The opinion of the Court was delivered by
O’HERN, J.In this capital case the State does not disagree that the death sentence must be vacated. Point XIII of the State’s brief summarizes the point:
Defendant contends that the trial court's charge during the penalty phase with respect to the jury’s consideration of mitigating factors was erroneous in that it required the jury to unanimously find an alleged mitigating factor before it could be considered. The State is constrained to agree that the court’s charge with respect to the need for juror unanimity before finding the existence of a mitigating factor violates the principles subsequently enunciated in Mills v. Maryland, 486 U.S. 367 [108 S.Ct. 1860], 100 L.Ed.2d 384 (1988), and State v. Bey II, 112 N.J. 123, 159-60 [548 A.2d 887] (1988). [Footnote omitted.]
Acknowledging that defendant’s death sentence with respect to the underlying murder must, therefore, be vacated, the State seeks to have us remand the matter for a new capital-sentencing proceeding unless the Court determines that the murder conviction does not establish death eligibility. In that circumstance, the State asks that the conviction of non-capital murder be affirmed and the case be remanded for resentencing on the murder count to a term of imprisonment, which for murder is between thirty years and life with a minimum of thirty years *229without possibility of parole. N.J.S.A. 2C:11-3b. Defendant, in addition to challenging the death sentence on the grounds that the trial court’s charge required juror unanimity with respect to the mitigating factors, asserts numerous challenges to the underlying convictions of murder and the related offenses. We find that the conviction for murder did not establish death eligibility, but affirm that conviction in all other respects. Pursuant to the State’s election so to proceed, we therefore remand the cause for imposition of the non-capital murder sentence.
I
The case arises from the brutal murder of a thirteen-year-old girl as she walked home from school. To her last moment, she fought against her stronger assailant. The marks she left on his body and the telltale presence of fibers drawn from his clothing, along with eyewitness testimony of fellow students, sealed the case against her assailant, defendant, Phillip Dixon.
As the young girl, Tanya, walked home from school on ' Friday afternoon, February 22, 1985, several of her fellow students saw defendant “on top of” her in an area of underbrush along a path between the children’s school and homes. (The area was in the Borough of Woodlynne, although the children attended Camden High School.) Although at first the other students thought that there might have been nothing more than an innocent encounter between the two, their suspicions deepened when Tanya did not soon arrive home. The children alerted Tanya’s mother of the fact that they had seen Tanya with defendant, an eighteen-year-old fellow student at the high school. Her mother went to defendant’s house in search of Tanya, but he was not home.
The police were informed of the missing child. An intensive search disclosed her body approximately one hundred yards from the place of the sighting by the students. Her partially-nude body had been dragged through the underbrush into a *230creek. Her body was lodged in the water underneath a car seat and other discarded refuse. Only a foot was showing above the surface of the water.
After his encounter with the victim, defendant went to his cousin’s home, where he changed his clothes. He claimed that he had been in a fight and called his brother to bring the change of clothes. Later that afternoon, defendant returned home, where his mother told him that the victim’s mother had been there asking about Tanya. Defendant again changed his clothes, and that evening went with his brother to Philadelphia. Later that night he went to his grandmother’s home in Hemp-stead, Long Island. Having been informed by the school children that Tanya had last been seen with defendant, the police put out an all-points alert for him. They soon learned that he was in Hempstead at his grandmother’s home. The Woodlynne police called the Hempstead police, who arrested defendant on Sunday afternoon, February 24, 1985.
Defendant gave an oral confession to the Hempstead police. A Hempstead officer summarized defendant’s statement as follows: On Friday afternoon defendant was walking with his mother to a local bank. He remembered that he needed money to see a movie later, so he returned home to get some money. While returning, he was walking along a path and saw a young girl. He decided to take her pocketbook. He chased her, grabbed her, and forced her down to the ground in a “weeded [sic] area,” at which point she was screaming and struggling. She eventually flipped onto her stomach and he straddled her with his knees. But she screamed as he tried to take her pocketbook. She looked at him and said “I know you, I’ve seen you.” As she continued to scream, he reached for “a spike or a nail” lying on the ground and hit her on the head with it. Defendant did not know why he struck the girl and could not remember the amount of pressure he used or whether the nail had penetrated the girl’s head. He said that the girl had been screaming “like in the movie ‘10 to Midnight.’ ” When the officer said that he had not seen the movie, defendant said, “[I]t *231was like in that movie when the girl in the movie kept screaming and she wouldn’t stop screaming and the guy stabbed her.”
A grand jury indicted defendant for the murder of Tanya as well, as for a variety of other offenses, including robbery, aggravated criminal sexual contact, hindering apprehension (by concealing her body, destroying evidence, and fleeing), and various other offenses.
At trial, the State produced the school children who had seen defendant with Tanya, on top of her, apparently engaged in a scuffle. They recalled that he wore a camouflage jacket. Another witness described seeing defendant drag Tanya into the woods towards the water. A fiber expert described the fibers found on her body as being identified with defendant’s cap. A sneaker imprint was found at the scene that matched the Nike sneakers seized at defendant’s cousin’s home. A pathologist said that the victim had been struck by a pointed object, that the blow to Tanya’s head had pierced her brain, and that death was inevitable from the blow, although she was probably alive when her body was submerged under the water. Two scenes or segments from the movie “10 to Midnight” were shown in which a perpetrator in dissimilar circumstances was stabbing a screaming young woman.
Defendant took the stand on his own behalf. He denied that he had committed the murder, asserting that he had never given the alleged confession and that on the date of the crime he had been wearing a blue Army jacket and red suede Puma sneakers. The Hempstead police officer who had interrogated him testified that defendant had told him that he had been wearing those clothes on Friday afternoon.
The jury convicted defendant of most of the counts but not the robbery count. At the sentencing proceeding, the State charged two aggravating factors: that the murder had involved torture, aggravated battery, or depravity; and that the murder had been committed to avoid apprehension. The jury unanimously found only two mitigating factors and imposed a sen*232tence of death. Defendant appeals to us as of right under Rule 2:2-1(a)(3).
II
Pretrial Issues
A. Composition of Grand and Petit Juries
Defendant contends that the Camden County jury-selection system violated his rights to a jury drawn from a fair cross-section of the community and to equal protection of law, and also violated his statutory right to a randomly-drawn representative jury, in violation of the provisions of N.J.S.A. 2A:70 and :71.
1. The Constitutional Issues
In evaluating the constitutional challenge, we briefly review the principles set forth in State v. Ramseur, 106 N.J. 123, 215-38, 524 A.2d 188 (1987). Without restating the principles in detail, we may summarize by stating that under the constitutional guarantees, selection of both grand and petit juries must be free from any taint of discriminatory purpose and the jurors must be drawn from pools that represent a “fair cross-section of the community.”
To prove either an equal-protection or a fair-cross-section claim, defendant must (1) identify a constitutionally-cognizable group, that is, a group capable of being singled out for discriminatory treatment; (2) prove substantial underrepresentation over a significant period of time; and (3) show discriminatory purpose either by the strength of his statistical showing or by showing the use of racially non-neutral selection procedures to support the inference of discrimination raised by substantial underrepresentation. State v. Ramseur, supra, 106 N.J. at 215-16, 524 A.2d 188.
In sum, we primarily focus on the cognizability of the group in question, the substantiality of the underrepresentation, and the possible causes of it.
*233For purposes of this appeal, we shall accept that the assertedly underrepresented minorities (Blacks, Hispanics, and Puerto Ricans) would meet the first prong of the test. We shall also accept, for purposes of this appeal, the statistics set forth in defendant’s brief with respect to the representation of those minorities. The data are as follows:
UNDERREPRESENTATION OF MINORITIES Camden County
DATA STANDARDS OF REPRESENTATIVENESS
Qual. Stat.
