Opinion of The Court by Justice ABRAMSON.

Errick Duncan appeals from a Judgment of the Jefferson Circuit Court convicting him of kidnapping, in violation of KRS 509.040; of first-degree sexual abuse, in violation of KRS 510.110; and of two counts of second-degree stalking, in violation of KRS 508.150. In accord with the jury's status determination and sentence recommendation, Duncan was sentenced as a first-degree persistent felon to concurrent terms of imprisonment totaling thirty years. He appeals to this Court as a matter of right.

Duncan was accused of a series of crimes that occurred in south Louisville in the fall of 2003. The Commonwealth alleged that he kidnapped fourteen-year-old SM, twice sexually abused her, sodomized her, and raped her. The jury found him guilty of kidnapping and of one count of sexual abuse, but could not reach a verdict on the other charges. The Commonwealth also alleged that Duncan stalked two other girls from the same neighborhood, thirteen-year-old JM and fourteen-year-old JH, and that he unlawfully imprisoned and menaced JH. The jury could not reach a verdict on the unlawful imprisonment and the menacing charges, but found Duncan guilty of stalking both girls. Duncan, who testified at trial, denied the charges and argued that he had been wrongly identified as the perpetrator.

On appeal, Duncan contends that his trial was rendered unfair when he was subjected to improper cross-examination and when, during her closing argument, the prosecutor misrepresented evidence concerning DNA test results. He also contends that one of the investigating officers was improperly allowed to bolster the victims' testimony; that under the kidnapping exemption statute he was entitled to a directed verdict on the kidnapping charge; and that JH's out-of-court identification of him as her stalker was tainted by a suggestive identification procedure and so should not have been admitted into evidence. We agree with Duncan that the Commonwealth misrepresented its DNA evidence and by so doing deprived Duncan of a fair trial, but reject his other claims of error.

RELEVANT FACTS

The Commonwealth relied on the three victims for proof that the crimes occurred. Each victim testified about encounters with a man in the neighborhood of Southern Middle School in Louisville in the fall of 2003. JM testified that on Halloween evening, while she was walking to a friend's house, a van pulled up beside her and the driver asked for directions. He then asked other questions that JM felt were inappropriate, such as her name and whether she had a boyfriend. The same man drove past her later that night as she was trick-or-treating. Soon thereafter she began to see the man parked in front of her house when she left for school in the morning. Finally, one day while she was visiting with two male friends, the same man, this time driving a white Lincoln, drove past them several times. The two boys confronted the man, asked why he was following JM, and wrote down his license number.

JH testified that one day in early November she and a friend were walking home from school when they noticed a man in a white van watching them. When the van approached them and the man began asking questions, the girls became frightened and ran home. A few days later, on November 5, 2003, JH was walking down an alley on her way home, when she again saw the white van. Moments later, a man grabbed her from behind, told her he had a gun, and warned her not to scream. JH testified that she freed herself by hitting the man with her elbow and ran home.

SM testified that on November 22, 2003, she was walking home from a friend's house when a man grabbed her from behind, pressed something hard against her back, told her he had a gun, and warned her not to scream. The man led her a short way down an alley, and then led her into a secluded area behind an abandoned house. There he ordered her to lower her pants and fondled her vagina. He then had her pull her pants up and led her through several city blocks, over a pedestrian bridge across the freeway, through an opening in a fence, and into a secluded area behind Southern Middle School. As they were walking they were approached by two people, but the man warned SM that he "would hate to have to shoot them." When they reached the area behind the school, the man put a toboggan over SM's face and again had her lower her pants. SM testified that he fondled her, orally sodomized her, and twice tried to rape her, although he was not able to penetrate her fully. He then forced SM to kiss him while he masturbated. Finally, he had SM adjust her clothes, removed the toboggan, and told her to leave without looking back. SM told her mother what had happened, and her mother summoned the police.

A principal issue at trial was the identity of the perpetrator. The Commonwealth's proof against Duncan included photo identifications by all three victims. JH and JM positively picked Duncan's photo from a photo array as the man who had stalked them, and SM picked out Duncan's photo, but told the officers that she was not positive. The police traced the license number JM had recorded to Duncan's white Lincoln. Duncan's girlfriend in the fall of 2003 testified that she had then owned a white van, which Duncan sometimes borrowed. JM identified both vehicles, and JH identified the van. Duncan's former girlfriend also authenticated two letters Duncan had sent to her from prison. In one letter he admitted having had sex with an underage girl, but claimed the encounter was consensual, and in the other he asserted that the girls making the other charges were "lying their asses off." Finally, the Commonwealth presented the results of DNA analyses performed on samples taken from the panties SM wore at the time of her assault. Those analyses indicated the presence of a male's DNA,1 and the partial profiles obtained matched, to the extent that they yielded results, the profiles obtained from Duncan.

