Gants, J.

A jury convicted the defendant Jesus Silva-Santiago of murder in the first degree on the theory of extreme atrocity or cruelty.1 Represented by new counsel on appeal, the defendant argues error in (1) the denial of his pretrial motion to suppress out-of-court photographic identifications; (2) the exclusion of evidence pertaining to the inadequacy of the police investigation; and (3) the judge’s instructions to the jury. We reject the defendant’s arguments, but conclude that he is entitled to reversal of his conviction and a new trial under G. L. c. 278, § 33E, because of misstatements of the evidence made by the prosecutor during his closing argument (to which the defendant’s trial counsel objected) that were not harmless in the circumstances of this case.

Background. On Saturday, June 28, 2003, the victim, Eugene Monteiro, joined his friends McKenzie Bruneau, Raymond Jenkins, Thomas Johnson, and David Thomas for an evening of drinking. With Thomas driving, they first visited a liquor store in the Jamaica Plain section of Boston, where they purchased some beer. They then drove to a bar in Brockton, where they drank and played some pool. A short time later, they decided to leave and go to another bar in Brockton, Mike’s Lounge (bar), which was located on the comer of Montello Street and Lincoln Street.

At about 11 p.m., the group parked behind the bar and walked *784to its main entrance. There were several people standing outside the entrance.2 Bruneau, who was “kind of drunk,” but alert, “glanced” at a man who stood outside the entrance. Because Jenkins was underage, the men were refused entry into the bar. While the victim and his friends were standing on the sidewalk outside the bar, the man Bruneau had seen outside the bar entrance earlier said, “No niggers allowed.”3 The victim responded, “What’s your problem?” The man ran away and returned to confront the victim, who was near a fire hydrant. The victim asked the man if he had a “toast” (handgun). The man lifted up his shirt, pulled out a handgun from his “waistline,” and shot the victim.4

The victim ran as the man repeatedly shot at him, and fell on an adjacent street not far from the front door of the bar. The man fired nine shots; three hit the victim. Two of the shots hit the victim in the chest and abdomen; another shot both entered and exited through the victim’s back.

After the first shot, Bruneau, Johnson, Jenkins, and Thomas ran in different directions. When the firing stopped, Bruneau, Johnson, and Jenkins returned to where the victim lay on the street. As Bruneau headed back to that area, he could hear sirens, and he saw somebody running down an alley next to the *785bar. Bruneau observed that the person running down the alley, whose face he could not see, had on “white sneakers with a red check.” He went over to the victim, who was alive and trying to breathe. The victim was bleeding from his chest and back.

Bruneau described the man he saw walking into the bar, whom he claimed was the shooter, as being “a little bit taller than [Bruneau5 and] light skinned .... He looked Spanish or Cape Verdean. ... He had some type of . . . red shirt, some red and white sneakers, all white with red checks on it [and] dark blue jeans . . . .” The man wore a red headband and a long gold chain. According to Johnson, the shooter wore a red and white shirt, a red bandanna, and red and white sneakers, and had short hair and a mustache. Jenkins recalled that the shooter wore a red headband, a red shirt, jeans, and a gold chain. A woman who was stopped at a red light on Montello Street observed part of the shooting. In a statement made to police two days later, she was unable to describe the shooter’s face, but observed that he appeared to be black, slim, and about five feet, ten inches tall; had short hair; and wore “baggy clothes” that included “a red shirt” and “blue jeans.”

At 11:30 p.m., officers of the Brockton police department received a dispatch call concerning shots fired outside the bar. Lieutenant Ken Williams and Sergeant Thomas Lafratta responded and, because the station was located only about two blocks from the bar, arrived within one minute. When they arrived, other officers were already at the scene. The victim was lying on the street and was unresponsive. The street was “fairly well lit,” with illumination from multiple street lamps on the side of the street opposite the bar. The victim was transported to a nearby hospital, where he later died as a result of gunshot wounds to his chest and abdomen.6

Police officers secured the crime scene, including the bar. *786Officers were posted at the bar’s entrance on Montello Street and at another entrance on Lincoln Street, preventing anyone from leaving or entering. Near a fire hydrant in front of the bar, and extending along the side of the street, police recovered nine discharged nine millimeter cartridge casings. A ballistics expert testified that the discharged cartridge casings had been fired from a nine millimeter semiautomatic pistol. The weapon was not recovered. When the bar was secured, there were about forty-two persons inside, including the defendant.7 The defendant was wearing what appeared to be white sneakers and a short-sleeved red shirt with gray sleeves. On both sides of the shirt, from its neck, a thin white stripe extended to the cuff of each sleeve. On the back of the shirt, the number “72” appeared in black with white stitching. The numbers were large and easily visible to the eye. On the front of the shirt, the lettering “*ecko unltd.” in navy blue embroidered letters appeared in the chest area, and above that, there' was a depiction of a rhinoceros within an oval covered by plastic. The defendant was not wearing a gold chain, a headband, or a bandanna.

Detective Williams separately escorted Bruneau and Johnson inside the bar to see if either man could identify the shooter. Neither man was able to identify the shooter, even though only two men inside the bar wore red shirts. The testimony, however, sharply differed as to what happened during this attempted identification procedure.8

*787The patrons inside the bar were videotaped when they were permitted to leave. See note 7, supra. From that videotape, a photograph (which was admitted in evidence) was generated depicting the defendant as he then appeared, wearing the red shirt above described. The videotape reflects that the defendant, as he was leaving, was asked to raise his arms and was frisked, apparently for weapons. There was no indication that he was carrying a weapon.

Tihani Pichardo, a fifteen year old patron of the bar that evening, testified that she and a girl friend (Christina Sabata) earlier had met with the defendant and two other friends of Sabata (one male, one female), went to a house party, and then went to the bar in Sabata’s female friend’s vehicle. Before the shooting, as she, Sabata, and one of the other young women stood in the doorway at the front of the bar, she saw the defendant and another man walk toward the parking lot. After some black males drove by, Pichardo, Sabata, and the other young woman went back inside the bar and went to the bathroom. When she left the bathroom and returned to the doorway of the bar, she saw “a whole bunch of people in a group in the parking lot.” She heard arguing and commotion, saw the victim running to the other side of the street, heard gunshots, and walked back inside the bar. She did not see who was shooting and did not see where the defendant was at the time of the shooting. When she returned inside, she smoked a cigarette in the bathroom and then went to the bar to get a couple of drinks. She later saw the defendant inside at the side of the bar near the bathrooms.

