Commonwealth v. Shepherd

NOTICE: All slip opinions and orders are subject to formal revision and are superseded by the advance sheets and bound volumes of the Official Reports. If you find a typographical error or other formal error, please notify the Reporter of Decisions, Supreme Judicial Court, John Adams Courthouse, 1 Pemberton Square, Suite 2500, Boston, MA, 02108-1750; (617) 5571030; SJCReporter@sjc.state.ma.us SJC-12405 COMMONWEALTH vs. RASHAD SHEPHERD. Essex. November 7, 2023. - February 22, 2024. Present: Budd, C.J., Gaziano, Kafker, & Wendlandt, JJ. Homicide. Felony-Murder Rule. Retroactivity of Judicial Holding. Practice, Criminal, Retroactivity of judicial holding, Instructions to jury, Argument by prosecutor, Questioning of witness by judge, Assistance of counsel, Capital case. Constitutional Law, Equal protection of laws. Evidence, Argument by prosecutor, Questioning of witness by judge, Hypothetical question. Jury and Jurors. Cellular Telephone. Indictment found and returned in the Superior Court Department on December 18, 2014. The case was tried before Richard E. Welch, III, J.; a motion for a new trial, filed on March 15, 2019, was considered by Timothy Q. Feeley, J.; a second motion for a new trial, filed on September 10, 2020, was heard by Kathleen M. McCarthy-Neyman, J.; and a third motion for a new trial, filed on February 7, 2022, was considered by her. Claudia L. Bolgen for the defendant. Kathryn L. Janssen, Assistant District Attorney, for the Commonwealth. Duke K. McCall, III, & Kayla Stachniak Kaplan, of the District of Columbia, Caitlin Glass, Joshua M. Daniels, & Vanessa M. Brown, for Boston University Center for Antiracist Research & others, amici curiae, submitted a brief. — 2 — Jessie J. Rossman & Isabel Burlingame, for American Civil Liberties Union of Massachusetts, Inc., amicus curiae, submitted a brief. WENDLANDT, J. In August 2014, Terrence Tyler, Monique Jones, and the defendant, Rashad Shepherd, hatched a plan to rob the victim, Wilner Parisse. The scheme involved Jones, who had a sexual relationship with the victim and frequently purchased marijuana from him, proposing a sexual tryst as a ruse to lure the victim into a vulnerable position, allowing Tyler and the defendant to enter the victim's apartment and to take the stash of marijuana they knew he kept in his bedroom closet. But in the early morning of August 16, 2014, when the three coventurers set their plot in motion, the victim was not the "easy mark" they had anticipated; he fought back. In the ensuing melee, the victim was shot once in the chest and killed. Based on the bullet's trajectory and Jones's retelling of the events, the prosecution theorized that the defendant was the shooter. Following a jury trial in April 2016, at which Jones testified pursuant to a cooperation agreement, the defendant was convicted of murder in the first degree on the theory of felony-murder, with attempted unarmed robbery as the predicate felony. He was sentenced to life without the possibility of parole. In this consolidated appeal, the defendant contends that our decision in Commonwealth v. Brown, 477 Mass. 805 (2017), — 3 — cert. denied, 139 S. Ct. 54 (2018), in which we abolished felony-murder as an independent theory of liability for murder in the first and second degrees, should extend to the defendant's case retroactively, despite our determination in Brown to apply our holding only prospectively -- a conclusion we have reaffirmed eight times. The defendant maintains that the determination to apply Brown only prospectively violates the equal protection principles of arts. 1 and 10 of the Massachusetts Declaration of Rights because the data show, inter alia, that use of felony-murder as an independent theory of liability for murder in the first degree disproportionately resulted in the incarceration of Black persons and that, as a result, more Black persons than white persons currently are serving a sentence of life without the possibility of parole for felony-murder. The defendant further urges that the trial judge gave erroneous jury instructions, that the judge's questioning of, and interactions with, certain witnesses biased the jury, and that he received ineffective assistance of counsel. Finally, the defendant asks this court to exercise its extraordinary authority pursuant to G. L. c. 278, § 33E, to grant him a new trial or to reduce the conviction to a lesser degree of guilt. Having carefully examined the record and — 4 — considered the defendant's arguments, we conclude that there is no reversible error and find no reason to disturb the verdict.1 1. Facts. a. The Commonwealth's case. The following facts are supported by the evidence presented at trial. i. Background. The victim shared an apartment on the second floor of a three-story apartment building in Lynn with his roommate and their two dogs. The victim sold marijuana from the apartment, including to Jones. The relationship between the victim and Jones had become sexual approximately six months prior to the shooting. The victim sold marijuana to Jones at a discount, and occasionally, Jones, who was unemployed, resold the marijuana at a profit. Jones and Tyler had known each other for at least a decade. They had previously dated and remained very close.2 In early August 2014, prior to the killing, Tyler had accompanied Jones to the victim's apartment; Tyler remained in 1 We acknowledge the briefs of amici curiae Boston University Center for Antiracist Research, Massachusetts Association of Criminal Defense Lawyers, Families for Justice as Healing, Felony Murder Elimination Project, Fred T. Korematsu Center for Law and Equality, National Council for Incarcerated and Formerly Incarcerated Women and Girls, Kat Albrecht, and The Sentencing Project; and American Civil Liberties Union of Massachusetts, Inc. 2 At the time of the shooting, Jones was dating Tyler's brother "D." Jones's sister had previously dated another of Tyler's brothers, Reginald Tyler, who was deceased when the shooting occurred. — 5 — Jones's vehicle while she purchased marijuana. After the sale, Tyler remarked that the victim would be "easy to rob," but Jones "brushed off" the comment. Tyler pressed the idea of robbing the victim several times thereafter, disclosing to Jones that Tyler had robbed the victim several years earlier. ii. The night before the shooting. At around 5 or 6 P.M. on August 15, 2014, the day preceding the shooting, Jones began drinking alcohol with a friend, who arrived at Jones's home in Lynn already intoxicated.3 Tyler called Jones to "hang out," and at approximately 11 P.M., Jones, accompanied by her friend, drove a rental vehicle to pick up Tyler and the defendant. Tyler and the defendant were friends. Jones had known the defendant for about four or five years, but she was not as close with the defendant as with Tyler. The four went to a restaurant in Lynn, where they would remain until approximately 1 A.M. When they arrived, Jones's friend went inside the restaurant, leaving Jones, Tyler, and the defendant in the vehicle. Tyler again broached the topic of 3 Jones also had smoked marijuana and later that evening would consume a few Percocet pills. — 6 — robbing the victim, emphasizing that it would be an "easy job"; this time, Jones agreed.4 Tyler suggested exploiting Jones's sexual history with the victim. They agreed that Jones would propose that she meet the victim at his apartment for a promised sexual tryst. Then, while the victim was in a vulnerable position, Tyler and the defendant would enter the apartment and take the victim's marijuana cache, which Jones knew he kept in his bedroom closet. The defendant was present during the formation of the scheme, but he remained silent. As agreed, Jones contacted the victim by text message, and she exchanged a series of text messages with him between 11:04 P.M and 1:03 A.M. Some of these text messages were drafted by Tyler, pretending to be Jones. Jones, or Tyler on her behalf, proposed a sex act, and the victim invited her to his apartment.5 Surveillance video footage from the restaurant shows the three coventurers there that evening; the defendant did not dispute that he was at the restaurant. The footage captures 4 Jones explained that she was "having a bad day," and was "aggravated" and "stressed" because several of her friends had been arrested and she had been blamed. 