*803A jury found appellant Ronald Jackson guilty of assault with a dangerous weapon. The sole question before this court is whether the trial court abused its discretion by admitting evidence that appellant used PCP eighteen hours before the attack without any accompanying expert testimony enabling the jury to evaluate whether Jackson was under the influence of the drug at the time of the assault. We reverse and remand for a new trial.
I. Factual Background
Appellant Ronald Jackson and Desmon Beasley had been best friends for over forty years. In 2014, Jackson moved into Beasley's apartment. At that time, Beasley weighed over 600 pounds and, as a result, had limited mobility. Jackson paid Beasley a small amount in rent and assisted Beasley with daily activities.
By November 2015, the situation between Beasley and Jackson had changed, and Beasley wanted Jackson to move out. Beasley had lost a considerable amount of weight and was better able to move around. He began to pursue a romantic relationship with his friend, Erika Williams, and found it difficult to do so with Jackson in the apartment. Additionally, Beasley was frustrated by Jackson's inability to maintain a steady job and had difficulty supporting both himself and Jackson on his Social Security income.
Tensions between Jackson and Beasley came to a head on November 13, 2015. That morning, Beasley told Jackson to leave the apartment and offered him money for a Metro fare. Beasley heard Jackson leave, and then went back to sleep in the bedroom with Williams. Later that afternoon, Williams woke up and asked Beasley to escort her to the bathroom. From the hallway, Beasley saw Jackson sitting on the couch in the living room, shucking clams and eating ice cream.
According to the evidence at trial, Beasley approached Jackson and demanded multiple times that he leave the apartment. Jackson did not respond to Beasley and, instead, stared at the television with a blank face and glassy eyes. Beasley grew upset that Jackson was not responding to him. Although Beasley and Jackson had never physically fought during their forty-year friendship, Beasley struck Jackson hard on the head. During the ensuing scuffle, Jackson picked up the knife he was using to shuck clams and swung it at Beasley's face, striking him in the eye. By all accounts, the attack was out of character for Jackson, who is normally a peaceful person. The attack left Beasley blind in his left eye.
II. Relevant Trial Testimony
A. Evidence of Drug Use
Prior to trial, the government moved in limine to admit evidence that Jackson used phencyclidine (PCP) on November 12, 2015, the night before the fight. The government contended that such evidence provided important context that would serve to explain Jackson's odd behavior, why Beasley wanted Jackson to leave the apartment, and why Beasley was in the living room when he was assaulted. Jackson opposed the motion, arguing that any evidence of PCP use was evidence of other crimes or bad acts which was unfairly prejudicial.
Judge Marisa Demeo ruled the evidence admissible. Citing cases which we discuss below, the court found that such testimony would provide context and serve to explain the witnesses' observations, beliefs, and behaviors. The court determined that there was a close temporal relationship between Jackson's use of PCP on November 12 and the attack the following afternoon.
*804After reviewing the proffered evidence, the court found that its probative value was not substantially outweighed by the danger of unfair prejudice. The court denied the government's request to admit testimony of Jackson's use of PCP prior to November 12, 2015.
At trial, both Williams and Beasley testified that they smelled PCP in the apartment prior to the assault. Williams testified that at around 8:00 p.m. on November 12, while she was at the apartment, she saw Jackson smoke a cigarette which had a "plastic" odor, similar to embalming fluid. Over objection, Williams testified that the cigarette smelled like PCP, an odor she recognized from walking past individuals smoking it on the streets. Williams further testified that Jackson began acting strangely after smoking the cigarette. She described Jackson's movements as being "disconjointed" and "demonic." Williams stated that the apartment smelled like PCP the following morning as well. Beasley, who had used PCP in the past, corroborated this testimony, stating that he smelled the odor in the apartment on November 13. Williams testified that because she felt uncomfortable around PCP, and "wasn't sure what was going to happen," she asked Beasley to walk her to the bathroom. Despite the testimony from Williams and Beasley, Jackson denied using PCP on November 12 or 13, or smelling it in the apartment either day.