Pop’n. Pool Absolute Comparative Sign.
GROUP_(%) (%) (%)_(%)_(sd)_
Blacks 13.37 9.57 3.8 28.42 2.27
Hispanics 3.27 .93 2.34 71.56 2.59
Puerto Ricans 2.65 .47 2.18 82.26 3.28
In Ramseur we explained the meaning of those various standards of representativeness. 106 N.J. at 219-27, 524 A.2d 188. We shall not repeat them here except in the simplest terms. Absolute disparity measures the difference between the proportion of the subject group in the general population and its proportion to the jury pool, that is, the qualified list. Comparative disparity builds on that figure by using the absolute-disparity figure and constructing a ratio to measure the magnitude of the disparity given the difference in population size. For example, a four-percent absolute disparity in a fifty-percent population represents an eight-percent comparative disparity, whereas that same four-percent absolute disparity in a population of eight percent is a fifty-percent comparative disparity. The comparative-disparity standard is more likely to register underrepresentation of smaller groups. The smaller the group, the more significant is any deviation. The third approach, the statistical significance test, measures the likelihood that aspects of the selection process do not operate randomly. That test is an entirely statistical analysis in which the statistician assumes *234absolute neutrality in the selection process and measures the statistical likelihood of deviation from the expected. In Ramseur, we gave the following example. The jury-selection process in which two groups are being compared can be likened to filling a box with a population of one thousand slips of paper of which six hundred are pink and four hundred are grey, and having someone randomly select a sample of one hundred slips. The expected number of pink slips would be sixty and the expected number of grey slips would be forty. That is, in any drawing there would be a sixty percent probability of drawing a pink slip and a forty percent probability of selecting a grey one. However, a statistician would not be surprised if the number of pink slips deviated from the expected. Statisticians measure that deviation by a formula that enables them to tell whether the result is so far from the expected as to demonstrate the result was not random or by chance. A complex mathematical formula determines that standard deviation. For example, in the case of the one-hundred-slip sample, a statistician expects that the standard deviation would be plus or minus 4.8 slips. (Calculations omitted, see Ramseur, supra, 106 N.J. at 221 n. 44, 524 A.2d 188.) If the result of a drawing were to yield only thirty pink slips, that would be approximately six standard deviations (6 X 4.8) away from the expected. A statistician assumes that a result more than two or three standard deviations from the expected would be suspect.
In this case the statistics submitted with respect to underrepresentation of minorities are not disturbing except with respect to the comparative disparity of Hispanies and Puerto Ricans. As we noted in Ramseur, “if these cases [dealing with under-representation of groups] have a common thread, it is that a comparative disparity well over fifty percent is strong evidence of underrepresentation cognizable under the sixth and fourteenth amendments.” 106 N.J. at 220, 524 A.2d 188 (quoting State v. Lopez, 107 Idaho 726, 733, 692 P.2d 370, 377 (Ct.App. 1984)).
*235With respect to the comparative disparity concerning under-representation of Hispanics and Puerto Ricans, we are satisfied that that does not demonstrate a constitutional underrepresentation because of the very small percentage of the population pool thereby represented. A comparative disparity of about fifty percent may or may not be adequate to show such under-representation, depending in part on the size of the group in question. State v. Lopez, supra, 107 Idaho at 733, 692 P.2d at 377. The absolute disparity becomes magnified when converted to a measurement of comparative disparity because of the smallness of the sample used.
The statistical-significance test confirms the inefficacy of the comparative-disparity figure. In the case of Hispanics and Puerto Ricans, the statistical significance of the discrepancy is very close to the two or three standard deviations from the expected and thus is not constitutionally suspect.
But even were those numbers themselves to straddle the borderline of substantial underrepresentation, we would look to the circumstances surrounding the statistical showing to determine its full constitutional import. Generally speaking, when jury-selection systems have been found to be constitutionally underrepresentative on the basis of statistical showings of underrepresentation, objective selection criteria such as voting registration and drivers’ licenses, as were used in this case, are not present. See State v. Ramseur, supra, 106 N.J. at 225, 524 A.2d 188. In addition, courts have looked at the time period over which violations are alleged. That inquiry goes to the existence of a history of exclusion. In this case, as in Ramseur, we have evidence based on only one survey of 500 jurors from a 1984 mailing. Not only does the survey fail to cover a significant period of time, but reforms were apparently made to the system the following year, rendering the data and resulting conclusions at least partly outdated.
Given the borderline nature of the disparities shown, we do not find inadequacies that rise to constitutional dimension.
*2362. The Statutory Issues
N.J.S.A. 2A:70-4 requires the assignment judge of a county to merge the lists of registered voters and holders of motor vehicle licenses to compile a single list from which all jurors shall be selected.
Defendant makes specific objections that he asserts demonstrate such a deviation from established statutory requirements that the indictment must be dismissed. For purposes of this appeal we shall accept defendant’s version of the record. (The State will be free in any collateral proceedings to dispute these assertions.)
(1) Elimination of Zip Codes. In the mailing of questionnaires before the spring of 1985, seven or eight zip codes were eliminated from the drivers’ license lists. The population within Camden County for those zip codes represented over one-fifth of the whole county population. The various jury-selection officials agreed that most of the zip codes were eliminated because they embraced areas that fell partially outside of Camden County.
(2) Failure to Eliminate Duplicates. This is the “match-merge” problem that we described in State v. Gerald, 113 N.J. 40, 130-31, 549 A.2d 792 (1988). The computers fail to recognize that some people use different names on voting or driver lists, resulting in two “Jones” at the same address.
(3) History File Exclusion. Once a juror in Camden has been excused for one reason or another, the name is put in a “history file.” Often a juror who may be disqualified at one time (as for pregnancy) could be later qualified, but insufficient attention was paid to the “history file” to restore names to the eligible list. The result, claims defendant, is that as many as two-thirds of the eligible jurors are in that category of ineligibility.
(4) Inactive-Voter File. The inactive-voter file was used to cull names from the jury list, but the fact that one does not vote does not mean that one is not eligible to serve as a juror. Although there may be some who are removed .from voting lists *237because they have died or moved, there was an insufficient explanation of why that practice should be followed, especially in the case of persons whose names appeared on the motor-vehicle lists.
(5) High No-Response Rate. Defendant points to inaction by the jury commissioners in the face of a persistently high no-response rate of up to 32.8%. That means that 32% of the people to whom questionnaires had been sent simply failed to return them or otherwise to show why they should be included or excluded from the juror lists. A court directive requires that a no-response rate above 15% be reported to the assignment judge.
(6) Improper Disqualification. Defendant claims that the non-retention of the received questionnaires of disqualified jurors made it impossible to resolve conflicting testimony regarding (1) medical disabilities regularly allowed without documentation; (2) exemptions for care of minor children; (3) exclusion of students who were county residents; and (4) exclusion of teachers even during summer months. That was described by defendant as making it very easy for county residents to avoid jury service and compounds the other defects set forth in his brief.
In State v. Ramseur, supra, 106 N.J. 123, 524 A.2d 188, we considered whether discrepancies in statutory or administrative directives with respect to jury selection necessitated the dismissal of an indictment. We repeat what we said there: “We do not construe our state’s statute as requiring dismissal of the indictment whenever the statutory commands are breached, regardless of the nature or effect of the violations or the intent of those who committed them.” Id. at 231-32, 524 A.2d 188. The judicial power to dismiss an indictment is not to be exercised except on the clearest and plainest grounds, and an indictment should stand unless manifestly deficient or palpably defective. State v. Wein, 80 N.J. 491, 501, 404 A.2d 302 (1979); State v. Weleck, 10 N.J. 355, 364, 91 A.2d 751 (1952). We said *238that violations of our jury-selection regulations should warrant a dismissal of an indictment only “where they substantially undermine the randomness and objectivity of the selection mechanism or cause harm to the defendant.” State v. Ramseur, supra, 106 N.J. at 232, 524 A.2d 188.
