Opinion concurring in part and concurring in the judgment filed by Senior Circuit Judge SENTELLE.
MILLETT, Circuit Judge:A jury convicted Dante Sheffield of unlawful possession of 100 grams or more of phencyclidine (PCP) with intent to distribute, in violation of 21 U.S.C. § 841(a) and (b)(l)(B)(iv). Based in part on its application of a career-offender enhancement, the *300district court sentenced Sheffield to 230 months in prison.
Sheffield challenges both his conviction and sentence, arguing that the district court erred in (i) denying his motion to suppress the PCP discovered during the search of a car in which he was a passenger, (ii) refusing to suppress statements he made following his arrest but before he received his Miranda warnings, (iii) admitting evidence of a decade-old drug conviction, (iv) denying his post-trial motion for independent testing of the drug evidence, and (v) applying the career-offender enhancement at sentencing. We affirm the judgment of conviction, but we reverse the district court’s imposition of the career-offender enhancement, vacate Sheffield’s sentence, and remand to the district court for resentencing.
I
A
On the evening of June 8, 2011, Metropolitan Police Department Detectives Christopher Smith and Michael Iannac-chione, along with two other officers, were driving in an unmarked police car through the 2300 block of 11th Street, N.W., in Washington, D.C. Detective Smith spotted Dante Sheffield, whom he and the other officers recognized from an earlier PCP investigation in the area. The officers then observed Sheffield and an unknown male enter a ear with tinted windows. One of the officers told Detective Smith that “he wanted to at least make a contact just to see who the [unjidentified * * * male was.” Supp. App. 7. Before they did so, however, they witnessed the car “pull[ ] * * * slightly forward and ma[ke] a sharp left without using a turn signal into' an alleyway[.]” Id at 8. The officers followed closely behind the car, which then made a right turn out of the alley “without using its signal.” Id. At that point, the officers initiated a traffic stop.
All four officers approached the vehicle, two on each side of the car. After asking the occupants to roll down the windows because of the tinting, Detective Smith observed a woman, Brande Dudley, in the driver’s seat, Sheffield in the passenger seat, and the unknown male, Anthony Grant, in the rear seat. Detective Smith later testified that he had detected a “faint” but “fresh” smell of marijuana on the passenger side. Supp. App. 10. In addition, Smith noticed “numerous air fresheners all [over] the vehicle,” “on the top, the bottom, the back, the front, all over the car.” Id. After asking for Dudley’s license, the officers asked the three occupants to get out of the vehicle.
Officers then searched the inside of the vehicle. Upon unlocking and opening the armrest console, Detective Iannacchione was immediately met with “a strong chemical odor” and found an eight-ounce lemon juice bottle, “which through [their] investigation was consistent with that of storing and packaging of PCP in large quantities.” Supp. App. 11-12. Detective Iannacchione opened the cap and noticed “a strong chemical [odor] consistent with that of PCP.” Id. at 12. At that point, the officers placed all three individuals under arrest.
Sheffield then asked Detective Smith “[w]hat are we getting arrested for?” Supp. App. 12. Smith responded that the arrest was for “[wjhat was in the car,” to which Sheffield responded “[e]verything is mine.” Id. After another detective began speaking privately with Brande Dudley, Sheffield “became more irritated and started yelling toward their direction for her not to say nothing, that they didn’t have a strong case, they got nothing on us, don’t say anything.” Hearing Tr. 15-16 (Sept. 16, 2011).
*301When the officers searched Grant incident to his arrest, they found a plastic bag in his right sock containing approximately 0.75 grams of marijuana.
B
1
The government indicted Sheffield on one count of unlawful possession with intent to distribute 100 grams or more of PCP, 21 U.S.C. § 841(a)(1) & (b)(l)(B)(iv). Before trial, Sheffield filed motions to suppress the physical evidence of the PCP and his statements made during his arrest.
The district court denied Sheffield’s motions to suppress. First, the court held that Dudley’s two turns made without signaling gave the officers probable cause to believe she had committed a traffic violation. The court rejected Sheffield’s argument that the traffic violations were mere pretext for a stop and search targeted at him because “the officers’ subjective motivations do not render unconstitutional a search that is otherwise justified by objective circumstances.” J.A. 70.
Second, the district court held that the search of the vehicle was lawful, citing inter alia “the smell of marijuana and the unusual number of air fresheners in the car[.]” J.A. 74. The court further held that the officers’ search of the locked armrest console was proper because there was a “ ‘fair probability’ that [the defendant] might have hidden additional drugs not necessary for his current consumption in areas out of plain sight, including the trunk of a ear” or an armrest console. Id. at 75 (quoting United States v. Turner, 119 F.3d 18, 20 (D.C. Cir. 1997)).
Third, with respect to Sheffield’s statements that everything in the car “is mine” and that “they [don’t] have a strong case, they’ve got nothing on us,” the district court ruled that Miranda warnings were not required for their admission. The court explained that the statements “were not made in response to a question posed by the officers, nor did the officers take any action to which defendant Sheffield’s response was required or expected.” J.A. 77.