B. Jackson's Claim of Self-Defense
At trial, Jackson testified that he stabbed Beasley in self-defense. He said that Beasley struck him so hard on the head that he "lost a little ... understanding of what was going on." Jackson stated that as he lay "sideways" on the couch, he felt Beasley "pin[ ] down" his left arm. Then, Beasley put his entire body weight on Jackson. Jackson testified that while Beasley was pushing into him with his whole weight, he was struggling to get off the couch, but "couldn't breathe and ... was in fear for [his] life." To defend himself, Jackson grabbed the clam shucker which was "right at [his] reach" and swung it at Beasley. It is undisputed that Jackson called the police after the fight and stated that he stabbed Beasley in the eye, but did not mention anything about doing so in self-defense. A character witness testified that Jackson's "reputation was exceptional for being a very peaceful person."
C. The Jury's Verdicts
Ultimately, the jury acquitted Jackson of aggravated assault while armed but convicted him of the lesser-included offense of assault with a dangerous weapon. This timely appeal followed sentencing.
III. Analysis
As a general rule, evidence of another crime is "inadmissible to prove disposition to commit" the charged offense. Drew v. United States , 331 F.2d 85, 89 (D.C. Cir. 1964). However, we have long recognized a distinction between other crimes evidence that is used to suggest propensity and that which is relied upon to prove the charged offense. Johnson v. United States , 683 A.2d 1087, 1096-98 (D.C. 1996) (en banc). In Johnson , we discussed three instances in which such evidence falls into the latter category: "where [it]: (1) is direct and substantial proof of the charged crime, (2) is closely intertwined with the evidence of the charged crime, or (3) is necessary to place the charged crime in an understandable context." Id. at 1098.
In Toliver v. United States , 468 A.2d 958, 960-61 (D.C. 1983), we held that evidence of another crime "is admissible when relevant to explain the immediate circumstances surrounding the offense charged." Id. at 960. Such evidence serves *805to provide the jury with the necessary context of the charged offense and to "complete the story of [that] crime." Id. (citation omitted). Indeed, because of its nature, "this limited class of evidence is not other crimes evidence because it is too intimately entangled with the charged criminal conduct" and falls outside of Drew . Id. Regardless of whether such evidence is admissible under Drew or Toliver / Johnson , it is still subject to a balancing of its probative value and prejudicial effect. Johnson , 683 A.2d at 1098-99 (adopting Federal Rule of Evidence 403 ).
A. Admissibility of PCP Evidence
We agree with the government that testimony regarding Jackson's use of PCP qualifies as Toliver/Johnson evidence. It was not offered for the impermissible purpose of showing propensity to commit crime, see Drew , 331 F.2d at 89-90, but "to explain the immediate circumstances surrounding" the attack, see Toliver , 468 A.2d at 960. The evidence served to provide context for the crime itself, to explain Jackson's behavior, and to test his memories and perceptions of the fight. Id.
However, our inquiry into the admissibility of the PCP evidence does not end there. As with all evidence, this testimony is inadmissible if its probative value is substantially outweighed by the danger of unfair prejudice. See Johnson , 683 A.2d at 1098-99. A wide "variety of factors" are considered in this analysis, "including the need for the evidence, the efficacy of alternative proof, and the degree to which the evidence probably will rouse the jury to overmastering hostility." United States v. Morton , 50 A.3d 476, 482 (D.C. 2012) (internal quotation marks omitted). We recognize that this evaluation is "quintessentially a discretionary function of the trial court, and [is] owe[d] a great degree of deference." Malloy v. United States , 186 A.3d 802, 811 (D.C. 2018) (quoting Holmes v. United States , 143 A.3d 60, 63 (D.C. 2016) ). However, we cannot defer to the trial court's judgment in this instance. As appellant contends, and we discuss further below, without the testimony of an expert, the probative value of the evidence was substantially outweighed by its prejudicial effect.
As a threshold matter, government counsel asserted for the first time at oral argument that appellant failed to preserve this argument. Appellee therefore argues that we must review only for plain error. We disagree. At a pretrial hearing on the government's motion to admit the PCP evidence, appellant asserted multiple times that it was "speculative." Appellant noted the lack of expert testimony, stating that it "would be one thing" if the government planned on calling "an expert dealing with PCP." He further addressed the gap in time between when Williams saw Jackson smoking PCP and the attack, stating that while Jackson may have used PCP prior to the attack, "it doesn't mean that he was under the influence at the time of the incident." Although much of this discussion was focused on the foundation for Williams' testimony, it was sufficient in our view to preserve the issue for appeal. See Hunter v. United States , 606 A.2d 139, 144 (D.C. 1992) ("Objections must be made with reasonable specificity; the judge must be fairly apprised as to the question on which he is being asked to rule.") (emphasis added). Accordingly, we review for abuse of discretion. See Muschette v. United States , 936 A.2d 791, 797 (D.C. 2007).
