unfairly prejudicial to the defendant. The prosecution appealed, however, and successfully secured an order from the Court of Appeals peremptorily reversing that determination. People v Lyles, unpublished order of the Court of Appeals, entered January 7, 2013 (Docket No. 313665). And with this appellate blessing in hand, the prosecution turned the defendant’s abuse of Louise into the overriding theme of his trial for murdering Melvin. 15 15 Before trial, the prosecution had moved to admit evidence regarding the defendant’s abuse of Louise under MRE 404(b) or MCL 768.27b (pertaining to “defendant’s commission of other acts of domestic violence”)—the latter of which, it bears noting, did not become available as a potential basis for admitting this evidence until 2006. See MCL 768.27b(6). The parties disputed whether the instant murder charge was an “offense involving domestic violence” such that it would fall within the scope of MCL 768.27b (thereby opening the door for the evidence of Louise’s abuse to be admitted under that rule). In its ruling, the trial court did not expressly decide that threshold question, instead concluding that the evidence was inadmissible under either the statute or MRE 404(b) because, per MRE 403, its “prejudicial [effect] . . . substantially outweigh[ed] any demonstrated relevan[ce]” it had to the charged offense. The court further noted, with respect to MRE 404(b), that the character inference raised by the evidence weighed against its admissibility. In peremptorily reversing this ruling, the order of the Court of Appeals did not address whether the prosecution’s offered evidence passed threshold admissibility standards under either MCL 768.27b or MRE 404(b), but it concluded that the trial court had “failed to make two distinct inquiries under the balancing test of MRE 403, which are whether introduction of the prior acts will be unfairly prejudicial and then to weigh the 11 As with the actual crime with which the defendant was charged, there was nothing in the way of physical evidence, records, reports, or other documentation to substantiate these uncharged acts of abuse. Nor was it even clear whether any such evidence ever existed. But as the majority summarizes, the prosecution’s witnesses had much to say about it. Melissa and Kimberly testified about the abuse at length, and other prosecution witnesses testified to little else; indeed, of the prosecution’s nine witnesses, only the two investigating officers did not speak on the topic. These other seven witnesses were all friends and family of the Kountzes, and together, they painted a very grim picture of the defendant’s relationship with Louise: one defined by physical, sexual, and verbal abuse, which was carried out with regularity and without remorse in full view of her family and neighbors, and which continued to manifest itself even after Louise had ended the relationship and the defendant had moved out of the house. probative value or relevance of the evidence against the unfair prejudice.” Lyles, unpub order at 1. The order reasoned that the “[a]dmission of the domestic violence acts against Louise Kantz [sic] will provide a full and complete picture of defendant’s history involving the victim and the occupants of Louise’s house, including Louise, and will tend to shed light on the likelihood that defendant is the perpetrator who broke into the house and stabbed his ex-girlfriend’s cousin.” Id. (quotation marks and citation omitted). The order continued: “Furthermore, the prior acts are not marginally probative nor is it likely that the evidence will be given undue or preemptive weight by the jury, especially if the trial court minimizes the prejudicial effect by properly instructing the jury.” Id. (quotation marks and citation omitted). Needless to say, the concluding condition of that rationale may have been overly optimistic. Regardless, despite the cursory and critical nature of this ruling, the defendant’s counsel did not seek our review of it then and did not further challenge the admission of this evidence at trial beyond renewing, for the record, the general MRE 403 challenge that the Court of Appeals’ order had already rejected. Nor did his appellate counsel challenge this ruling, the subsequent admission of the domestic-violence evidence at trial, or trial counsel’s effectiveness with respect to this evidence, in the direct appeal of his conviction. 12 None of these witnesses, however, could—or even attempted to—explain how or why the defendant’s treatment of Louise might have led the defendant to kill Melvin that night. There was no testimony, for instance, that Melvin had ever tried to intervene or even object to the abuse, that he had ever become the target of it himself, or that he had otherwise gotten entangled in the defendant’s relationship with Louise at any point. 16 No, the driving purpose of all this testimony was not to show why or how the defendant killed Melvin, but instead to show that he could have killed Melvin, that he was capable of it, because that sort of violence and brutality was in his character. It’s just the kind of man he was; he had, as the prosecution told the jurors, a “tendency towards violence,” and these accounts of his abuse of Louise were “highly probative” of that tendency and thus “the fact that he committed” Melvin’s murder. And from its opening remarks through to its closing ones, the prosecution made clear to the jurors that this was an inference upon which they could, and should, hang their hats—in evaluating the circumstantial proofs it had presented, and in determining whether the defendant should 16 To the contrary, Melvin was described as younger than the defendant, “small” and “on the frail side” relative to him, and “scared” by his alleged violent conduct, fleeing the room when it would occur. This was in contrast to Melissa and some others, who testified to attempting to stop the defendant at times and getting into altercations with him in the process. There was also some testimony, as the majority notes, that the defendant and Melvin’s sister—who testified at trial and had lived in Louise’s house at some point well prior to Melvin moving in—didn’t get along, resulting in the sister moving out and the defendant and Louise later arguing about whether she could move back in. Melissa provided the sole testimony about the details of that argument, saying that it ended with the defendant expressing his anger by getting a knife from the kitchen and stabbing a mattress. There was no testimony or other indication, however, that Melvin was involved in any of the apparent ill will between his sister and the defendant or that the defendant attributed any of it to him. 13 be convicted. As the prosecution asked the jurors in closing, “Do you really think [these witnesses] all want to come in here and waste their time to get somebody who wasn’t a vicious abuser?” 