Luis Ortiz, another patron of the bar, recounted that, minutes before the shooting, and after the victim and his friends had left the bar, three women and a man who was not the defendant went outside the bar. Ortiz had seen the defendant before the young women went outside and after the shooting, but not when the shots were fired.

On July 7, 2003, Trooper Diane Lilly, Detective Lieutenant Joseph Mason, and Trooper Robert Dateo of the Massachusetts State police, along with Detective Williams of the Brockton police department, went to the victim’s mother’s house to show Johnson and Bruneau an array containing eight photographs. Trooper Lilly explained to Johnson and Bruneau that they were *788going to be shown a photographic array. Then Trooper Lilly and Detective Williams met with Johnson in one room, while Lieutenant Mason and Trooper Dateo met (at the same time) in a different room with Bruneau. The same photographic array was shown to Johnson and Bmneau. Each man depicted in the array had similar features, including a mustache and short hair.

Trooper Lilly informed Johnson that he was going to be shown a series of eight photographs on one page that were numbered consecutively one through eight.9 Trooper Lilly asked Johnson to look at the photographs carefully, and to look at each photograph. She instructed Johnson to let her know if any of the photographs depicted the shooter. Trooper Lilly informed Johnson that he was not “obliged” to select a photograph. The following colloquy then occurred between the prosecutor and Trooper Lilly during her direct examination:

The prosecutor: “Could you tell us, please, what Thomas Johnson said or did when you showed him that array?”

The witness: “He looked at the array and he pointed to photograph number seven and said that that person was the most similar to the shooter that he saw that night. He said it was the same fade haircut, as he called it. And he said that although he was similar, he could not make a positive identification.”

The prosecutor: “After he said that to you, did you see him do anything?”

The witness: “He was — while he was speaking, he was tapping number seven [which was a photograph of the defendant] with his finger.”

During his direct examination, Johnson contended that he had made a positive identification of the defendant when shown the “photospread.”10 He did not recall stating that he had been *789unable to make a positive identification, and was impeached with his grand jury testimony in which he admitted that he had been unable to say positively that the person he identified had been involved in the shooting.

Meanwhile, in a different room Lieutenant Mason inquired of Bruneau whether he would be able to recognize and identify the shooter. Lieutenant Mason handed Bruneau the photographic array and asked him to take his time, to look at each photograph in the array, and to let him know if he recognized the shooter. The following colloquy then occurred between the prosecutor and Lieutenant Mason:

The prosecutor: “Could you describe for us, please, what Mr. Bruneau said or did when you handed him the array?”

The witness: “He took the array. And he appeared to look at each photo. I could see his eyes scanning across the photos. And he then said, ‘That’s him. Number seven. That looks just like the guy.’ ”

The prosecutor: “What did you do at that point?”

The witness: “At that point I asked —■ basically for clarification, I asked him if he meant that that’s the individual who he saw shoot [the victim].”

The prosecutor: “What did he say?”

The witness: “He said, ‘Yes.’ ” [Objection overruled.]

Bruneau then circled the number seven under the photograph of the defendant on the array and signed his name.

During his direct examination concerning the identification, Bruneau recalled that, when shown the array and asked if he recognized the shooter, he told the trooper: “I said I recognized the shooter but I told them I had a glance. I didn’t really see his face but I had got a glance at the man. And I had . . . told them that number seven looked like — a lot like the shooter.”

On July 10, 2003, Trooper Dateo met Jenkins at a predetermined location and Jenkins got into Trooper Dateo’s unmarked automobile. Trooper Dateo instructed Jenkins that he would be showing him an array containing eight photographs, and that, *790after viewing it, if he recognized anyone who might be the shooter, he should say so. He then displayed the array to Jenkins. The defendant’s photograph within the array had been moved from position number seven to position number two in order to prevent a tainted identification due to witness collusion. Trooper Dateo testified as follows:

The prosecutor: “What did [Jenkins] do or say when you showed him that photo array?”

The witness: “I handed him the photo array, Ray Jenkins. He examined it, looked at number two, and he told me that number two was more familiar — number two looks most familiar. ‘He is the one that looks most familiar,’ I believe is the way he said it.”

The prosecutor: “Most familiar as whom, did he say?”

The witness: “The shooter.”

Jenkins then circled the number two as well as the photograph above it, and initialed and dated the array.

Jenkins testified that when shown the array and asked if anyone looked familiar, he replied, “ ‘Yeah, I guess so.’ I mean, yeah, I said, ‘Yes.’ ” Jenkins confirmed that he initialed the array and circled the photograph in position two, as well as the number two on the array.11

The next day, July 11, police arrested the defendant at his apartment. Pursuant to a search warrant, police recovered the red shirt that the defendant had been wearing on the night of the shooting. Police also recovered a pair of white “Nike” high-top sneakers on which the “Nike swoosh” logo was red.

The parties stipulated that the sneakers seized from the defendant’s apartment were tested for the presence of human blood, and that the result of that testing was negative. In addition, the defendant’s red shirt was tested for the presence of blood and gunshot residue, and the results of that testing also were negative.

The defendant did not testify. His trial counsel called two witnesses: Samuel Perez, a patron at the bar on the evening of *791the shooting, and Amy Douglass, an assistant professor of psychology at Bates College in Maine and an expert in the field of eyewitness identification. Perez testified that he knew the defendant “[f]rom the street,” and saw the defendant talking with three girls in a comer of the bar near the bathroom at the time of the shooting.

Professor Douglass testified that three procedures recommended by the United States Department of Justice in its publication Eyewitness Evidence, A Guide for Law Enforcement (1999), were not followed by police in this case. First, the witnesses were not told before viewing the photographic arrays that the “culprit” may or may not be present in the array. Without such an instruction, she testified, a witness tends to assume that the “culprit” is present and proceeds to make a choice from the array. Second, the witnesses should have been asked how confident they were in their identifications immediately after making them, because some research shows that a witness’s confidence in the accuracy of his identification grows once he learns that the police believe he made the correct identification. Third, the identification procedure should be fully documented, preferably by a videotape or audiotape recording.

Professor Douglass also identified two other recommended procedures that were not employed in this case. First, she testified that many experts recommend that the photographs of potential suspects be displayed sequentially, with the witness determining whether a single photograph is familiar before turning to the next photograph, rather than simultaneously, with the photographs of all the suspects displayed together on one page. She noted that the simultaneous method is “a very effective strategy” if the “culprit” is in the array, but the strategy is “problematic” if the “culprit” is not present in the array because it may lead to a false identification if the witness attempts to ascertain which photograph is “the closest” to the “culprit.” Second, she testified that most experts recommend a “double-blind procedure” in which the person showing the photographs to the witness does not know the identity of the suspect. She explained that a double-blind procedure eliminates the risk that the person showing the photographs will influence the witness’s identification, intentionally or unintentionally.