B. The Need for an Expert
In assessing the need for an expert, "[t]he determinative question ... is whether the jurors are 'just as competent as the expert to consider and weigh the evidence and draw the necessary conclusions.'
*806" Minor v. United States , 57 A.3d 406, 415 (D.C. 2012) (emphasis omitted) (quoting Adams v. United States , 502 A.2d 1011, 1021-22 (D.C. 1986) ). The evaluation of juror competence may vary with the issue being addressed. For example, while we permit lay witnesses to offer opinions on whether an individual appeared to them to be under the influence of drugs, see Harris v. District of Columbia , 601 A.2d 21, 24-26 (D.C. 1991), and allow jurors to weigh the credibility of that testimony without the aid of an expert, assessing the effect of drug use upon testimonial capacity "seems to require information beyond that ordinarily attributable to the average juror," Durant v. United States , 551 A.2d 1318, 1328 n.12 (D.C. 1988) (quoting J. Weinstein & M. Berger, 3 Weinstein's Evidence ¶ 607[04], 607-61)).
A review of our cases reveals that expert testimony ordinarily is required when the key question is whether and how PCP continues to affect a particular individual hours after ingestion. See Robinson v. United States , 50 A.3d 508, 525 (D.C. 2012) (error to exclude expert testimony proffered by defense "about the general effects of PCP within 24 hours of use" because it would help explain to the jury whether "PCP [smoked several hours before the time in question] was still affecting [a witness]."); Kigozi v. United States , 55 A.3d 643, 651-53 (D.C. 2012) (stating that where the crucial issue was "whether [murder victim] was actively under the influence of PCP at the time of the shooting so as to undermine the reliability of his dying declaration," defense counsel was ineffective in deciding to rely on what the "average juror would know about PCP and its general effects ... without knowing what an expert would be able to contribute to the jury's understanding of the evidence") (internal quotations omitted); Durant , 551 A.2d at 1328 (holding that an expert was necessary to determine whether an individual who had an unspecified amount of PCP in his urine the day after he allegedly committed a crime was under the influence of the drug at the time of the crime); United States v. Roy, 114 Daily Wash. L. Rptr. 2481, 2491 (Burgess, J.) (Super. Ct. Dec. 1, 1986) (requiring expert testimony to determine how long evidence of PCP use can remain in the body). Relying on our precedent, we conclude that the analysis the jury was asked to conduct in this instance was not within "the ken of the average layperson." Blakeney v. United States , 653 A.2d 365, 369 (D.C. 1995) (brackets and citation omitted).
The lack of expert testimony regarding the effects of PCP and how long they may last impacts both probative value and prejudicial effect. In this case, without an expert to assist the jury, the probative value of the PCP evidence was low. An eighteen-hour gap existed between when Williams saw Jackson smoking a PCP-laced cigarette at 8:00 p.m. on November 12 and the attack, which occurred at 2:00 p.m. the following afternoon. This gap in time, coupled with the general uncertainty about the effects of PCP on a particular individual, leaves too much room for speculation. We see several questions attributable to the lack of expert testimony, including: the amount of time it takes a PCP high to dissipate; whether repeated use of PCP can create a higher level of tolerance, thus reducing the drug's effect on a particular individual; and whether Jackson was under the influence of PCP at the time of the attack. See Coates v. United States , 558 A.2d 1148, 1154 (D.C. 1989) (recognizing that the effect of narcotics depends on a variety of factors). Because of these unresolved questions we conclude that "the government provided too slim a reed to support a conclusion that [appellant] was under the influence of PCP at the time" of *807the attack. Durant , 551 A.2d at 1328. Expert testimony was therefore required to "impart[ ] probative value" to the evidence of Jackson's PCP use. Id.
C. The Government's Arguments Regarding the Evidence's Probative Value
The government attempts to persuade us that the PCP evidence was "highly probative" for two reasons: (1) to explain why Williams wanted Beasley to escort her to the bathroom and (2) to explain the "shocking" attack and Jackson's odd behavior the night before. We are not convinced.
