¶1 Ruibal petitioned for review of the court of appeals' judgment affirming his conviction for second degree murder. Over defense objection and without taking evidence or making any findings as to reliability, the trial court admitted expert testimony to the effect that the victim's injuries in this case demonstrated "overkill," a formal term describing multiple injuries focused on one area of the victim's body, which includes blows about the head and face that are numerous *591 and extensive, indicating that the assailant likely had either a real or perceived emotional attachment to the victim. Relying on case law from several other jurisdictions, a treatise dealing with related kinds of injuries, and the witness's own experience with autopsies involving similar injuries, the court of appeals concluded that the expert opinion was sufficiently reliable and that the trial court had implicitly found as much by granting the prosecution's proffer.

¶2 Because the trial court made no specific finding that the theory of "overkill" espoused by the witness was reliable, nor was the reliability of that theory either supported by evidence in the record or already accepted in this jurisdiction, its admission amounted to an abuse of discretion. Because there was, however, overwhelming evidence of the defendant's guilt quite apart from the expert testimony, the error was necessarily harmless. The judgment of the court of appeals is therefore affirmed.

I.

¶3 George Ruibal was charged with and convicted of second degree murder for the beating and strangulation death of the woman with whom he was living at the time. He was sentenced to forty years in the custody of the department of corrections.

¶4 It was undisputed at trial that the victim's body was discovered lying on a couch in the couple's apartment, on a Monday, by the defendant and a co-worker, from whom the defendant had gotten a ride home and whom he had invited in to see the couple's new apartment. The co-worker immediately called 911, and upon arrival, the responding officer could see that the victim had bruises and scratches on her face. An autopsy determined that the victim died from closed head injuries-specifically a subdural hemorrhage, a subarachnoid hemorrhage, and a brain contusion-due to blunt force associated with manual strangulation, both of which were estimated to have occurred many hours before she died. The coroner also testified that the victim was covered in contusions and abrasions: eight or nine contusions on her head and face; multiple abrasions on her face and neck; and as many as fifty contusions and another twenty abrasions on her torso, arms, and legs.

¶5 The prosecution presented extensive evidence tending to show that the defendant had beaten the victim in their apartment and left her unattended to die. In addition to the medical evidence of the injuries themselves, the prosecution demonstrated inconsistencies in the defendant's accounts of his and the victim's movements on the weekend preceding discovery of the body and presented evidence, including documentary and other physical evidence, making it unlikely that events could have transpired as he asserted. Among other things, the defendant's account of the victim's having been attacked by a stranger while on a shopping trip to a nearby grocery store on Saturday night and yet continuing to function doing household chores over the weekend was challenged by expert testimony as very unlikely, given the severity of her injuries. Similarly, other physical evidence like the absence of blood on the victim's hands and the defendant's DNA under her fingernails and on her collar, as well as scratches and scabs on the defendant's nose and knuckles, remained unexplained by the defendant's account. In addition, his account of rarely speaking with the victim from work as an explanation for not calling to check on her was contradicted by phone records and the testimony of co-workers, who also recounted the turbulence of the defendant's relationship with the victim, including prior incidents of domestic violence.

¶6 Finally, the prosecution presented the testimony of the defendant's cellmate to the effect that the defendant confessed to strangling the victim and admitted that he brought someone home with him to witness his supposed discovery of the body. The cellmate also testified that the defendant said he was angry because he had to sell his truck to pay bills while the "ungrateful" victim sat at home drinking. While it appeared that the cellmate could possibly have accessed a newspaper article reporting the crime, officer testimony also indicated that the article in question did not mention two important details in the cellmate's story-that the defendant referred to the victim as "Baby," which was corroborated by the co-worker upon discovery *592 of the body, and the fact that the victim's family had pushed police to continue to investigate the crime. 1

¶7 Although the defendant did not testify, he presented a theory of defense involving an alternate suspect through his prior statements, other witnesses, the cross-examination and arguments of his counsel, and his theory-of-the-case instruction. According to the defendant's theory, another man, J.D., had beaten the victim somewhere outside their apartment. The prosecution introduced the defendant's version of events, through his interviews with the police, to the effect that the victim left the apartment around 8:30 p.m. Saturday with a twenty dollar bill the defendant had given her to buy milk at Albertsons. When the victim returned to the apartment several hours later, she had bruises and scratches on her face. Although she initially wanted to be taken to the hospital, the defendant attempted, through phone calls, but was unable to find her a ride. On Sunday, the victim made the bed, ate breakfast, walked around, and said she no longer wanted to go to the hospital. When the defendant left for work on Monday morning, the victim was on the couch. He had no contact with the victim until he returned home with his co-worker. He explained that the victim never called him during the day, and he did not call her because it would upset her.

