concurring in part and dissenting in part.
{102 At the jury trial in this very serious criminal case, the trial court permitted a witness to testify, over defendant's objection, that defendant was "very guilty-looking" immediately after the commission of the crime alleged.1 Yet, the witness could not even articulate the behavior that led to this characterization. In my view, a court should mever admit testimony by a witness that a *489defendant was "very guilty-looking."2 Such testimony is irrelevant, inadmissible under CRE 701, unfairly prejudicial, and may render unreliable criminal convictions in which such evidence has been admitted. The error in the admission of this evidence was not harmless and requires the reversal of defendant's conviction and a new trial. Accordingly, I respectfully dissent from Part III of the majority opinion. 3
I. "Very Guilty-Looking" Testimony
A. Relevance-CRE 401
{103 The threshold inquiry for all evidence is relevance. CRE 402. Evidence that is relevant generally is admissible. Id. Irrelevant evidence is always inadmissible, unless no objection is made to it. Id. Evidence is relevant if it has "any tendency to make the existence of any fact that is of consequence to the determination of the action more probable or less probable than it would be without the evidence." CRE 401. Some types of evidence are so obviously relevant that no further inquiry or showing is required as a condition of admissibility. For example, eyewitness testimony regarding the commission of a crime or a voluntary confession to a crime by a defendant is so obviously relevant that no showing need be made to admit such evidence. But the relevance of other types of evidence is not so obvious and may depend upon the validity of intermediate propositions; in other words, some evidence may be relevant only if certain propositions underlying the evidence are valid. Such is the case here.
1 104 Opinion evidence by a witness that a defendant was "very guilty-looking" would be relevant only if there were a commonly accepted definition of the matrix of human expressions, body language, and actions that demonstrate guilt. This definition, or understanding, may be supplied by the law, by common experience, or perhaps by social science. Without such a common understanding, the opinion by the witness that defendant looked guilty has no probative value whatsoever.
1105 I simply do not know what it means to be "very guilty-looking," and the record in this case is devoid of any information upon which the trial court or this court could determine that a common understanding or definition of what it means exists.4 The witness struggled to explain what she meant by the term "very guilty-looking"; that defendant's "hands were behind his back," "he was looking down and looking up," and "he looked like a little kid or a dog would ... [wihen they have done something wrong" was the extent of her explanation. And importantly, the prosecution did not introduce any evidence that these behaviors are commonly understood indications of guilt.
1106 In Salcedo v. People, the Colorado Supreme Court held that testimony by an " 'expert in the area of narcotics interviews'" that the defendant's behavior and characteristics conformed to the profile of a drug courier was inadmissible as substantive evidence of guilt. 999 P.2d 883, 886-37 (Colo.2000). The supreme court explained that "the prosecution offered no foundational evidence that a defendant's conformity to [the expert's] drug courier profile is a reliable indication of a person's guilt," especially because "[the behaviors and characteristics in [the expert's] drug courier profile are not unique to drug smugglers." Id. at 889. "Indeed, our review of cases addressing use of drug courier profiles indicates that ... few if any [air travel] passengers are entirely above suspicion. Drug courier profiles are broad in their sweep and sometimes appear depen*490dent on seemingly contradictory behaviors and characteristics." Id.
€107 The supreme court concluded that, due to "[tlhe lack of evidence in the record indicating that conformity to [the expert's] drug courier profile is a reliable indicator of guilt," the expert's testimony was irrelevant: that drug couriers displayed certain commonplace behaviors and characteristics did not make it more probable that a person displaying those behaviors and characteristics was a drug courier. Id. at 839-40.
{108 I believe a similar analysis applies here. There may be some actions under some cireumstances that some people might construe to mean that the person at issue was "very guilty-looking," but I am aware of nothing in the law, in social science, or in common experience that creates anything even close to a consensus as to when certain actions, expressions, or body language make a person "very guilty-looking." Moreover, to the extent that there might be any such common understanding, it almost certainly varies by culture.
{109 Given the explosive nature of such testimony-after all, the whole purpose of a criminal trial is to determine whether the defendant is guilty or not guilty-a court should not permit a jury to hear such testimony unless it has some degree of confidence that this type of evidence is based on a generally accepted understanding of what a "very guilty-looking" person looks like. Reducing the complex of factors that may be characteristic of "very guilty-looking" persons to what a little kid or a dog looks like when it realizes it has done something wrong is insufficient to support the admission of such explosive evidence in a criminal trial. Even assuming that there is a common understanding regarding what characteristics children or dogs exhibit after they have been discovered misbehaving, there could be many reasons an adult might display some of these characteristics that have nothing to do with criminal guilt.