Duncan presented a denial defense. He claimed to have had no encounters with any of the victims. He established that SM and JH, when first describing the man to the police, had guessed him to be short, in the neighborhood of five feet, eight inches tall, when in fact Duncan is over six feet tall. He elicited testimony from the forensic witnesses to the effect that the DNA analyses did not identify him as the source of the DNA found in SM's panties, but rather as one of an undetermined number of potential sources. And he testified that the letter to his former girlfriend referred not to SM but to a seventeen-year-old girl he had been seeing at the time and who, he mistakenly believed, was the source of the rape charges. As noted, although the jurors did not accept Duncan's defense, neither were they all convinced that he had committed the more serious of the crimes charged.

Duncan maintains that his trial was rendered unfair in several ways. His principal complaint concerns instances of what he characterizes as prosecutorial misconduct. He contends that he was cross-examined unfairly when the prosecutor asked him to "explain" the DNA evidence and to characterize the testimony of other witnesses as "lying" or as "wrong." He also contends that during her closing argument one prosecutor misrepresented the DNA evidence by stating that it established that he was the source of the male DNA recovered from SM's panties. We begin with these alleged instances of prosecutorial misconduct.

ANALYSIS

I. Duncan's Trial Was Rendered Unfair by Prosecutorial Overreaching.

As Duncan correctly notes, prosecutorial misconduct can assume many forms, including improper questioning and improper closing argument. Brown v. Commonwealth, 313 S.W.3d 577 (Ky.2010); State v. Singh, 259 Conn. 693, 793 A.2d 226 (2002). If the misconduct is objected to, we will reverse on that ground if proof of the defendant's guilt was not such as to render the misconduct harmless, and if the trial court failed to cure the misconduct with a sufficient admonition to the jury. Where there was no objection, we will reverse only where the misconduct was flagrant and was such as to render the trial fundamentally unfair. Barnes v. Commonwealth, 91 S.W.3d 564 (Ky.2002); Partin v. Commonwealth, 918 S.W.2d 219 (Ky.1996).

A. During His Cross-Examination of Duncan, the Prosecutor Misrepresented Key Aspects of the Commonwealth's Evidence.

Toward the end of his lengthy cross-examination, one prosecutor asked Duncan, "Your story today is that JM was lying. JM has to be lying in order for your story to be true, right?" Defense counsel objected on the ground that JM could have testified honestly but mistakenly. Although the court did not sustain the objection, it did advise the prosecutor to "clear the matter up." The prosecutor then asked, "For your story to be right, JM has to be wrong. Right?" He then proceeded to ask the same question with respect to SM and JH, with respect to the detectives, and with respect to the DNA evidence, i.e., "The DNA has to be wrong for your story to be true. Isn't that correct." When Duncan protested that he did not understand the DNA evidence, the prosecutor held up the exhibit which showed the results of the DNA analysis, had Duncan acknowledge, point by point, the instances where the DNA from SM's panties was consistent with his DNA, and then asked again: "The DNA has to be wrong for your story to be right. Isn't that true?" To none of these latter questions did Duncan's counsel object. Duncan contends on appeal, nevertheless, that these questions were flagrantly improper and rendered his trial unfair. With respect to the DNA questions, we agree.

In Moss v. Commonwealth, 949 S.W.2d 579, 583 (Ky.1997), we reiterated what had long been the rule in this state and what is overwhelmingly the majority rule in other jurisdictions when we emphasized that "[a] witness should not be required to characterize the testimony of another witness ... as lying." See also Liggett v. People, 135 P.3d 725 (Colo.2006) (collecting cases). Not only do such questions invade the jury's role as sole determiner of credibility and unfairly require the witness to disparage another witness, but they are usually misleading, for, as in this case, mistake rather than either witness's dishonesty could account for disparities in their testimonies. The prosecutor's question to Duncan as to whether his story did not imply that JM was lying was a blatant Moss violation, and Duncan's objection should have been sustained. That violation alone, however, would not entitle Duncan to relief, for following the objection the prosecutor asked instead whether Duncan's version of events did not imply that the victims' versions were wrong rather than dishonest, and that question at least clarified for the jury that Duncan's defense did not hinge on a conclusion that the victims had fabricated their allegations.