With respect to the selection of petit juries, we recognize that in some circumstances the special fundamental role played by the petit jury in our system of criminal justice may call for reversal of the conviction because of improper selection procedures, even in the absence of a showing of prejudice. State v. Kociolek, 23 N.J. 400, 406, 129 A.2d 417 (1957); State v. Wagner, 180 N.J.Super. 564, 567, 435 A.2d 1190 (App.Div.1981). In Kociolek, the trial court had employed one of its regular criminal panels, contrary to a statute requiring the special drawing of a forty-eight-member panel prior to trial for murder. A majority of the Court agreed that that violation alone would have required reversal, even absent a showing of actual prejudice. The Court described the statutory requirement as “a course of procedure so imperative in expression as to bar waiver by the defendant or his counsel where the indictment is for treason or murder.” 23 N.J. at 406-07, 129 A.2d 417.
State v. Wagner, supra, involved an even more blatant violation of the statutory provisions governing the selection of petit juries that required jurors’ names to be “[d]rawn from the box, one at a time.” 180 N.J.Super. at 567, 435 A.2d 1190 (quoting N.J.S.A. 2A:74-1). The trial court simply seated the first fourteen jurors to enter the courtroom and replaced excused jurors by calling on other jurors seated in various rows by “going right down the row.” Id. at 566, 435 A.2d 1190. Those are such plain and blatant violations of law that they cannot be condoned.
But what we have here are inefficiencies rather than invidiousness. There is no suggestion whatsoever that any of the failings were in any way intended to undermine the randomness of the jury-selection process. Underrepresentation is not de*239fined by comparing the representation that “would have occurred randomly” with the actual representation of the group on the jury. Rather, “underrepresentation must be proved, by comparing the proportion of the group in the total population to the proportion called to serve as grand jurors.” Castaneda v. Partida, 430 U.S. 482, 494, 97 S.Ct. 1272, 1280, 51 L.Ed. 2d 498, 510 (1977). In addition, although defendant is indeed entitled to have his challenge based on the array and not the individual panel chosen, we note simply that of the twelve-person panel actually chosen for the trial and penalty phases, two of the five men were black and two of the seven women were black. Of the four alternates, all women, one was black.
Although the procedures used obviously implicated the randomness of the selection process, there is no showing that they substantially undermined the randomness principle. When, as here, the purpose of the official’s action was not impermissibly to exclude members of a cognizable group, the asserted statutory violations do not call for a dismissal of the indictment or a reversal of the conviction. Defendant interprets the opinion of the court below as not calling for a correction or improvement in the noted deficiencies, but we are certain that the motion judge who now sits as the assignment judge intended his opinion to have that effect. When he said that the “Camden County Selection Process had to be corrected,” but that the “deviations from the New Jersey Jury Selection Statutes directives do not rise to a level constituting grounds for relief,” we are certain he intended that changes be made.
In sum, although the noted deficiencies may be considered somewhat more than merely technical, here, as in State v. Long, 119 N.J. 439, 469, 575 A.2d 435 (1990), and State v. Gerald, supra, 113 N.J. at 131, 549 A.2d 792, there was no evidence that the grand- and petit-jury panels, as composed, were not representative of the community, nor was there any suggestion that the independence of the grand or petit jury had been compromised. None of these practices reduced a minority’s representation to “impotence” or “restrict[ed] unreason*240ably the possibility that the petit jury will comprise a representative cross-section of the community.” State v. Gilmore, 103 N.J. 508, 529, 511 A.2d 1150 (1986). The methods used were in no sense an attempt to undermine or inject invidious discrimination into randomness. The goal was an objectively random list that reached substantially all eligible sectors of the juror pool.
B. The Validity of the Confession
Defendant contends that he was not given full and effective warning of his constitutional right to remain silent as he was not informed that he had a continuing opportunity to exercise that right, and he was never told that he could terminate any interrogation. In addition, he contends that although he was not informed of his right to terminate questioning, the police had reason to believe he was attempting to exercise that right when, after having answered questions for a time and after having agreed to make a written statement, he refused to sign anything. The Hempstead police officers testified to a marked change in defendant’s attitude following a telephone conversation with his grandmother, after which he was “not cooperative” and became “somewhat hostile.” Defendant contends that his actions constituted assertion of his right to remain silent and, as part of their obligation scrupulously to honor defendant’s fifth-amendment rights, the police should have made an effort to ascertain the significance of the distinct change in attitude. Defendant asserts also that he requested counsel at the station house, a request that would have prevented the police from questioning him before counsel was provided. That claim was not raised below and is part of a claim of ineffective assistance of counsel, which we discuss in Point III(F) hereof, infra at 259-262, 593 A.2d at 284-285.
When a suspect makes a statement that arguably amounts to an assertion of Miranda rights, and the interrogating agent recognizes that the statement is susceptible of that construction, police questioning concerning the crime should immediately cease and the officer then should inquire of the suspect *241about the correct interpretation of the statement. State v. Bey II, supra, 112 N.J. at 136, 548 A.2d 887.
Each side presents a different version of the interrogation. As noted, on Sunday afternoon, February 24, 1985, Detective McLaughlin of the Nassau County police department was notified of a teletype from the Woodlynne, New Jersey, police department. The teletype indicated that there was a warrant for defendant’s arrest and set forth a description of defendant and his probable location. Based on the information contained in the teletype, McLaughlin proceeded to the specified location.
When McLaughlin arrived at the apartment house, he found the front door open and other police officers already present. McLaughlin spoke privately with defendant’s uncle, Robert Newbill, explaining that defendant was sought in connection with a homicide. Defendant surrendered to McLaughlin, who handcuffed him, informed Newbill that defendant was being taken to headquarters, and left the apartment with defendant. McLaughlin testified that he informed defendant of his rights while en route to the police station: that he had the right to remain silent, that anything he said could be used against him, that he had a right to an attorney, and that if he could not afford one, an attorney would be provided for him. At police headquarters, defendant was taken to an office in the basement. Detective Martin Alger of the Homicide Squad was notified. Newbill came to the police station. McLaughlin informed him that defendant was talking to homicide detectives. Newbill asked if he could see defendant and was told that he could not at that time. According to the police, Newbill made no mention of securing an attorney.
Detective Alger conducted the interrogation. He informed defendant of his rights from memory. He did not produce a card, and he concedes that he did not specifically inform defendant that he had a right to terminate questioning at any time. He asked defendant if he could speak with him without having an attorney present. Defendant agreed to speak to him. At *242about 6:00 p.m., defendant provided the narrative of the offense that we set forth as summarized by a Hempstead police officer in our recital of the facts.
The trial court, after considering the totality of the circumstances, properly concluded that defendant’s statements were voluntarily made after a knowing and intelligent waiver of his constitutional rights and were not the product of interrogation that resulted in the overbearing of defendant’s will. See State v. Bey II, supra, 112 N.J. at 134-35, 548 A.2d 887 (State must establish the voluntariness of the confession beyond a reasonable doubt).