17 The defendant did not take the stand, but an investigating officer testified to statements the defendant made upon his arrest in 2012. 18 In those statements, the defendant denied much of what was asked of him, including that he had stabbed anyone in 1983 or had ever threatened Louise, and while he acknowledged leaving the state after he and Louise broke up, he explained it was out of fear of her and her family. 19 Although 17 Similarly, the prosecution, in closing, pointed to testimony regarding the defendant’s military service, which the defendant had offered to cast doubt on the time line of certain events offered in the prosecution’s proofs, and argued to the jurors that such service “doesn’t mean you get to come home and beat women, and slam hoods on their head, and rape them in front of their children.” The prosecution then stressed that “[t]here is one person responsible for that [abusive conduct], and one person only, and that’s the defendant. Because he wanted to be in control, he wanted to be the boss. Just like there’s only one person who is responsible for driving that [knife] into Melvin’s heart that night, the defendant.” And while the prosecution would later note to the jurors that “this case is not about Louise,” it was only after telling the jurors that “each and every person has to account for their misdeeds in life,” that “the defendant needs to be held accountable right here, right now for what he did,” and that “the only blind eye anyone is asking you . . . to turn is to the horrible things that this man did, to just let it slide, [because] it won’t help Louise.” 18 As the majority notes, the officer testified that, when he initially approached the defendant, stated his purpose, and asked the defendant’s name, the defendant provided an alias. This testimony, however, was not corroborated by the officer’s written report of the encounter, which did not indicate an alias (despite there being a specific location on the report to do so), or by the defendant’s subsequent statement to police, which did not address it at all. 19 The proofs at trial did not make clear precisely when the defendant left the state: sometime after he moved out of Louise’s home but before the murder occurred; immediately after the murder; or sometime later, after he had been identified as a suspect and/or a warrant was issued for his arrest. The prosecution suggested that he left town 14 he himself didn’t testify, the defendant did present evidence to counter the prosecution’s character attack that had come to dominate his trial; this came in the form of reputation and opinion testimony regarding his general character for peacefulness, and was the extent of the proofs. 20 In particular, he presented the testimony of Kim Harden. Harden was, at the time of trial, a long-time school principal; she was younger than the defendant and had known him her whole life. She testified that she had grown up in the same neighborhood as the defendant and had spent time with him on a daily basis throughout those years until she left for college in 1980. After that, she would return home whenever school wasn’t in session and would again see the defendant on a daily basis, if he was around. He was, according to Harden, “like a cousin to [her],” someone she looked up to “as a mentor, [a] role model”; he was “the reason why [she] finished college,” and his home was “like [her] second home.” And based on all the time she spent with him and what she knew of him, she opined that he was—including at the time relevant to this immediately after the murder based on testimony from Melissa and Kimberly that they never saw the defendant after the murder and that certain harassing behavior toward the house, which they attributed to the defendant, had ceased once the murder occurred. Neither Melissa nor Kimberly, however, testified to actually witnessing the defendant engage in this behavior, or to seeing him at all since an altercation at the house in October 1983. The investigating officer at the time, meanwhile, testified that he began attempting to locate the defendant upon issuance of the arrest warrant in early February 1984 (a little over a month after the murder); after failing to determine the defendant’s whereabouts from speaking with his family members, the officer ultimately learned that the defendant had “gone down south” at some unspecified time. 20 Indeed, this was the only form of affirmative evidence the defendant was permitted to offer as to his good character; unlike the prosecution, his proofs could not include specific instances of past conduct. See MRE 405; People v Williams, 134 Mich App 639, 642; 351 NW2d 878 (1984). 