First, we do not see why there was a need to explain, with or without the aid of an expert, why Beasley was present in his own living room. Moreover, while Williams testified that she felt "uncomfortable" because of the PCP use the night before, she did not express those concerns to Beasley. Beasley believed that Williams asked for an escort because "she was kind of inebriated." Because Williams' uneasiness around PCP had no bearing on the interaction between Beasley and Jackson, it was not probative with respect to Jackson's claim of self-defense.
The government's other reason - explaining appellant's shocking attack - is undercut by the lack of proof to support the desired conclusion. It may well be that many violent attacks are committed because the assailant is under the influence of PCP. However, as discussed, the evidence that Jackson was acting under the influence of PCP at the time of the attack was largely speculative. Williams' testimony that she saw Jackson smoking PCP eighteen hours beforehand is the only concrete evidence of his PCP use. Although there was testimony describing a PCP odor the following day, it is not clear whether that smell lingered from the night before or provided circumstantial evidence of more recent use by Jackson.1 Importantly, while the government characterizes the fight as a "shocking" attack, the initial aggressor was Beasley, not Jackson.
D. The PCP Evidence's Prejudicial Effect
Turning to the prejudicial effect, we further conclude that without expert testimony, the PCP evidence was prejudicial. Evidence of an individual's drug use may inherently have a prejudicial effect. See, e.g., Rogers v. United States , 419 A.2d 977, 981 (D.C. 1980) (considering evidence that a witness is a drug user to be "inflammatory"). In many instances, this effect is minimal, and would not be so substantial as to preclude a trial court from admitting such evidence. Here, however, this prejudicial effect was compounded by the government's failure to present expert testimony describing how an individual may act while on PCP. The jurors apparently were forced to rely on their own (perhaps quite varied) impressions of the drug. See Harris , 601 A.2d at 24 (recognizing that "drugs other than alcohol can produce a confusing array of symptoms") (quoting State v. Rifkin , 140 Vt. 472, 476, 438 A.2d 1122, 1124 (1981) ). Those impressions may be colored by stories in the news describing individuals on PCP as violent and impulsive. But see Kigozi , 55 A.3d at 653 (holding that, even assuming a jury would know about the "general effects of PCP," it would not be able to determine how a particular individual's actions *808could have been affected by the drug). Furthermore, had an expert testified, Jackson would have been able to cross-examine her and perhaps could have presented evidence to rebut her conclusions. In the circumstances of this case, Jackson could not know what the jurors' assumptions may have been, and did not have an opportunity to address them appropriately.
For the reasons discussed above, the trial court exercised its discretion erroneously in admitting evidence of Jackson's PCP use. But this error does not necessarily establish an abuse of discretion. We must also consider "whether the impact of that error requires reversal." Johnson v. United States , 398 A.2d 354, 367 (D.C. 1979). "It is when both these inquiries are answered in the affirmative that we hold that the trial court 'abused' its discretion." Id.
E. Harmless Error Analysis
In order to affirm, we must be satisfied "with fair assurance, after pondering all that happened without stripping the erroneous action from the whole, that the judgment was not substantially swayed by the error." Kotteakos v. United States , 328 U.S. 750, 765, 66 S.Ct. 1239, 90 L.Ed. 1557 (1946). An error is harmless if we "find it highly probable that that error did not contribute to the verdict." Clark v. United States , 593 A.2d 186, 192 (D.C. 1991) (internal quotations omitted). This decision cannot be based on a "mere hunch that the case would have ended with the same verdict." Id. Rather, "we must look to the closeness of the case, the centrality of the issue affected by the error, and any steps taken to mitigate the effects of the error." Id. at 193. After analyzing all relevant facts and inferences, we cannot say with fair assurance that admitting the PCP evidence without testimony from an expert did not substantially sway the jury's verdict.
We agree with the government that there were valid reasons why the jury may have disbelieved Jackson's version of events. To start, although Jackson described a fight in which he "fear[ed] for [his] life," there was little evidence of a struggle. Neither Jackson nor the apartment appeared to be "disheveled" in any way, and Williams did not hear "any kind of tussling" while she was in the bathroom. Furthermore, Jackson's testimony at trial differed from his initial statements to emergency personnel and the police. Perhaps even more detrimental to his defense, Jackson's attestations in his application for a civil protection order conflicted with both the version of events he told police and what he testified to at trial.