Defendant maintains that when his demeanor changed following the conversation with his grandmother, the officers should have made an effort to ascertain the significance of the distinct change in his attitude instead of simply ignoring that change and continuing their interrogation. The police testified that defendant became “very tense, very agitated,” and that he refused to write anything. It is not reasonable to construe defendant’s behavior as a cutoff of questioning. Here defendant was twice informed of his rights. Included in those warnings was the information that if defendant spoke with the detectives, anything he said could be used against him. Defendant knew that he had been arrested for homicide. That he was not specifically informed that he was subject to the death penalty does not disqualify the confession. Defendant was not a juvenile whose age is relevant to a voluntariness determination. Nor was it fatal that he was not told in so many words that he had the right to terminate questioning at any time, although obviously it is preferable that such a warning be specifically stated. It is the substance of the warning, however, that counts. Duckworth v. Eagan, 492 U.S. 195, 109 S.Ct. 2875, 106 L.Ed.2d 166 (1989); State v. Melvin, 65 N.J. 1, 13-14, 319 A.2d 450 (1974). Defendant was plainly told that he had the right to remain silent and not to incriminate himself.
*243C. Voir Dire
Defendant contends that the jury voir dire in this case was so inadequate that he was denied his right to a fair trial by an impartial jury. In addition, defendant contends that the court’s preliminary voir dire instructions informed the jurors of what they needed to say in order to avoid sitting in this ease.
Before prospective jurors were questioned, they were required to complete questionnaires that provided the court with information about their personal lives or their familiarity with defendant, counsel, or witnesses; their ability to consider impartially the evidence presented; and their general ability to serve as jurors.
Defendant contends that by instructing the jurors how they would become disqualified, the court created a danger that some of the potential jurors would feign an excusable position on the death penalty in order to avoid serving on this capital jury. He objects in particular to a portion of the preliminary voir dire in which the court summarized the Adams-Witt test:
You are disqualified only if your view is so broad and firmly held that you will not be able to follow my instructions as to the law at the close of the trial. In short, your views concerning the death penalty may disqualify you only if they would prevent or substantially impair your ability to decide this case fairly and impartially based on the evidence presented to you in this courtroom, and in accordance with the law which the court will instruct you.
See Wainwright v. Witt, 469. U.S. 412, 105 S.Ct. 844, 83 L.Ed.2d 841 (1985), and Adams v. Texas, 448 U.S. 38, 100 S.Ct. 2521, 65 L.Ed.2d 581 (1980).
Defendant argues that in State v. Williams II, 113 N.J. 393, 550 A.2d 1172 (1988), the trial court gave the potential jurors virtually the identical preliminary voir dire instruction as provided in this case and the Court emphatically disapproved. We stated there that an instruction that effectively tells the juror what answers could lead to automatic excusal and what responses would avoid excusal “unwisely put the potential juror in the position of determining whether he or she met the legal requirements to serve on a jury.” Id. at 412, 550 A.2d 1172.
*244Defendant contends that of the sixty-three prospective jurors who were questioned about their death-penalty attitudes, nine were excused “due to their scruples against execution,” claiming that they “said what the court told them it had to hear for them to be excused due to their death-penalty views.”
In addition, defendant asserts that the overall quality of the voir dire was insufficient to enable counsel to exercise their peremptory challenges, suggesting that the responses to the generalized questions of the court resulted in virtually no information that would distinguish one prospective juror from another. Defendant asserts that because the court did not ask open-ended follow-up questions, it was impossible for the court below, and it remains impossible for this Court, to recognize the differences in the opinions of the separate jurors. Specifically, not one of them was asked whether the sexual assault and murder of a teenage girl would be a circumstance that would influence the juror concerning the imposition of the death penalty.
In short, defendant contends that there was no “thorough and searching inquiry” by the trial court into each individual’s attitude concerning the death penalty and that the jurors were inadequately screened concerning publicity about the case. See Williams II, supra, 113 N.J. at 413, 550 A.2d 1172 (death qualification process in capital-murder prosecution requires “thorough and searching” inquiry into “jurors’ opinions and biases”).
Our independent review of the voir dire reveals that as a whole the voir dire was sufficiently probing in its attempt to weed out any prospective jurors who indicated through their answers that the facts of this ease might impair their ability to decide defendant’s guilt or innocence or to determine the proper sentence.
Again we note that this voir dire, like others that we have seen, acquired a rhythm of its own, as the jurors’ attitudes became more apparent to court and counsel. There is one *245quality to this voir dire that differs from others: it displayed much more limited participation by counsel in the voir dire process. As is often the case, the trial court conducted almost all of the questioning. In State v. Moore, 122 N.J. 420, 585 A.2d 864 (1991), we reviewed various voir dire methods and explained the processes that we believe best produce death-qualified jurors, particularly encouraging judges to be open to the suggestions of counsel. Id. at 455-56, 585 A.2d 864 (citing State v. Long, supra, 119 N.J. 439, 575 A.2d 435).
We believe, however, that the process here enabled counsel to select a fair and impartial jury. Jury selection took place over a six-day period. The court employed the “struck-jury system” in which it qualified approximately fifty prospective jurors, thus allowing each side to exercise peremptory challenges from a qualified pool. The first step in the process was a general orientation given to each day’s panel. In that orientation the court described the nature of the case, the age of the victim, the charges against defendant, and the two-step process in capital cases.
At the time of this trial, the United States Supreme Court had recently decided Wainwright v. Witt, supra, 469 U.S. 412, 105 S.Ct. 844, 83 L.Ed.2d 841. In that case, the Court directed that the proper standard for determining when a prospective juror may be excluded for cause because of his or her views on capital punishment is whether those views would “prevent or substantially impair the performance of his duties as a juror in accordance with his instructions and his oath.” Id. at 424, 105 S.Ct. at 852, 83 L.Ed.2d at 851-52. The trial court here gave the jurors a brief summary of that ruling, now objected to by defendant. However, the court had made it clear to the jurors that “I’ve described our effort in a very general fashion, because it will assist you in understanding our inquiry with respect to your views concerning the imposition of the death penalty.”
*246Following that general orientation, the jurors were excused to complete a questionnaire before returning to the courtroom for voir dire. That questionnaire is set forth in the appendix to defendant’s brief and asks a number of specific questions of the jurors. When the jurors had been returned individually to the courtroom, the court’s pattern was to review each of the questions in summary fashion with the juror and to question the jurors with respect to any problematic answers on the questionnaire. In its general orientation, the court had made it clear to the jurors that there are “no right or wrong answers” to the questions. It was, as he said, an attempt to obtain qualified jurors.
In the course of interrogating each juror, the court encouraged candor and frankness, expressing these types of sentiments to the jurors: “Tell me straight out * * *. Tell me like it is”; “I’m looking for honest, frank and candid answers”; that he would like it “straight from the shoulder”; and in the case of a disqualified juror who had expressed some reservations about the presumption of innocence, “I compliment you for being candid with me.”
The trial court did not employ a rigid slot-defining format such as that described in State v. Williams II, supra, 113 N.J. at 414, 550 A.2d 1172 (automatic life, automatic death, all others), but very often asked a completely open-ended question of jurors focusing on whether they had any attitudes or opinions at all concerning the imposition of the death penalty. It did not attempt to force the jurors into any mode.
The court’s rulings with respect to the qualifications of jurors were even-handed. The court denied a prosecution attempt to disqualify for cause a juror who, having once been a defendant himself, had voted in a prior jury service to acquit; excused for cause a juror who expressed a view that he might be “swayed by ages”; let sit a juror who said he would have voted against capital punishment in a referendum; removed a juror who expressed “doubt” about the presumption of innocence; worked *247hard to calm down a somewhat distraught juror who felt overcome by the complexity of the proceedings and qualified that rather straight-talking person to sit on the jury.