5 Previously that evening, the victim, his roommate, and the roommate's six year old son were at the apartment, playing a board game until approximately 10 or 11 P.M. The roommate and his son retired into the roommate's bedroom and fell asleep shortly thereafter. — 7 — Tyler, who wore his hair in long dreadlocks, entering the restaurant just prior to 12:15 A.M. The defendant, who wore a baseball cap, a light-colored hooded sweatshirt, darker pants, and light-colored sneakers, entered the restaurant shortly after Tyler. Jones entered the restaurant at approximately 12:26 A.M., and at 12:35 A.M., the defendant and Jones engaged in a conversation. The footage shows Jones and the defendant walking away from the restaurant together at 12:39 A.M. The prosecution introduced cell site location information (CSLI) data, which indicated that, at 12:42 A.M., the defendant's cellular telephone connected to a cellular tower covering an area that included the restaurant. Telephone records show that the victim sent Jones a text message at 1:03 A.M., apparently perturbed that Jones had not yet arrived. In response, the defendant and Jones called Tyler four times between 1:08 and 1:12 A.M. Shortly thereafter, Tyler rejoined the defendant and Jones, and the three coventurers, along with Jones's friend, got into Jones's vehicle. iii. The botched robbery. After leaving the restaurant, Jones, Tyler, Jones's friend, and the defendant drove to the victim's apartment and parked nearby. While Jones's friend, who was intoxicated, was asleep in the front passenger seat, the three coventurers rehashed the plan. — 8 — After exchanging telephone calls with the victim at 1:15 and 1:22 A.M., Jones then left Tyler, the defendant, and her slumbering friend in the vehicle. Tyler and the defendant had planned to wait in the vehicle for twenty minutes to allow Jones time to execute the first stage of their plot. Jones entered the exterior door of the victim's apartment building. She climbed the back staircase leading to the back door of the victim's apartment, which led to the kitchen. She left the doors unlocked. To her surprise, she found the victim already partially undressed in his bedroom, which was located off the kitchen. She stalled to give Tyler and the defendant time to execute the next stage of the plan. Jones excused herself to the bathroom, which was located adjacent to the kitchen. Call logs show that she placed a telephone call to Tyler at approximately 1:32 A.M.; Tyler told Jones that he and the defendant were on their way. The surveillance video footage, while grainy, appears to capture two men, dressed like the defendant and Tyler had been in the restaurant surveillance video footage, waiting outside a vehicle.6 It also shows that, at approximately 1:35 A.M., the 6 One man, inferably the defendant, is wearing a bulky light-colored top, light-colored shoes, and darker pants. The other man, inferably Tyler, has longer hair, light-colored shoes, and patterned pants. The appearance of the two men is consistent with the appearance of the defendant and Tyler in the restaurant footage. — 9 — two men cross the street in the direction of the victim's apartment, consistent with Jones's testimony concerning the scheme and its execution. The footage shows the defendant making movements that the prosecutor suggested indicated that he was "securing a gun in his waistband." Meanwhile, in the apartment, Jones returned to the bedroom. The victim locked the bedroom door behind her. Realizing the locked door would stymie the plan to take the victim's marijuana stashed in his bedroom closet, at 1:36 A.M. Jones sent Tyler a text message: "He just locked the door. So I'm[] [g]oing to act like [I] have a play[.] Wait." Jones asked the victim to get her a drink, and when he opened the bedroom door, he encountered Tyler. Tyler and the victim immediately began fighting in the kitchen. The defendant stood at the threshold of the back door, watching. Grappling and exchanging blows with the victim, Tyler pushed the victim back into the bedroom, and they crashed into a dresser.7 The victim grabbed a baseball bat and swung it at Tyler, who retreated to the kitchen, as the victim advanced. In the kitchen, Tyler charged the victim, tackling him to the floor. In the ensuing scrum, the victim bit Tyler's finger, and Tyler screamed for the defendant to help. 7 Jones was sitting on the bed at this time. — 10 — Jones grabbed her clothes and pocketbook and ran from the bedroom, past the men fighting in the kitchen, and into the bathroom. Moments later, she heard "one or two" gunshots.8 Leaving the bathroom, Jones found the victim lying on the kitchen floor; he was bleeding. She saw Tyler fleeing out the back door. At trial, based on the bullet's trajectory and Jones's testimony that the defendant had been standing by the back door, the Commonwealth's theory was that the defendant had fired the gun, killing the victim. Jones also fled. She gathered her belongings and ran to her vehicle; in her panic, however, she left her cellular telephone on the victim's bed. She drove some distance, and then stopped. She evicted her friend9 from the vehicle. At that time, Tyler approached Jones's vehicle; his hand was bleeding from the bite wound the victim had inflicted. The two fled to Boston. Tyler's blood, confirmed by deoxyribonucleic acid analysis, subsequently was found on the 8 Physical evidence showed that the victim died of wounds from a single bullet, and officers recovered only one bullet and shell casing. See discussion infra. 9 By then, the friend had finally roused from her stupor and asked Jones what had transpired. The friend, whom Jones had testified was intoxicated during the relevant events, could not be located and did not testify. — 11 — exterior handle of the rear passenger's side door and on the interior driver's side door frame of the vehicle. Call logs show that the defendant spoke with Tyler by cellular telephone at approximately 1:44 A.M., shortly after the shooting. CSLI data indicated that the defendant's telephone connected to a cellular tower covering an area that included the victim's residence when he placed this call to Tyler. In the next two hours, as call logs show, the defendant placed three unsuccessful telephone calls to Jones, whose cellular telephone was still at the victim's apartment. He also placed several calls to Jones's friend and to Tyler. iv. The aftermath and investigation. At around 1:45 A.M., the victim's neighbor placed a 911 call, reporting a shooting, and Lynn police department officers were dispatched to the area. Around this time, the victim's roommate awoke to the sound of his and the victim's dogs10 barking. He found the victim lying in a pool of blood on the kitchen floor and flagged down one of the responding officers. Officers entered the apartment and unsuccessfully administered first aid to the victim. Minutes later, responding emergency medical technicians pronounced the victim dead. 10The victim and the roommate kept their two dogs in a spare bedroom when they entertained guests. — 12 — Officers located Jones's cellular telephone on the victim's bed, which at 1:51 A.M. showed an incoming call from a caller identified as "City," the defendant's nickname. Officers identified a spent cartridge casing in the hallway by the back door of the victim's apartment, the location where, according to Jones, the defendant had been standing just prior to the shooting. A bullet also was recovered; subsequent analysis showed that the bullet had passed through the victim's chest, aorta, and left lung, killing him within seconds. The bullet then exited the victim's body, crossed the kitchen, passed through a window screen, and lodged into a neighboring building. The bullet's path was consistent with the firearm being discharged from the back door where Jones had testified the defendant was standing. No identifiable prints were recovered from the scene or from the cartridge casing, and the firearm was not recovered. By 8 P.M. that day, August 16, Jones had learned that officers wanted to interview her; she complied, arriving intoxicated at the police station. She told officers that she had been in bed with the victim when three masked white men had entered and shot the victim. When it became apparent to Jones — 13 — that the officers found her story to be not credible, she terminated the interview.11 By early September 2014, Jones retained counsel and recanted her story. She reported instead that she, Tyler, and the defendant had conspired to rob the victim. She entered into a cooperation agreement in which she agreed, inter alia, to testify at the defendant's trial;12 in exchange, prosecutors agreed to recommend that she receive a sentence of from five to seven years for her role in the victim's killing. In October, the defendant was arrested in Boston, and later Tyler was apprehended in Florida.13 b. The defendant's case. The defense centered on attacking Jones's credibility and intimating that the third coventurer was not the defendant. The defendant did not testify; instead, the defense relied primarily on crossexamination, casting Jones as a "coldhearted killer," 11Later, she told Tyler that she thought the officers did not believe her tale. 12She also agreed to testify before the grand jury and at Tyler's trial. 