15 case—“a peaceful person” who “didn’t believe in violence” and who had affirmatively counseled her against violent responses to trouble. She further testified that this was his reputation in the community as well at that time—a community which included the Kountzes (whom she also knew). And while, over these years, she had seen the defendant with Louise as well as with other girlfriends on various occasions, she had never seen him verbally or physically abuse any of them, nor had she heard anything from anyone to that effect; she had only ever seen him treat them well, including Louise. He was, Harden said, “always nice and kind” to Louise in the four or five times she saw them together, and she didn’t ever know him to be any other way. The prosecution did not cross-examine her. 21 B. THE COURT’S INSTRUCTIONS TO THE JURORS As is typical, the parties’ presentation of the case was bookended by the court’s instructions to the jurors—which, as the court stressed at both ends, provided the jurors with “the law that you’re to apply to the facts as you find the facts to exist” and which “[y]ou must take . . . as I give it to you.” As is also typical, the court instructed the jurors on the elements of the charged offenses and on an assortment of other standard matters common to criminal trials—including, as the majority notes, general advisements about 21 As the majority summarizes, the defendant also presented testimony from two other witnesses. First, his sister testified to some basic biographical information about him, such as his history of employment and military service. Then, a former girlfriend, who had lived in his and the Kountzes’ neighborhood and had dated him for roughly a year at one point, testified that (1) she had had a “[v]ery nice” relationship with the defendant, free from any verbal or physical abuse, and (2) during the time period relevant to this case, she had never heard anything in the neighborhood about the defendant being verbally and physically abusive to his girlfriend at that time. 16 the meaning of “reasonable doubt” and the jury’s role as the finder of fact (and with it, the arbiter of credibility). The court, however, wove amidst these standard instructions some of its own making, including the one that launched this case on its long appellate odyssey. A few of these bear particular mention. First, the court instructed the jurors regarding the prosecution’s voluminous testimony about the defendant’s abuse of Louise. Before doing so, the court had discussed with the parties the proper instruction for this evidence, with the parties ultimately agreeing that M Crim JI 4.11a (regarding “Evidence of Other Acts of Domestic Violence”) would be best. This model instruction provides: (1) The prosecution has introduced evidence of claimed acts of domestic violence by the defendant for which [he / she] is not on trial. (2) Before you may consider such alleged acts as evidence against the defendant, you must first find that the defendant actually committed such acts. (3) If you find that the defendant did commit those acts, you may consider them in deciding if the defendant committed the [offense / offenses] for which [he / she] is now on trial. (4) You must not convict the defendant here solely because you think [he / she] is guilty of other bad conduct. The evidence must convince you beyond a reasonable doubt that the defendant committed the alleged crime, or you must find [him / her] not guilty.[22] This, however, was not the instruction the court then provided to the jurors. Instead, the court instructed: You have heard a lot of evidence, testimony that the respondent committed some improper acts against Louise Kountz, who was at the time his girlfriend. They had that type of relationship. It appears from the 22 Brackets appearing in quotations of the Michigan Criminal Jury Instructions in this opinion appeared in the original. 17 testimony that that testimony of his improper acts against someone is not the decedent and who is accused of murdering, that testimony you must very carefully consider it for certain purposes. You may only think about whether that evidence intends to show that the defendant had a motive to kill Andrew Weathers, the decedent, and also that it tends to show to you that the complete picture of the defendant’s history as it relates to the part of it that Melvin, the decedent Andrew Weathers. You’ve heard that testimony involving Louise Kountz, the deceased Andrew Weathers, Melvin, the deceased --- now, Louise Kountz died, I guess, just four or five years ago and of natural --- not in this murder, but she was living at this address at 88 Louise Street in Highland Park, and that there were other people living at that address. And that the other people have testified as to improper acts that the defendant committed not against Andrew Weathers, but against Louise Kountz for the most part. You must carefully balance that testimony of improper acts so not to prejudice yourself against this defendant and this murder, to only that he had a motive and he was in this environment, the decedent Andrew Weathers. Now, I mentioned motive, in that bit of the instructions as to those improper acts committed, if you find that they were, in fact, committed, against, if you will, Louise Kountz, who was --- well, you know the relationship between her and the defendant. That those acts were for motive of the defendant to kill someone else, Melvin, not Ms. Kountz. This, it seems, was an amalgam between the agreed-upon M Crim JI 4.11a and M Crim JI 4.11, the model instruction for evidence admitted under MRE 404(b) 23—with substantial 23 M Crim JI 4.11 provides: (1) You have heard evidence that was introduced to show that the defendant committed [a crime / crimes / improper acts] for which [he / she] is not on trial. (2) If you believe this evidence, you must be very careful only to consider it for certain purposes. You may only think about whether this evidence tends to show: [Choose one or more from (a) through (g):] 18 liberties taken by the court as to both. For instance, while the court advised the jurors that they “must carefully balance that testimony of improper acts so [as] not to prejudice yourself against this defendant and this murder,” it did not expressly advise them, as both M Crim JI 4.11 and 4.11a would have, that they “must not convict the defendant here because you think [he] is guilty of other bad conduct.” 