Duncan maintains that he should not have been asked to characterize other witnesses as "wrong" any more than as "lying." This appears to be a question we have not addressed before. Other courts have distinguished the two types of questions: "Asking a witness whether a previous witness who gave conflicting testimony is `mistaken' highlights the objective conflict without requiring the witness to condemn the prior witness as a purveyor of deliberate falsehood, i.e., a `liar.'" United States v. Gaind, 31 F.3d 73, 77 (2nd Cir. 1994). In denying plain error relief in a case, like this one, in which during cross-examination the defendant was asked if several witnesses were mistaken about particular details of their testimony, details that conflicted with the defendant's version of events, the United States Court of Appeals for the First Circuit cited the Second Circuit's distinction between "mistake" and "lie," and held that "[w]hether this avoidance [of the "L" word] would suffice in all situations, we need not decide now. As [the defendant] did not object in the district court to these questions ... our review is limited to plain error. Clearly that standard was not transgressed." United States v. Gaines, 170 F.3d 72, 82 (1st Cir.1999). Likewise here, Duncan did not preserve this issue by proper objection at trial. We are limited, therefore, as noted above, to asking whether the prosecutor engaged in flagrant misconduct and if so whether it rendered Duncan's trial fundamentally unfair. Since Duncan's defense was precisely that the three victims had mistakenly identified him, it assuredly did not amount to flagrant misconduct for the prosecutor to ask him if they had to be mistaken for his version of events to be true.

The prosecutor's asking whether the detectives had to be mistaken is more troubling, because the prosecutor did not specify what in particular about their testimonies conflicted with Duncan's version in such a way that either they were mistaken or Duncan's version could not be right. We need not decide whether that failure made the question improper, however, for even if it did the question did not render Duncan's trial fundamentally unfair.

The prosecutor's DNA question, however—"The DNA has to be wrong for your story to be right. Isn't that true?"—was altogether more egregious. To understand why, it is necessary to consider in some detail the Commonwealth's DNA evidence. That evidence was introduced through the testimonies of two lab supervisors employed by Orchid Cellmark, a company specializing in DNA analysis. One of those witnesses, William Watson, the supervisor of Orchid Cellmark's Nashville facility, testified that among other items to be analyzed, the Commonwealth sent to it cuttings from SM's panties, a blood sample from SM, and a blood sample from Duncan. According to Watson, the lab performed what has become the standard PCR/STR analysis of these samples.

As background, we note some basic principles pertaining to DNA analysis from the COMMITTEE ON DNA FORENSIC SCIENCE: AN UPDATE, NATIONAL RESEARCH COUNCIL, THE EVALUATION OF FORENSIC DNA EVIDENCE (1996) ("NRC Update") discussed extensively in State v. Bander, 150 Wn.App. 690, 208 P.3d 1242 (2009). Nuclear DNA is a long, spiral molecule found in every human cell except red blood cells. NRC Update. Most of this long molecule is identical from person to person, but along its length are segments where variations occur. Id. These varying segments are referred to as polymorphic sites, and the possible variants, generally only a few for each polymorphic site are referred to as alleles. Id. In all but sex cells, a person has two alleles at each site, one inherited from each parent. Id. When several sites are tested and the alleles determined, the pattern of alleles is referred to as a profile. Id. Several tests targeting different polymorphic sites have been marketed, but in 1994 Congress established a set of thirteen such sites as a national standard for use in the Combined DNA Identification System (CODIS). Roberts v. United States, 916 A.2d 922 (D.C.2007). Watson did not testify that those were the thirteen sites his lab tested, but it appears from his report filed with the Commonwealth's discovery that they were.

In any event, PCR/STR testing refers to a test targeting a particular kind of polymorphic site, sites at which the alleles are short pieces of DNA repeated some number of times (Short Tandem Repeats or STR). Roberts, supra. The alleles are distinguished by the number of repeats. Id. PCR, short for Polymerase Chain Reaction, refers to a testing process in which the DNA from a sample is isolated; the polymorphic sites to be analyzed are chemically marked and copied millions of times; and then, from that amplified genetic material, the alleles present at each site are identified. Bander, supra. By virtue of its "amplifying" capacity, PCR has made possible the analysis of very small DNA samples, but even with PCR it sometimes happens that an analysis will fail to identify some or any of the alleles at a particular site. Id. When some sites yield full results and some do not, the incomplete pattern of alleles is sometimes referred to as a partial profile. Commonwealth v. Linton, 456 Mass. 534, 924 N.E.2d 722 (2010).