Granted that specific questions related to the facts of this case might have elicited more responsive answers from the jurors, it does not appear that any jurors hesitated to inform the court of any reservations they had on account of the age or characteristics of the victim. One said that his four-and-one-half-year-old child would, in fact, affect his judgment, and that person was excused from jury service. In the process of the voir dire, court and counsel agreed that the simplest approach to reinterrogation of a juror was to ask the juror briefly to leave the courtroom and then have the court confer with counsel about further questions. On one occasion when the prosecution asked for further inquiry from a prospective juror, who was a lawyer, the court stated, “I’ll chat with her.” In short, there is every indication that the court would have been receptive to questions from counsel. This counsel, like counsel in State v. Marshall, 123 N.J. 1, 586 A.2d 85 (1991), neither sought nor believed in an exhaustive voir dire process, seeming to wish that the court not overemphasize the qualification process. Prosecution and defense counsel rarely asked for follow-up questions. When they did, the court invariably acceded. The struck-jury panel ran the gamut of professions and occupations, including a farmer, an unemployed person, a school-bus driver, and three lawyers, one of whom had been a clerk for a United States Supreme Court Justice. As we note in the jury-selection discussion, the jury that was finally selected consisted of eight whites and four blacks, five men and seven women.
Hence, we conclude that the questioning was sufficiently calculated to produce a fair and unbiased jury. Several of the jurors questioned did state that they had heard about this case or read about it. Any of those who had preconceived notions of guilt were excused. Most said they could remember only bare details and would wait to hear the evidence in the courtroom. *248It seems to us that most jurors approached the case in a quite fair-minded way. We are satisfied that the information furnished to counsel and the impressions drawn with respect to each of the jurors as they presented themselves enabled counsel to select a fair and impartial jury.
Ill
Trial Issues
A. Use of the movie, “10 to Midnight”
At trial, the State sought to introduce into evidence portions of the film “10 to Midnight.” The relevancy proffered by the State was related to the statement attributed to defendant by the Hempstead police officers that the killing was “like in the movie ‘10 to Midnight.’ ” That film has been described in a review as a “police and murder thriller” and a “propaganda piece that argues against laws that let brutal slayers escape with insanity pleas.” N.Y. Times, Mar. 13, 1983, § 1, at 62, col. 5.
The record includes the two clips from the film that were admitted into evidence. The trial court viewed the film with counsel and concluded that those two clips, which involved stabbings of women, might aid the jury in three respects: (1) to establish what defendant was talking about or what he meant by the reference; (2) by demonstrating that there was indeed such a movie, which would give credibility to the officers’ accounts; and, finally, (3) to establish the motive that defendant had stabbed the victim because she was screaming.
The court relied, in its preliminary ruling, on Evidence Rule 7(f), the keystone of the Rules of Evidence, which provides that unless evidence is specifically excluded elsewhere, it is admissible if relevant. Relevant evidence is defined as evidence “having any .tendency in reason to prove any material fact.” Evid.R. 1(2).
*249Defense counsel objected to the admission of the evidence on the basis that it was not relevant, and that even if it were relevant, it should be excluded under Evidence Rule 4, which allows courts to exclude evidence if it finds that its probative value is substantially outweighed by the risk that its admission will create substantial danger of undue prejudice or of confusing the issues or misleading the jury.
The court ruled that the evidence was admissible. When the clips were shown, the court told the jury that the evidence “may aid or assist you in corroborating or believing whether the defendant ever made reference to [the movie ‘10 to Midnight’],” and may “give you some flavor as to what he had in mind.”
Obviously in some cases there might be a marginal relevance between an extrinsic reference in a statement and a material fact at issue in a trial. Probably the most notable example is the case of John Hinckley, in which Hinckley’s defense psychiatrist proffered that he had been influenced by the movie “Taxi Driver” to attempt to assassinate President Reagan in order to ingratiate himself with a woman. See L. Caplan, The Insanity Defense in the Trial of John W. Hinckley, Jr., 77 (1984). Another such incident occurred in the Charles Manson trial in which his bizarre interpretation of the Beatle song “Helter Skelter” may have contributed to the overall motivation for his conduct. People v. Manson, 61 Cal.App.3d 102, 132 Cal.Rptr. 265, 278 (1976), cert. denied, 430 U.S. 986, 97 S.Ct. 1686, 52 L.Ed.2d 382 (1977). The problematic use of such an evidentiary device, at least by the State, is compounded in a capital case by the bifurcated nature of the proceedings. Any relevance that type of evidence might have in the guilt phase of a capital trial would be so overwhelmed by prejudice in the sentencing phase by extraneous factors irrelevant to capital sentencing that the admission of such evidence would almost inevitably require reversal of a death sentence. See State v. Williams II, supra, 113 N.J. at 451, 550 A.2d 1172 (jury may not base death sentence on considerations that are constitutionally impermissi*250ble or totally irrelevant to the sentencing process); State v. Rose, 112 N.J. 454, 548 A.2d 1058 (1988) (admission of repetitive and highly-inflammatory evidence of defendant’s past conduct without careful and precise limiting instruction compelled reversal of death sentence).
For even if the jury is instructed, as here, to use the portions of the film shown only for purposes of establishing (1) that defendant used those words, and (2) what he had in mind when he was making the reference, there is no realistic way for the jury to separate the voyeuristic aspects of the film from the evidentiary purpose. The character referred to in the movie is a psychopathic serial killer who repeatedly engages in sadistic killings of young women. He and his victims are depicted in the nude. The final message of the film is that the only way to protect society from such people is for the Charles Bronson-type character to shoot them to death lest the law afford them some defense.
Given the structured nature of capital sentencing, we have repeatedly emphasized that juries are not to be subjected to extraneous factors that may influence a jury’s verdict in a way neither contemplated nor authorized by statute. The subliminal message of the film here surely tainted the penalty phase. With respect to the guilt phase of the trial, however, given that the two most objectionable inferences cited by defendant, “what is he like,” “how can we protect society from this type,” have so little relevance to the guilt-phase issues, we find that the potential for prejudice from the film clips is so slight as not to warrant a reversal. The background evidence (the reference to the movie) was of such marginal significance that it was not even in the written notes of the Hempstead police officer who took down the statement. Its anecdotal value is peripheral to the State’s case.
We might be inclined to reverse the guilt convictions if this were a case of less-than-overwhelming evidence of guilt. See Carter v. Rafferty, 621 F.Supp. 533 (D.N.J.1985) (introduction *251of evidence of racial animus into case was so potentially prejudicial that it affected outcome of ease), aff'd, 826 F.2d 1299 (3d Cir.1987), cert. denied, 484 U.S. 1011, 108 S.Ct. 711, 98 L.Ed. 2d 661 (1988). It is the very marginality of the evidence that establishes the lack of prejudice. We find it inconceivable that this evidence would have swayed a jury from the essential facts of the case and the overwhelming evidence of guilt — multiple disinterested eyewitnesses, physical evidence in the form of shoeprints, fibers, and blood, and, finally, the confession of defendant.
B. The Gerald Issue. Does the jury’s verdict establish death eligibility?
Although we have often visited this issue, there has been a persistent misapprehension that in the process we are second-guessing juries about the jury’s findings of intent. The public obviously does not understand what we are doing. Even lawyers who comment on our opinions seem to confuse the question of the court’s role with that of the jury. Hence, we repeat again the basics.
We regret carrying pre-Code terminology forward into the discussion, but there may be no other way to get the interested readers to understand the issue. Under pre-Code law, no person could be sentenced to death without a jury finding of first-degree murder. At common law, prior to the adoption of the Code of Criminal Justice, there were two forms of murder: first-degree murder and second-degree murder. Only the former subjected a defendant to the death sentence. First-degree murder was characterized by the concept of “deliberate and premeditated” murder. Second-degree murder constituted essentially all other murders and was characterized by the presence of “malice aforethought,” which could be inferred by a jury from the actor’s infliction of serious bodily injury on a victim. When the new Code of Criminal Justice was enacted, there was no need to distinguish between first- and second-degree murder. The sentence for each was the same. L. 1979, *252c. 178 (thirty years’ imprisonment). The Code included both forms of murder in its general definition of murder. N.J.S.A. 2C:ll-3 (criminal homicide constitutes murder when the actor “purposely” or “knowingly” causes death or serious bodily injury resulting in death).