24 (a) That the defendant had a reason to commit the crime; (b) That the defendant specifically meant to ___________; (c) That the defendant knew what the things found in [his / her] possession were; (d) That the defendant acted purposefully—that is, not by accident or mistake, or because [he / she] misjudged the situation; (e) That the defendant used a plan, system, or characteristic scheme that [he / she] has used before or since; (f) Who committed the crime that the defendant is charged with. (g) [State other proper purpose for which evidence is offered.] (3) You must not consider this evidence for any other purpose. For example, you must not decide that it shows that the defendant is a bad person or that [he / she] is likely to commit crimes. You must not convict the defendant here because you think [he / she] is guilty of other bad conduct. All the evidence must convince you beyond a reasonable doubt that the defendant committed the alleged crime, or you must find [him / her] not guilty. 24 On direct appeal, the defendant raised a very cursory challenge to this instruction, which the Court of Appeals panel addressed briefly in its first unpublished opinion, after concluding the defendant was entitled to a new trial based on the instructional error now before us. The panel observed that “although the trial court’s instructions on other acts evidence may not have been so inadequate as to warrant reversal, they were certainly imperfect” and should, on re-trial, be corrected to adhere to the model instructions. Lyles, 19 At this point, the prosecution asked to approach the bench, but the court preferred to finish its instructions first. Ultimately, the court instructed the jurors on various other aspects of the prosecution’s proofs, including that (1) “evidence that the defendant moved away . . . does not prove his guilt” but may reflect “a consciousness of guilt,” and the jurors “must decide whether the evidence is true and if true, whether it shows that the defendant had a guilty state of mind”; (2) “[t]he prosecutor does not have to prove that the defendant had a reason to commit the alleged crime, she only has to show that the defendant actually committed the crime or that he meant to do so”; and (3) “you may use the [prosecution’s] identification testimony alone to convict the defendant as long as you believe the testimony and you find that it proves beyond a reasonable doubt that the defendant was the person who committed the crime.” After the court finished its instructions, counsel for both sides raised a number of concerns and omissions—including the court’s failure to provide the agreed-upon instruction under 4.11a, and its failure to instruct the jury on the defendant’s character evidence under M Crim JI 5.8a(1), which provides: (1) You have heard evidence about the defendant’s character for [peacefulness / honesty / good sexual morals / being law-abiding / (describe other trait)]. You may consider this evidence, together with all the other evidence in the case, in deciding whether the defendant committed the crime with which (he / she) is charged. Evidence of good character alone may sometimes create a reasonable doubt in your minds and lead you to find the defendant not guilty.[25] unpub op at 6. The panel also briefly rejected two other, similarly cursory challenges the defendant had raised. Id. The defendant did not seek our review of any of these rulings. 25 M Crim JI 5.8a provides in full: 20 The court then recalled the jury and rattled off further instructions. It did not revisit its instruction regarding the domestic-violence evidence, but—as we well know at this point, and in apparent response to defense counsel’s request for M Crim JI 5.8a(1)— it did provide the following instruction: [Y]ou’ve heard the testimony of --- about witness’ [sic] truthfulness. You may consider this evidence together with all other evidence in the case in deciding whether you believe the testimony of the witness, in[] deciding how much weight to give to that witness. The prosecutor has examined some of the defendant’s character witnesses as to whether or not they heard anything bad about the defendant. You should consider such cross- (1) You have heard evidence about the defendant’s character for [peacefulness / honesty / good sexual morals / being law-abiding / (describe other trait)]. You may consider this evidence, together with all the other evidence in the case, in deciding whether the defendant committed the crime with which (he / she) is charged. Evidence of good character alone may sometimes create a reasonable doubt in your minds and lead you to find the defendant not guilty. [(2) The prosecutor has cross-examined (one / some) of the defendant’s character witnesses as to whether they had heard anything bad about the defendant. You should consider such cross-examination only in deciding whether you believe the character witnesses and whether they described the defendant fairly.] [(3) The prosecutor also called witnesses who testified that the defendant does not have the good character described by the defendant’s character witnesses. This evidence can only be considered by you in judging whether you believe the defendant’s character witnesses and whether the defendant has a good character for (describe trait). It is not evidence that the defendant committed the crime charged.] As the instruction’s Use Notes make clear, and as the parties here don’t dispute, only Subsection (1) of this instruction applied to this case, as the prosecution did not cross- examine “the defendant’s character witnesses as to whether they had heard anything bad about the defendant” and did not call any witnesses to rebut them. 21 examination only in deciding whether or not you believe the character witness and whether they described the [defendant] fairly. The prosecutor also has called witnesses who have testified that the defendant did not have good character of the other acts. This, of course, wasn’t M Crim JI 5.8a(1), and neither it nor any other instruction made any mention of the defendant’s character evidence or its potential legal effect on the jurors’ assessment of reasonable doubt. Instead, it was another court-fashioned amalgam, seemingly derived from M Crim JI 5.8 (“Character Evidence Regarding Credibility of Witness”)—an instruction that the parties neither requested nor had reason to request, given the evidence presented. 