It also happens, frequently in the course of forensic DNA analyses, that a sample will prove to contain a mixture of DNA from more than one person. A mixture is apparent when the analysis discloses more than two alleles at any given site. Bander, supra. Two samples, a known sample, say, such as Duncan's in this case, and an unknown sample such as that recovered from SM's panties, are said to match if the profiles they yield are identical, i.e., if the alleles present at each site in both samples are the same. Id. If an unknown sample yields only a partial profile, but that partial profile is identical, as far as it goes, with a known sample, the two are sometimes said to partially match. Id. If the unknown sample is a mixture, with more than two alleles per site, the known sample is said to match or the source of that sample to be a possible contributor to the mixture if the known sample's alleles are among those present at each site in the mixture. Id. If the unknown sample is a mixture and yields only a partial profile, the known sample is said to partially match it, or again to come from a possible contributor to the mixture, if the known sample's alleles are present at each complete site of the unknown sample and are either present or potentially present at each site for which there are incomplete results. Id.

The significance of a match or a partial match depends upon how rare or how common the profile or partial profile is. Population geneticists have amassed data bases for the commonly tested polymorphic sites, such as those in the CODIS standard, which enable them to estimate the frequency with which the various alleles occur in the general United States population and in certain subsets of the population. Roberts, supra. Those estimates then allow the calculation of the frequency with which a given profile occurs. Id. That frequency is often expressed as the "random match probability" or the odds that an unrelated person chosen at random from the reference population would have the given profile. Id. For large profiles, such as those based on the twenty-six alleles of the thirteen CODIS sites, the random match probability is frequently vanishingly small, one chance in billions or trillions or quadrillions. In such cases a very strong inference arises that the matching known and unknown samples came from the same source. Young v. State, 388 Md. 99, 879 A.2d 44 (2005). For smaller profiles, however, those based on partial matches, say, at only a few sites, the odds of a random match can be much higher and the inference that the source of the known sample was also the source of the unknown sample much weaker. Commonwealth v. Mattel, 455 Mass. 840, 920 N.E.2d 845 (2010).

In this case, for example, Watson testified that the sample recovered from SM's panties yielded a mixture of DNA from at least three persons, two of whom were males. Although he did not specify the results of the analysis for each site tested, he indicated that the results were partial, with several sites yielding incomplete results. SM and Duncan were both possible contributors to the mixture, he testified, but the odds of choosing a possible contributor at random from the reference population (Watson did not state whether he was referring to the general population or to a subset) were better than one in three. Because these results did not have much inferential force, Watson recommended to the Commonwealth that it have the sample recovered from SM's panties and the sample obtained from Duncan retested at Orchid Cellmark's Dallas facility, where a different test, a so called Y-STR test, could be performed.

The director of Orchid Cellmark's Dallas facility, Cassie Johnson, testified that the Commonwealth agreed to the further testing and submitted samples from SM's panties, from Duncan, and from SM's step-father. Johnson explained that she tested the samples twice, the first time in 2004 and then again in 2005 after a new, expanded version of the test came out. Like the test Watson's lab performed, the tests Johnson oversaw employed PCR to amplify certain STR polymorphic sites, and then sought to identify the alleles present at those sites. The difference was that the sites Johnson's lab targeted are located on the Y-chromosome, a portion of the DNA molecule that occurs only in men and that is passed on intact from father to son. Bander, supra. This male-only test would screen out any female contribution to the mixture present on the panties, and thereby, it was hoped, would provide more resolution for the analysis of the male contribution. Johnson testified that because the Y-chromosome comes only from the father, only one allele occurs at each site, and the pattern of alleles is the same for father and son and hence for any man in the same paternal lineage. As a result, Johnson testified, Y-STR testing cannot identify a particular individual, and a match between samples indicates that not only the source of the known sample, but others in his lineage, were possible sources of the unknown sample.

The first test Johnson's lab performed targeted ten sites and from the sample isolated from the panties yielded results for seven of them. At all seven of those sites, Duncan's alleles occurred. In 2005, Johnson's lab analyzed the samples again, this time with a new version of the test that targeted seventeen Y-chromosome sites. For the sample recovered from SM's panties, the test yielded results at ten of the seventeen. At two of those sites two alleles were detected, indicating that the sample contained a mixture of DNA from two men. At all ten of the sites where results were obtained, Duncan's alleles occurred. At the two sites where an additional allele was detected, that allele matched the step-father's corresponding allele. At four other of the ten result-yielding sites Duncan's and the step-father's alleles were the same. Johnson concluded that Duncan and the step-father were both potential contributors to the mixture, and neither Duncan nor any male in his paternal lineage could be excluded as a source of the DNA recovered from SM's panties. Johnson was not asked for and did not offer any testimony regarding the statistical significance of her results. The jury was told that Duncan could not be ruled out, but it was not told that anyone else could be either.