' 110 Without some showing by the prosecution that the behavior described by the witness is more likely to manifest in those who have committed criminal acts than those who have not, I cannot agree with the majority that the witness's description of defendant's behavior formed a proper basis for testimony that defendant was "very guilty-looking." Cf id. at 840 ("We conclude that before drug courier profile evidence can be considered logically relevant to whether a person conforming to the profile is a drug courier, the prosecution must demonstrate that the behavior and characteristics that constitute the profile are relatively unique to drug couriers. The prosecution made no such demonstration in this case.").
111 Accordingly, I conclude that, at least . in the context of evidence admitted against a defendant during a criminal trial, a witness's lay testimony that the defendant was "very guilty-looking" does not have any tendency to make it more probable that the defendant committed the criminal act at issue and therefore is not relevant under CRE 401.
B. Lay Opinion-CRE 701
T 112 In addition to the testimony's lack of relevance under CRE 401, I disagree with the majority that the testimony was properly admitted under CRE 701. For a variety of reasons, primarily based upon considerations of reliability and relevance, lay witnesses at common law were prohibited from testifying as to opinions. Only expert witnesses could testify as to opinions. This common law rule has been codified, with some exceptions, in CRE 701. Under limited cireumstances, a lay witness may express an opinion that is "(a) rationally based on the perception of the witness, (b) helpful to a clear understanding of the witness'[s] testimony or the determination of a fact in issue, and (c) not based on scientific, technical, or other specialized knowledge within the scope of Rule 702." Id. Testimony that a defendant was "very guilty-looking" does not meet the requirements of CRE 701 that the opinion be "rational" and "helpful to a clear understanding of the witness'[s] testimony."
1113 Such testimony is rational only if there is some common, accepted, and generally accurate understanding of the human activities that make someone "very guilty-looking." As discussed above, the record is devoid of such evidence. And such testimony *491is not helpful because, as previously mentioned, the term "very guilty-looking" has no established meaning. A person who is very guilty-looking to one person may be viewed by another observer as being depressed, sad, or uncomprehending. Consequently, the witness's testimony here that defendant was "very guilty-looking" did not provide any useful information to the jury regarding her perception of defendant's demeanor after the alleged assault.
C. Unfair Prejudice-CRE 403
114 Even if I were to believe that the witness's opinion was "rational" and "helpful" as required by CRE 701, the testimony still should have been exeluded under CRE 403 because its probative value (which I consider to be either nonexistent or extremely minimal) was far outweighed by its prejudicial impact. The probative value of the evidence was minimal precisely because of what I discussed above: There is no common understanding as to the complex of factors that makes one "very guilty-looking." The prejudicial impact is obvious under the facts of this case: A witness identified with defendant (and particularly a witness who might be expected to "side" with defendant) told the jury that she thought he was "very guilty-looking." Contrary to the majority's analysis, the leap from a defendant being "very guilty-looking" immediately after the commission of the alleged crime to a determination of legal guilt is but a short jump.
D. Case Law
(115 In a different, but somewhat analogous context, the Colorado Supreme Court has instructed us that counsel may not argue to a jury that a witness is a liar, no matter how strong the evidence may be of untruthfulness. Domingo-Gomez v. People, 125 P.3d 1048 (Colo.2005). The rationale for the supreme court's holding was that
[slome words or analogies by their very nature resonate more powerfully in the heart and minds of the jury. They evoke strong reactions in jurors and take them down the path towards a conviction where the evidence does not necessarily lead. The word 'lie' is such a strong expression that it necessarily reflects the personal opinion of the speaker. When spoken by the State's representative in the courtroom, the word 'lie has the dangerous potential of swaying the jury from their duty to determine the accused's guilt or innocence on the evidence properly presented at trial. Thus, it is improper for a lawyer to assert his opinion that a witness is lying. He can argue to the jury that they should not believe a witness, but he should not call him a liar.
Id. at 1050 (citation, internal quotation marks, and alterations omitted). If counsel cannot argue that a witness is a "liar" or has "lied" in a criminal case, it is difficult for me to imagine that the admission of the much more damning testimony that was admitted in this case would be permissible.