26 The court didn’t simply swap 5.8a for 5.8, however. The court modified the last sentence of M Crim JI 5.8(3) to refer not to 26 M Crim JI 5.8 provides in full: (1) You have heard evidence about the character of [name witness] for truthfulness. You may consider this evidence, together with all the other evidence in the case, in deciding whether you believe the testimony of [name witness] and in deciding how much weight to give that testimony. [(2) The prosecutor has cross-examined (one / some) of the defendant’s character witnesses as to whether they had heard anything bad about the defendant. You should consider such cross-examination only in deciding whether you believe the character witnesses and whether they described (him / her) fairly.] [(3) The prosecutor also called witnesses who testified that the defendant does not have a good character for truthfulness. This evidence can only be considered by you in judging the believability of the defendant’s testimony. It is not evidence that (he / she) committed the crime charged.] As with M Crim JI 5.8a, there is seemingly no dispute that the second two subsections of M Crim JI 5.8 are inapplicable here. The same holds true for the first subsection of 5.8, as neither party provided character evidence regarding the truthfulness of any witness who testified at trial. 22 testimony about the defendant’s bad character for truthfulness but instead to testimony about his bad character “of the other acts”—hearkening the jury back once again to the evidence of the defendant’s abuse of Louise and the character inference to be drawn therefrom, but lending no further guidance as to how that inference could be considered. Defense counsel again objected, but—after a seemingly prescient but unheeded aside by the trial court that judges “get in trouble when they ad lib”—the instructions went unchanged and the jurors went forward with their deliberations, 27 ultimately returning the guilty verdict now before us once again. 28 III. WHAT AFFIRMATIVELY APPEARS With “the entire cause” examined, then, what “affirmatively appears”? This case was a pure contest of credibility and character. Though charged with the murder of Melvin, the defendant’s trial was more about his abuse of Louise—and what that uncharged misconduct said about the kind of person he was. Did the defendant kill 27 Puzzlingly, the jurors were not provided with a written copy of the court’s instructions and, when they requested as much during deliberations, the court, without explanation, only agreed to give them a copy of the two instructions it provided on the elements of each charged offense (first-degree and second-degree murder). 28 What I have summarized in this opinion is but a glimpse of the trial court’s apparent difficulties in instructing the jurors and carrying out other procedural matters in this trial; a full cataloging is not needed here, but the record well reflects them, and they are troubling. At oral argument before this Court, the prosecutor suggested that they might be attributable to the fact that the judge who ended up presiding over this trial was a former probate-court judge who was standing in and had not previously presided over a murder case. How exactly these circumstances came to pass, and whether they may account for the court’s difficulties, the record doesn’t show—nor does any of that matter. The jurors, the defendant, and Mr. Weathers deserved better from our courts than what the record here betrays, particularly with a matter as grave as murder on the line. 23 Melvin? By the time of trial, there simply wasn’t any hard evidence that either side could offer to guide the jurors on that question. There were 30-year-old memories and beliefs and inferences, but no substantiation, and no lucid reason why the defendant would have snuck into Louise’s house in the dead of a random winter’s night for no purpose other than to plunge a knife through Melvin’s chest while he slept. So the prosecution asked the jurors a different question: Was the defendant a murderer? Was he the sort of man that had such malice and violence inside him? And the prosecution focused intently on proving that he must be—because, just look what he did to Louise. Of course, there wasn’t any hard evidence on this secondary question, either, and so the parties found the friends and memories they could. The Kountz family found more, but the defendant found some, and each presented to the jurors what they had. From these proofs, the jurors were tasked with determining the defendant’s guilt. They were given legal rules for making that assessment, rules which were to shape and bind how they could, as a matter of law, answer the questions posed to them and determine whether the defendant’s guilt of the crime charged had been proved beyond a reasonable doubt. Rules which would, in turn, assure us, as appellate courts, that the decision reached was not merely one of passion or bias, but of law. And rules which we correspondingly, and duly, presume the jurors follow with rigorous, near-panacean precision. See, e.g., People v Breidenbach, 489 Mich 1, 13; 798 NW2d 738 (2011); People v Dupree, 486 Mich 693, 711; 788 NW2d 399 (2010); People v Graves, 458 Mich 476, 486; 581 NW2d 229 (1998); People v Abraham, 256 Mich App 265, 279; 662 NW2d 836 (2003); People v Crawford, 232 Mich App 608, 619; 591 NW2d 669 (1998) 24 (“A court must instruct the jury so that it may correctly and intelligently decide the case.”). The court correctly instructed the jurors on the role these rules must play in their verdict. But the court then botched the rules themselves—namely, by failing to inform the jurors that, as a matter of law, the defendant’s character evidence could in itself give rise to a reasonable doubt that would foreclose