*792Finally, Professor Douglass explained the phenomenon of “unconscious transference,” where an identification is made because the person looked familiar to the witness, but not because that person necessarily is the “culprit.” She said it was possible that this phenomenon occurred in this case because “the witnesses saw the defendant both in the bar and in the videotape,” and therefore the defendant looked familiar to them.12

Douglass acknowledged on cross-examination that, in the only field study comparing the simultaneous method of identification with the sequential method, Report to the Legislature of the State of Illinois: The Illinois Pilot Program on Sequential Double-Blind Identification Procedures (2006) (Illinois study), there were more false positives using the sequential method than the simultaneous method. She also confirmed that the results of the Illinois study revealed that, when the simultaneous method was used, the wrong person was identified only three per cent of the time.13

Approximately three hours after the case was submitted to the jury, the jury submitted a note to the judge indicating a “complete standstill.” After discussing the note with trial counsel, the judge excused the jury for the day. The following morning, the judge concluded that the jury had not engaged in due and thorough deliberation14 and, over the objection of the defendant, *793gave an instruction similar to that approved in Commonwealth v. Rodriquez, 364 Mass. 87, 100-101, 102 (1973) (Appendix B) (referencing and approving supplemental instruction proposed by the American Bar Association “when the jurors appear to be running into difficulty reaching a verdict”). See Commonwealth v. O’Brien, 65 Mass. App. Ct. 291, 297 (2005).

Around 12:45 p.m., the jury sent another note to the judge expressing their inability to reach unanimity. The note concluded, “All members of this jury strongly feel that our individual positions are not going to change with further discussion.” As a result, the judge, who noted that the circumstances had changed and that the jury had been deliberating for six and one-half hours over two days, gave the so-called Tuey-Rodriquez charge. See Commonwealth v. Rodriquez, supra at 101-102 (Appendix A); Commonwealth v. Tuey, 8 Cush. 1, 2-3 (1851). The jury continued to deliberate until shortly after 4 p.m., when the jury returned their verdict.

Discussion. We first take up the three claims of error raised by the defendant.

1. The photographic identifications. The defendant contends that the judge should have allowed his motion to suppress the out-of-court photographic identifications made by Bruneau, Johnson, and Jenkins, on the ground that the photographic arrays used were “unnecessarily suggestive and conducive to irreparable mistaken identification.” In support of his motion, the defendant pointed to the fact that he was the only person depicted in the photographic array with a so-called “fade” haircut. The defendant further argued that the identification procedure was flawed. The defendant also moved to suppress any subsequent in-court identification, maintaining that it would be tainted by the inadmissible out-of-court identifications.15

After an evidentiary hearing, the judge denied the motion, concluding, in a handwritten margin indorsement, that “the methods used by the police to elicit the identification in each instance were not so impermissibly suggestive as to give rise to a very substantial likelihood of misidentification.” The judge did not issue a written memorandum of decision and did not make find*794ings of fact. Presumably, she credited the Commonwealth’s witnesses, Troopers Lilly and Dateo, who were the only witnesses who testified at the evidentiary hearing. The defendant filed a motion for reconsideration, which the judge denied.16

The description at the motion hearing concerning the procedures that preceded each identification was similar to the evidence at trial, so we will not describe it separately. At the evi-dentiary hearing, however, Trooper Lilly described how she compiled the photographic array and why she included the defendant’s photograph. Trooper Lilly, along with other officers, reviewed the videotape recording of the patrons leaving the bar after the shooting and saw that the clothing the defendant wore that evening matched that given in the general description of the shooter. As a result, a photograph of the defendant, obtained from a Federal probation officer, was included in a photographic array that was compiled by a bureau of the Plymouth County sheriff’s department. The array contained eight photographs on a single sheet of paper.

At the evidentiary hearing, Trooper Lilly stated that she had last attended training in identification procedures within the past five years. She was not aware of any written procedures of the State police for ensuring the accuracy of photographic identifications, although she indicated that, on the back of a folder they use, “there is a gist of what they would prefer you to cover.” Trooper Lilly stated that there was no attempt to record the identification procedure she conducted with Johnson because it was not a practice that the State police used for identifications obtained through photographic arrays.

Trooper Dateo also stated that he did not attempt to record Jenkins’s identification. Nor did he ask Jenkins about the level of certainty of his identification.

“It is undeniable that a defendant has a due process right to identification procedures meeting a certain basic standard of fairness.” Commonwealth v. Dougan, 377 Mass. 303, 316 (1979). “Where a defendant alleges that witness identifications arise from unnecessarily suggestive circumstances, the ‘defendant has the burden to prove, by a preponderance of the evidence, that the *795witness was subjected by the State to a pretrial confrontation . . . “so unnecessarily suggestive and conducive to irreparable mistaken identification” as to deny the defendant due process of law.’ ” Commonwealth v. Odware, 429 Mass. 231, 235 (1999), quoting Commonwealth v. Otsuki, 411 Mass. 218, 232 (1991). In considering whether identification testimony should be suppressed, the judge must examine “the totality of the circumstances attending the confrontation to determine whether it was unnecessarily suggestive.” Commonwealth v. Odware, supra, quoting Commonwealth v. Otsuki, supra at 232-233.

We reject the defendant’s contention that the arrays used were unnecessarily suggestive because the photograph of him therein impermissibly distinguished him from the other men included in the array by his so-called “fade” haircut. While we “disapprove of an array of photographs which distinguishes one suspect from all the others on the basis of some physical characteristic,” Commonwealth v. Melvin, 399 Mass. 201, 207 n.10 (1987), we conclude, based on our own review of the arrays shown to the three eyewitnesses, that the men depicted therein possessed reasonably similar features and characteristics, including the style and length of their hair.

We next consider the defendant’s argument that the identification procedures used here were unnecessarily suggestive because the witnesses may have seen the defendant inside the bar on the night of the shooting. The defendant did not make this argument to the motion judge and did not elicit any evidence regarding it at the evidentiary hearing on the motion to suppress. As such, the argument cannot be a basis for reversing the motion judge. See Commonwealth v. Rivera, 441 Mass. 358, 367 (2004), quoting Commonwealth v. Ramos, 402 Mass. 209, 216 (1988) (“Evidence adduced at trial but not before the motion judge . . . cannot be determinative of the propriety of the motion judge’s decision”).17

The defendant submits that the identification procedure was *796unduly suggestive because the police did not “observe recognized procedures for guarding against error.” The defendant more specifically points out that the photographs were shown simultaneously rather than sequentially; the troopers who presented the arrays knew at that time that the defendant was a suspect18; there was no instruction that the shooter may or may not be in the array, or that the investigation would continue irrespective of whether a photograph was identified; and there was no recording made of the identification procedure.19 In support of his argument, the defendant relies on various “studies,” including studies conducted by the offices of the district attorney in Suffolk, Middlesex, and Norfolk Counties, and proposals made by the American Bar Association and the Massachusetts Bar Association.