As we explained, however, in State v. Gerald, supra, 113 N.J. 40, 549 A.2d 792, when the death penalty was superimposed on the Code of Criminal Justice in 1982, no specific reference was made to which of the two forms of murder would be death eligible. However, the legislative history of the Act and constitutional concerns convinced the Court that it was only the intentional killing that was to be subject to death eligibility. As the sponsor of the bill stated, using the familiar forms known to common-law lawyers in New Jersey, a defendant faces death-penalty proceedings only after having been “found guilty unanimously and beyond a reasonable doubt of first degree murder, willful, premeditated murder.” Capital Punishment Act: Hearings on S. 112 Before the Senate Judiciary Committee at 1 (1982), quoted in State v. Gerald, supra, 113 N.J. at 90, 549 A.2d 792. Any other interpretation would present a possible constitutional problem.
Hence, we ruled that part of a capital trial must be the determination by a jury whether defendant had committed what the Code now refers to as “knowing or purposeful murder” (capital/first-degree murder). If required by the evidence, a jury must consider in the alternative whether defendant purposefully or knowingly caused death, or purposefully or knowingly caused serious bodily injury that resulted in death (SBI/second-degree murder), only the former rendering the defendant death eligible. Id. at 69, 549 A.2d 792.
The misunderstanding that has arisen over our application of the Gerald doctrine stems from confusing the question of apparent guilt with the question of how guilt is to be decided. The question that has been asked about our review of capital cases is how it can be that one who had repeatedly stabbed a *253victim, State v. Jackson, 118 N.J. 484, 572 A.2d 607 (1990), or one who had bludgeoned a victim, State v. Harvey, 121 N.J. 407, 581 A.2d 483 (1990), cert. denied, — U.S.-, 111 S.Ct. 1336, 113 L.Ed.2d 268 (1991), or even one who had shot a victim, State v. Pennington, 119 N.J. 547, 575 A.2d 816 (1990) and State v. Clausell, 121 N.J. 298, 580 A.2d 221 (1990), can be thought not to have intended the death of the victim.. The question answers itself. But the problem is not with the question; the problem is with who gets to answer the question. Under our system of justice only a jury that knows the difference between the two forms of murder and the question that it must decide may answer that question.
As appellate judges, our job is to ask: did the jury, with proper instructions, answer the question that establishes death eligibility? It seems plain to us in this case that the jury did not. We intend no criticism of court or counsel when we note that in this case the jury was specifically instructed that it could find defendant guilty of murder if it found that he intended to inflict on the victim serious bodily injuries that resulted in death. The court instructed the jury that “criminal homicide constitutes murder when the actor knowingly causes death or serious bodily injury resulting in death.” After defining “knowingly,” the court explained:
In other words, if, for example, a person shot the victim only with a purpose to cause serious bodily harm but death does, in fact, occur, you may find that the accused knew that what he did would, as a matter of practical certainty, result in death. Then he is guilty of murder as if he shot the victim with a conscious object of killing.
This charge plainly embraced SBI or what used to be called second-degree murder.
Although legal charges are at the core of a fair trial, one could consider attempting to rehabilitate the verdict were it not that the SBI/second-degree charge played a prominent part in the State’s prosecution. For although a strong case can be made that points unerringly to an intent to kill based on the deliberate submersion of the victim’s body, in his closing remarks to the jury, the prosecutor, quite understandably, direct*254ed the jury’s attention to the earlier stabbing as the basis for murder. Recall that in his confession defendant said that he did not know if Tanya were alive or dead. Thus, the prosecutor said to the jury:
Obviously, ladies and gentlemen, people don’t go around saying I’m going to commit a knowing murder. They don’t do^that. Instead what do they do? They commit the murder. And you can infer from the acts of the individual himself circumstantially what it was that that individual intended to do. So that you can deduce or you can figure out from the way someone kills somebody that they intended to or that they knew that when they did it it would cause serious bodily injury resulting in death or death itself.
The State thus presented the jury with at least two plausible bases for the conviction, the intentional infliction of serious bodily injury resulting in death (SBI/second-degree) or the intentional killing by drowning (capital/first-degree). The court’s charge to the jury did not inform it that only the latter form of murder was death eligible. When a jury verdict can rest on one of two available bases, no court can presume which basis it is. Bollenbach v. United States, 326 U.S. 607, 613, 66 S.Ct. 402, 405, 90 L.Ed. 350, 355 (1946). The error was not harmless because there was evidence in this case that could have sustained an SBI/second-degree murder verdict. We do not suggest that such a verdict was likely, but merely that if the jury returned that verdict, the court could not reject it.
Insistence on that fundamental requirement of jury findings in the face of overwhelming evidence of guilt may appear to some to be too great a price to pay by a society beset by rising crime. But there is an enduring value to the principle that only juries can decide our guilt or innocence. If people had to choose the liberties they would give up, we think that the right to have a jury determine their guilt would be among the last surrendered. It is that right that we are enforcing. We are not second-guessing juries. Juries, properly instructed, not judges, decide whether the murder is capital or non-capital. This jury was not instructed that it had to make that distinct determination. There can be no substitute for that jury verdict.
*255C. Was it error not to have charged a felony murder based on attempted sexual assault?
Defendant was charged with capital murder, aggravated criminal sexual contact, robbery, and felony murder with robbery as the predicate felony. According to defendant, the gist of the State’s case was that defendant had tried to rob and sexually attack his victim and during the commission of those crimes he killed her. The jury acquitted defendant of robbery and felony murder in the course of robbery. During deliberations, the jury asked whether they could find defendant guilty of felony murder based on a felony other than robbery. The court responded that robbery is “the only felony that you address your attentions to.” Defendant contends that the evidence warranted instruction on felony murder with attempted sexual assault as the predicate felony. A conviction of felony murder, as opposed to capital murder, would not render defendant eligible for the death sentence.
The difference between aggravated criminal sexual contact and attempted sexual assault is that the former is proved by showing an unauthorized “touching,” N.J.S.A. 2C:14-1d, whereas the latter requires proof of an attempt of “an act of sexual penetration.” N.J.S.A. 2C:14-2c. In the context of a felonious sexual crime, felony murder can be based only on a felony involving penetration or attempted penetration. N.J.S.A. 2C:11-3a(3). Defendant suggests that the record strongly indicates that he was attempting to commit the act of sexual penetration on the victim and that he stabbed the victim in an attempt to force her to submit to such an attack, but that the stabbing proved more serious than intended, causing death. In sum, defendant says that this was an attempted sexual assault requiring submission of a felony-murder charge.
Of course we agree that “the court ordinarily has a supervening responsibility to charge the jury concerning any version of the offense ‘clearly indicated’ by the evidence to require proper jury consideration.” State v. Grunow, 102 N.J. 133, 148, 506 A.2d 708 (1986) (quoting State v. Choice, 98 N.J. 295, 299, 486 *256A.2d 833 (1985)). But the trouble with the analysis in a case like this, in which defendant did not request such a charge, is that it puts the court in an impossible position. Defendant was not charged with sexual penetration or attempted sexual penetration. What if the jury had convicted defendant of intentional murder and then had premised a sentence of death on an attempted sexual penetration, viewed as an aggravating factor? We could well envision that if the court on its own motion had submitted that felony to the jury and the jury had not only found attempted sexual penetration but predicated a death sentence on it, what a troublesome issue it would be to review on appeal. See State v. Zola, 112 N.J. 384, 548 A.2d 1022 (1988) (Handler, J., concurring and dissenting) (proof of sexual penetration was not sufficient to have been submitted to the jury as aggravating factor in capital-murder case), cert. denied, 489 U.S. 1022, 109 S.Ct. 1146, 103 L.Ed.2d 205 (1989).