The district court separately granted the government’s motion to admit a record documenting Sheffield’s conviction in 2002 for possession with intent to distribute PCP, pursuant to Federal Rule of Evidence 404(b).
2
At trial, Detectives Smith, Iannacchione, and a third officer involved in the stop and arrest all testified to the circumstances of the traffic stop, the discovery of the PCP, and Sheffield’s statements at the time of his arrest. In addition, the jury heard testimony from the law enforcement officials who transported, stored, and tested the PCP found in the car, including “two arresting officers, who observed and seized the lemon juice bottle at the scene of the traffic stop; three other officers who established the chain of custody of the lemon juice bottle, and detailed their handling, storage, and documentation of the evidence; and a DEA forensic chemist, who testified as to his testing of the PCP evidence.” J.A. 132. The jury learned that photographs were taken of the lemon juice bottle and that the quantity and weight of the liquid was measured. Id. In addition, the jury heard that, because a DEA regulation prohibits the agency from accepting more than 28.35 grams of PCP for testing, Officer Joseph Abdalla “separated approximately one ounce of the drug evidence into a vial and gave it to [the DEA chemist]” for testing. Id. at 133. The remaining PCP was never tested; though it was introduced at trial. The jury subsequently found Sheffield guilty of possessing with intent to distribute 100 or more grams of PCP.
*302Following trial, Sheffield filed a “motion to test drugs not submitted to DEA.” J.A. 121. He argued that he wanted the test “to ensure that the drugs introduced at trial are the same drugs seized on June 8, 2011, and to ensure that the measurements performed by the Metropolitan Police are accurate.” Id. at 123.
Treating it as a motion for a new trial in light of newly discovered evidence under Federal Rule of Criminal Procedure 33, the district court denied the motion. The court explained that testing the remaining liquid would not lead to new evidence that could produce an acquittal because (i) five officers “testified as to the smell, seizure, documentation, field testing, and storage of the drug evidence,” and Sheffield never argued that such testing was erroneous or that the witnesses were not credible, J.A. 136-137; (ii) measurement of the PCP six months after Sheffield’s arrest would “provide[] minimal probative value” because PCP evaporates over time, id. at 137; and (iii) testing the remaining PCP would not address whether the drugs presented at Sheffield’s trial were the same drugs seized at his arrest, id. at 138.
Sheffield was sentenced to 230 months of imprisonment and 96 months of supervised release. In calculating that sentence, the district court applied a career-offender enhancement under the United States Sentencing Guidelines.
II
Sheffield first challenges the admission of the PCP discovered during the search of the car. We decide de novo whether the police had probable cause both to stop the car and to search it. See, e.g., United States v. Burroughs, 810 F.3d 833, 839 (D.C. Cir. 2016). However, we review the district court’s fact findings for clear error, giving “due weight to inferences drawn from those facts and to the court’s determinations of witness credibility.” United States v. Brown, 334 F.3d 1161, 1164 (D.C. Cir. 2003) (quotation marks omitted).
A
Given the district court’s factual findings, we hold that the officers had probable cause to stop the car in which Sheffield was riding. “As a general matter, the decision to stop an automobile is reasonable where the police have probable cause to believe that a traffic violation has occurred.” Whren v. United States, 517 U.S. 806, 810, 116 S.Ct. 1769, 135 L.Ed.2d 89 (1996). Even minor violations of traffic law may justify a stop. See, e.g., id. at 808, 116 S.Ct. 1769 (turning without signaling and speeding); United States v. Williams, 773 F.3d 98, 103 (D.C. Cir. 2014) (failure to wear a seatbelt).
Here, the officers had probable cause to believe that the driver of the vehicle had violated a mandatory traffic regulation because the driver twice failed to signal a turn, in contravention of. D.C. Municipal Regulations, Title 18 § 2204.3. Section 2204.3 provides that “[n]o person shall turn any vehicle * * * from a direct course or move right or left upon a roadway without giving an appropriate signal * * * if any other traffic may be affected by the movement.” In this case, the district court found that Dudley, the driver of the car in which Sheffield was riding, first “turn[ed] sharply into an alley without using a turn signal,” and then after leaving the alley “ma[de] another right turn without a turn signal” when another car was behind her. J.A. 63, 69. Under Whren, those violations of D.C. traffic law provided probable cause to stop the vehicle, see 517 U.S. at 808-809, 819, 116 S.Ct. 1769.
Sheffield does not deny that those traffic infractions occurred, but argues that their *303use by the police was mere pretext for the officers’ true motivation, which was to stop him. That no traffic citation was ever issued, Sheffield argues, is “clearly indicative of the intent to target Dante Sheffield via this ever so slight traffic violation.” Pet. Br. 29.
Binding precedent forecloses that argument. The test for probable cause is an objective one, focusing on whether the stop was reasonable. See Whren, 517 U.S. at 811-813, 116 S.Ct. 1769. Accordingly, “the constitutional reasonableness of traffic stops [does not] depend[] on the actual motivations of the individual officers involved,” even when those motivations are “admitted.” Id. at 813-814, 116 S.Ct. 1769.