We have long recognized that “[e]yewitness identification of a person whom the witness had never seen before the crime or other incident presents a substantial risk of misidentification and increases the chance of a conviction of an innocent defendant.” Commonwealth v. Jones, 423 Mass. 99, 109 (1996). See Commonwealth v. Johnson, 420 Mass. 458, 465 (1995) (“mistaken identification is believed widely to be the primary cause of erroneous convictions”).20 We have also recognized that eyewitness identification may be highly probative of who did (and did not) commit a crime, and that the exigencies of police work *797sometimes require police to employ less than perfect identification procedures. Commonwealth v. Austin, 421 Mass. 357, 362 (1995). Indeed, while disfavored, we have found that one-on-one showup identifications, whether in-person or photographic, may be admissible, even though they are inherently suggestive. See id. at 361; Commonwealth v. Venios, 378 Mass. 24, 29 (1979). Our focus has been not only whether the identification procedure is suggestive and conducive to irreparable mistaken identification, but also whether it is unnecessarily so.

We have yet to conclude that an identification procedure is unnecessarily suggestive unless it is administered by a law enforcement officer who does not know the identity of the suspect (double-blind procedure), recognizing that it may not be practicable in all situations. At the same time, we acknowledge that it is the better practice because it eliminates the risk of conscious or unconscious suggestion. Here, while the troopers who administered the photographic arrays knew (or may have known) that the defendant was a suspect, there was no evidence that the troopers who presented the photographic arrays signaled a particular response to, or otherwise attempted to influence, any of the eyewitnesses.21 In these circumstances, we do not conclude that the absence of a double-blind procedure by itself renders an identification procedure inadmissible in evidence as unnecessarily suggestive and conducive to irreparable mistake. The absence of such a procedure is properly a matter of the weight of the identification evidence, as it was here, rather than of admissibility. See Commonwealth v. Odware, supra at 236, quoting Commonwealth v. Melvin, supra at 208 (“The degree of suggestiveness present in [these] identification[s] went to the weight, not the admissibility, of this evidence and it is for the jury to assess the infirmities of [these] identification[s]”).

What is practicable in nearly all circumstances is a protocol to be employed before a photographic array is provided to an eyewitness, making clear to the eyewitness, at a minimum, that he will be asked to view a set of photographs; the alleged wrong*798doer may or may not be in the photographs depicted in the array; it is just as important to clear a person from suspicion as to identify a person as the wrongdoer; individuals depicted in the photographs may not appear exactly as they did on the date of the incident because features such as weight and head and facial hair are subject to change; regardless of whether an identification is made, the investigation will continue; and the procedure requires the administrator to ask the witness to state, in his or her own words, how certain he or she is of any identification. See United States Department of Justice, Eyewitness Evidence: A Guide for Law Enforcement 19, 31-32, 33-34 (1999). Not only would such a protocol provide important information to the eyewitness that may reduce the risk of a misidentification, but adhering to it would permit the law enforcement officer following the protocol to testify more accurately and with greater precision as to what the witness was told prior to the identification. Here, no formal protocol was used, and Trooper Lilly was aware only of information on the back of a folder providing the “gist of what they would prefer you to cover.” We decline at this time to hold that the absence of any protocol or comparable warnings to the eyewitnesses requires that the identifications be found inadmissible, but we expect such protocols to be used in the future. See Commonwealth v. Diaz, 422 Mass. 269, 273 (1996) (warning “that the time may come when recording in places of detention . . . will be mandatory if a statement obtained during custodial interrogation is to be admissible”). See also Commonwealth v. DiGiambattista, 442 Mass. 423, 441-442 (2004).

We have not declared the simultaneous display of photographs in an array to be so unnecessarily suggestive and conducive to irreparable mistaken identification as to render inadmissible any identification arising from such a procedure. We are aware of the debate among scholars and practitioners whether the sequential showing of photographs leads to greater accuracy, and of the empirical evidence that is still developing as to which practice is preferable.22 While that debate evolves, the choice of a simultaneous rather than a sequential display of photographs *799shall go solely to the weight of the identification, not to its admissibility.

We recognize the importance of accurately documenting the identification procedure and the actual identification (or failure to identify), but also recognize that recording such a procedure is often not practicable and do not require such a recording as a prerequisite to admissibility. We note here that the troopers appeared to have reported verbatim (or nearly verbatim) the words used by the eyewitnesses in making their identifications, and conclude that their careful reporting is sufficient in the circumstances.23

In the end, having carefully considered the totality of the circumstances, we conclude that the eyewitness identifications here were not so unnecessarily suggestive and conducive to irreparable mistaken identification as to require suppression of the identification evidence. Nor do we conclude that the jury were unable reasonably to assess the weight of the eyewitness evidence, especially because they heard the expert testimony of Professor Douglass detailing the alleged flaws in the identification procedure. There was no error in the denial of the motion to suppress the identifications.24

2. Exclusion of evidence regarding the inadequacy of the police investigation. During the cross-examination of Trooper Lilly, defense counsel sought to elicit that the trooper had received information from Adolpho Acevito, an individual who was present at the bar on the night of the shooting, that Wanito Mercado, a man “wanted for a couple of other homicides,” shot the victim that night. Defense counsel proffered that she then *800intended to elicit from the trooper that, despite receiving this lead, the trooper never showed Mercado’s photograph to any of the witnesses or followed up on the information in any way.

The Commonwealth objected, arguing that defense counsel was seeking to get hearsay information before the jury that was offered for its truth. Defense counsel countered that she was offering the information for two purposes: as “third party culprit information” and “to impeach the investigation.” The judge found the proffered information to be hearsay, and asked defense counsel if she intended to offer any witness who would testify that Mercado was the shooter. Defense counsel admitted that Acevito, although he was in the bar that night, did not see Mercado shoot anyone. The judge ruled that the information defense counsel wished to elicit was “unreliable hearsay,” because it derived from a person who did not see the shooting. The judge foreclosed defense counsel from exploring whether Trooper Lilly had indeed received the information regarding Mercado and had failed to take reasonable steps to investigate this lead. Defense counsel made an offer of proof that Acevito had told her investigator that the “word of mouth in the community was that [Mercado] had done it and was wanted for a couple of other homicides,” and that she wished to use this information to “impeach the investigation of this officer.”