a finding of guilt. Instead, the jurors heard about a phantom cross-examination of the defendant’s witnesses regarding his bad character and about the “other acts” testimony the prosecution had offered to prove that bad character (and through it, his guilt). As is beyond dispute at this point, this was a mistake; the defendant was entitled to have the jurors informed of the potential legal significance of his evidence of good character, not to have them needlessly reminded of his alleged bad character. The panel detailed why this mistake, given both its nature and the weight and strength of the untainted evidence otherwise presented to the jurors, was not harmless to the jurors’ verdict and thus required a new trial under Lukity. Lyles (On Remand), unpub op at 3-6. The majority, however, seems to think that this mistake simply couldn’t have been that bad. I don’t see why—or how the panel reversibly erred in applying Lukity to conclude otherwise. First, I agree with the panel that “the nature of th[is] error” was, in a word, big. It deprived the jurors of instruction on whether and how they could consider the only affirmative evidence the defendant could and did offer on what became the core question at his trial: whether he was the kind of man who would commit the crime charged. And to make matters worse, the instruction instead gestured the jurors toward the 25 prosecution’s evidence to that effect. It thus strikes me as perfectly apt to say, as the panel did, that the court’s misinstruction “eviscerated the significance of [the] defendant’s character evidence” and, with it, his “best defense to th[e prosecution’s] evidence of his past violence,” 29 id. at 3, 5—thereby tainting the jurors’ consideration of that evidence in reaching a verdict on his guilt. 29 In criticizing the panel’s reasoning on this point, the majority calls out this language, but incompletely. It says the panel concluded “that ‘[d]efendant’s best defense . . . was his evidence regarding his peaceful character,’ ” when in fact, as noted above, the panel’s full statement was more targeted—that his “evidence regarding his peaceful character” was his “best defense to [the prosecution’s] evidence of his past violence” and was “rendered largely ineffective by the trial court’s failure to properly instruct the jury on its use.” Lyles (On Remand), unpub op at 5 (emphasis added). The panel elsewhere spoke more broadly regarding the defense’s strategy at trial, which was “to deny responsibility for the murder,” “argue[] that there was no evidence tying [him] to” it, and “bolster the credibility of [those] denials, and . . . refute the prosecution’s evidence of [his] past violence, by introducing evidence that [he] was a peaceful individual.” Id. at 4. I see no flaw in any of these characterizations. The majority also questions why, if the defendant’s character evidence were so important to his defense, his counsel didn’t raise it in her opening or closing arguments to the jurors. The record before us doesn’t make that clear. Maybe she was counting on the court’s jury instructions to do the important work with respect to this evidence, or had some other strategic reason. (If that were so, however, one might have expected an even more strenuous objection by counsel to the court’s failure to correct its misinstruction.) Or maybe—and as seems more likely, all things considered—she simply mishandled this key aspect of the defendant’s case, just like the trial court did. That is a question for another day (and one to be answered under a different standard of prejudice than what governs here, see Strickland v Washington, 466 US 668, 694; 104 S Ct 2052; 80 L Ed 2d 674 (1984)). In any event, I fail to see how counsel’s handling of this evidence in arguments, deficient or not, affects whether the evidence, on the record before us, did constitute the “best defense” the defendant had to the prosecution’s attack on his character, which was central to his trial. To the contrary, counsel’s failure to argue this evidence to the jurors only exacerbated the harm from the court’s failure to instruct them on the evidence and the potentially dispositive legal effect it could have on their determination of guilt. 26 But how much harm could there really have been from this? Not much, the majority says. The jurors, after all, heard and were permitted to consider the defendant’s character evidence; the error here was “only” a misinstruction about that evidence. Given their presumptive import to a verdict, however, jury instructions, and errors therein, can’t be brushed aside quite so easily; indeed, the harm from an instructional error can be so inherently pervasive as to categorically require a new trial, no matter how cleanly the evidence may have come in. See, e.g., People v Cain, 498 Mich 108, 117 n 4; 869 NW2d 829 (2015) (listing “defective reasonable-doubt instructions” among the errors that have been deemed to be structural). And here, while the jurors were presented with the defendant’s character evidence, they were given no legal guidance, no rule, regarding whether and how they could consider this evidence in reaching a verdict. More specifically, and most importantly, they were not told that the evidence could itself legally determine that verdict. This is no small thing. Attempting to make it smaller, the majority points to other instructions the jurors did receive—that “[a] reasonable doubt is a fair, honest doubt growing out of the evidence or the lack of evidence,” and that they, as jurors, must consider and weigh all the evidence in reaching a verdict. These are generic rules—fundamental and important, to be sure, but not necessarily adequate in themselves to meet the instructional needs of every case. If they were, it seems there would never be any need to provide, nor any harm from failing to provide, the many, more specific instructions that build them out— that inform the jury how they can consider and weigh the particular types of evidence in a given case, and how that evidence can and cannot