That means that, contrary to Sheffield’s argument, “ulterior motives [cannot] invalidate police conduct that is justifiable on the basis of probable cause to believe that a violation of law has occurred.” Whren, 517 U.S. at 811, 116 S.Ct. 1769; see also United States v. Washington, 559 F.3d 573, 575 (D.C. Cir. 2009) (traffic stop for running a stop sign was reasonable notwithstanding evidence of “aggressive traffic patrols” that “use[d] routine traffic stops to try to detect and prevent drug and gun crimes”). Indeed, in United States v. Bookhardt, 277 F.3d 558 (D.C. Cir. 2002), we upheld a stop even though the officers did not have probable cause to stop the defendant for the reason they gave — that the defendant was driving with an expired license, id. at 564 — because the officers as an objective matter did have probable cause to stop the defendant for a reason they did not give — reckless driving, id. at 565-566.
Accordingly, the two undisputed signaling violations committed in front of the officers’ car provided an objectively reasonable basis for the officers to believe that the driver had violated the traffic laws, and that in and of itself provided probable cause to stop Dudley’s car.
B
We likewise hold that, given the district court’s factual findings, probable cause existed to search the car after the stop.
1
To begin with, the government argues that Sheffield lacks Fourth Amendment “standing” to challenge the search of the vehicle because he was just a passenger in a car that did not belong to him. But that argument is forfeited because the government failed to raise it in district court.
To be clear, Fourth Amendment “standing” is not really a “standing” inquiry at all. Ordinarily in federal cases, “standing” refers to the jurisdictional requirement that a plaintiff have a sufficient stake in the outcome of the case to be entitled to litigate it. See Susan B. Anthony List v. Driehaus, — U.S. -, 134 S.Ct. 2334, 2341, 189 L.Ed.2d 246 (2014). Standing, in that context, is a critical component of the Constitution’s case-or-controversy requirement, and as such may not be waived or forfeited. See United States v. Cotton, 535 U.S. 625, 630, 122 S.Ct. 1781, 152 L.Ed.2d 860 (2002) (“[D]efects in subject-matter jurisdiction require correction regardless of whether the error was raised in district court.”).
Fourth Amendment “standing,” by contrast, has nothing to do with jurisdiction. Fourth Amendment standing instead limits the assertion of Fourth Amendment rights to those who have an individualized expectation of privacy in the searched property: “[Suppression of the product of a Fourth Amendment violation can be successfully urged only by those whose rights were violated by the search itself, not by those who are aggrieved solely by the *304introduction of damaging evidence.” Alderman v. United States, 394 U.S. 165, 171—172, 89 S.Ct. 961, 22 L.Ed.2d 176 (1969). In other words, “Fourth Amendment jurisprudence * * * does not countenance the assertion of another’s right to be free'from unreasonable searches and seizures.” United States v. Caicedo-Llanos, 960 F.2d 158, 161-162 (D.C. Cir. 1992); see also Plumhoff v. Rickard, — U.S. -, 134 S.Ct. 2012, 2022, 188 L.Ed.2d 1056 (2014) (“Fourth Amendment rights are personal rights which may not be vicariously asserted.”) (quotation marks and alterations omitted).
So understood, Fourth Amendment standing is merely an aspect of the substantive merits of a Fourth Amendment claim, inquiring whether the party invoking the Amendment has a privacy interest that was invaded. As the Supreme Court has explained, it would serve no “useful analytical purpose to consider this principle a matter of standing, distinct from the merits of a defendant’s Fourth Amendment claim.” Rakas v. Illinois, 439 U.S. 128, 138-139, 99 S.Ct. 421, 58 L.Ed.2d 387 (1978); see id. at 140, 99 S.Ct. 421 (The “definition of [Fourth Amendment] rights is more properly placed within the purview of substantive Fourth Amendment law than within that of standing.”).
As a non-jurisdictional principle of substantive law, Fourth Amendment “standing” is subject to ordinary rules of waiver and forfeiture, as a number of circuits have held. See United States v. Golson, 743 F.3d 44, 55 n.9 (3d Cir. 2014); United States v. Moss, 963 F.2d 673, 676 (4th Cir. 1992); United States v. Price, 54 F.3d 342, 345-346 (7th Cir. 1995); United States v. Dewitt, 946 F.2d 1497, 1499 (10th Cir. 1991); see also United States v. Gonzales, 79 F.3d 413, 419 (5th Cir. 1996) (government forfeits Fourth Amendment standing argument where defendant provides facts in district court supporting an inference of standing and government did not raise the issue); United States v. Noble, 762 F.3d 509, 527-528 (6th Cir. 2014) (government may forfeit Fourth Amendment standing argument, but appeals court may review argument for plain error).