13 On a recorded call that the defendant placed to his then girlfriend from jail on October 30, 2014, the defendant learned of Tyler's arrest. The defendant told the girlfriend that Jones and Tyler were going "to blame this whole shit on me," and that "[Tyler] shouldn't have even went on the run in the first place." — 14 — "unemployed . . . drug dealer," and unreliable narrator who needed to rob the victim to fund her "lifestyle."14 The defense also maintained that because Jones and Tyler were much closer with each other than with the defendant, he was too far removed from them to be brought into their scheme and that Jones concocted the defendant's involvement to secure a deal, which itself gave Jones a motive to lie in exchange for a lighter sentence. See note 15, infra. The defense additionally attempted to undermine the prosecution's forensic evidence, in particular the CSLI analysis, emphasizing the limitations of the technology to locate precisely a cellular telephone. The defense also presented testimony from one of the victim's neighbors. The neighbor testified that, on the night of the shooting, he heard arguing between a man and a woman in the victim's apartment, and one or two gunshots; thereafter, he saw a woman fleeing the scene but did not see the defendant. 2. Procedural history. In December 2014, a grand jury indicted the defendant on charges of murder in the first degree, G. L. c. 265, § 1; home invasion, G. L. c. 265, § 18C; and armed 14The defense elicited testimony that Jones had no income but had substantial expenses. The defense used this information to paint Jones as a drug dealer with a motive to rob the victim and also to suggest that she was the shooter. — 15 — assault with intent to rob, G. L. c. 265, § 18 (b).15 Following a jury trial in April 2016, the defendant was convicted of felony-murder in the first degree with attempted unarmed robbery as the predicate felony; he was acquitted of the other charges. The trial occurred prior to our September 2017 decision in Brown. In March 2019, the defendant filed a motion for a new trial, arguing that he received ineffective assistance of counsel. The defendant did not provide an affidavit from trial counsel. He presented an affidavit from a CSLI expert, who raised questions regarding the reliability of the CSLI evidence presented at trial.16 The motion was denied in October 2019 by a judge (second judge) who was not the trial judge, the trial judge having retired. In September 2020, the defendant filed a second motion for a new trial, claiming, inter alia, ineffective assistance of counsel because trial counsel did not use certain information to impeach Jones's credibility. Again, the defendant did not 15Following a separate jury trial, Tyler was convicted of felony-murder in the first degree in March 2016, and received a mandatory sentence of life in prison. His appeal is pending before this court. See Commonwealth vs. Tyler, SJC-12836. Jones received a sentence of from five to seven years as part of her cooperation agreement. 16The defendant also submitted his own affidavit and an affidavit from the defendant's original appellate counsel. — 16 — submit an affidavit from trial counsel. Following a nonevidentiary hearing, the motion was denied in June 2021 by a third judge. In February 2022, the defendant filed a third motion for a new trial, arguing, inter alia, that the decision not to apply Brown retroactively violated equal protection principles. In August 2022, the third judge denied this motion. The defendant's timely appeals from the denials of his motions were consolidated with his direct appeal. 3. Discussion. In this consolidated appeal, the defendant raises four categories of claimed errors, which we address in turn. "We review the defendant's consolidated appeal pursuant to G. L. c. 278, § 33E, assessing preserved issues according to the appropriate constitutional or common-law standard and unpreserved issues for a substantial likelihood of a miscarriage of justice." Commonwealth v. Fernandes, 492 Mass. 469 , 474 (2023). In analyzing the denial of a motion for a new trial, we examine the motion judge's conclusions "to determine whether there has been a significant error of law or other abuse of discretion" (citation omitted). Id. at 474-475 . Where, as here, the motion judges did not preside at trial and did not conduct evidentiary hearings, "we regard ourselves in as good a position as the motion judge[s] to assess the trial record" — 17 — (citation omitted). Commonwealth v. Kirkland, 491 Mass. 339 , 346 (2023). a. Retroactive application of Brown. On appeal, the defendant first maintains that principles of equal protection embodied in the Massachusetts Declaration of Rights require that our decision in Brown, in which we abolished felony-murder as an independent theory of criminal liability, be applied to his conviction retroactively.17 i. Brief background of felony-murder. Until 2017, Massachusetts recognized the doctrine of felony-murder as "an independent theory of liability for murder," permitting a defendant to be convicted of murder in the first or second degree without requiring that the jury also find that the defendant acted with malice. See Brown, 477 Mass. at 807-808 . Instead, the felony-murder doctrine imposed "criminal liability 'on all participants in a certain common criminal enterprise if a death occurred in the course of that enterprise.'" Id., at 822 , quoting Commonwealth v. Watkins, 375 Mass. 472, 486 (1978), 17The defendant does not assert arguments under the Federal Constitution. "Our 'review of an equal protection claim under the Massachusetts Constitution is generally the same as the review of a Federal equal protection claim, . . . although we have recognized that the Massachusetts Constitution is, if anything, more protective of individual liberty and equality than the Federal Constitution.'" Commonwealth v. Roman, 489 Mass. 81 , 86 (2022), quoting Commonwealth v. Freeman, 472 Mass. 503 , 505 n.5, (2015). 18 S.C., 486 Mass. 801 (2021). "'The effect of the felony-murder rule,' both for principals and accomplices, '[was] to substitute the intent to commit the underlying felony for the malice aforethought required for murder.'" Brown, supra at 822-823 , quoting Commonwealth v. Hanright, 466 Mass. 303, 307 (2013). In Brown, we abrogated felony-murder as an independent theory of liability. Although the felony-murder rule was constitutional, Brown, 477 Mass. at 807 , a majority of the court concluded that the doctrine was of "questionable" historical provenance, that developments in our joint venture and constructive malice jurisprudence had undermined the common-law pillars of the doctrine, and that the doctrine "erode[d] 'the relation between criminal liability and moral culpability,'"18 id. at 826-833 (Gants, C.J., concurring), quoting Commonwealth v. Matchett, 386 Mass. 492 , 503 n.12, 507 (1982). After Brown, a felony-murder conviction requires proof of actual malice;19 18We limited felony-murder to its "statutory role under G. L. c. 265, § 1, as an aggravating element of murder, permitting a jury to find a defendant guilty of murder in the first degree where the murder was neither premeditated nor committed with extreme atrocity or cruelty but was committed in the course of a felony punishable by life imprisonment." Brown, 477 Mass. at 825 (Gants, C.J., concurring). In this opinion, we use the term "felony-murder" to refer to the pre-Brown, independent theory of liability unless otherwise indicated. 