{116 Each of the Colorado eases relied upon by the majority to reach the conclusion that the testimony was properly admitted under CRE 701 is clearly 5In Elliott v. People, the challenged testimony was that the defendant "was getting ready to hit the owner of the store." 176 Colo. 873, 377, 490 P.2d 687, 689 (1971). Getting ready to hit someone may, in many cases, be a reasonable characterization of actions taken by a person. If a person raises his or her fist to another person, it is reasonable to believe that the person was getting ready to hit the other person. And, importantly, other people who heard a witness describe someone as "getting ready to hit" another person likely would have a frame of reference from their own experiences that would provide them with an understanding of what the witness meant by the statement. Such conduct is completely different than the amorphous concept of being "very guilty-looking," because that concept lacks a commonly-understood meaning.
*492{117 Similarly, in People v. Farley, a division of this court permitted a counselor who had assisted the police in investigating a crime to testify that the victim was "in a state of shock." 712 P.2d 1116, 1118-19 (Colo.App.1985). I submit that being in a state of shock is something that may have a common meaning or understanding to a lay person and thus a lay juror. In any event, it does not carry with it the tremendous potential of prejudice of the testimonial evidence admitted in this case. The determination of guilt is the fundamental purpose of a criminal trial. Although the fact that the victim was in a state of shock might have consequence in a particular criminal case, testimony that someone was in shock is not equivalent to testimony that someone looked guilty of the crime charged.
{118 People v. Rubamowitz, also relied upon by the majority, merely supplies the noncontroversial proposition that "[the requirement of CRE 701) as to the testimony's helpfulness to the jury is satisfied when a witness'[s] expression of his opinion provides the jury with his [or her] overall impression and brings the particulars into focus." 678 P.2d 45, 47 (Colo.App.1983) (internal quotation marks omitted). Therefore, Rubanowitz does not offer any guidance regarding the specific issue raised here: whether the witness's testimony that defendant was "very guilty-looking" actually satisfies the requirement that it "provide{ ] the jury with [her] overall impression and bring[ ] the particulars into focus."
{119 The out-of-state cases relied upon by the majority, none of which are binding upon this court, are also either distinguishable or were wrongly decided. In State v. Braxton, 352 N.C. 158, 581 S.E.2d 428, 444 (2000), a police officer testified that the defendant "looked guilty" "based upon his observation that, as [the] defendant saw [the officer] approaching, [the] defendant immediately raised his hands." While I would prophylac-tically prohibit all such "looks guilty" testimony, the factual predicate for the admission of the evidence in Braxton is far sounder than in this case. Holding one's hands in the air when a police officer approaches might, in some contexts, be reasonably characterized as exhibiting evidence of guilt (although if a defendant holds his or her hands in the air to avoid a potentially fatal encounter with an armed officer, the probative value of such testimony would be, like in this case, minimal).
{120 Moreover, in Braxton, the witness was a police officer. The jury will almost always know that police officers are part of the prosecution "team"; they are not generally perceived to be impartial witnesses. Conversely, in this case, the witness who testified that defendant was "very guilty-looking" was (presumably) a friend of defendant who did not previously know any of the other participants at the party. This presumed impartiality was repeatedly emphasized by the prosecutor during closing argument. In substance, if not words, the prosecutor told the jury that if this friend of defendant believed that defendant looked extremely guilty, how could the jury not similarly conclude that to be the case.
121 Neither People v. Jackson nor People v. Meza, two unpublished cases relied upon by the majority, ask or answer the critical, threshold question that the majority also does not address: what it means to say that a person is "very guilty-looking." In Jackson, the defendant claimed that his trial counsel provided ineffective assistance of counsel when he failed to object to testimony by a trial witness that the defendant appeared to be guilty of second degree murder because he avoided contact with his friends and failed to act upset by the victim's death. No. 304168, 2013 WL 276054, at *2 (Mich.Ct.App. Jan. 24, 2018) (unpublished opinion). The only analysis on this issue was that "[the witness's] testimony regarding [the defendant's behavior] was rationally based on his perceptions of [the defendant], thus it was 'proper" and "any error [by] defense counsel in failing to object to or for eliciting such testimony does not warrant relief." Id.
1122 In Meza, a witness testified that when he looked into the eyes of the defendant he saw "guilt." No. C©O67992, 2012 WL 2924540, at *4 (Cal.Ct.App. July 19, 2012) (unpublished opinion). The appellate court refused to decide the admissibility of the "guilt" statement stating, "[this argument *493was forfeited ... because [the] defendant failed to make it in the trial court." Id. Additionally, the court's statement in a footnote that the testimony would not have been objectionable as improper opinion testimony if the defendant had objected at trial because the witness "essentially described [the] defendant's behavior as being consistent with guilt," id. at *4 n. 2, does not provide any analysis regarding the testimony's helpfulness to the jury.