Two decades ago, this court indicated otherwise, allowing the government to raise a Fourth Amendment standing objection for the first time on appeal. In United States v. Caicedo-Llanos, this court held that “we are powerless to rule on Fourth Amendment rights which do not belong to the parties before us,” 960 F.2d at 162. Two other circuits also ruled that Fourth Amendment standing could be raised for the first time on appeal. See United States v. Bouffard, 917 F.2d 673, 677 (1st Cir. 1990); United States v. Smith, 621 F.2d 483, 489 n.3 (2d Cir. 1980).1
Since those decisions, the Supreme Court’s ruling in Minnesota v. Carter, 525 U.S. 83, 119 S.Ct. 469, 142 L.Ed.2d 373 (1998), intervened and “expressly rejected” treating Fourth Amendment standing like jurisdictional standing, making clear that the question of a defendant’s reasonable expectation of privacy must be analyzed like a substantive merits issue, id. at 87, 119 S.Ct. 469.
Carter thus confirms our jurisdiction to decide this case on the merits be*305cause Fourth Amendment standing is merely a merits inquiry. Beyond that, the dispute over the government’s right to raise Sheffield’s “standing” argument for the first time on appeal is of no practical consequence in this case. As Caicedo-Lla-nos recognized, defendants always bear the burden of establishing that the government violated a privacy interest that was protected by the Fourth Amendment, 960 F.2d at 162. Here, Sheffield has failed in that task because probable cause existed for the search.2
2
While the Fourth Amendment generally requires the police to obtain a warrant for a search, motor vehicles are different both because of their mobility and the government’s extensive regulation of their use. See Pennsylvania v. Labron, 518 U.S. 938, 940, 116 S.Ct. 2485, 135 L.Ed.2d 1031 (1996) (noting that an automobile’s “ready mobility” and its “pervasive regulation” justify an “automobile exception to the Fourth Amendment’s warrant requirement”) (quotation marks omitted). Accordingly, officers generally may search a car if it “is readily mobile and probable cause exists to believe it contains contraband.” United States v. Maynard, 615 F.3d 544, 567 (D.C. Cir. 2010) (quotation marks omitted). “If probable cause justifies the search of a lawfully stopped vehicle, it justifies the search of every part of the vehicle and its contents that may conceal the object of the search.” United States v. Ross, 456 U.S. 798, 825, 102 S.Ct. 2157, 72 L.Ed.2d 572 (1982).
In this case, the smell of marijuana, in conjunction with other evidence of drug use, provided probable cause to believe the vehicle contained drug contraband, which in turn supported a search of the car’s compartments. In United States v. Turner, 119 F.3d 18 (D.C. Cir. 1997), we held that “the smell of burnt marijuana emanating from the car, * * * pieces of torn cigar paper arrayed around [the defendant], and [a] ziplock bag of green weed material found on the floor behind [the] seat” justified a search “elsewhere in the car, including its trunk.” Id. at 20. Such evidence “established] a fair probability that [the defendant] might have hidden additional drugs not necessary for his current consumption in areas out of plain sight, including the trunk of the car.” Id. (quotation marks omitted).
In this case, the district court found that one officer smelled the “faint” scent of “fresh marijuana,” and saw an abnormally large number of air fresheners throughout the car, which in the officers’ experience was consistent with efforts to disguise narcotics odors. J.A. 73. Finding no clear error in those fact findings, we hold that the officers had probable cause. See United States v. Ortiz, 669 F.3d 439, 445 (4th Cir. 2012) (holding that “the presence of multiple air fresheners in [a] vehicle” was a relevant factor in establishing probable cause to search a vehicle). In addition, that evidence “established] a fair probability that [an occupant] might have hidden additional drugs not necessary for his current consumption in areas out of plain sight,” Turner, 119 F.3d at 20, justifying the search of the armrest console. The district court therefore properly admitted the PCP evidence.
Ill
Sheffield next challenges the admission into evidence of his statement to *306police officers at the time of his arrest that “[everything is mine” in the car and that “they [don’t] have a strong ease, they’ve got nothing on us.” He argues that those statements should be suppressed because he had not yet received his Miranda warnings, see Miranda v. Arizona, 384 U.S. 436, 86 S.Ct. 1602, 16 L.Ed.2d 694 (1966). Because the statements were not made during an interrogation or in response to any police questioning, however, the protections of Miranda do not apply, and the statements were lawfully admitted.
“Miranda warnings are required where a suspect in custody is subjected to interrogation.” United States v. Vinton, 594 F.3d 14, 26 (D.C. Cir. 2010) (quotation marks omitted). Sheffield’s statements, however, were not made in response to any police “interrogation” within the meaning of Miranda. “Interrogation” is “either express questioning or its functional equivalent.” Rhode Island v. Innis, 446 U.S. 291, 300-301, 100 S.Ct. 1682, 64 L.Ed.2d 297 (1980). The functional equivalent of express questioning is “any words or actions on the part of the police (other than those normally attendant to arrest and custody) that the police should know are reasonably likely to elicit an incriminating response from the suspect.” Id. The words or coercive pressure must be “above and beyond that inherent in custody itself.” Id. at 300, 100 S.Ct. 1682.
There is no dispute that Sheffield was not being questioned at the time of the statements. Rather, unrefuted testimony establishes that Sheffield himself started talking on his own, asking “[w]hat are we getting arrested for?” Supp. App. 12. Detective Smith responded not with a question, but with the matter-of-fact statement: “[w]hat was in the car.” Id. Without anything further from the officers, Sheffield then declared: “Everything is mine.” Id.