prove guilt beyond a reasonable doubt. That’s the sort of information imparted by M Crim JI 5.8a(1), and neither these general 27 instructions, nor any others the jurors heard, provided it. We presume the jurors follow the rules, not that they already know them. And if anything, the rules the court did provide to the jurors only made the harm from this omission worse. The jurors did, after all, receive instructions akin to M Crim JI 5.8a(1); they simply were given only the ones that governed the prosecution’s proofs. The defendant, they were told, could be convicted based on the prosecution’s “identification testimony alone,” if they believed it proved his guilt beyond a reasonable doubt. His move out of state, on the other hand, couldn’t alone prove his guilt, but it could be taken to show his “consciousness of guilt.” And he could be convicted, they also learned, without any proof of motive. If they were looking for motive, however, they could consider the evidence of the defendant’s “improper acts” against Louise, which could also be considered to show “th[e] environment” Melvin was in, “the complete picture of the defendant’s history” in that regard, and—as suggested by the court’s subsequent misinstruction—“that the defendant did not have good character.” They were told to “carefully balance that testimony . . . so [as] not to prejudice yourself against this defendant and this murder,” but they were not told (per the model instructions) that they “must not convict the defendant here solely because you think [he] is guilty of other bad conduct.” As the majority notes, the propriety of these instructions (or, for that matter, of the evidence to which they pertained) isn’t before us here. But as part of the “entire cause,” their relationship with the instant error is. And these (mis)instructions made that error’s harm all the more palpable. Amidst all this exposition on the legal meaning and effect of the prosecution’s evidence, not a word about that of the defendant’s—as if his evidence had none. 28 No matter, says the majority, because “[e]ven if the jury had been properly instructed, it is not more likely than not that [the] defendant’s character evidence would have been sufficient to create a reasonable doubt, given the prosecution’s evidence[.]” The court’s misinstruction, however, cannot be so readily isolated from the jurors’ assessment of the proofs, given how fundamentally it compromised the jurors’ ability to weigh those proofs in reaching their verdict; they simply did not know that the defendant’s evidence could, legally speaking, be enough in itself to defeat the prosecution’s case. To have this legal effect, the defendant’s evidence only had to cast a reasonable doubt over the prosecution’s proofs in the eyes of one juror. Why could the defendant’s evidence not have done so? What “weight and strength,” to use Lukity’s terms, did the “untainted evidence” possess that prevented it? Let’s review that evidence: The recounted memories of a shadow in a dark house that reminded Louise’s teenage daughters, in that moment of violence, of the man who had replaced their father and beat their mother. The recounted memories of the girls’ frantic announcement of that suspicion to their neighbors right afterward. A few other memories about that night 30 years ago and the man’s suspected place within it, recounted solely by the daughter who hated him and ensured he was now being brought to trial. The recounted memories of a phone call before that night in which the defendant said he’d “get” Melvin—though no one could say why. The defendant’s flight from town at some point, either because he did “get” Melvin or because he feared the Kountzes were out to get him. Was this evidence enough, as a matter of law, to sustain the defendant’s conviction? We haven’t been asked that question, but perhaps. Does it leave ample room for a juror, upon hearing testimony about the defendant’s peaceful character, to 29 reasonably doubt his guilt? Absolutely. But, the majority continues, that good-character evidence was “extremely weak” and “far outweighed” by the bad-character evidence of the defendant’s abuse of Louise. I lack the majority’s confidence in that assessment. The prosecution certainly produced more pages of testimony on the topic than the defendant did, but that testimony does not strike me as clearly or inherently more objective or reliable. The jurors may very well have seen strength in the prosecution’s numbers and credited the accounts of abuse provided by Louise’s friends and family. They also may have discounted those memories, in whole or in part, as too old, too unsubstantiated, or too distorted by a long-festering contempt for the defendant and a desire that he be punished somehow, some way, for whatever he’d done to Louise. Likewise, the jurors may have found little import in Kim Harden’s memories of a man whom she knew long ago, but whom she didn’t really see with Louise all that much—or they may have found some objectivity and reliability in the opinions of this long-time school principal who had known the defendant very well but wasn’t mixed up in the strife between him and the Kountzes. The record doesn’t betray what the jurors, together or separately, might have thought about all this character evidence—just that they were only instructed by the court on how to think about the prosecution’s bad-character evidence in determining guilt. Twice. And even if the jurors did, as the majority does now, credit the prosecution’s character evidence over the defendant’s, what does that tell us about the harm to their verdict from the court’s misinstruction about that evidence? The jurors may have believed the defendant was a “vicious abuser” of Louise, or even a bad person overall, and a proper instruction on the defendant’s evidence may not have changed these 30 impressions. But the jurors were not tasked with rendering a verdict on any of that. Rather, to convict the defendant, they still had to get from this other-acts evidence, and any impressions