1123 The majority states that three out-of-state cases that have excluded "looked guilty" evidence provide no analysis or reason for the exclusion, implying that the courts rendering those decisions did not have a principled basis to reach their conclusions. See United States v. Fabel, No. CR06-041L, 2007 WL 318984 (W.D.Wash. Jan. 30, 2007); State v. Baldwin, 86 Kan. 1, 12 P. 818, 326 (1886); State v. Yarber, 102 Ohio App.3d 185, 656 N.E.2d 1822, 1826-27 (1995). However, that is not a fair characterization of those cases.
(124 Yarber was an ineffective assistance of counsel case in which defense counsel failed to object to improper leading questions, failed to establish a coherent strategy, failed to question the alleged victim concerning inconsistencies in his statements or even argue the inconsistencies during closing argument, and appeared "confused" throughout the trial. Yorber, 656 N.E.2d at 1324-25. One of the few coherent objections that defense counsel did make was to the testimony by a witness that the defendant "looked guilty." Id. at 1826. That objection was sustained by the trial court, but not addressed by the appellate court, presumably because the defendant did not argue on appeal that the testimony, despite the sustained objection, influenced the jury's verdict.
1125 In Baldwin, an "expert ... skilled and experienced in detecting crime from the appearance of those charged with it," testified that, based upon his experience in dealing with criminals, it was his opinion that the defendant had the appearance of being a guilty man. 12 P. at 826. The trial court sustained an objection to this question, and the Kansas Supreme Court held the opinion "very properly excluded" by the court. Id.
1 126 Even though it is true that the courts did not provide any analysis of their holdings in this regard, I would look at these cases a little differently than the majority. To me, it is entirely possible that those courts thought it was so obvious that such testimony should not be admitted in a criminal trial that no further explication was necessary. While my analysis may be more nuanced, my conclusion is the same.
1127 An unpublished case from the Kentucky Supreme Court, Martinez v. Commonwealth, also excluded "looked guilty" evidence. No. 2008-SC-000082-MR, 2009 WL 2706958 (Ky. Aug. 27, 2009). In Martinez, a police detective testified that the defendant reacted to a police interrogation "in a manner which indicated guilt." Id. at *5. The court held that the admission of this testimony was error under Kentucky's version of CRE 701 because the detective "had no personal knowledge that [the defendant] was guilty, only a belief, and his belief [the defendant] was guilty was not useful to helping the jury understand [the] ... statements" that the defendant made during the interrogation. Id. at *6. The same could be said here: The witness's belief that defendant was "very guilty-looking" was not useful to helping the jury understand defendant's conduct at the time of the incident, and the admission of her testimony therefore was error.
E. Reversible Error
(( 128 Because I believe that the trial court abused its discretion in admitting the witness's testimony that defendant was very "guilty-looking," it is necessary to determine whether its admission constituted reversible error. While there probably was sufficient evidence, under the applicable standard of review on a sufficiency challenge, to support defendant's conviction with or without this challenged testimony, this is not a case of overwhelming evidence of guilt.
129 Although the observations of the two adult witnesses demonstrate reprehensible . conduct by defendant (and perhaps could have supported a conviction on other charges), those observations do not prove the elements of the erime of which defendant was *494convicted: sexual assault on a child. As in many cases involving sexual assaults on young children, there was no physical evidence. Also, as noted by the majority, the victim did not testify at trial that defendant committed any offenses against her. 'To the contrary, at trial, the victim could not even identify defendant and could not remember whether any of the events she had previously related to investigators actually occurred.
{130 Given this record, I cannot conclude that the error in the admission of the "very guilty-looking" testimony was harmless.6 Therefore, a new trial is required.
II, Victim Reaction Testimony
{131 Whether the victim reaction testimony (the victim's father's testimony that the victim said thinking about "it" made her feel sick) was properly admitted is a closer question. While the majority's analysis regarding this issue is persuasive, ultimately I cannot agree. Even if this testimony were admissible under a hearsay exception, I conclude that the evidence should have been excluded under CRE 403.