Two ways in which a defendant may seek to raise a reasonable doubt about his guilt are to elicit evidence that the crime was committed by a third person (third-party culprit evidence) and that the police investigation was deficient (inadequate police investigation evidence). These defenses are sometimes offered simultaneously and often confused with each other, but they are logically (and legally) distinct.

Third-party culprit evidence is “a time-honored method of defending against a criminal charge.” Commonwealth v. Rosa, 422 Mass. 18, 22 (1996). “A defendant may introduce evidence that tends to show that another person committed the crime or had the motive, intent, and opportunity to commit it.” Commonwealth v. Lawrence, 404 Mass. 378, 387 (1989), quoting Commonwealth v. Harris, 395 Mass. 296, 300 (1985). We have given wide latitude to the admission of relevant evidence that a person other than the defendant may have committed the crime *801charged. “If the evidence is ‘of substantial probative value, and will not tend to prejudice or confuse, all doubt should be resolved in favor of admissibility.’ ” Commonwealth v. Conkey, 443 Mass. 60, 66 (2004), quoting Commonwealth v. Keizer, 377 Mass. 264, 267 (1979). Yet, this latitude is not unbounded. The limitations are twofold. First, because the evidence is offered for the truth of the matter asserted — that a third party is the true culprit — we have permitted hearsay evidence that does not fall within a hearsay exception only if, in the judge’s discretion, “the evidence is otherwise relevant, will not tend to prejudice or confuse the jury, and there are other ‘substantial connecting links’ to the crime.” Commonwealth v. Rice, 441 Mass. 291, 305 (2004), quoting Commonwealth v. O’Brien, 432 Mass. 578, 588 (2000). Second, the evidence, even if it is not hearsay, “must have a rational tendency to prove the issue the defense raises, and the evidence cannot be too remote or speculative.” Commonwealth v. Rosa, supra. Each of these hmitations recognizes that the admission of feeble third-party culprit evidence poses a risk of unfair prejudice to the Commonwealth, because it inevitably diverts jurors’ attention away from the defendant on trial and onto the third party, and essentially requires the Commonwealth to prove beyond a reasonable doubt that the third-party culprit did not commit the crime. See id.

Evidence that the police investigation was inadequate is equally admissible, but quite different in the inference it invites. While the inference to be drawn from third-party culprit evidence is simply that someone else committed the crime, the inference that may be drawn from an inadequate police investigation is that the evidence at trial may be inadequate or unreliable because the police failed to conduct the scientific tests or to pursue leads that a reasonable police investigation would have conducted or investigated, and these tests or investigation reasonably may have led to significant evidence of the defendant’s guilt or innocence. A jury may find a reasonable doubt if they conclude that the investigation was careless, incomplete, or so focused on the defendant that it ignored leads that may have suggested other culprits. See Commonwealth v. Bowden, 379 Mass. 472, 486 (1980) (“The fact that certain tests were not conducted or certain police procedures not followed could raise a reasonable doubt as to the *802defendant’s guilt in the minds of the jurors”). See also Commonwealth v. Phinney, 446 Mass. 155, 165 (2006) (“Defendants have the right to base their defense on the failure of police adequately to investigate a murder in order to raise the issue of reasonable doubt as to the defendant’s guilt. . .”).

Because a defendant’s entitlement to present this defense was first articulated by this court in Commonwealth v. Bowden, supra, this has become commonly referred to as the Bowden defense, and generally is focused on the failure of the police to have conducted reasonably expected scientific tests, which was the focus of the defense in Bowden. The defense, however, has not been limited to the failure to conduct scientific tests. We have recognized that the failure of the police to investigate leads concerning another suspect is sufficient grounds for a Bowden defense. See Commonwealth v. Phinney, supra at 166; Commonwealth v. Reynolds, 429 Mass. 388, 391-392 (1999). In Commonwealth v. Phinney, supra, we recognized that police reports detailing the incriminating words and conduct of another suspect would be admissible to show that the police were on notice of the suspect, but failed to investigate his possible involvement in the murder. Similarly, in Commonwealth v. Reynolds, supra at 391, it was error when the judge at a murder trial failed to admit testimony that two “lieutenants” of James “Whitey” Bulger’s crime organization were present at the pub where the victim was last seen, and that two confidential informants had reported to the police that the victim had been fighting over money with Bulger’s organization and recently had a violent argument with Bulger’s lieutenants.

For two reasons, information regarding a third-party culprit, whose existence was known to the police but whose potential involvement was never investigated, may be admissible under a Bowden defense even though it may not otherwise be admissible under a third-party culprit defense. First, under a Bowden defense, the information may be offered simply to prove that the police knew of the possible suspect and failed to take reasonable steps to investigate it. Therefore, it is not hearsay, because it is offered not to show the truth of the matter asserted, but simply to show that the information was provided to the police. Because it is not hearsay, it need not meet the standard we set to admit *803hearsay evidence regarding a third-party culprit, most formidably, the substantial connecting links. That is why we concluded in the Reynolds case that the police detective could testify to what the confidential informants had told him about the Bulger lieutenants’ motive and opportunity to kill the victim, even though the confidential informants’ information may itself not have been first hand.

Second, the risk of prejudice to the Commonwealth from the admission of evidence of a Bowden defense is less, because the police are able to explain what they did to determine that the suspect was not guilty of the crime. See Commonwealth v. Reynolds, supra at 391 n.1. In contrast to the third-party culprit defense, where evidence may be admitted regardless of whether the police knew of the suspect, third-party culprit information is admissible under a Bowden defense only if the police had learned of it during the investigation and failed reasonably to act on the information.25

This does not mean that defense counsel may always be able to admit third-party culprit evidence as part of a Bowden defense even when it would otherwise not be admissible under a third-party culprit defense. The judge would first need to conduct a voir dire hearing to determine whether the third-party culprit information had been furnished to the police, and whether the probative weight of the Bowden evidence exceeded the risk of unfair prejudice to the Commonwealth from diverting the jury’s attention to collateral matters.

Here, the judge was mistaken in ruling that, because the proffered hearsay would not be admissible under a third-party culprit defense, it could not be admissible under a Bowden defense. The evidence potentially could have been admissible if (1) Acevito had indeed told the trooper that the word on the street was that Mercado was the shooter and that he was wanted for “a couple of other homicides,” and (2) the judge determined that the probative weight of the evidence was greater than the risk of unfair *804prejudice. The judge should have conducted a more extensive voir dire hearing, to clarify both whether there was credible evidence that Trooper Lilly had indeed been told this information and, if so, whether the “word on the street” had sufficient indicia of reliability that a jury reasonably could have found that a competent police investigation would have investigated this lead. The latter may have been established if the source of the street information was ascertained and determined to be reliable, if there was other information that potentially linked Mercado to the crime, or if Mercado met the physical description of the shooter or otherwise looked like the defendant.