of the defendant it conjured, to the unanimous conclusion that he was guilty, beyond a reasonable doubt, of murdering Melvin. The instructions traced the legal paths the jurors could take through the prosecution’s proofs to that conclusion. We don’t, and can’t, know which path they each may have taken. But we do know none of them was told that the defendant’s evidence could provide its own path to a different outcome. And what assurance is there in the record that, had the jurors been apprised of this path— provided it as a rule—none would have followed it? That, ultimately, is the core of our inquiry here: how reliably can we say, on the record before us, that these 12 jurors would have unanimously agreed on the defendant’s guilt for murdering Melvin, had the court not misinstructed them on the reasonable doubt that could, as a matter of law, arise from the defendant’s character evidence alone. We, of course, are not those jurors, and harmlessness review does not require or invite us to be; what matters is the reliability of their verdict, not our own agreement with it. 30 Did 30 See, e.g., People v Mateo, 453 Mich 203, 221; 551 NW2d 891 (1996) (discussed approvingly in Lukity, and explaining, inter alia, that “courts analyzing preserved error in terms of their view regarding whether the defendant is guilty have been wrong,” as “[t]he defendant’s right to a fair trial by jury requires that preserved error be reviewed in terms of its effect on the factfinder”). Accord Sullivan v Louisiana, 508 US 275, 279; 113 S Ct 2078; 124 L Ed 2d 182 (1993) (“Harmless-error review looks, we have said, to the basis on which the jury actually rested its verdict. The inquiry, in other words, is not whether, in a trial that occurred without the error, a guilty verdict would surely have been rendered, but whether the guilty verdict actually rendered in this trial was surely unattributable to the error. That must be so, because to hypothesize a guilty verdict that was never in fact rendered—no matter how inescapable the findings to support that verdict might be—would violate the jury-trial guarantee.”) (quotation marks and citation omitted). 31 the defendant kill Melvin? I don’t know. Will a new set of jurors, properly instructed, conclude beyond a reasonable doubt that he did? I don’t know that, either. But even if I knew those things, they wouldn’t answer the question before us. Here is what I do know—what affirmatively appears to me, after examining this entire cause: The jurors were asked to assess the defendant’s guilt for a 30-year-old murder on the basis of unsubstantiated memories and character inferences drawn from sources of questionable reliability. These proofs teemed with doubt—a doubt which the defendant encouraged through the presentation of his own character evidence. The court provided the jurors with rules they must, and presumptively did, follow in measuring that doubt and deciding whether they could, as a matter of law, still convict the defendant in spite of it. But the court only provided rules for the prosecution’s proofs. It did not instruct the jury that the defendant’s proofs—his evidence of good character, the only evidence he presented—could, alone and as a matter of law, create a reasonable doubt and foreclose a finding of guilt. Instead, the jurors were reminded again about the prosecution’s evidence on that central issue at trial while the defendant’s proofs went wholly without mention, as if they didn’t warrant it, legally or factually. Did this failure to inform the jurors of the potentially dispositive legal effect of the defendant’s evidence undermine the reliability of their guilty verdict? Is it more probable than not that, had this rule been given to the jurors, it would have made a difference to at least one of them in how they viewed that evidence? That one of them would have realized that, when it came to their legal determination of guilt, the defendant’s evidence could be worth not only something, but everything? That the doubt it generated could be, in itself, the “reasonable” sort countenanced by the law to foreclose a finding of guilt? And that, 32 because of this, the defendant—for whatever else he did or whatever else he was—should and could not be convicted under the law for the 1983 murder of Andrew “Melvin” Weathers? On these proofs, in this cause, I say yes. The court’s error was not harmless. A new trial is needed to ensure that any verdict on this charge is reliable under the law. 31 This was also the conclusion of the Court of Appeals panel. Its work on this question was careful and correct. It did what we asked, and well. We should leave its work in place. We should allow a new trial to go forward. We should deny leave. I dissent. Bridget M. McCormack David F. Viviano Richard H. Bernstein 31 Indeed, if the court’s error as to M Crim JI 5.8a(1) doesn’t require relief in this case, I struggle to think of a case in which such an error would. At oral argument, the prosecutor was similarly stumped on this point. Of course, that wasn’t of particular concern to the prosecutor, as she had urged us—in both of this case’s trips to this Court— to dispense with the instruction entirely, as legally unnecessary and improper. Both times, we declined this invitation. The model instruction comports with the law in this state, and a party is entitled to have it presented to the jury when the evidence supports it. The court’s failure to provide it here was indisputably erroneous and subject to review for harmlessness. But, with the majority’s ruling here, I am not sure if any force is left to that legal conclusion, or if this Court now has effectively done, through harmlessness review, what the prosecutor really wanted us to do. What meaning is there to assigning error, after all, if there can never be any consequence to that error? Or, as this Court has elsewhere wondered, “What is the point in applying a harmless error rule if the error is always going to be harmless?” People v Francisco, 474 Mich 82, 91 n 10; 711 NW2d 44 (2006). 33