So too for Sheffield’s statement that “they [don’t] have a strong case.” He made that statement after another officer spoke to Brande Dudley in a separate conversation to which Sheffield was not even a party.
Sheffield argues that the officer’s statement to Ms. Dudley that she would be taken back to the station and that she might lose her vehicle was intended to elicit a response from Sheffield. Pet. Br. 34. But such a commonplace explanation given to a third party, which was not made in a manner designed for Sheffield to even overhear, cannot amount to an interrogation of Sheffield. In United States v. Morton, 391 F.3d 274 (D.C. Cir. 2004), after officers arrested the defendant and while transporting her to the police station, the defendant “expressed concern over what would happen to her vehicle,” id. at 275. Officers said “her vehicle would be impounded,” that “she had been arrested for a serious charge,” and that “she might not be getting out as quickly as she thinks.” Id. In response, the defendant said “her lawyer would help her beat the charge, and when she did get out, she would be back down in the same area riding around with another gun that she kept at her home.” Id. at 276 (quotation marks omitted). We held that the officers’ statements — about impounding the vehicle, her arrest on a serious charge, and that she “might not be released as quickly as she thought — were directly responsive to what Morton had said and were not reasonably likely to elicit an incriminating response.” Id.
That same answer applies here, especially where one of the officer’s statements was not even made to Sheffield. Instead, in both instances, it was Sheffield who “initiated the conversation * * * and concede[d] that [the officer] did not ask h[im] any questions.” Morton, 391 F.3d at 276.
*307For those reasons, the statements were not the product of a custodial interrogation; no Miranda warnings were required; and the statements were properly admitted at trial.
IV
Sheffield argues that the district court improperly allowed the government to admit “other crimes” evidence under Federal Rule of Evidence 404(b). Specifically, he objects to the jury being told that, “[i]n February 2002, Defendant Dante Sheffield was convicted of a possession with intent to distribute Phencyclidine, PCP in the District of Columbia Superior Court.” Supp. App. 52. We agree that the district court erred in admitting that evidence, but the error was harmless.
We review the district court’s decision to admit evidence under Rules 403 and 404(b) for an abuse of discretion. See Henderson v. George Washington University, 449 F.3d 127, 132-133 (D.C. Cir. 2006); United States v. Cassell, 292 F.3d 788, 792 (D.C. Cir. 2002).
“Convictions are supposed to rest on evidence relevant to the crime charged, not on evidence of other, unrelated bad acts suggesting nothing more than a tendency or propensity to engage in criminality.” United States v. McGill, 815 F.3d 846, 878 (D.C. Cir. 2016). Consequently, Rule 404(b) prohibits the admission of “Evidence of a crime, wrong, or other act * * * to prove a person’s character in order to show that on a particular occasion the person acted in accordance with' the character.” Fed. R. Evid. 404(b)(1). However, such evidence may “be admissible for another purpose, such as proving motive, opportunity, intent, preparation, plan, knowledge, identity, absence of mistake, or lack of accident.” Fed. R. Evid. 404(b)(2).
In addition to Rule 404(b)’s specific limitations on the admission of evidence of prior bad acts, “Federal Rule of Evidence 403 permits a court to exclude otherwise-relevant evidence ‘if its probative value is substantially outweighed by a danger of unfair prejudice, confusing the issues, misleading the jury, undue delay, wasting time, or needlessly presenting cumulative evidence.’ ” McGill, 815 F.3d at 880 (quoting Fed. R. Evid. 403).
When it comes to evidence of prior drug dealings, we have recognized that “[a] defendant’s hands-on experience in the drug trade cannot alone prove that he possessed drugs on any given occasion.” United States v. Crowder, 141 F.3d 1202, 1208 n.5 (D.C. Cir. 1998). But “it can show that he knew how to get drugs, what they looked like, where to sell them, and so forth.” Id. Said another way, “Evidence of a defendant’s experience in dealing drugs * * * may be a ‘brick’ in the ‘wall’ of evidence needed to prove possession.” Id. Thus, the type of evidence the government introduced here — that of Sheffield’s prior PCP dealing — would generally be permissible to show that Sheffield had the requisite knowledge and intent to possess and distribute the PCP the officers found in the ■ armrest console.