In State v. Choice, supra, 98 N.J. at 299, 486 A.2d 833, we emphasized that a trial court has no obligation to charge the jury on a particular offense when not requested by the parties unless the facts “clearly indicate” the appropriateness of the charge. In a capital case a court cannot lightly submit to a jury an uncharged predicate felony that may result in the imposition of a sentence of death. In any event, this matter arises as plain error. To be reversible error it must have been capable of bringing about an unjust result. Defendant asserts only that the basis for reversal is that it “would have given the jury the option of convicting defendant of a non-capital murder.” Because the sentence for felony murder is the same as the sentence for knowing or purposeful murder where the death penalty is not imposed, there is no prejudice to defendant in the circumstances of this case.
D. Was it error to have submitted the crime of aggravated criminal sexual contact based on the presence of a deadly weapon?
The indictment charged only the commission of aggravated criminal sexual contact based on the commission of a *257robbery. The interrelated provisions of the sexual-offense provisions of the Code of Criminal Justice, N.J.S.A. 2C:14-1 to -8, set forth a variety of factors that determine the degree of criminal culpability in different settings. We give but a shorthand summary of the provisions. Among the factors to be considered are the nature of the act, e.g., whether it involved penetration or other contact; the relationship between the parties; and age differentials. Recall that in this case the victim was found partially unclothed and that defendant was charged not with penetration but with sexual contact of the victim.
Criminal sexual contact is elevated for grading and sentencing if it is done during the commission of another crime or involves the use of a weapon. In this case the indictment charged defendant with aggravated sexual contact based on the commission of a robbery.
At the close of the State’s case, defendant moved for an acquittal on all counts. The court questioned the factual basis for the aggravated criminal sexual contact in that the language of the indictment failed to specify whether it occurred during the course of the robbery or while defendant was armed. The court consulted with counsel, reviewed pretrial proceedings, and noted that the prosecutor’s pretrial representation to defendant was that the sexual contact had been committed during “either the course of the commission of the robbery or the possession of the weapon.” The State argues that, having found no prejudice to defendant from lack of notice, the court properly concluded that there was sufficient evidence to support an aggravated criminal sexual contact charge based on either the robbery or the fact that defendant was armed.
We agree with the State’s position on that issue. As a general rule a criminal defendant may not be convicted for an offense not charged in the indictment, but a trial court may instruct a jury on lesser-included offenses of the crime charged in the indictment on the prosecutor’s request when there is a *258rational basis for the charge and when the defendant consents. State v. Sloane, 217 N.J.Super. 417, 423, 526 A.2d 226 (App. Div.1987), rev’d on other grounds, 111 N.J. 293, 544 A.2d 826 (1988). Because the indictment fairly apprised defendant of the charge (after all, the offense is the sexual contact, the other factors grade it) and the pretrial discovery revealed that the State’s theory was based on either the possession of the weapon or the robbery, giving sufficient notice to defendant to defend against the charge, there is no constitutional infirmity. See State v. Talley, 94 N.J. 385, 466 A.2d 78 (1983).
E. Did comments made by the prosecutor deprive defendant of his right to a fair trial?
Defendant contends that his right to a fair trial was violated by comments made by the prosecutor during the proceedings. Specifically, defendant asserts that it was highly improper for the prosecutor to speculate that young witnesses were testifying as though they themselves feared they might have been defendant's victim; to describe defendant as a “big man” among “lesser thugs,” referring to a statement made by defendant at the Hempstead police station; to use dehumanizing and degrading epithets about defendant as being a cold person who had “a lack of feeling, a lack of any remorse, a complete lack of concern about his own predicament”; to suggest that defendant had an obligation to prove that his picture had been in the paper when defense counsel cross-examined witnesses about whether their eyewitness accounts had been influenced by the picture; to speculate that there was blood from defendant under the victim’s fingernails; to suggest to the jury that it conduct its own scientific experimentation without adequate evidentiary guidelines; and, finally, that the prosecutor undermined defendant’s right to be present at his own trial by commenting on the fact that he had sat in the courtroom and “he knows what they’ve all said to you and he knows how his story’s got to add up.”
*259We are satisfied that those asserted improprieties were not of such a nature as to deprive defendant of a fair trial. Although generally limited to commenting on the evidence and to drawing any reasonable inferences supported by the proofs, a prosecutor may nonetheless make “a vigorous and forceful presentation of the State’s case.” State v. Bucanis, 26 N.J. 45, 56, 138 A.2d 739, cert. denied, 357 U.S. 910, 78 S.Ct. 1157, 2 L.Ed. 2d 1160 (1958). In each of the instances cited the prosecutor’s comments were sufficiently related to the scope of the evidence before the jury. Although the blood under the victim’s fingernails could not be identified as defendant’s blood, there were scratches and other marks on defendant’s face that could warrant an inference that the blood had come from defendant. The reference to defendant as a “big man” among “thugs” also had some tangential reference to the evidence offered that defendant explained to the other inmates in the Hempstead lockup that he was going to the head of the line because he had committed a murder. The comments with respect to defendant’s demeanor on testimony, taken in their entirety, did not suggest that defendant had any burden to disprove the State’s case. The State suggests that the comments about defendant’s presence in the courtroom at least nominally addressed the credibility of defendant’s comments when he could not explain inconsistencies in the differing accounts. Finally, the suggestion that the jury itself examine the fibers did not bolster any expert’s testimony in any way but was more or less a suggestion of a common-sense observation by the jurors that could not have misled the jury.
F. Ineffective assistance of counsel
In a recent series of capital cases, we have set forth the standards for determining whether a capital defendant has received the effective assistance of counsel at both phases of a trial. State v. Oglesby, 122 N.J. 522, 585 A.2d 916 (1991) (Handler, J., concurring); State v. Savage, 120 N.J. 594, 577 A.2d 455 (1990); State v. Davis, 116 N.J. 341, 561 A.2d 1082 (1989). Suffice it to state that capital cases present issues of *260extraordinary complexity, and to be considered reasonably competent capital counsel obviously requires a very broad grasp of principles of both capital jurisprudence and constitutional criminal procedure.
We will state the issue as defendant presents it. This case was characterized by a change of counsel in the course of trial. At the inception of the proceedings, defendant was represented by attorneys from the Camden County Public Defender’s office. In addition to filing motions challenging the jury arrays in Camden County and the constitutionality of the death penalty, original counsel obtained experts to assist them in challenging the jury-selection system. Defense counsel also sought to question the sufficiency of the evidence supporting aggravating factor e(4)(c) (battery/torture/depravity), and to require the prosecutor to state the specific facts on which he relied in asserting aggravating factors.
In appellate counsel’s view, the most crucial challenges for the defense to meet and successfully defend against were defendant’s alleged confessions and the eyewitness testimony of several children. Defense counsel prepared pretrial challenges to the admissibility of the confessions and the reliability of the identifications. However, shortly before trial was to begin, defendant’s family obtained the services of a private attorney, who was not admitted to practice in New Jersey, but was substituted as counsel for defendant four weeks before the trial commenced. This was his first capital case in New Jersey.
Defendant has asserted a variety of challenges to the effectiveness of his retained counsel. Among them are his ineffectiveness during the jury voir dire, a point that we have discussed in part 11(C) hereof, an issue primarily related, however, to death qualification, an issue mooted by our disposition of this ease. As we noted in that discussion, defense counsel took a very low profile in the conduct of the jury voir dire. We need not debate whether that was tactical. We are satisfied it did not interfere with the selection of an impartial jury.