Without that voir dire, we must determine based on the record before us whether the judge abused her discretion by excluding the proffered testimony.26 Trooper Lilly did not recall receiving any information that another suspect possibly could have been the shooter in this case. When defense counsel asked the trooper whether she recalled interviewing Acevito, the trooper answered, “Possibly.” When asked for a proffer, defense counsel declared that Acevito had told her investigator that the word of mouth in the community was that Mercado was the shooter and was wanted for a couple of other homicides; she did not proffer that Acevito had told this to Trooper Lilly. In view of the inadequate defense proffer, we conclude that there was no abuse of discretion.27

Nor did defense counsel’s proffer demonstrate that the probative weight of this Bowden evidence exceeded the risk of unfair prejudice. The “word on the street” carries no indicia of reliability by itself, and defense counsel did not bolster it by show*805ing that the “word” came from a percipient witness to the shooting or that Mercado was physically similar to the defendant and therefore worthy of further investigation by the police to eliminate him as a suspect.

3. The judge’s slip of the tongue in charging the jury. After her final instructions on the law and before sending the jury to deliberate, the judge instructed the jury on how to complete the verdict slip. In so doing, she erroneously stated: “If you determine that the defendant is guilty of murder in the second degree, you must also determine under which theory or theories of first degree murder that you find him so guilty” (emphasis added). Trial counsel did not object to what was plainly a slip of the tongue, so the error has consequence only if it created a substantial likelihood of a miscarriage of justice.28 See Commonwealth v. Tolan, ante 634, 651 (2009); Commonwealth v. Oliveira, 445 Mass. 837, 842 (2006).

There was not even a remote likelihood of a miscarriage of justice based on this error for three reasons. First, the jury found the defendant guilty of murder in the first degree, and therefore did not reach the question of murder in the second degree. Second, in reviewing the charge in its entirety, no reasonable juror could have been misled by the judge’s single misstatement. See id. at 844-845. Immediately thereafter, the judge twice correctly explained that, if the jury found the defendant guilty of murder in the first degree, they had to be unanimous as to each theory that they found. Third, the verdict slip, to which the judge had referred and which the jury later received, clearly set forth that, to convict the defendant of murder in the first degree, the Commonwealth was required to prove at least one of two theories of guilt: deliberate premeditation or extreme atrocity or cruelty. It also expressly gave the jury the choice of acquittal.

4. Review pursuant to G. L. c. 278, § 33E. Pursuant to G. L. c. 278, § 33E, we also consider errors not raised by the defendant on appeal that arise from the prosecutor’s closing argument.

a. The prosecutor’s closing argument. At the conclusion of the prosecutor’s closing argument, defense counsel objected to *806the prosecutor’s attempt to explain how Braneau and Johnson could have entered the bar with the police after the shooting and failed to identify the defendant. The prosecutor told the jury:

“Doesn’t it make sense that maybe, just maybe, they weren’t in a position, given their frame of minds, to calmly look around the bar that had about thirty-five, forty, forty-five people in it to try to make an identification? Doesn’t it make sense that even if they did, even if they saw everybody in the bar and saw the shooter, they were maybe too scared to identify him, given what they had just seen?”

This argument was error, as was the judge’s overruling of defense counsel’s objection to it. There was no evidence that Braneau and Johnson had identified the shooter in the bar but were too frightened to inform the police of his presence there. Nor is this a fair inference from the evidence, because, if they had seen the shooter in the bar, they could have informed the police after they had left the bar, without any risk that the shooter would have pulled his gun and shot them on the spot.

Defense counsel also timely objected to the prosecutor’s characterization in closing argument of Pichardo’s testimony as to where the defendant was at the time of the shooting, contending that Pichardo had testified that she did not know where the defendant was when the shots were fired. The prosecutor suggested to the jury:

“Pichardo tells us that she is in the doorway when the shooting starts. She tells us not only is the defendant outside the bar at the time of the shooting, but she also puts him right over here by the fire hydrant when she hears this arguing and commotion, as she puts it, just prior to the shots being fired. Think about that. This is where the shooting occurs. We know it not only from the testimony of the eyewitnesses, but that’s where we find all the shell casings.”

The judge could not recall whether Pichardo had put the defendant near the fire hydrant when the shooting commenced, and said she had no notes on it.29 She promised simply to instruct the *807jury that what counsel said in closing argument was not evidence, which she included as part of her final instructions to the jury.

The prosecutor was mistaken in his characterization of Pichar-do’s testimony. As described earlier, Pichardo testified that, before the shooting, she stood in the doorway of the bar and saw the defendant walk toward the parking lot. The black males drove by the bar; before they approached the bar (apparently referring to the victim and his friends), she went back inside the bar to the bathroom. After she left the bathroom, which she estimated to be ten to fifteen minutes later, she returned to the doorway of the bar, and it was then that she heard the gunshots. She testified that, just before the shooting started, she saw “a whole bunch of people in a group in the parking lot.” When asked, “Did you see where [the defendant] was at that point in time?” she answered, “No, I didn’t.” In short, Pichardo did not testify that she saw the defendant in the parking lot at the time of the shooting; she saw him there roughly ten to fifteen minutes earlier. Indeed, she testified that she did not see him among the group in the parking lot at the time of the shooting. The prosecutor’s mischaracterization in closing argument of a critical point of Pichardo’s testimony was error, as was the judge’s failure to correct it.30

b. Harmless error analysis. Having identified two errors of consequence in the prosecutor’s closing argument, we now must determine whether these errors, viewed collectively, require reversal of the conviction. Whether such an error is reversible “depends on our consideration of ‘(1) whether the defendant seasonably objected; (2) whether the error was limited to collateral issues or went to the heart of the case; (3) what specific or general instructions the judge gave the jury which may have mitigated the mistake; and (4) whether the error, in the circumstances, possibly made a difference in the jury’s conclusions.’ ” Commonwealth v. Perez, 444 Mass. 143, 151 (2005), quoting Commonwealth v. Kater, 432 Mass. 404, 422-423 (2000). We address each factor.