But even general rules have their limits, and “evidence of a prior conviction is subject to analysis under Rule 403 for ' relative probative value and for prejudicial risk of misuse as propensity evidence,” Old Chief v. United States, 519 U.S. 172, 182, 117 S.Ct. 644, 136 L.Ed.2d 574 (1997). In •this case, the Rule 404(b) evidence was a conviction that occurred a decade before the offense conduct at issue in this case. Moreover, by telling the jury only about the fact of a decade-old conviction, the evidence bore little evidentiary relevance to the question of Sheffield’s knowledge in 2011 about.such matters as how to acquire *308or to market PCP. Cf. Crowder, 141 F.3d at 1208 n.5. Under those circumstances, the staleness of this conviction reduces its probative value such that it is “substantially outweighed” by the danger of unfair prejudice already inherent in the admission of prior-bad-act evidence. See United States v. Bigesby, 685 F.3d 1060, 1065 (D.C. Cir. 2012) (explaining that the “gap between [a] conviction and the [offense conduct] limited the conviction’s probative value”); United States v. Pettiford, 517 F.3d 584, 590 (D.C. Cir. 2008) (noting the risk of unfair prejudice inherent in “the admission of prior possession-with-intent-to-distribute evidence”).3
“Wrongly admitted evidence, however, does not always compel reversal.” McGill, 815 F.3d at 886. The error is harmless if we “can say that the error did not affect the jury’s verdict.” United States v. Watson, 171 F.3d 695, 700 (D.C. Cir. 1999). While Sheffield has failed to make any argument regarding harmlessness, the burden here is on the government to show that its error “should not upset the trial court’s determination.” Shinseki v. Sanders, 556 U.S. 396, 410, 129 S.Ct. 1696, 173 L.Ed.2d 532 (2009); see also United States v. Olano, 507 U.S. 725, 734, 113 S.Ct. 1770, 123 L.Ed.2d 508 (1993) (in a “harmless error” inquiry, “the Government * * * bears the burden of persuasion with respect to prejudice”).
Admission of the stale conviction was harmless for three reasons. First, the bad-act evidence “was neither so dramatic nor compelling as to rivet the jury’s attention on [Sheffield’s] bad character[.]” United States v. Brown, 597 F.3d 399, 405 (D.C. Cir. 2010). The evidence instead was a single-sentence paper admission, without any of details of the prior crime being aired before the jury.
Second, “the district court took caution to guard the space between the permissible and impermissible inferences by instructing the jury to consider the evidence only for its proper purpose.” United States v. Mitchell, 49 F.3d 769, 777 (D.C. Cir. 1995). Here, the district court instructed the jury both after the earlier drug conviction was admitted at trial and during the final jury instructions that:
If you find that Mr. Sheffield was previously convicted of possession with intent to distribute PCP, you may use this evidence only for the limited purpose of determining whether the Government has proved beyond a reasonable doubt that Mr. Sheffield intended to possess [the] PCP found in this case knowingly and on purpose and not by mistake or accident.
You may not use this evidence for any other purpose. Mr. Sheffield is only on trial for the crimes charged. You may not use this evidence to conclude that Mr. Sheffield has a bad character or that he has a criminal personality. The law does not allow you to convict a defendant simply because you believe he may have done bad things not specifically charged as crimes in this case.
Supp. App. 74-75. And during final jury instructions, the district court reiterated the purpose of the Rule 404(b) evidence:
*309You have heard evidence by way of a stipulation that Mr. Sheffield was previously convicted of possession with intent to distribute PCP. * * * You may use this evidence only for the limited purpose of determining whether the Government has proved beyond a reasonable doubt that Dante Sheffield intended to possess the PCP found in his case knowingly and on purpose and not by mistake or accident.
You may not use this evidence for any other purpose, Dante Sheffield is only on trial for the crime charged.
You may not use this evidence to conclude that the defendant has a bad character or that he has a criminal personality. The law does not allow you to convict a defendant simply because you believe he may have done bad things not specifically charged as crimes in this case.
Id. at 124. Absent evidence to the contrary, the jury is presumed to have followed that instruction, and Sheffield offers no such contradictory evidence. Brown, 597 F.3d at 406.
Third and finally, “[t]he most significant factor that negates the error’s impact is the weight and nature of the evidence against [the defendant].” United States v. Williams, 212 F.3d 1305, 1311 (D.C. Cir. 2000). Here, the one-sentence stipulation about an earlier drug crime “formed a small part of what was otherwise an overwhelming case against” Sheffield, McGill, 815 F.3d at 886, and that amply evidenced his knowledge and intent. The jury heard (i) the officers’ testimony that Sheffield was a passenger in a car that contained PCP, (ii) testimony about Sheffield’s statement that “[everything is mine” in the car, (iii) Dudley’s testimony that Sheffield had earlier taken her car for 30-45 minutes and that there was no PCP in her ear before then, and (iv) an extensive recounting of the transportation and testing of the PCP, see infra Part V. In addition, the jury heard legitimate Rule 404(b) evidence that Sheffield dealt PCP in 2009, the admission of which Sheffield does not challenge. That Rule 404(b) evidence bore on his motive, intent, and knowledge to deal in PCP. Taken together, all of that properly admitted evidence rendered harmless the mistaken admission of the 2002 conviction.
Y
Turning to the post-trial stage of the proceedings, Sheffield challenges the district court’s denial of his motion for independent testing of the portion of the seized PCP that was not submitted to the DEA for analysis. The district court treated that as a motion for a new trial because the motion sought to challenge the jury finding that Sheffield was guilty of “unlawfully, knowingly and intentionally possessing] with intent to distribute” PCP in “the amount of * * * 100 grams or more,” J.A. 12,120.