¶8 It was undisputed that a man identified as J.D. lived in the area and had been contacted by the police shortly after midnight on that Sunday morning, near Albertsons. A woman who testified that she had a prior relationship with J.D. was also permitted to testify that J.D. had a history of being violent towards her and in one instance had even strangled her. Although it appeared to be contradicted by surveillance footage and the discovery of the twenty dollar bill still in the victim's pocket, an Albertsons' manager also testified that she saw the victim buy milk that night sometime between 9:30 and 10 p.m.

¶9 In light of the defendant's theory that the victim had been attacked by a stranger, the prosecution presented the expert testimony of a second forensic pathologist expressing, among other things, conclusions about the relationship between a killer and victim that he opined could be reliably inferred from what he referred to as "overkill." Over the objection of the defense and without taking evidence about the reliability of the theory, the trial court permitted the witness to offer an expert opinion to the effect that the victim's injuries in this case demonstrated "overkill," a formal term describing multiple injuries focused on one area of the victim's body, which includes blows about the head and face that are numerous and extensive, indicating that the assailant likely had either a real or perceived emotional attachment to the victim.

¶10 Following his conviction, the defendant included among his assignments of error in the intermediate appellate court the admission of this expert opinion. Relying on case law from several other jurisdictions, a treatise dealing with related kinds of injuries, and the witness's own experience with autopsies involving similar injuries, the court of appeals concluded that the expert opinion was sufficiently reliable and that the trial court had implicitly found as much by granting the prosecution's proffer. Upon rejecting the defendant's other assignments as well, the court of appeals, with one member of the panel dissenting, affirmed his conviction.

¶11 We granted the defendant's petition for certiorari review solely on the question whether the expert opinion concerning "overkill" was erroneously admitted without a specific finding that the principles upon which it was based were reliable.

II.

¶12 In People v. Shreck , 22 P.3d 68 (Colo. 2001), we extended our prior holdings beyond what we had at times referred to as "experienced-based specialized knowledge," Brooks v. People , 975 P.2d 1105 , 1114 (Colo. 1999), to make clear that CRE 702, rather than the Frye 2 "general acceptance" standard, governs *593 the admissibility of all expert testimony in this jurisdiction, including testimony based not only on technical or other specialized knowledge, but even on novel scientific devices and processes involving manipulation of physical evidence. Shreck , 22 P.3d at 74, 78 . In doing so, however, we also explained in detail the nature of the CRE 702 inquiry and articulated the obligations of trial courts prior to admitting expert evidence pursuant to this rule. Id. at 77-78 . We there held that the trial court's inquiry should be broad in nature and take into consideration the totality of the circumstances of each specific case, focusing on both the reliability and relevance of the evidence. Id. at 77 . In light of the wide range of factors that may be considered in any individual case and the liberal nature of the standard, we imposed upon trial courts admitting evidence pursuant to CRE 702 an obligation to first determine and make specific findings on the record, not only as to the reliability of the scientific principles upon which the expert testimony is based and the qualifications of the witness giving that testimony, but also the usefulness of such testimony to the jury, including specific findings with regard to the court's obligation pursuant to CRE 403 to ensure that the probative value of the evidence would not be substantially outweighed by any of the countervailing considerations enumerated in the rule. Id. at 70, 77-78 .

¶13 As we have also made clear, whether making those determinations will require an evidentiary hearing outside the presence of the jury will ultimately depend on whether the record can support the court's findings without doing so. People v. Rector , 248 P.3d 1196 , 1201 (Colo. 2011). Depending upon the extent to which the reliability of the scientific principles at issue has already been determined or is not disputed at all, for example, further evidence of their reliability may not be required. See id. at 1201 ; cf. Kumho Tire Co. v. Carmichael , 526 U.S. 137 , 152, 119 S.Ct. 1167 , 143 L.Ed.2d 238 (1999) (admonishing trial courts to "avoid unnecessary 'reliability' proceedings"). Similarly, while we have indicated that it would be preferable in light of discovery and endorsement requirements to make these determinations prior to trial, it may be necessary, especially with regard to such matters as the incremental probativeness of proffered expert testimony, to withhold ruling until later. Rector , 248 P.3d at 1200 n.5. With regard to the requirement for specific findings concerning a determination of the reliability and relevance of evidence to be admitted pursuant to CRE 702, with record support, we have, however, been unwavering. See, e.g. , id. at 1200 (noting that although a trial court has discretion to determine whether a challenge to expert testimony warrants a Shreck analysis, where a proper challenge has been raised, a trial court "is required to issue specific findings" as to relevance and reliability under CRE 702 ); Estate of Ford v. Eicher , 250 P.3d 262 , 266 (Colo. 2011) (stating a trial court "is required to issue specific findings regarding its analyses" as to relevance and reliability); City of Aurora ex rel. Util. Enter. v. Colo. State Eng'r , 105 P.3d 595 , 612 (Colo. 2005) ("The court must issue specific findings as it performs its CRE 702 and 403 analyses.").