*261Several of the allegations of ineffective assistance of counsel during the guilt phase of the trial cannot, however, be resolved on this record. Most of the objections we consider to be tactical in nature, for example, that defense counsel failed to make opening statements on behalf of his client in either the guilt or penalty phases.1 At least one aspect, however, will require further consideration in post-conviction relief proceedings, namely, whether defendant’s failure to call two witnesses at the Miranda hearing was the product of a strategic decision. We noted in State v. Davis, supra, 116 N.J. at 357, 561 A.2d 1082, that matters of tactics and strategy are virtually unassailable when they are based on a proper understanding of the law and evaluation of all the facts in a case. In this case, one of the most damaging pieces of evidence against defendant was his confession. Before trial counsel was substituted to represent defendant, the Public Defender’s office was preparing for a lengthy Miranda hearing, anticipating producing witnesses to testify in defendant’s behalf, specifically defendant’s uncle, Robert Newbill, and his grandmother, Thelma Dixon. Those witnesses were going to be called to testify to the circumstances surrounding defendant’s arrest in New York. Newbill would testify that defendant told the police in his presence that he (defendant) would not speak without a lawyer present. That information was given to defendant’s new defense team, whose members allegedly never bothered to interview the witnesses. At the Miranda hearing, the witnesses were not called.
Had defendant invoked the right to counsel, then all further police questioning would have had to cease in the absence of his initiating further interrogation. That issue was raised in a motion for a new trial, but the record is inadequate to disclose what reasons of tactics and strategy motivated counsel not to *262call the witnesses. He might have thought they were unreliable. He concentrated his Miranda challenge on the unlikelihood that the police would have failed to obtain written Miranda waivers if the warnings were given. In State v. Savage, supra, 120 N.J. 594, 577 A.2d 455, a similar trial-counsel failure — failure even to consider a psychiatric defense — necessitated a post-trial hearing at which trial counsel explained his rationale for the decision. The Court found that rationale totally unacceptable. It failed to demonstrate minimal competence; it was not a strategic decision, but a total lack of a decision. 'Defendant suggests that permissible trial strategy can never include the failure to conduct a substantial investigation into any of defendant’s plausible lines of defense.
We find this record inadequate to evaluate the ineffective-assistance-of-counsel issue with respect to the confession. A post-conviction relief proceeding will have to address that issue. See State v. Savage, supra, 120 N.J. at 609-12, 577 A.2d 455 (at post-trial hearing, trial court considered factors that went into decision not to assert mental disease or defect in defense); United States v. Schaflander, 743 A.2d 714 (9th Cir.1984) (court reviewed allegations of ineffective assistance of counsel in post-conviction relief proceeding), cert. denied, 470 U.S. 1058, 105 S.Ct. 1772, 84 L.Ed.2d 832 (1985). Those proceedings must determine whether the non-production of witnesses was the result of tactics or of neglect and, if the latter, whether it prejudiced defendant. At that proceeding, counsel should also address defense counsel’s failure to cross-examine the State’s forensic expert on the fiber samples. Even if it were neglect, however, the possibility of any prejudice on that account seems most unlikely. Finally, there has been no showing that a Wade hearing could in any way have aided defendant.
G. Were the non-capital sentences excessive?
Dixon was convicted in counts five, six, and nine of hindering his own apprehension for murder by hiding the victim’s body in a nearby creek, by washing the clothing he *263wore at the time of the murder, and by leaving the state. The court imposed three consecutive maximum sentences of five years, two and one-half without parole, on each hindering conviction, for a total of fifteen years, seven and one-half without parole.
Defendant contends that imposition of the maximum term with the maximum period of parole ineligibility on each count is excessive, as is imposition of three consecutive terms. He asserts that the sentences violate the specific standards enunciated by this Court in State v. Yarbough, 100 N.J. 627, 498 A.2d 1239 (1985), cert. denied, 475 U.S. 1014, 106 S.Ct. 1193, 89 L.Ed.2d 308 (1986), to guide courts in the fashioning of consecutive sentences.
Because the matter must be remanded for resentencing on the murder count, we defer resolution of those sentencing issues until reconsideration of the entire sentence, including the murder count. See State v. Moore, 113 N.J. 239, 310, 550 A.2d 117 (1988) (leaving sentences to discretion of trial, court following disposition of murder count). We note, however, that this case does not fit within the exception to State v. Yarbough, supra, 100 N.J. at 647, 498 A.2d 1239: “[E]ven within the general parameters that we have announced there are eases so extreme and so extraordinary that deviation from the guidelines may be called for.” In State v. Moore, supra, the defendant committed an incredibly bestial cult killing in which the murdered victim had been beaten, tortured, subjected to sexual abuse, cuffed to a hook on the kitchen wall, chained to a bathtub at night, and was neither clothed nor fed. A case such as that occasions departure from the Yarbough guidelines. See also State v. Louis, 117 N.J. 250, 566 A.2d 511 (1989) (even in rare cases justifying departure from “two longest” rule, other Yarbough guidelines remain relevant). The State suggests that the three counts need not merge and we agree with that conclusion. Each offense required proof of a different fact. However, the fact that the three offenses do not merge does not mean that the principles of Yarbough do not apply.
*264H. Other sentencing-phase issues
In view of the disposition that we make, we are not required to resolve other sentencing-phase issues raised by defendant. In particular, defendant argues that submission to the jury of the c(4)(f) aggravating factor (killing to escape detection) was precluded because in pretrial proceedings the State had specifically relied upon the robbery as the underlying predicate offense for charging this factor. Although the prosecutor emphasized in his closing arguments that the basis for the jury’s finding of this factor could be the sexual-contact conviction, in its charge to the jury the trial court did not specify a basis for submitting the c(4)(f) factor. Either basis raises the serious concerns noted by Justice Handler’s dissent. Trial fairness is “a categorical imperative” for death-penalty prosecutions. State v. Williams, 93 N.J. 39, 61, 459 A.2d 641 (1983). No one can face capital sentencing without adequate notice of the charges against him or her. See Lankford v. Idaho, — U.S. -, 111 S.Ct. 1723, 114 L.Ed.2d 173 (1991) (vacating court-imposed death sentence when State has not given notice that it would seek death penalty).
If the jury based its finding of the killing-to-avoid-detection factor on the basis of the sexual contact, there is a serious due-process issue. On the other hand, if the jury based its finding of this factor on the previously specified offense of robbery, defendant would seek to assert a double-jeopardy violation because the jury had acquitted defendant of the underlying robbery. The State argues that the obvious basis was the sexual-contact offense and that because defendant knew from the prosecutor’s remarks that that was the basis for submission of the factor, there is no lack of notice and no due-process violation. We need not resolve the issue in this case, but we caution courts concerning the need to observe Lankford’s admonitions about fair notice in presenting capital cases to the sentencing jury.
*265IV
To sum up, the right to trial by jury includes the right to have a jury decide the essential elements of an offense. This jury was not instructed to decide which of the two forms of murder defendant had committed. That finding is necessary to determine death eligibility. The error was not harmless because there was evidence in this case that would have sustained an SBI/second-degree murder verdict. A court can no more dispense with the fundamental right of trial by jury in a capital case than in any other case.
The death sentence is vacated and the matter remanded to the Law Division for sentencing on the murder conviction in accordance with this opinion.
Appellate counsel also questions trial counsel’s failure to object on discovery grounds to the admission of the movie, "10 to Midnight." Because we consider the evidence not to have been unduly prejudicial, this failure would not, in itself, be grounds for reversing the conviction.