Given this complete lack of evidence regarding the significance of Duncan's partial match, it was a gross misrepresentation of Johnson's testimony for the prosecutor to suggest that the "DNA has to be wrong" for Duncan's version of events to be right. The DNA evidence was consistent with a scenario in which Duncan was the perpetrator. Given Johnson's very limited testimony, however, (the failure to establish through statistics or otherwise the significance of the finding) it was also consistent with a scenario in which any other man on the planet was the perpetrator, and thus did not need to be wrong for Duncan's testimony to be right. The prosecutor compounded the impropriety by having Duncan acknowledge, one by one, all ten sites at which his profile matched the partial profile obtained from the panty sample. The significance of those matches is precisely what the expert failed to establish, and by suggesting that those matches were either "wrong" or conflicted with Duncan's testimony the prosecutor invited the jury to be its own expert—to make inferences that it was not qualified to make and which amounted to pure speculation. This was a flagrant abuse of cross-examination that, given the aura of conclusiveness that surrounds DNA evidence, rendered Duncan's trial manifestly unfair.

B. During Closing Argument, the Prosecutor Again Misrepresented the Import of the Commonwealth's DNA Evidence.

This conclusion is underscored by portions of the prosecutor's closing argument to which Duncan objected. Twice during that portion of her2 argument based on Johnson's testimony, the prosecutor went from quoting Johnson's conclusion that Duncan could not be excluded as a source of the DNA in SM's panties, to insisting that "not excluded" means "included," to "the bottom line: What is Errick Duncan's DNA doing in SM's panties?" Counsel objected both times the prosecutor reached this "bottom line" assertion, and following the second objection the court admonished the jury to the effect that while the evidence had not established "a direct, absolute match," both sides were allowed to argue reasonable inferences from the evidence. Duncan maintains that the prosecutor's argument went beyond "reasonable inference" and had the effect, like that of the improper cross-examination, of misrepresenting Johnson's testimony. For the reasons discussed above, we agree.3

The problem is not that the evidence failed to establish a match between Duncan's profile and the profile obtained from the sample. The evidence established a match, or a partial match, at ten of the seventeen tested sites. But missing from the Commonwealth's proof was any testimony establishing the significance of that partial match. Johnson's testimony that Duncan could not be excluded as a source of the panty DNA said nothing at all about how likely or unlikely it was for such a partial match to occur, and most assuredly it did not say that Duncan was the source. By asking the jury to infer on the basis of Johnson's testimony that he was, the prosecutor sought to wring from that testimony a conclusion it could not reasonably yield.

This concern was reflected in the 1992 report of the National Research Council's Committee on DNA Forensic Science. That report, titled DNA Technology in Forensic Science, emphasized that "[t]o say that two patterns match, without providing any scientifically valid estimate (or, at least, an upper bound) of the frequency with which such matches might occur by chance, is meaningless." Id. at 301. It is "meaningless," or, at most, of marginal value, for precisely the reason noted above: although the evidence of a partial match established that Duncan could not be excluded as a potential source of the panty DNA, without further evidence putting the match in some context of significance, statistical or otherwise, it does not establish that anyone can be excluded and so merely includes Duncan among the rest of the world as a potential source. For this reason, several courts have held that DNA "match" or "non-exclusion" evidence is inadmissible without reliable accompanying evidence as to the likelihood that the test could or could not exclude other individuals in a given population. Without the accompanying evidence, these courts note "the jury have no way to evaluate the meaning of the result." Commonwealth v. Mattel, 920 N.E.2d at 856 (collecting cases).

In Sholler v. Commonwealth, 969 S.W.2d 706 (Ky.1998), this Court rejected that per se exclusionary approach and held that bald DNA "match" or "non-exclusion" evidence was admissible as circumstantial evidence akin to blood type evidence:

Id. at 710. We adhere, with some reluctance, to that holding today, but emphasize the following qualifications.

Matching blood types provide weak circumstantial evidence that the defendant was the source of the crime scene blood. The rarer the blood type the slightly better the evidence, but no blood-type evidence standing alone would serve to identify the defendant beyond a reasonable doubt as the source of the crime scene blood. The potential probative value of DNA evidence is vastly more powerful. Although theoretically a DNA match over a large number of polymorphic sites, such as the thirteen sites of the CODIS standard, does not conclusively identify a defendant as the source of crime scene DNA, when the odds of a random match begin to vanish in the mists of the unimaginably small, as they often do in such cases, for all practical purposes the defendant is identified, and that evidence, standing alone, would support such a finding. In such cases, the Maryland Court of Appeals has held, the expert need not couch his or her testimony in the theoretical language of mathematics, but may testify directly that the match proves, to a reasonable scientific certainty, that the defendant is the source of the crime scene DNA. Young v. State, supra.