19The Commonwealth must now show "one of the three prongs of malice: that [the defendant] intended to kill or to cause grievous bodily harm, or intended to do an act which, in the circumstances known to the defendant, a reasonable person would — 19 — constructive malice inferred from commission of the predicate felony no longer suffices. See Brown, supra at 825 (Gants, C.J., concurring). The new rule, we determined, would apply only to trials commenced after our decision in Brown, recognizing that "a felony-murder case might have been tried very differently if the prosecutor had known that liability for murder would need to rest on proof of actual malice." Brown, 477 Mass. at 834 (Gants, C.J., concurring).20 Since then, we have declined to apply our decision retroactively on at least eight occasions, including once in the face of an equal protection challenge. See Commonwealth v. Pfeiffer, 492 Mass. 440 , 453-454 (2023); Commonwealth v. Cheng Sun, 490 Mass. 196 , 224 (2022); Commonwealth v. Duke, 489 Mass. 649 , 658 n.5 (2022); Commonwealth v. Tate, 486 Mass. 663 , 674 (2021); Commonwealth v. Chesko, 486 Mass. 314 , 326-327 (2020); Commonwealth v. Martin, 484 Mass. 634 , 644-646 (2020), cert. denied, 141 S. Ct. 1519 (2021) (equal protection); Commonwealth v. Bin, 480 Mass. 665 , have known created a plain and strong likelihood that death would result." Brown, 477 Mass. at 825 (Gants, C.J., concurring). 20In Brown, we also exercised our discretion, pursuant to G. L. c. 278, § 33E, to reduce the defendant's conviction to murder in the second degree because he was involved in only the "remote outer fringes" of the crime. Brown, 477 Mass. at 824 , quoting Commonwealth v. Rolon, 438 Mass. 808, 824 (2003). — 20 — 681 (2018); Commonwealth v. Phap Buth, 480 Mass. 113, 120 , cert. denied, 139 S. Ct. 607 (2018). ii. Equal protection. Because the defendant was tried prior to our decision in Brown, its holding did not apply to him; instead, his trial proceeded under the felony-murder rule, which as we stated supra, was constitutional. See Brown, 477 Mass. at 807 . He asks us to revisit our decision to apply Brown only prospectively, contending that the court's decision not to apply Brown retroactively offends the guarantees of equal protection. In support of his argument, the defendant, who is Black, relies on the racial and ethnic demographics of individuals currently serving life without the possibility of parole for felony-murder.21 Specifically, he asserts that Black persons are overrepresented in the population of those serving life without the possibility of parole for felony-murder when compared to the population of white persons serving the same sentence. According to the data collected by the defendant's appellate counsel,22 of the 108 inmates currently incarcerated for murder in the first degree on a felony-murder theory, 59.25 percent are 21The Commonwealth does not "take issue with the tenor or the accuracy of the defendant's statistics per se," although it points to several methodological flaws. 22 These data are from December 1, 2021. — 21 — Black, while 17.59 percent are white. By contrast, the data show that 32.51 percent of those serving life without the possibility of parole for murder based on a malice theory are Black persons, while 43.65 percent are white persons. Thus, the defendant calculates that "more than three times as many Black people . . . are sentenced to first-degree felony murder as compared to [w]hite people," while "[r]oughly 1.34 times as many [w]hite people . . . are sentenced to first-degree malice murder as compared to Black people." The data further show that, of all Black persons serving life without the possibility of parole, eighteen percent are doing so because of a conviction of murder in the first degree on a theory of felony-murder; by comparison, of all white persons serving life without the possibility of parole, only 4.6 percent are doing so for murder in the first degree on a theory of felony-murder. And, while Black persons comprise 29.9 percent of the total population serving any sentence at Department of Correction (DOC) facilities,23 they comprise 59.25 percent of those serving life without the possibility of parole for felony-murder; by comparison, white persons comprise forty percent of the total DOC population and 17.5 percent of those 23These data, supplied by the defendant, on the total "criminally sentenced persons" are from May 1, 2022. — 22 — serving life without the possibility of parole for felonymurder.24 The data, the defendant contends, evince structural racism, racial disparities in prosecutors' use of discretion in charging decisions and plea offers, and implicit bias. He urges us to apply the decision in Brown to his case to correct these societal and prosecutorial ills. We review the defendant's constitutional challenge de novo. See Fernandes, 492 Mass. at 479. To begin, the decision in Brown comports with equal protection's essential mandate that 24The defendant also presents data that show that 82.4 percent of those serving life without the possibility of parole for felony-murder are people from historically disadvantaged racial and ethnic groups. Additionally, 56.34 percent of those serving life without the possibility of parole for murder based on a malice theory are from historically disadvantaged racial and ethnic groups. From these data, the defendant concludes that "people of color" are roughly 1.5 times overrepresented in the felony-murder population serving life without the possibility of parole as compared to the demographic percentage breakdown of the malice murder population serving life without the possibility of parole. He also compares the racial and ethnic makeup of those serving life without the possibility of parole for felony-murder with that of the over-all population of incarcerated persons, of whom sixty percent are from historically disadvantaged groups; 29.9 percent are Black, and forty percent are white. We have noted that such "lump[ing] together" of members of various racial and ethnic groups is not proper when conducting an equal protection analysis (citation omitted). See Commonwealth v. Lopes, 478 Mass. 593 , 600 n.5 (2018). See also Commonwealth v Prunty, 462 Mass. 295 , 307 n.17 (2012); Gray v. Brady, 592 F.3d 296, 305-306 (1st Cir.), cert. denied, 561 U.S. 1015 (2010) ("minorities," "African Americans," and "Hispanic" jurors not same "cognizable group"). The defendant presents no argument to deviate from our prior jurisprudence in this regard. — 23 — "all persons similarly situated should be treated alike." Moore v. Executive Office of the Trial Court, 487 Mass. 839 , 848 (2021), quoting Doe v. Acton-Boxborough Regional Sch. Dist., 468 Mass. 64, 75 (2014). This is because our decision to apply Brown only prospectively treated all persons serving life without the possibility of parole for felony-murder alike -- that is, regardless of race or ethnicity (or other suspect classification) none of those incarcerated for felony-murder received the benefit of our abolishment of the felony-murder doctrine. Such a "neutral" decision, even if it "'has a disproportionately adverse effect upon a racial minority[,]' is unconstitutional 'only if that impact can be traced to a discriminatory purpose.'" Commonwealth v. Grier, 490 Mass. 455 , 469 (2022), quoting Personnel Adm'r of Mass. v. Feeney, 442 U.S. 256, 272 (1979). Discriminatory purpose requires that the State "selected or reaffirmed a particular course of action at least in part 'because of,' not merely 'in spite of,' its adverse effects upon an identifiable group." Feeney, supra at 279 . No such discriminatory purpose underlies the decision to apply Brown only prospectively. More specifically, in Brown, we recognized that the abolition of the felony-murder doctrine "clearly involved a change in the common law of felony-murder." Martin, 484 Mass. at 645. We also affirmed that the pre-Brown — 24 — felony-murder rule itself was constitutional. See Brown, 477 Mass. at 807 . Accordingly, as we have explained, because there was no constitutional requirement that the new rule be applied retroactively, "we [were] free to declare that our new substantive law shall be applied prospectively." Martin, supra. Cf. Commonwealth v. Galvin, 466 Mass. 286, 290 (2013), superseded on other grounds as recognized by Commonwealth v. Beverly, 485 Mass. 1 , 5 (2020) (newly enacted penal statute is presumptively prospective and repeal of statute shall not affect any punishment incurred before repeal takes effect). And, as discussed infra, the decision was not arbitrary. Nor did our decision to apply Brown only prospectively burden a fundamental right. The defendant has no right, fundamental or otherwise, to retroactive application of new common-law rules, so long as the rule pursuant to which he was convicted was, as here, constitutional. See Martin, 484 Mass. at 645. And, while