Here, timely objections to both closing argument errors were made immediately after the prosecutor’s closing argument. *808Neither objection was sustained. The only instruction the judge gave that may have mitigated the error was her reminder to the jury in her final instructions that “the closing arguments of the lawyers are not a substitute for the evidence. They are only intended to assist you in understanding the evidence and the respective contentions of the parties.” The judge did not focus on any statement in the prosecutor’s closing argument when she provided this guidance, so the jury were not warned to be careful in comparing their memory of Pichardo’s testimony with the attorneys’ characterization of it.

The errors very much went to the heart of the case. The most troubling weakness of the prosecution case was that Bruneau and Johnson, immediately after the shooting, accompanied by police, walked into the bar where the defendant was present for the purpose of determining whether the shooter was inside, and did not identify the defendant as the shooter, even though he was one of only two males wearing a red shirt (indeed, a distinctive red shirt). By suggesting to the jury that they had identified the defendant as the shooter but were too frightened to say so, the prosecutor sought to limit the damage from their apparent failure to make any identification inside the bar.

Pichardo’s testimony provided the strongest corroboration of the eyewitness identifications, because she placed the defendant outside in the parking lot at or about the time the victim and his friends arrived at the bar. Pichardo knew the defendant, so her identification of him was not seriously in question. The prosecutor, however, told the jury that Pichardo had placed the defendant near the fire hydrant where the shooting occurred at the time of the shooting, even though Pichardo testified that she had not seen where the defendant was at the time of the shooting. In short, he transformed into inculpatory testimony the exculpatory part of her testimony as to whether Pichardo had seen the defendant among the crowd that gathered in the parking lot at the time of the shooting.

We cannot say with assurance that the closing argument errors, considered together in the totality of the circumstances, could not have influenced the jury to convict. See Commonwealth v. Coren, 437 Mass. 723, 733 (2002). The jury plainly struggled with the evidence, twice declaring that they were deadlocked *809before ultimately reaching a verdict. The case hinged mainly on identification evidence and the testimony of a fifteen year old girl. Professor Douglass’s testimony would have sensitized the jury to the danger of inaccurate identification testimony, if they did not already appreciate that danger. The identifications of the defendant from the photographic arrays were equivocal, even without considering the failure of Bruneau and Johnson to identify the defendant in the bar after the shooting. Bruneau was the only one of the three to make a positive identification, and he qualified his identification with statements to the effect that he only got a “glance” at the shooter and “didn’t really see his face.” Johnson said the defendant’s photograph was the most similar to the shooter; he could not make a positive identification. Jenkins said only that the defendant’s photograph in the array looked “most familiar” to the shooter.

The presence of the defendant in the bar after the shooting was also two-edged for the prosecution. While it certainly demonstrated that he was there that evening, it reasonably raised the logistical question as to how the defendant could have shot the victim; ran down the alley beside the bar; got rid of his gun, chain, and headband (which were never located by the police); and returned inside the bar, all within the short time before the police arrived at the bar and secured it as part of the crime scene.

The seizure of the defendant’s shirt and sneakers did not strengthen the case against the defendant. The defendant did not contest the fact that he owned the shirt and sneakers seized from his apartment. Both of those items tested negative for the presence of blood, and the shirt tested negative for the presence of gunpowder residue. While in their descriptions of the shooter, Johnson, Bruneau, and Jenkins consistently described the shooter as having worn a red shirt, no one, including the driver who passed by and observed part of the shooting, mentioned the very distinctive features of the red shirt worn by the defendant that night — a fairly large rhinoceros in the center of the front of the red shirt, and a large “72” on the back.

Given the underwhelming strength of the other evidence, the testimony of Pichardo may have been pivotal, especially if the jury mistakenly accepted the prosecutor’s characterization of her testimony as placing the defendant at the precise spot where the *810shooter first fired. We cannot say with assurance that the jury would have accurately recalled her testimony and discounted the prosecutor’s mischaracterization of it. Indeed, after closing arguments, the judge could not recall precisely what Pichardo said as to the defendant’s whereabouts at the time of the shooting, and her notes shed no light on the issue. While the collective memory of twelve jurors often is better than that of a single judge, we cannot say with assurance that it was here. As a result, we cannot say with assurance that the jury’s verdict would have been the same had the prosecutor not made these errors in his closing argument.

Conclusion. We reverse the defendant’s conviction of murder in the first degree, set aside the jury’s verdict, and remand the case to the Superior Court for a new trial.

So ordered.

The Commonwealth had proceeded also under a theory of deliberate premeditation, which the jury rejected.

Before the grand jury, Bruneau stated that he saw two males and one female. At trial, he testified that there was one male and two females. There was other trial testimony on this point. Johnson testified that two males and three females were at the front door. Jenkins recalled simply that there were people outside. Thomas did not testify at trial.

The victim and his friends were black. Johnson testified that the man made the racial epithet as they were entering the bar, not as they were leaving. Jenkins recalled that, as they headed into the bar, words were exchanged indicating that they were not going to be allowed inside, and further words were exchanged as the group left the bar.

Johnson’s memory of events differed somewhat from Bruneau’s. Johnson recalled that, as they were walking back to their automobile, the man asked the victim, “What are you looking at?” The victim asked the same question of the man. The man said he would be back, walked to the front door of the bar, and quickly returned to where the victim was standing with his friends, standing face to face with the victim. When the victim asked, “What you got — the toast?” the man pulled a handgun from his waistband and fired at the victim, at which point the others fled.

Jenkins recalled that, as he and his friends were returning to the automobile, the man lifted up his shirt, revealing a pistol. After the victim asked him what he was going to do with it, the man laughed, pulled the pistol out, and started shooting at the victim.

Bruneau was not asked to state his height for the record.

The forensic pathologist who performed the victim’s autopsy, Dr. James Weiner, had retired by the time of trial. Consequently, a different forensic pathologist, Dr. Faryl Sandler, testified at trial concerning the victim’s injuries and cause of death. Dr. Sandler’s testimony was based on her review of Dr. Weiner’s autopsy report, photographs taken during the autopsy, a police report, and hospital records. The defendant’s trial counsel did not cross-examine Dr. Sandler or contest that the victim’s gunshot wounds caused his death. Contrast Commonwealth v. Nardi, 452 Mass. 379, 383, 384-395 (2008).

The testimony varied concerning the number of people inside the bar. The police, however, recorded on videotape the interior of the bar and each patron’s eventual departure therefrom. It is apparent from this videotape, which was entered in evidence, that approximately forty-two persons were inside and then left the bar.

According to Detective Williams, before he separately brought Bruneau and Johnson inside the bar, he instructed that each was to look around inside for the shooter. With Bruneau, they walked together to the middle of the bar. Bruneau became emotional, screaming that his friend had just been shot outside. After only five or six seconds inside, Detective Williams escorted him outside. With Johnson, virtually the same thing happened. Detective Williams walked him to the middle of the bar, Johnson became emotional and screamed that his friend had been shot outside, and they departed after only five or ten seconds inside.