In conjunction with this vast probative potential, DNA evidence is also subject to vast misunderstanding and misuse. Although at times highly probative, it can also, as this case illustrates, be much more modestly probative or hardly probative at all. Where its significance falls along that spectrum is a matter dependent upon expert testimony, since lay jurors are not qualified to make the assessment on their own. Statistical evidence, if possible, is the best way to convey the strength of DNA evidence, but statistics may not always be available, and where it is not the expert may use other valid means to give some idea of the extent to which the DNA evidence narrows the field of possible sources.4 Whatever avenue is chosen, the Commonwealth must abide by the limitations of its own proof and not make claims that its DNA evidence is more probative than the expert's testimony has shown it to be. In particular, if, as in this case, the Commonwealth relies on mere evidence of a partial match and "non-exclusion" without any other evidence of the match's significance, it may not, as it did here, ask whether the DNA evidence is "wrong" because it contravenes the defendant's version of events. Similarly, the Commonwealth may not, by underscoring the fact of the partial match, invite the jury to speculate that the match is actually more significant than the expert testified, or that evidence of the partial match by itself is sufficient to identify the defendant as the source of the crime scene DNA. These are not reasonable inferences from bare "non-exclusion" evidence and so are not fair game either for cross-examination or for closing argument. Because the Commonwealth failed to present any evidence of the significance of Duncan's partial DNA match other than the expert's "non-exclusion" testimony, its suggestion during Duncan's cross-examination and during closing argument that the DNA evidence pin-pointed Duncan was highly improper and, given the immense weight jurors are apt to accord DNA evidence, rendered Duncan's trial manifestly unfair. We are obliged, accordingly, to reverse Duncan's conviction and to remand for additional proceedings. Because Duncan's other claims of error raise issues that could recur at a retrial, we shall briefly consider them as well.

II. Duncan Was Not Exempt From the Charge that He Kidnapped SM.

In pertinent part, KRS 509.040 provides that "[a] person is guilty of kidnapping when he unlawfully restrains another person and when his intent is: ... (b) To accomplish or to advance the commission of a felony." Many felonies, however, either inherently or as a matter of common occurrence, involve the restraint of the victim in some manner. Forcible rape, sodomy, and sexual abuse, for example, the offenses Duncan was accused of committing against SM, all involve restraining the victim in conjunction with and so as to accomplish the unlawful sexual contact. To distinguish kidnapping from the restraint that is part-and-parcel of another offense, KRS 509.050, the kidnapping exemption statute, provides in pertinent part that

We have construed this statute narrowly and have held that for the exemption to apply the restraint will have to have been "close in distance and brief in time ... If the victim is restrained and transported any substantial distance to or from the place at which the crime is committed or to be committed, the offender will be guilty of an unlawful imprisonment offense as well." Timmons v. Commonwealth, 555 S.W.2d 234, 241 (Ky.1977). We have required, moreover, that one seeking to invoke the exemption strictly satisfy all three of the statutory requirements: (1) that the underlying criminal purpose was the commission of a crime defined outside KRS Chapter 509; (2) that the interference with the victim's liberty occurred immediately with or incidental to the commission of the underlying intended crime; and (3) that the interference with the victim's liberty did not exceed that which is ordinarily incident to the commission of the underlying crime. Hatfield v. Commonwealth, 250 S.W.3d 590 (Ky.2008). The trial court, rather than the jury, determines whether the exemption applies, Arnold v. Commonwealth, 192 S.W.3d 420 (Ky.2006) (citing Calloway v. Commonwealth, 550 S.W.2d 501 (Ky.1977)), and we review that determination under the abuse of discretion standard. Id.

Duncan contends that his restraint of SM is indistinguishable from the alleged sexual offenses because she was restrained for no reason other than the commission of those offenses and because the restraint was no longer than necessary to carry out the sexual assault. This last contention is absolutely meritless. Even granting that Duncan's purpose was the commission of an offense defined outside KRS Chapter 509, and that he thus satisfied the first statutory requirement, his restraint of SM clearly exceeded what could be deemed merely incidental to the sex offenses.

According to SM, Duncan first forced her from the street, where she was walking, down an alley and then into a secluded area behind an abandoned house where he sexually abused her. In Mitchell v. Commonwealth, 908 S.W.2d 100 (Ky.1995), a woman was carried by two attackers from the street into a backyard, where the attackers raped her. We held that the trial court properly refused to apply the exemption statute to the restraint involved in moving her from the street to the backyard. It is arguable, therefore, that Duncan's forcible removal of SM from the street to the area behind the abandoned house was not an exempt restraint. But see Hatfield, supra (removal of murder victim from church parking lot to secluded area behind the church deemed merely incidental to the murder and so exempt). Regardless, the additional restraint Duncan imposed when he forced SM to walk for five to ten minutes through several blocks to the area behind the school was neither brief in time nor short in distance, for the purposes of the exemption statute, and exceeded what was merely incidental to the alleged sexual offenses. The trial court did not abuse its discretion by ruling that the exemption statute did not apply.