in some sense the decision not to apply Brown retroactively touches on a liberty interest (to be free of the physical constraint of incarceration), a fundamental right is burdened "only where State action significantly interfere[s] with the fundamental right at issue, not simply where State action involves a fundamental right" (quotations and citation omitted). Commonwealth v. Roman, 489 Mass. 81 , 86 (2022) (concluding fundamental right to be free from physical restraint — 25 — implicated but not interfered with where statute granted criminal defendants in District Court procedural defenses not available to defendants in Superior Court). We may prospectively change "our substantive common law of murder . . . much like the Legislature may do when it revises substantive criminal statutes." Martin, 484 Mass. at 645. "All prospective [law making] must have a beginning date, and . . . [t]he mere fact that some persons were at some later date governed by a law more favorable to them than the law which applied to the defendant is insufficient to strike down an otherwise valid [law]" (quotations and citation omitted). Commonwealth v. Freeman, 472 Mass. 503, 507 (2015). To conclude otherwise "would be either to eradicate all new [laws] or to make them all retroactive." Commonwealth v. Purdy, 408 Mass. 681, 685 (1990). Because the determination to apply Brown only prospectively was not borne out of discriminatory animus and neither implicates a fundamental right nor draws a suspect classification, it would violate equal protection only if it were not "rationally related to the furtherance of a legitimate [S]tate interest" (citation omitted). Roman, 489 Mass. at 86. Our decision to apply Brown only prospectively readily passes rational basis review. We reasoned that prosecutors might have tried felony-murder cases very differently if proof of actual — 26 — malice were then a required element. See Brown, 477 Mass. at 834 (Gants, C.J., concurring). See also Pfeiffer, 492 Mass. at 453; Martin, 484 Mass. at 645-646 (reaffirming wisdom of prospective application of Brown and noting unfairness of retroactive application where defendant was shooter and jury were not instructed that they had to find malice, but "likely would have found that the defendant acted with malice"). For this reason, and because the pre-Brown rule was constitutional, we determined not to apply Brown retroactively. Such reasoning continues to be valid. To be sure, the data show that the existing population of persons serving life without the possibility of parole for felony-murder convictions is comprised of more Black persons than white persons. Perforce, any prospective narrowing of the crime's scope would leave a population of inmates that was comprised of more Black persons than persons who are white. The defendant does not allege that we made our decision in Brown prospective because of this effect. Nonetheless, the defendant urges us to revisit our equal protection jurisprudence to allow for "disparate impact alone" to constitute an equal protection violation. The defendant calls on us to correct structural racism, prosecutorial discretion in charging decisions, and implicit bias that the defendant contends results in more Black persons than white — 27 — persons serving life without the possibility of parole for felony-murder by reversing course and applying Brown retroactively. He urges: "This [c]ourt has the opportunity to redress part of the systemic racism and implicit bias within the court system that has resulted in the egregious racial disparity in persons serving felony murder [life without the possibility of parole]." In other words, the defendant urges us to apply Brown retroactively because of race. Far from showing that our decision resulted in disparate racial treatment, however, the data demonstrate that our decision eliminated a theory of first-degree murder that may have disproportionately affected Black persons.25 Given the 25Notably, the data are not supported by analysis from an expert, such as a statistician, who might provide the court with an assessment of the data's statistical significance. See, e.g., Lopez v. Commonwealth, 463 Mass. 696 , 712 n.20 (2012), quoting Tinkham, The Uses and Misuses of Statistical Proof in Age Discrimination Claims, 27 Hofstra Lab. & Employment L.J. 357, 358 (2010) ("Standard statistical analysis in discrimination cases generally takes the unprotected group and compares the treatment of that group to the treatment of the protected group to determine whether there is a statistically significant difference. . . . Differences, if any, can be measured in terms of absolute numbers, standard deviations or percentages"); Jones v. Boston, 752 F.3d 38, 43-45 (1st Cir. 2014) (noting explanatory power of expert analysis of statistical significance and standard deviations in employment disparate impact case); McReynolds v. Sodexho Marriott Servs., Inc., 349 F. Supp. 2d 1, 21-22 (D.D.C. 2004) (noting usefulness of regression analyses). See also Watson v. Fort Worth Bank & Trust, 487 U.S. 977, 996-997 (1988) (sampling types of infirmities that may emerge when "facially plausible statistical evidence" is scrutinized, including small or incomplete data sets, inadequate techniques, and unsuitable control groups). — 28 — disparities in incarcerated persons relative to the over-all population of such persons within the Commonwealth, the same data underlying the defendant's argument here could be marshalled for nearly any change in the law that result in more defendant-friendly rules.26 There being no supportable distinction between any such changes and the defendant's present claim, we decline his invitation to employ race (or ethnicity) in this manner in our decision making as to whether to apply a new criminal rule retroactively. At bottom, although couched as an equal protection claim based on our decision in Brown, the defendant's actual objection is a claim of selective prosecution. Under the tripartite selective prosecution test, however, 26See Massachusetts Sentencing Commission, Selected Race Statistics 2 (Sept. 27, 2016), https://www.mass.gov/files /documents/2016/09/tu/selected-race-statistics.pdf [https: //perma.cc/3TAF-2VUE] (in 2014, Massachusetts incarcerated people who are Black at 7.9 times the rate of people who are white). See also E.T. Bishop, B. Hopkins, C. Obiofuma, & F. Owusu, Criminal Justice Policy Program, Harvard Law School, Racial Disparities in the Massachusetts Criminal System 1 (Sept. 2020), https://hls.harvard.edu/content/uploads/2020/11 /Massachusetts-Racial-Disparity-Report-FINAL.pdf [https://perma .cc/W5KA-MX3R] (same). As of 2020, according to the data presented by amicus American Civil Liberties Union of Massachusetts, Inc., the Commonwealth's population was sixty-nine percent white and 6.8 percent Black; overall, thirty-one percent of the Commonwealth’s population identified as nonwhite. — 29 — "the defendant bears the initial burden to 'present evidence which raises at least a reasonable inference of impermissible discrimination, including evidence that a broader class of persons than those prosecuted violated the law, . . . that failure to prosecute was either consistent or deliberate, . . . and that the decision not to prosecute was based on an impermissible classification such as race, religion, or sex'" (quotations omitted). Commonwealth v. Bernardo B., 453 Mass. 158, 168 (2009), quoting Commonwealth v. Franklin, 376 Mass. 885, 894 (1978). If a defendant makes this prima facie showing, "the Commonwealth must rebut that inference of discrimination." Commonwealth v. Robinson-Van Rader, 492 Mass. 1 , 17 (2023). "Because a claim of selective prosecution is a collateral attack on prosecutorial decision-making, a degree of rigor is demanded to balance such claims against the presumption of prosecutorial regularity." Bernardo B., 453 Mass. at 168 . Here, the defendant, in essence, asks us to sidestep this required rigor by crafting a new standard for retroactive application of new rules to target essentially the same conduct that the selective prosecution framework already addresses. We decline to do so. b. Jury instructions. The defendant next contends that certain jury instructions were erroneous. In giving instructions, "[a] trial judge has the duty to state the applicable law clearly and correctly." Commonwealth v. Wall, 469 Mass. 652, 670 (2014). "In assessing the sufficiency of the — 30 — jury instructions, we consider the charge in its entirety, to determine the 'probable impact, appraised realistically . . . upon the jury's factfinding function.'"27 Id., quoting Commonwealth v. Batchelder, 407 Mass. 752, 759 (1990). See Commonwealth v. Denis, 442 Mass. 617, 621 (2004) ("In examining a claim of error in jury instructions, we do not look at individual phrases taken out of context; rather, we consider the instructions viewed as a whole"). i. Cooperating witness instruction. The defendant asserts that the judge's instruction concerning the jury's evaluation of the testimony of a cooperating witness did not comply with the requirements of Commonwealth v. Ciampa, 406 Mass. 257, 266 (1989). Because trial counsel timely objected,28 we examine whether any error was prejudicial. See Commonwealth v. 27We have encouraged trial judges to follow the model jury instructions. See, e.g., Commonwealth v. Bonner, 489 Mass. 268 , 285 (2022); Commonwealth v. Howard, 479 Mass. 52, 61 (2018). But we also have affirmed that a judge need not use particular words in giving an instruction "so long as the charge, as a whole, adequately covers the issue." Commonwealth v. McGee, 467 Mass. 141, 154 (2014), quoting Commonwealth v. Daye, 411 Mass. 719, 739 (1992). 