¶14 While it may be inferable from an adequate record of the trial court's awareness of the applicable standard and its admission of the expert's testimony that it considered the expert evidence both reliable and relevant, the requirement for specific findings is not satisfied by this inference alone. In light of the broad range of expertise governed by the rule and the necessarily non-specific nature of the factors governing the reliability, relevance, and incremental probativeness of expert opinion in any given case, the requirement for specific findings is imposed as a means of ensuring meaningful review of this broadly discretionary decision. In the absence of these specific findings, or a record not only supporting admission but virtually requiring it or precluding any reasonable dispute as to the basis of the court's admission, the trial court must be considered to have abused its discretion in admitting expert testimony.

¶15 Although the record in this case indicated merely that the witness was a forensic pathologist and the trial court made only a general finding that his testimony helped put the nature and type of the victim's injuries into context for the jury, we need not address the adequacy of those findings as to *594 the qualifications of the expert or relevance of his testimony because the trial court made no findings, and the record was virtually devoid of support, concerning the reliability of the scientific principles underlying the theory and interpretation of "overkill." The witness relied on a single treatise as support for the theory of "overkill," which even he did not accept as generally authoritative, and which, in any event, defined "overkill" far too narrowly to be applicable to the injuries inflicted in this case or to support the essential inference, drawn by the expert in this case, of an emotional relationship between the victim and killer. 3 Similarly, although the witness testified that he had performed many autopsies himself and knew "who confessed to doing what," he failed to offer even anecdotal, much less empirical, evidence supporting his conclusion that beatings like the one in this case were likely committed by someone with an emotional connection to the victim. Finally, neither the appellate courts of this jurisdiction nor those of any other jurisdiction have yet accepted as reliable the theory or interpretation of "overkill" advanced by the witness.

¶16 Of the handful of reported cases in which the concept of "overkill" is analyzed, apparently none has found the theory reliable for purposes of the expert testimony analysis required by Rule 702. See , e.g. , State v. Lenin , 406 N.J.Super. 361 , 967 A.2d 915 , 925-26 (2009) (rejecting expert testimony that overkill showed "interpersonal aggression" because, among other things, no New Jersey case had ruled on the reliability of crime scene or behavioral analysis); cf. State v. Hebert , No. 2010 KA 0305, 2011 WL 2119755 , at *10 (La. Ct. App. Feb. 11, 2011) (unpublished opinion holding defense failed to establish the reliability of its theory that extreme wounding, or overkill, can be used to determine whether an assailant was suffering from psychosis, according to the standard for admitting expert testimony under Daubert 4 ); State v. Wright , No. 0801010328, 2009 WL 3111047 , at *8 (Del. Super. Ct. Sept. 14, 2009) (unpublished opinion rejecting as insufficiently reliable for admission under Delaware Rule of Evidence 702 testimony that overkill indicates "personalized anger and sustained aggression or rage"). Of the one published and two unpublished cases relied on by the court of appeals below, none purported to find the theory of overkill sufficiently reliable for admission as expert testimony. See Richardson v. State , 83 S.W.3d 332 , 339, 350 (Tex. Ct. App. 2002) (rejecting assertion that jury's verdict of murder rather than sudden passion was against the great weight and preponderance of the evidence where among other testimony the medical examiner testified that crimes of passion are generally overkills with dozens and dozens of stab wounds ); State v. Suttle , No. A-2417-08T3, 2011 WL 2314474 , at *5 (N.J. Super. Ct. App. Div. June 10, 2011) (unpublished opinion rejecting claim of collateral estoppel bar to second verdict that defendant acted purposely or knowingly, where first jury may have found crime of passion based on prosecutor's argument that crime was personal and overkill); People v. Varela , No. B197473, 2008 WL 2764578 , at *5 (Cal. Ct. App. July 17, 2008) (unpublished opinion rejecting claim of prosecutorial misconduct for evoking expert testimony that stabbing was not overkill, on grounds that prosecutor did not evoke opinion that stabbings were not committed in heat of passion).

III.

¶17 Although the trial court therefore abused its discretion in admitting expert testimony of "overkill," reversal of the defendant's conviction is nevertheless not warranted in this case. Error in the trial process does not warrant the reversal of a conviction if it can be shown to be harmless. People v. Summitt , 132 P.3d 320 , 327 (Colo. 2006). Even a properly objected-to trial error will be disregarded as harmless whenever the *595 error did not substantially influence the verdict or affect the fairness of the trial proceedings. James v. People , 2018 CO 72 , ¶ 19, 426 P.3d 336 , 341. The strength of properly admitted evidence supporting the verdict is one important consideration when evaluating such error. Crider v. People , 186 P.3d 39 , 43 (Colo. 2008). If that evidence overwhelmingly demonstrates the defendant's guilt, the error must be disregarded as harmless. Pernell v. People , 2018 CO 13 , ¶ 25, 411 P.3d 669 , 673 ; Summitt , 132 P.3d at 327 .