Despite these inconsistencies, we cannot say that the evidence against Jackson was overwhelming. It was the government's burden to disprove self-defense beyond a reasonable doubt. Williams v. United States , 90 A.3d 1124, 1128 (D.C. 2014). By all accounts, Beasley was the initial aggressor, and he struck Jackson without warning. It is also undisputed that Beasley weighed almost 500 pounds at the time of the attack - nearly three times more than Jackson weighed. That information alone made a claim of self-defense plausible.
It may be reasonable for a man attacked by someone three times his size to believe it "necessary to protect himself from imminent bodily harm." But when a jury is led to believe that Jackson was high on PCP at the time of the attack, this self-defense theory is undermined. It is likely that the jury viewed the stabbing as the product of a drug-fueled rage rather than Jackson's reasonable effort to protect himself. Such a conclusion would be entirely valid if founded *809on evidence, but not when it is based on speculation.
Furthermore, "[t]he issue to which the error related was an important one." Clark , 593 A.2d at 193. In making such determinations, we often look at the emphasis a prosecutor places on the evidence in question. See Allen v. United States , 837 A.2d 917, 923 (D.C. 2003) (quoting United States v. DeLoach , 504 F.2d 185, 192 (D.C. Cir. 1974) ). Specifically, we have noted that "[a] prosecutor's stress upon the centrality of particular evidence in closing argument tells a good deal about whether the admission of the evidence was meant to be, and was, prejudicial." Morten v. United States , 856 A.2d 595, 602 (D.C. 2004) (internal quotations omitted).
Throughout closing arguments, the government relied on the PCP evidence to explain Jackson's behavior. The government's theory was that "in an explosive moment of anger, the defendant, who had been acting erratically for about more [sic] than twelve hours from November 12 into November 13," stabbed Beasley. Multiple times, the government described Jackson as a "peaceful, regular guy" until he smoked PCP and his "demeanor actually changed." According to the government, after smoking the PCP, Jackson "snapped, and in an explosive moment of anger, stabbed [Beasley] in the eye."
The government tied the fact that Jackson smoked PCP on the night of November 12 to the attack the following afternoon, stating that "what happens on November 12th ... explains and ... puts into context the sequence of events that happened on November 13th. So, you can't evaluate November 13th without November 12th." The prosecutor commented several times that Jackson was acting strangely for at least twelve hours after smoking the PCP.2 But the testimony at trial does not corroborate this assertion. Beasley and Williams were not with Jackson for the entire eighteen-hour period between when he smoked the PCP and the fight the following afternoon. Accordingly, the witnesses did not testify-nor could they-that Jackson was acting strangely for that entire period of time.
Given the government's reliance on Jackson's use of PCP and, as we discussed above, the likelihood that such evidence would undermine Jackson's self-defense claim, the erroneously admitted testimony was directly related to the key issue before the jury. Finally, we note that the trial judge took no steps to mitigate the effects of this error. See Clark , 593 A.2d at 193.
The government argues that any error was harmless because "the evidence that appellant used PCP was almost certain to come before the jury anyway" to "assess [Jackson's] perception of the events and his ability to remember them." But, had the evidence been admitted on cross-examination of Jackson, it would have been admitted for the purpose of impeaching his ability to accurately perceive and recall events. It would not have been admitted, as it was here, to explain his behavior.
Moreover, using the evidence to test Jackson's credibility may not have avoided the need for an expert. In order for evidence of an individual's drug use to be admissible for impeachment purposes, a sufficient foundation must be laid to show that the individual "in fact was under the influence of drugs at the relevant time." Durant , 551 A.2d at 1326. See also Rogers , 419 A.2d at 981. It is not obvious that the *810use of PCP eighteen hours "prior to the events in question" was "probative of impaired perception and memory" and could have been properly admitted on cross examination without illumination from an expert. Durant, 551 A.2d at 1326. Likewise, this evidence would be equally inadmissible to impeach the account of Jackson's reputation by the character witness. The PCP testimony can only rebut Jackson's claim to have a peaceful nature if it can be established that the attack on November 13 was a result of using PCP. As we have reiterated throughout this opinion, without relying on expert testimony, such a conclusion is too speculative. Accordingly, we conclude that the trial court's error was not harmless.
IV. Conclusion
For the reasons discussed above, we reverse appellant Jackson's conviction for assault with a dangerous weapon and remand this case to the trial court for further proceedings consistent with this opinion.
It is so ordered .