28Contrary to the Commonwealth's contention, trial counsel's request to "re-instruct on the model jury instruction on the cooperating witness" because the trial judge "narrowed or diminished some of the instructions to the detriment of the defense" sufficiently highlighted the nature of the objection, and the judge considered and rejected the request. See Commonwealth v. Costa, 88 Mass. App. Ct. 750 , 754 n.6 (2015) (objection preserved where "the trial judge considered the objection fully"). — 31 — Teixeira, 490 Mass. 733 , 742 (2022); Commonwealth v. Meuse, 423 Mass. 831, 832 (1996). "When a prosecution witness testifies pursuant to a plea agreement containing a promise to tell the truth, and the jury are aware of the promise, the judge should warn the jury that the government does not know whether the witness is telling the truth." Meuse, 423 Mass. at 832 . The judge should also "specifically and forcefully tell the jury to study the witness's credibility with particular care." Ciampa, 406 Mass. at 266 . "[I]f the prosecutor has vouched for that witness's credibility, such a failure to instruct is reversible error." Meuse, supra. "Vouching can occur if an attorney expresses a personal belief in the credibility of a witness . . . or if an attorney indicates that [the attorney] has knowledge independent of the evidence before the jury verifying a witness's credibility." Ciampa, supra at 265 . Here, the Commonwealth's key witness, Jones, testified pursuant to a cooperation agreement. The prosecutor briefly elicited on direct examination that Jones entered into a cooperation agreement and was receiving a reduced sentence in exchange for her testimony against the defendant. The prosecutor did not elicit that the agreement was contingent on Jones telling the truth; nor did the prosecution admit a copy of the agreement in evidence. The prosecutor neither expressed her — 32 — personal belief in Jones's credibility nor suggested that she possessed special knowledge of Jones's truthfulness. Instead, in her closing argument, the prosecutor urged the jury to believe Jones based on specific evidence that corroborated her testimony. Contrast Commonwealth v. Meuse, 38 Mass. App. Ct. 772, 774 (1995), S.C., 423 Mass. 831 (1996) (prosecutor emphasized in closing argument that if cooperating witness was "not telling the truth, we have an army of police that can go out and corroborate every detail he is giving us. If he gives us one wrong detail . . . we will not show up for sentencing. That’s the leverage we have . . ."). However, on cross-examination, after trial counsel suggested that Jones was being untruthful to secure her deal, Jones responded: "I wouldn't make up a story. It was an agreement to be honest a hundred percent or there's no agreement in place." Cf. Commonwealth v. Chaleumphong, 434 Mass. 70 , 7475 (2001) (officer's testimony about methods of confirming truthfulness of cooperating witness was not vouching where testimony was extracted by defense's cross-examination). After the close of evidence, the judge instructed the jury that it should "treat [Jones's] testimony with particular care because you know she has received a benefit from the Commonwealth." While the judge did not caution that the Commonwealth "could not know whether [Jones] was telling the — 33 — truth," see Meuse, 423 Mass. at 832 , he emphasized that the jury were the sole ultimate arbiters of witnesses' credibility, and that in evaluating credibility, they could take into account bias and whether "a witness has something to win or lose by their testimony." See Ciampa, 406 Mass. at 266 . See also Commonwealth v. Grenier, 415 Mass. 680, 687 (1993) ("The judge's instruction on credibility, including references to witnesses' interests in the outcome of the case and to their possible bias, was sufficient in the circumstances"). Although, in view of Jones's characterization of her obligation to tell the truth under the cooperation agreement, it may have been preferable for the judge also to specify that the prosecution had no special method of determining Jones's truthfulness, these circumstances, combined with the vigorous cross-examination of Jones that elicited her prior inconsistent statements, lead us to conclude with fair assurance that the omission did not sway the jury's decision. See Commonwealth v. Cruz, 445 Mass. 589, 591 (2005). ii. "Lifestyle" commentary. The defendant additionally challenges the instruction to the jury that "[w]e're not here to judge someone's lifestyle; be it the alleged victim . . . be it a witness, be it anybody involved here" (emphasis added). He asserts that the instruction impermissibly bolstered Jones's credibility and was prejudicial because the defense relied on attacking Jones's lavish lifestyle relative to her income. — 34 — In conducting a trial, a judge may not "express an opinion on the credibility of particular witnesses," or "instruct the jury that they must draw particular inferences from the evidence." Commonwealth v. Sneed, 376 Mass. 867, 870 (1978). Here, the instruction neither conveyed the judge's views of Jones's credibility nor ordered the jury to ignore evidence linked to her lifestyle when evaluating credibility. Rather, the judge was clear that instead of "judg[ing] someone's lifestyle," the jury must "coolly and calmly sift through evidence" and "draw reasonable inferences.""29 Furthermore, the judge repeatedly reaffirmed that the jury were the ultimate arbiters of credibility determinations. Although it may have been prudent to avoid altogether the use of the defense's chosen phrase, "lifestyle," the judge did not err. iii. Hypotheticals. The defendant also maintains that the trial judge gave hypotheticals that too closely tracked the facts of the case or that aligned the judge with the victim and prosecution. A judge generally may employ hypotheticals to 29In this regard, the judge's instruction conveyed the essence of the model instructions on the role of the jury, which state that jurors "must be completely fair and unbiased" and should "not let [their] emotions, any kind of prejudice, or [their] personal likes or dislikes influence [them] in any way." Superior Court Model Jury Instructions, Final Charge Script 5 (Nov. 2023). See Instruction 2.120 of the Criminal Model Jury Instructions for Use in the District Court (Sept. 2022) (jurors should not "allow bias . . . to interfere with [their] ability to fairly evaluate the evidence"). — 35 — explain concepts to the jury. See, e.g., Denis, 442 Mass. at 621-622, 624-625 ; Commonwealth v. Moses, 436 Mass. 598, 604-605 (2002). But in doing so, the judge must "not improperly comment on the . . . evidence or offer his opinion regarding the defendant's guilt." Moses, supra at 605 . Additionally, a judge should not offer a "hypothetical that too closely tracks the facts of the defendant's case." Commonwealth v. Gumkowski, 487 Mass. 314 , 331 (2021). We further have discouraged "examples in which hypothetical individuals commit crimes" (citation omitted). Id. Here, immediately prior to the introduction of a recording of a telephone call made by the defendant from jail, see note 13, supra, the judge cautioned the jury not to let the defendant's pretrial detention bias them. The judge then stated that if he were arrested, he "would hope [his] wife would come . . . make [his] bail," and