¶18 Here the erroneously admitted evidence was limited to expert opinion to the effect that the nature of the victim's injuries made it likely her assailant was someone with an emotional connection to her rather than a stranger. In addition to the inference of anger or passion to be intuitively drawn, in the absence of another likely explanation, from such a beating, the prosecution presented an abundance of physical, documentary, and testimonial evidence making it highly unlikely not only that the victim would have been physically capable of functioning over the ensuing days, as claimed by the defendant, but also that his account was contradicted and therefore untrue for a host of other reasons. It was the testimony of two forensic pathologists that it would have been medically unlikely the victim could have walked the distance from the grocery store with her injuries, much less that she could have made the bed, eaten breakfast, or walked around the apartment the following day. Admissible expert forensic testimony also noted the likelihood that unless she were immediately rendered unconscious by the beating, her hands would have shown blood from reflexively grasping at her wounds. Beyond the medical implications derived from the physical nature of her injuries, otherwise unexplained physical evidence-such as scratches on the defendant's face and his DNA discovered under the victim's fingernails and near her throat-strongly suggested his involvement in a physical altercation with the victim.

¶19 In addition to the physical evidence, however, the defendant's ever-changing account was contradicted in a number of key respects by both testimonial and documentary evidence. Telephone records contradicted the defendant's account of attempting to call for medical help on Saturday night, as well as his initial explanation for not calling the victim on Monday during the day. Co-workers recounted telephone calls indicating the defendant's turbulent relationship with the victim, as well as the defendant's suspicious request for a co-worker to see his new apartment, leading to discovery of the victim's body, and the defendant's subsequent request of that co-worker to tell a particular version of the discovery. Finally, the defendant's cellmate contradicted his entire account by testifying that he actually confessed to the killing, providing details that were not publicly available.

¶20 Assuming evidence of the defendant's alternate suspect theory was properly admitted at all, it offered little to question this powerful case against the defendant. Although the trial court did not have the benefit of our clarification in People v. Elmarr , 2015 CO 53 , ¶ 34, 351 P.3d 431 , 440, we have long made clear that because of the danger of confusing the jury, even a showing of motive and opportunity on the part of a third, uncharged person is insufficient to assert an alternate suspect theory. People v. Mulligan , 193 Colo. 509 , 568 P.2d 449 , 456-57 (1977) ; see also People v. Salazar , 2012 CO 20 , ¶ 21, 272 P.3d 1067 , 1073 ; People v. Flowers , 644 P.2d 916 , 920 (Colo. 1982). In this case, even evidence of motive and opportunity was extremely weak, if not entirely lacking.

¶21 The only evidence of opportunity-that the victim was in the same area near the same time as the alternate suspect-came from the defendant himself and an Albertsons employee who believed she saw someone matching the victim's description buying milk on the night in question. Apart from medical evidence indicating the strong likelihood that after sustaining such severe injuries the victim could not have walked the distance from Albertsons, video footage from the store was unable to substantiate the employee's observation, and the presence of a twenty dollar bill in the victim's pocket appeared to dispel any suggestion that she ever purchased milk with the money reportedly *596 given her by the defendant for that purpose. With regard to motive, robbery and sexual assault were clearly ruled out, and although there was testimony to the effect that the alternate suspect had been abusive in the past to a domestic partner, there was no evidence of pattern, a unique modus operandi, or any prior connection to the victim providing a motive for attacking her. Not only was there no direct evidence, either physical or testimonial, but in fact no meaningful circumstantial evidence of contact between the two.

¶22 In the face of such overwhelming evidence of the defendant's guilt, we cannot find even a reasonable possibility that the outcome of the trial would have been different but for the expert's opinion concerning a likely real or perceived emotional relationship with the victim. See Krutsinger v. People , 219 P.3d 1054 , 1058 (Colo. 2009) (making clear that the "reasonable possibility" standard for constitutional error is a more onerous harmless-error standard than the "substantially influence" standard for non-constitutional error).

IV.

¶23 Because the trial court made no specific finding that the theory of "overkill" espoused by the witness was reliable, nor was the reliability of that theory either supported by evidence in the record or already accepted in this jurisdiction, its admission amounted to an abuse of discretion. Because there was, however, overwhelming evidence of the defendant's guilt quite apart from the expert testimony, the error was necessarily harmless. The judgment of the court of appeals is therefore affirmed.

JUSTICE GABRIEL does not participate.