We review a district court’s denial of a motion for new trial for an abuse of discretion. See United States v. Johnson, 519 F.3d 478, 487 (D.C. Cir. 2008). A district court may grant a new trial on the ground of newly discovered evidence if (i) the evidence was discovered after the trial; (ii) the movant was diligent in attempting to procure the evidence; (iii) the evidence is material and not merely cumulative or impeaching; and (iv) if admitted, the evidence would probably produce an acquittal. See United States v. Pettiford, 517 F.3d 584, 591 (D.C. Cir. 2008).
The district court reasonably concluded that testing the remaining liquid in the lemon juice bottle was unlikely to produce an acquittal. The jury heard ample, detailed, and unrebutted testimony describing the transportation and testing *310procedures that were used for the PCP evidence admitted at trial, and thus jurors had a full opportunity to evaluate the reliability of the determination that the bottle contained PCP and the amount. For example, the jury heard that, after the search of the car, one detective took photographs of the bottle of PCP, and another detective then “physically picked up the lemon juice bottle and placed it in a bag and transported it back to the narcotic and special investigations division for processing.” Supp. App. 41. At the police station, that detective gave the lemon juice bottle to Officer Joseph Abdalla, “the inhouse officer that processes liquid PCP when large amounts are seized.” Id. at 42.
The jury also heard Officer Abdalla testify about how he tested and processed the PCP. He first photographed the lemon juice bottle and then conducted a field test by sticking a tester directly into the lemon juice bottle’s liquid, which came back positive for PCP. Trial Tr. 201-202 (Nov. 29, 2011). Officer Abdalla then weighed -the PCP using two different measurement techniques. Id. at 211-212. He determined that the lemon juice bottle contained approximately 195.9-198.7 grams of liquid. Supp. App. 96. Officer Abdalla next undertook “the remediation process,” in which he took “a bulk amount of the liquid * * * [and] remov[ed] one ounce from that bulk substance, plac[ing] it into a small glass vial and then * * * submitted [it] to the Drug Enforcement Administration chemist for analysis.” Id. at 85-86. Officer Abdalla explained that extracting out a single-ounce vial was necessary because the DEA lab would not “accept any more than a one ounce sample from any seizure of PCP.” Id. at 87.
After that, the jury heard testimony about the DEA’s testing process. Officer Abdalla gave the one-ounce vial to a detective, who placed it in a heat-sealed package for submission to the DEA lab. The detective placed the vial “into a secure evidence property box * * * where it [was] then * * * transported to [the] DEA lab for analysis.” Supp. App. 48. At the DEA, Richard Isaacs, a forensic chemist, tested the liquid in the vial and determined that the vial contained 25.9 grams of liquid that was 16.9% PCP. Isaacs also testified that, when he received the vial, the heat seal was intact.
Finally, at the beginning of trial, the remaining PCP was made available in the courthouse for Sheffield and his attorney to inspect. The lemon juice bottle was ultimately admitted into evidence, along with five photographs of the bottle, its contents, and its weight, all of which were taken the night of Sheffield’s arrest.
Notably, Sheffield did not challenge any of that testimony at trial. He never argued, for example, that the testing process was flawed, that those witnesses were not credible, or that the chain of custody was interrupted.
The jury, in sum, considered the unre-butted “testimony of five Metropolitan Police Department officers who testified as to the smell, seizure, documentation, field testing, and storage of the drug evidence,” viewed “photographs of the lemon juice bottle and its contents that were taken shortly after the evidence was seized” that indicated “the weight of the liquid, its quantity, and the amount separated for DEA testing,” and heard “the testimony of the DEA forensic chemist who tested the liquid sample and identified the liquid as PCP.” J.A. 136-137. Given that extensive trial record addressing the handling and testing of the liquid found in the lemon juice bottle and the scientific determinations that it was PCP and of its amount, the district court did not abuse its discretion in concluding that further testing of the remaining seven ounces of PCP had no *311realistic prospect of producing an acquittal.
Other factors governing motions for a new trial reinforce the district court’s judgment. To begin with, the evidence Sheffield’s motion sought could have been obtained before the district court issued its judgment. Specifically, Sheffield could have sought independent testing of the PCP before trial, but he did not. Resp. Br. 48. Furthermore, to the extent that Sheffield seeks to double check the measurements the various officers performed, those goals would not be served by testing the remaining seven ounces of liquid now because it is undisputed that “some of the remaining seven ounces of PCP [would have] evaporated over time despite being sealed in a plastic bottle and a heat-sealed evidence bag.” J.A. 129.
Finally, Sheffield suggests that, under California v. Trombetta, 467 U.S. 479, 104 S.Ct. 2528, 81 L.Ed.2d 413 (1984), the Due Process Clause requires that he be allowed to test the remaining PCP to maintain the “fundamental fairness” of the proceedings. Sheffield failed to make any due-process argument below, so we review for plain error. See United States v. Pryce, 938 F.2d 1343, 1350 (D.C. Cir. 1991). And Sheffield has offered no authority- — -let alone plain authority — for the proposition that a defendant has the right under the Due Process Clause to test drugs after failing to challenge the chain of custody or testing procedures during trial, or to avail himself of a pretrial testing opportunity.
For all of those reasons, the district court did not abuse its discretion when it denied Sheffield’s motion to test the PCP evidence.