that "people with means" can generally "make bail." The judge added "just because someone can't make bail, you can't hold that against them. . . . [T]hat would be very unfair." While the judge's reference to his wife's assistance was better left unsaid, the instruction, as a whole, was not error.30 30Similarly, the defense complains that in explaining the unarmed robbery charge, the judge used himself as an example victim, thereby equating himself with the victim, and thus the — 36 — Additionally, after instructing the jury to weigh Jones's testimony with "particular care" in view of her cooperation agreement, see part 3.b.i, supra, the judge gave one example of how the jury could assess credibility. He told the jury that if he said, "what a miserable, wet rainy day," but they could see that it was sunny, the jury could conclude that he is "crazy" because they have "contrary evidence." The defendant contends that this statement instructed the jury to disbelieve Jones's testimony only if they had direct contrary evidence. But the judge did not convey that only direct evidence can lead the jury to disbelieve testimony. Rather, he gave it as one example of how the jury could assess credibility; he urged them to use their "common sense" and to draw "reasonable inferences." The defendant also asserts that the judge erred in connection with a hypothetical the judge employed to illustrate joint venture liability. In it, the judge and his "crazy" and "dumb" brother-in-law conspired to rob a bank. Contrary to the defendant's contention, the outlandish hypothetical did not "closely mirror[] the circumstances of the defendant's case" or "emphasize the prosecution's theory of the case" -- a death prosecution. This claim is too attenuated and speculative to constitute error. — 37 — resulting from a botched drug heist.31 See Gumkowski, 487 Mass. at 332. The judge made clear that he was using a hypothetical illustratively and emphasized that the jurors were the sole arbiters of the facts. See Moses, 436 Mass. at 605 . No reasonable juror would have been misled by the judge's example. c. Trial judge's conduct. The defendant additionally claims that the trial judge prejudicially injected himself into the proceedings. i. Questioning of witnesses. The defendant first points to the judge's questioning of witnesses.32 "A judge may properly question a witness, even where to do so may 'reinforce the Commonwealth's case, so long as the examination is not partisan in nature, biased, or a display of belief in the defendant's guilt.'" Commonwealth v. Carter, 475 Mass. 512, 525 (2016), quoting Commonwealth v. Festa, 369 Mass. 419, 422 (1976). Although we have expressed concerns with an "overspeaking judge," see Commonwealth v. Campbell, 371 Mass. 40, 45 (1976), "[t]here exists no quantitative test for determining whether the 31The defendant also argues that the example highlighted the defendant's decision not to testify and suggested that the defendant was "dumb" or "crazy." We disagree. Indeed, the judge specifically instructed the jury to draw no inferences from the defendant not testifying and not to take cues from the judge. 32Here, the defendant complains that the judge asked witnesses a total of 146 questions, including sixty to Jones. — 38 — judge has gone beyond the bounds which the law imposes," Commonwealth v. Dias, 373 Mass. 412, 416 (1977), S.C., 402 Mass. 645 (1988). The judge's actions are to "considered in the context of the entire trial." Festa, supra at 423 . Here, the judge's questioning did not interfere "with counsel's ability to put on a full defense." See Commonwealth v. Sylvester, 388 Mass. 749, 751-752 (1983). And, while some of the questions clarified facts that, in turn, benefited the Commonwealth, none showed bias or favor toward the prosecution; rather, the judge's questions were directed either to clarifying information or to mitigating the risk of the jury making unfairly prejudicial inferences. In the circumstances, while it would have been better for the judge to interject his questions less frequently, we discern no error in the questions he asked. ii. Banter with witnesses. The defendant also argues that the judge improperly engaged in extraneous social conversation with Commonwealth witnesses, which, he contends, enhanced those witnesses' likability and demonstrated partiality to the Commonwealth. In particular, the judge bantered with the victim's roommate about a board game, asked about a forensics witness's broken leg, and thanked the telephone records custodian for traveling from afar.33 33Additionally, the defendant complains that the judge "plac[ed] his finger on the scales of justice in favor of the — 39 — "Although we discourage gratuitous remarks by judges," Commonwealth v. Mello, 420 Mass. 375, 392 (1995), the judge's "folksy" mannerism, even if error, did not result in a substantial likelihood of a miscarriage of justice, see Commonwealth v. Lucien, 440 Mass. 658, 664-665 (2004). None of the remarks displayed partiality toward the prosecution or the witnesses beyond the normal bounds of affability and courtesy. Indeed, the judge displayed a similar chattiness with jurors during voir dire but had little opportunity to do the same with defense witnesses, as the defense called only one witness. Moreover, the judge instructed the jury not to take any cues from him in assessing credibility. See id.; Mello, supra. d. Ineffectiveness of counsel. The defendant also asserts that the motion judges abused their discretion in denying his motions for a new trial because trial counsel provided ineffective assistance. "When evaluating ineffective assistance of counsel claims in connection with the direct appeal of a conviction of murder in the first degree, 'we review for a substantial likelihood of a miscarriage of justice . . . .'" Kirkland, 491 Mass. at 346, quoting Commonwealth v. Don, 483 Commonwealth" by asking the prosecutor to help display a jury instruction chalk -- rather than asking a court officer -- and by asking the prosecutor for her opinion on the instructions. These actions do not alter our conclusion. 40 Mass. 697 , 704 (2019).34 "In conducting this review, we accord tactical decisions of trial counsel due deference" and reverse only if counsel's decisions were "manifestly unreasonable" (quotation and citations omitted). Kirkland, supra. "'[O]nly strategy and tactics which lawyers of ordinary training and skill in the criminal law would not consider competent' rise to the level of manifestly unreasonable." Id., quoting Commonwealth v. Kolenovic, 471 Mass. 664, 674 (2015), S.C., 478 Mass. 189 (2017).35 i. Lack of CSLI expert. The defendant argues that trial counsel provided ineffective assistance by failing to retain a CSLI expert. "There is no requirement that trial counsel always present expert or documentary evidence to support an argument, especially where other evidence is presented to support it." Commonwealth v. Hensley, 454 Mass. 721, 736 (2009). Here, trial 34This is because the "statutory standard of [G. L. c. 278, § 33E,] is more favorable to a defendant than is the constitutional standard for determining the ineffectiveness of counsel" (citation omitted). Commonwealth v. Martin, 467 Mass. 291, 316 (2014). 35Where a claim asks us to speculate on the strategic decision-making of trial counsel, the absence of an affidavit from trial counsel is significant. See, e.g., Commonwealth v. Gonzalez, 443 Mass. 799 , 809 n.10 (2005) ("It is significant that there is no affidavit from trial counsel to inform us of his strategic reasons for these decisions"); Commonwealth v. Vasquez, 55 Mass. App. Ct. 523, 533 (2002) ("Conspicuously absent in these circumstances is an affidavit from defense trial counsel"). — 41 — counsel made a strategic decision to rely on cross-examination of the Commonwealth's expert. See Commonwealth v. Sena, 441 Mass. 822, 827-829 (2004) (no ineffective assistance for not calling expert where counsel effectively cross-examined Commonwealth's expert).36 On cross-examination, trial counsel effectively elicited that the Commonwealth's CSLI evidence could provide no more than an approximate location of the defendant's cellular telephone; counsel evoked that the cellular telephone plausibly could have connected to cellular towers further from the telephone's location based on any number of factors, including call volume and physical obstructions.37 36The defendant also maintains that trial counsel provided ineffective assistance because, in her opening statement, she told the jury that cellular telephone records would exculpate the defendant. Although a failure to deliver on a promise of key evidence may constitute ineffective assistance, see Commonwealth v. Duran, 435 Mass. 97, 109 (2001), here, counsel delivered on her promise. The telephone records were admitted in evidence through the Commonwealth's cellular telephone records custodian witness. And, in closing argument, trial counsel argued that because the defendant called Jones near the time of the shooting, this indicated that the defendant was not with Jones and therefore was not the third coventurer. This was not a manifestly unreasonable tactic. See Fernandes, 492 Mass. at 492 (defense counsel reasonably suggested that calls between defendant and codefendant showed that they were not together). 