III. JH's Pre-trial Identification of Duncan Was Not Tainted by an Unduly Suggestive Procedure.

Duncan next contends that JH's pretrial identification of him was tainted by a suggestive identification procedure and that under the Due Process Clause of the United States Constitution evidence of her identification should therefore have been suppressed. The admissibility of identification evidence for dueprocess purposes is governed by a two-part test. First, the court determines if the identification procedures were "impermissibly suggestive." Simmons v. United States, 390 U.S. 377, 384, 88 S.Ct. 967, 19 L.Ed.2d 1247 (1968). If they were not, then the admission of evidence based thereon does not violate the Due Process Clause, and the inquiry is at an end. If the procedures were unduly suggestive, then the court moves to the second step of the test and determines whether, in light of the totality of the circumstances, the suggestive procedures created "a very substantial likelihood of irreparable misidentification." Simmons, 390 U.S. at 384, 88 S.Ct. 967. See also Neil v. Biggers, 409 U.S. 188, 93 S.Ct. 375, 34 L.Ed.2d 401 (1972) (outlining factors to be considered in totality of circumstance analysis); Manson v. Brathwaite, 432 U.S. 98, 97 S.Ct. 2243, 53 L.Ed.2d 140 (1977) (affirming the totality-of-the-circumstances approach and permitting admission of the evidence unless the circumstances make it likely that the evidence is not reliable). We review the trial court's underlying factual findings for clear error and its ruling on the admissibility of the identification evidence under the abuse of discretion standard. King v. Commonwealth, 142 S.W.3d 645 (Ky.2004).

Here, JH testified that on November 1, 2003, when she was fourteen-years old, a man driving a white van pulled up along side her and the friend with whom she was walking and asked them their names. Frightened, the two girls ran home. A few days later, on November 5, 2003, as she was walking home through an alley by herself, a man grabbed her from behind, told her he had a gun, and warned her not to scream. JH freed herself by hitting the man with her elbow and ran home, but as she was wrenching herself free, she faced the man and caught a clear, albeit brief, look at him. She testified that about a month later police officers twice showed her an array of six photographs of men matching the description she had given of her attacker. The first time she tentatively picked out two men, one of whom was Duncan, as resembling the man who grabbed her. She remembered her attacker as wearing glasses but none of the men in the photo array wore glasses. A few days later the police presented her with a second array of six photographs, this one with a more recent photograph of Duncan, one in which he and the other men included in the array are wearing glasses. From that array, JH positively identified Duncan as her assailant.

Duncan sought to suppress JH's identification and argued that showing JH his photo the second time was unduly suggestive. The trial court, disagreed. It ruled that the identification procedures were not unduly suggestive, and so admitted JH's identification evidence without going to the second step of the test and considering whether there was a "substantial likelihood" that JH had misidentified Duncan. On appeal, Duncan renews his objection to the second photo array as unduly suggestive.

An identification procedure is suggestive when it tends to focus attention on a single individual. United States v. Montgomery, 150 F.3d 983 (9th Cir.1998); St. Clair v. Commonwealth, 140 S.W.3d 510 (Ky.2004) (showing witness mug shots of two suspects and no other photos impermissibly suggestive); Moore v. Commonwealth, 569 S.W.2d 150, 153 (Ky.1978) ("The display ... of a single mug shot ... unaccompanied by any other pictures, was unnecessarily suggestive."). Repeatedly showing the picture of an individual can also have such a focusing effect. Simmons, 390 U.S. at 383, 88 S.Ct. 967. Duncan insists that the appearance of his photo in both arrays rendered the second array impermissibly suggestive because by virtue of the repetition his second photo was "emphasized." This precise question has not previously been before us, but federal courts addressing the issue have held that while "the fact that a defendant's photo is the only one which occurs in several photo arrays shown to the same witness could be unduly suggestive," it is not necessarily so. Stewart v. Duckworth, 93 F.3d 262, 265 (7th Cir.1996) (citing United States v. Bagley, 772 F.2d 482 (9th Cir. 1985)). In holding that this scenario was not unduly suggestive in the case before it, the Stewart Court noted that eleven days separated the presentation of the two arrays and that the two photos of the defendant were taken four years apart and looked very different. In one the defendant looked thin and wore long hair, while in the other he was heavier and had short hair.