132 Unlike the majority, I conclude that the record is unclear as to what precisely the victim was reacting to when she "felt sick to her stomach." Like the jury, we do not know the content of the counseling sessions. It is entirely possible, indeed likely, that matters discussed at the counseling session were very disturbing. Thus, it is entirely possible that the victim's sickness could have 'had everything to do with the content of the counselling sessions and nothing to do with defendant's actions. As a result, this evidence does not survive a proper CRE 408 analysis; its probative value is far outweighed by its likelihood to confuse the jury. I therefore believe the trial court abused its discretion in admitting this evidence, and for the reasons discussed above regarding the quantity and quality of the evidence presented in this case, this error was not harmless.
(133 But even if the erroneous admission of the victim reaction evidence does not in itself require reversal of the conviction, when combined with the erroneous admission of the "very guilty-looking" testimony, these errors together require reversal under the cumulative error doctrine. See Oaks v. People, 150 Colo. 64, 66-67, 371 P.2d 448, 446 (1962) (explaining the cumulative error doctrine).
IIL Judicial Bias
1134 Lastly, I address the claim of judicial bias. I agree with the majority that there is nothing in the record that supports a finding or conclusion that the trial court was prejudiced against defendant. The transeripts of all of the proceedings before the trial court demonstrate the opposite: The trial court was fair and impartial and, in the instances challenged by defendant, was simply trying to be helpful. t
135 That said, however, a judge must not only be impartial in fact, but also must not engage in any actions that would cause a reasonable person to conclude that there was any favoritism for or animosity against any party. See C.J.C. 1.2 ("A judge ... shall avoid impropriety and the appearance of impropriety."); C.J.0. 1.2 emt. 5 ("The test for appearance of impropriety is whether the conduct would create in reasonable minds a perception that the judge violated this Code or engaged in other conduct that reflects adversely on the judge's ... impartiality.").
11136 A trial court risks crossing that very important line when it suggests to one side or another how that party may resolve difficult evidentiary issues. It is the responsibility of the parties, both of whom in this case were represented by counsel, to make appropriate objections and to determine how to properly offer the evidence they wish to present. -It is the job of the trial court to *495rule on evidentiary objections; it is not the job of the trial court to instruct or suggest to - represented parties how evidence may be . admitted and how evidentiary barriers may be surmounted. CJ C.J.0. 2.2 emt. 4 ("It is not a violation of [the Code of Judicial Conduct] for a judge to make reasonable accommodations to ensure pro se litigants the opportunity to have their matters fairly heard.").
IV. Conclusion
1137 The jury found defendant guilty. For many good reasons, in our system of justice, that determination is committed to a jury, not judges. But it is the responsibility of judges to apply the rules of evidence to ensure that the information upon which jurors make these determinations is reliable. If the information made available to the jurors does not meet the requirements of the rules of evidence, the resulting decision by the jurors is not sufficiently reliable to deprive a person of his or her life or liberty.
1138 Accordingly, I respectfully dissent. Defendant's conviction should be reversed and the case remanded for a new trial.
. Defendant objected twice to this evidence, first, in a motion in limine which the trial court denied and, second, during the presentation of the evidence before the jury.
. In my view, the admission of such testimony always constitutes error. Whether the erroneous admission of such testimony requires a new trial depends upon application of the harmless error rule (when a contemporaneous objection is made) or the plain error rule (when no contemporaneous objection is made).
. As discussed below, I also respectfully dissent from Part IV of the majority opinion. However, I agree with the majority that the trial court did not err by refusing to sanction the prosecution for the discovery violation, and accordingly I join Part II of its opinion.
. The rules of evidence do not apply to preliminary questions concerning the admissibility of evidence except those with respect to privileges. CRE 104(1). However, the record still must reflect the information used by the trial court in making a determination of admissibility.
. While I believe the cited decisions of this court are distinguishable, to the extent that any of the cited decisions of a prior division of this court are inconsistent with my analysis, I would not follow those decisions. See, eg., People v. Smoots, 2013 COA 152, 20, - P.3d -- (One division of the court of appeals is "not obligated to follow the precedent established by another division.").
. The fact that the witness testified that defendant remained silent and did not defend himself when the victim's friend's father made accusations against him and ordered him to leave does not change my harmlessness analysis. In my opinion, remaining silent in the face of accusations is significantly different from confessing, and defendant may not have defended himself for reasons other than because he was guilty. Cf. People v. Quintana, 665 P.2d 605, 611 (Colo.1983) ("Due to the many possible explanations for the defendant's post[-larrest silence, ... evidence of his failure to make a statement to the arresting officers was so ambiguous and lacking in probative value as to be inadmissible as substantive evidence" of guilt.).