37The affidavit submitted by the defendant's posttrial expert showed, at most, that the possible area from which the call could have been placed was somewhat larger than the already sizable area the Commonwealth's expert proffered; significantly, the larger area still encompassed the victim's home. Contrast Commonwealth v. Baker, 440 Mass. 519, 528-529 (2003) (counsel — 42 — ii. Telephone records custodian's testimony. The defendant also contends that trial counsel provided ineffective assistance by not objecting to the cellular telephone records custodian's qualifications to testify regarding how cellular telephone towers function. Assuming, arguendo, that the expert was unqualified as to that subject matter, the error does not raise a substantial likelihood of a miscarriage of justice. As discussed supra, the CSLI data merely corroborated an otherwise strong case against the defendant, which also included surveillance video footage that placed the defendant close to the victim's home shortly before the shooting, as well as call logs indicating that the defendant was in communication with Jones and Tyler. iii. CSLI exhibits. The defendant next faults trial counsel for not objecting to the admission of two maps derived from CSLI data that placed the defendant's cellular telephone in the vicinity of the victim's home and the restaurant. Such charts derived from CSLI data, for which a proper foundation is laid, are admissible. See Bin, 480 Mass. at 679-680 (judge did not abuse discretion in admitting computer-generated map police officer created to plot CSLI data). See also Commonwealth v. Carnes, 457 Mass. 812, 825 (2010) ("Summaries of testimony are failed to introduce expert who would have rebutted "the only physical evidence used by the Commonwealth to link" defendant). — 43 — admissible, provided that the underlying records have been admitted in evidence and that the summaries accurately reflect the records"). Therefore, counsel's lack of objection was not manifestly unreasonable. iv. Murder in the second degree instruction. The defendant asserts that trial counsel should have sought an instruction on felony-murder in the second degree. Where "the defendant's trial strategy was to present an all-or-nothing choice to the jury," not requesting an instruction on an available lesser included crime is not manifestly unreasonable. Commonwealth v. Roberts, 407 Mass. 731, 737-739 (1990). Here, the primary defense was that the defendant did not participate in the robbery and that Jones fabricated her testimony. The choice to forgo the instruction on second degree murder was not manifestly unreasonable. v. Adequacy of preparation. The defendant argues that his trial counsel inadequately prepared for trial. Among the duties of counsel are the duties "to consult with the defendant on important decisions and to keep the defendant informed of important developments in the course of the prosecution." Strickland v. Washington, 466 U.S. 668, 688 (1984).38 Counsel 38Prior to trial, the defendant met with trial counsel in person at least six times and was able to speak with her by telephone "numerous times through [his] incarceration." Without more, the defendant's claim of a failure to communicate is — 44 — also has a duty to conduct an independent investigation of the facts. Commonwealth v. Duran, 435 Mass. 97, 102 (2001). See Commonwealth v. Saferian, 366 Mass. 89, 96 (1974). To establish ineffective assistance of counsel, a defendant must identify with particularity how any investigation that counsel failed to conduct would have benefited the defense. Duran, supra at 103 . "Speculation, without more, is not a sufficient basis to establish ineffective representation." Id. The defendant contends that trial counsel failed to give him certain discovery materials in a timely manner, causing him to underestimate the strength of the Commonwealth's case. The defendant does not explain how earlier access to discovery material would have altered his strategy. The defendant further maintains that trial counsel failed to contact, call, and prepare two neighbors (one of whom testified), as well as Jones's friend and Tyler. He does not identify any noncumulative, material exculpatory testimony that the two neighbors could have supplied; Jones's friend was unavailable to testify; and Tyler was himself a defendant in a parallel case for the same crime, see note 15, supra. In the absence of an affidavit from trial counsel, we reject the claim unsupported. See Martin, 484 Mass. at 643-644 (no ineffective assistance where counsel visited defendant six times and defendant failed to articulate how more contact would have affected strategy or verdict). — 45 — that a failure to call these witnesses was not a strategic choice. Nor will we speculate as to what these witnesses might have said.39 vi. Firearm and drugs seized from Jones's home. The defendant also argues that trial counsel failed to seek to introduce information that shortly before the shooting, police officers had seized a firearm and "crack" cocaine from Jones's apartment and had arrested her boyfriend. "[I]mpeachment of a witness is, by its very nature, fraught with a host of strategic considerations to which we will, even on [G. L. c. 278, § 33E,] review, still show deference" (citation omitted). Commonwealth v. Valentin, 470 Mass. 186, 190 (2014). Therefore, "a claim of ineffective assistance based on failure to use particular impeachment methods is difficult to establish." Commonwealth v. Fisher, 433 Mass. 340, 357 (2001). "This is particularly so where [trial counsel] conducted a thorough impeachment . . . ." Valentin, supra at 191 . The crux of the defense was that the defendant was not involved in the robbery; Jones's motivation for the robbery and her prior involvement with drugs and firearms have little 39See Commonwealth v. McWilliams, 473 Mass. 606, 621 (2016) ("[A] motion judge may reject a defendant's self-serving affidavit as not credible" [citation omitted]); Commonwealth v. Rice, 441 Mass. 291, 304 (2004) (without affidavit from trial counsel, defendant's assertions are speculative). — 46 — bearing on whether the defendant also participated in the robbery. Moreover, trial counsel vigorously cross-examined Jones; she raised serious questions regarding Jones's version of events and elicited that Jones initially had lied about the robbery, that Jones needed money, and that Jones had an incentive to testify against the defendant. Further expounding on Jones's motivation for the robbery was unlikely to have influenced the jury's decision.40 e. Review under G. L. c. 278, § 33E. The defendant also asks us to apply Brown retroactively to his case as a matter of fairness pursuant to review under G. L. c. 278, § 33E.41 Unlike in Brown, however, the defendant here was not in the "remote outer fringes" of the scheme that led to the victim's death. See Brown, 477 Mass. at 824 (reducing felony-murder verdict from first degree to second degree where defendant's involvement was limited to supplying firearm and clothing used in robbery). We discern no error warranting relief under G. L. c. 278, § 33E. 40The defendant also contends that Jones's posttrial arrest in November 2021 for possessing an illegal firearm showed that the pretrial firearm seizure was critical to the defense. However, this posttrial development has no bearing on decisions trial counsel made at trial. 41The defendant contends that he preserved the issue of whether felony-murder continued to provide an independent theory to liability. — 47 — 4. Conclusion. The defendant's conviction of murder in the first degree is affirmed. The orders denying the defendant's first, second, and third motions for a new trial are also affirmed. So ordered.