Here, three days separated JH's viewing of the two arrays, not as long a period as in Stewart, but a substantial gap nonetheless. The two photos of Duncan, moreover, are markedly different. They were taken more than three years apart, and in the earlier one Duncan appears noticeably younger, is not wearing glasses, and has a full goatee. In the more recent one, Duncan's face is fuller, he is wearing glasses, and he has a mustache but no beard. Given the significant differences between the two photos, we are convinced that JH was not shown what amounted to a mere repetition of the first photo, and that fact plus the amount of time between the two viewings convince us that the photo identification process was not unduly suggestive. The trial court did not abuse its discretion by so ruling.

IV. The Trial Court Did Not Err by Allowing a Police Detective to Express a Modest, Experience-Based Opinion.

During his cross-examination of the lead detective in the case, Detective Mark Fulmore, Duncan elicited testimony to the effect that SM and JH both described their attacker as not very tall, as perhaps around five-foot-eight inches tall, whereas Duncan is six-foot-one. On redirect examination, the Commonwealth asked Detective Fulmore if in his experience it was unusual for witnesses to describe a suspect's height inaccurately. Duncan objected on the ground that the detective was not an expert in eye-witness identifications and so was not qualified to render an opinion. Overruling the objection, the trial court ruled that if a proper foundation were laid the officer could testify about his experience. Asked about his experience, the detective stated that he had been interviewing witnesses for twenty-one years and that it was common for a witness to be mistaken about a suspect's height. Duncan contends that the trial court abused its discretion under KRE 702 by admitting what amounted to an unreliable expert opinion. We disagree.

Under KRE 702, an expert may testify concerning "scientific, technical, or other specialized knowledge" if such testimony "will assist the trier of fact to understand the evidence or to determine a fact in issue." A witness may be qualified by "knowledge, skill, experience, training, or education," to give such testimony, and we have held that where the knowledge required is neither extensive nor complex, an officer's opinion may be admissible on the basis of training and experience. Dixon v. Commonwealth, 149 S.W.3d 426 (Ky.2004) (narcotics investigator allowed to opine that notations on slip of paper referred to drug transactions). Here, Detective Fulmore did not purport to be an expert in eye-witness psychology nor did he claim that certain types of witnesses or witnesses in certain situations were more or less apt than others to estimate a suspect's height inaccurately. In essence, he testified only that his experience confirmed the common sense notion that people are generally poor judges of height. As such the problem with his testimony, if there was one, was not that it was unreliable, but that it was not calculated to "assist the trier of fact." As the United States Court of Appeals for the First Circuit has explained, "[ejxpert testimony on a subject that is well within the bounds of a jury's ordinary experience generally has little probative value. On the other hand, the risk of unfair prejudice is real. By appearing to put the expert's stamp of approval on the government's theory, such testimony might unduly influence the jury's own assessment of the inference that is being urged." United States v. Montas, 41 F.3d 775, 784 (1st Cir.1994). Whether Detective Fulmore's opinion could have been elicited during his direct examination, is a question that is not before us.

In fact, his opinion was elicited only in response to Duncan's questions meant to discredit the victims' identifications. Generally, "when a witness has been impeached, the party introducing him may introduce evidence to corroborate his testimony or support his credibility." Samples v. Commonwealth, 983 S.W.2d 151, 154 (Ky.1998) (citation and internal quotation marks omitted). Accordingly, as the Montas Court noted, "[e]xpert testimony may be used `on some occasions to explain even non-esoteric matters, when the defense seeks to discredit the government's version of events as improbable.'" 41 F.3d at 784 (quoting from United States v. Taylor, 18 F.3d 55, 59 (2nd Cir.1994)). The trial court here did not abuse its discretion under KRE 702 by allowing Detective Fulmore to testify on redirect examination that in his extensive experience witnesses commonly misgauge a suspect's height.

CONCLUSION

In sum, having failed to establish the significance of its DNA evidence beyond the expert's testimony that it did not exclude Duncan as a possible source of the DNA recovered from SM's panties, the Commonwealth grossly misrepresented that evidence when, during its cross-examination of Duncan and during its closing argument, it asserted that that evidence identified Duncan as the source and had to be "wrong" if Duncan's testimony was to be believed. Given the immensely persuasive effect DNA evidence tends to have, the Commonwealth's misuse of it rendered Duncan's trial manifestly unfair and necessitates that his conviction be reversed. Duncan's other claims of error are not well-taken. At a retrial, Duncan is not exempt from the charge that he kidnapped SM, JH's pretrial identification of him is not subject to suppression on the ground that it resulted from an unduly suggestive procedure, and should Duncan attack the victims' identifications of him by eliciting that they initially described their attacker as a much shorter person, the Commonwealth's investigator can counter that in his experience witnesses commonly misjudge a suspect's height. Accordingly, we reverse the Judgment of the Jefferson Circuit Court and remand for additional proceedings.

All sitting. All concur.