Bruneau testified that he “walked around” the interior of the bar, in which there were “a lot of people,” but did not see the shooter. Johnson stated that, when he was escorted inside the bar, he stayed in one spot and looked around at each person, but was unable to pick anyone out. Neither said anything about becoming so emotional that he had to leave.

During her cross-examination, Trooper Lilly stated that she had included the defendant’s photograph in the array because, according to the descriptions made by witnesses, his clothing was similar to that worn by the shooter. For that reason, she considered the defendant a suspect.

Specifically, Johnson testified that he “pointed out the defendant,” because the photograph of the defendant “look[ed] just like [the shooter]." Johnson went on to make an in-court identification of the defendant.

During his cross-examination, Jenkins testified that he told Trooper Dateo that he was not “one hundred per cent sure” about his identification.

During Professor Amy Douglass’s cross-examination, she conceded that the defendant’s trial counsel told her that the witnesses who had identified the defendant observed the defendant in the bar and saw him on the videotape. There was no evidence that Jenkins, Johnson, or Bruneau had ever seen the videotape.

Professor Douglass noted that, in the Report to the Legislature of the State of Illinois: The Illinois Pilot Program on Sequential Double-Blind Identification Procedures (2006) (Illinois study), the sequential procedures were double-blind while the simultaneous procedures were not. It was not elicited at trial that many of the identifications in the Illinois study were conducted by a live lineup rather than a photographic array. Illinois study, supra at iv, 24. Illinois courts prefer a live lineup identification over a photographic identification when a suspect is in custody and a live lineup is therefore possible. See People v. Holiday, 47 Ill. 2d 300, 307 (1970). Nor was it elicited at trial that an identification was determined to be a “false” identification under the Illinois study if the witness identified one of the individuals selected by computer to fill out the lineup or array; an identification was not determined to be false if the witness identified the suspect. Illinois study, supra at 29-30.

The judge made her conclusion based on “the number of witnesses in this case, given that it was a four day case, and that [the jurors] only deliberated for three hours, during which they ate lunch as well [as] watched the [videotape].”

Of the three witnesses, Johnson was the only one who made an in-court identification at trial.

The judge who ruled on the motion to suppress the identifications was not the trial judge.

We add that the argument would not prevail even if it had been preserved. There is no evidence that Bruneau, Johnson, or Jenkins saw the defendant inside the bar when they briefly first entered and were asked to leave because Jenkins was underage. Only Bruneau and Johnson (not Jenkins) were taken inside the bar after the shooting. While there was testimony that these two witnesses did look around the bar (and did not see the shooter), there was also testimony, from a police officer, that the attempt to have them look for a *796suspect therein was futile because they grew so upset inside and, within seconds, had to be escorted outside. In these circumstances, it was not unnecessarily suggestive to include the defendant’s photograph in the array. The risk that these witnesses confused someone they had seen on the night of the shooting with their memory of the shooter, which Professor Douglass described as the danger of unconscious transference, was made known to the jury through her testimony.

Trooper Lilly plainly understood that the defendant was a suspect because she placed his photograph within the array; the evidence does not reflect whether Lieutenant Mason or Trooper Dateo also knew that the defendant was a suspect.

The Commonwealth does not contest these points.

“Of the first 163 exonerations secured through the use of DNA [deoxyribonucleic acid] evidence, for example, we know that seventy-seven per cent of the convictions were the product of mistaken eyewitness identifications.” Commonwealth v. Martin, 447 Mass. 274, 293 & n.4 (2006) (Cordy, J„ dissenting), citing Mistaken Eyewitness Identifications, The Innocence Project (2006).

We recognize that there may not be evidence of any signaling even when it occurs, because the law enforcement officer may not be consciously aware that she is sending signals and the signals may be so subtle that the eyewitness may not understand them to be signals.

Compare Klobuchar, Improving Eyewitness Identifications: Hennepin County’s Blind Sequential Lineup Pilot Project, 4 Cardozo Pub. L. Pol’y & Ethics J. 381,410-413 (2006) (field data from pilot project favor blind sequential *799method of identification), with Malpass, A Policy Evaluation of Simultaneous and Sequential Lineups, 12 Psychol., Pub. Pol’y, & L. 394 (2006) (simultaneous lineups superior to sequential lineups under most conditions). See also Thompson, What Price Justice? The Importance of Costs to Eyewitness Identification Reform, 41 Tex. Tech. L. Rev. 33, 47-48 (2008) (reform groups split on whether sequential procedures should be required).

A preidentification protocol would diminish the need for a recording if the police were to follow the protocol in every identification procedure.

In view of our conclusion, we need not address the defendant’s contention that the subsequent in-court identification was tainted by the out-of-court identification. Also, where the denial of the defendant’s motion to reconsider is based on the same arguments he advanced concerning the denial of his motion to suppress the identifications, we need not separately address the propriety of the denial of his motion to reconsider.

The defense articulated in Commonwealth v. Bowden, 379 Mass. 472, 485-486 (1980) (Bowden defense), therefore, is a two-edged sword for the defendant, because it opens the door for the Commonwealth to offer evidence explaining why the police did not follow the line of investigation suggested by the defense. See Commonwealth v. Lodge, 431 Mass. 461, 467-468 (2000); Commonwealth v. Reynolds, 429 Mass. 388, 391 n.1 (1999).

While the exclusion of third-party culprit evidence is of constitutional dimension and therefore examined independently, see Commonwealth v. Conkey, 443 Mass. 60, 66 (2004), the exclusion of evidence of a Bowden defense is not constitutional in nature and therefore is examined under an abuse of discretion standard, see Commonwealth v. Mayfield, 398 Mass. 615, 629 (1986).

By contrast, in Commonwealth v. Reynolds, 429 Mass. 388, 390 (1999), the police detective who obtained the third-party culprit information from the confidential informants testified at a pretrial hearing that he had passed on this information to the detective leading the investigation of the murder of which Reynolds was convicted. Similarly, in Commonwealth v. Phinney, 446 Mass. 155, 165-167 (2006), there were police reports reflecting receipt of the third-party culprit information.

We assume that this error arose from a slip of the tongue, but recognize that it may also have arisen from a stenographic error of transcription, which may explain the absence of any objection.

The prosecutor claimed to be certain of his memory, telling the judge, “I remember the testimony specifically. She put him right in the comer.”

The judge did not have available to her a transcript of Pichardo’s testimony, as we have on appeal.