VI
Lastly, Sheffield challenges the district court’s imposition of a sentence enhancement for being a career offender. The district court invoked a Sentencing Guidelines provision to increase Sheffield’s sentence based on a determination that he “ha[d] at least two prior felony convictions of either a crime of violence or a controlled substance offense.” U.S.S.G. § 4Bl.l(a). Specifically, the government introduced evidence that Sheffield had previously been convicted of possession with intent to distribute PCP and attempted robbery in the District of Columbia. Applying the career-offender enhancement increased Sheffield’s Guidelines offense level from 26 to 37, which in turn moved his original Guidelines range of 92 to 115 months up to 360 months to life in prison. The district court ultimately sentenced Sheffield to 230 months in prison.
Sheffield argues that the career-offender enhancement was erroneously applied because his attempted robbery conviction does not qualify as a crime of violence under Sentencing Guideline 4Bl.l(a). Although we typically review de novo a district court’s determination that a defendant’s “conviction qualified as a crime of violence,” In re Sealed Case, 548 F.3d 1085, 1090 (D.C. Cir. 2008), Sheffield did not raise this objection in the trial court. To prevail, then, Sheffield must demonstrate plain error by showing that (i) an error occurred; (ii) it was a plain, clear, or obvious error; (iii) the error affected Sheffield’s substantial rights; and (iv) the error also seriously affected the fairness, integrity, or public reputation of judicial proceedings. See Puckett v. United States, 556 U.S. 129, 135, 129 S.Ct. 1423, 173 L.Ed.2d 266 (2009). As the government now concedes, see Gov’t 7/5/16 Letter at 1, the district court’s application of the career-offender enhancement was plain error, and Sheffield is entitled to a new sentencing free from any such enhancement.
*312To be considered a career offender, Sheffield’s prior conviction for attempted robbery in the District of Columbia had to qualify as a “crime of violence” under the Sentencing Guidelines. The Guidelines define “crime of violence” as a state or federal offense that is punishable by imprisonment for a term exceeding one year, and that: “(1) has as an element the use, attempted use, or threatened use of physical force against the person of another, or (2) is burglary of a dwelling, arson, or extortion, involves use of explosives, or otherwise involves conduct that presents a serious potential risk of physical injury to another.” U.S.S.G. § 4B1.2(a). The first clause is referred to as the “elements clause.” The “otherwise involves” portion of the second clause is known as the “residual clause.” See Welch v. United States, — U.S. -, 136 S.Ct. 1257, 1261, 194 L.Ed.2d 387 (2016).
Unfortunately, the district court never specified which clause of the “crime of violence” definition it believed applied to Sheffield’s attempted robbery conviction. Nor did the Presentence Report or the government’s sentencing memorandum. See Gov’t Sentencing Mem. ¶ 7. At the sentencing hearing, the district court stated only that “your 2007 conviction for attempted robbery * * * qualifies you under the Guidelines as a career offender under the Guideline Section 4Bl.l(b).” Sentencing Tr. 12.
The record, however, leaves only the residual clause as a possible basis for finding that attempted robbery constituted a crime of violence. That is because the government “carries the burden of proving any facts that may be relevant in sentencing.” United States v. Price, 409 F.3d 436, 444 (D.C. Cir. 2005). The government, however, introduced no evidence into the district court record providing any basis for specifically determining that the attempted robbery was a crime of violence under the elements clause. The sentencing occurred prior to the Supreme Court’s decision in Descamps v. United States, — U.S. -, 133 S.Ct. 2276, 186 L.Ed.2d 438 (2013). Accordingly, the only way the district court could have classified Sheffield’s conviction for attempted robbery as a crime of violence would have been by reference to additional documentation — which the Government admittedly did not submit. See In re Sealed Case, 548 F.3d at 1089-1090. By default, the only thing left for the district court to rely on was the residual clause.
The district court committed plain error by concluding that Sheffield’s attempted robbery conviction supported an enhanced sentence. In Johnson v. United States, — U.S. -, 135 S.Ct. 2551, 192 L.Ed.2d 569 (2015), the Supreme Court held that the identically worded residual clause in the Armed Career Criminal Act (“ACCA”), 18 U.S.C. § 924(e)(2)(B), was unconstitutionally vague, 135 S.Ct. at 2563; see also Welch, 136 S.Ct. at 1262-1263. To be fair to the district court, we note that Johnson came out after sentencing in this case. But “as long as [an] error [is] plain as of * * * the time of appellate review * * * the error is ‘plain’ within the meaning of [Federal Rule of Criminal Procedure 52(b) ].” Henderson v. United States, — U.S. -, 133 S.Ct. 1121, 1124-1125, 185 L.Ed.2d 85 (2013).4
While this case involves the Sentencing Guidelines rather than ACCA, the government agrees with Sheffield that Johnson’s rationale equally requires resentencing in this direct appeal. See Gov’t Supp. Br. 4 n.3 (“[T]he government has consistently *313conceded that the residual clause of the career offender guideline is unconstitutionally vague[.]”). That concession makes ample sense; This court has repeatedly noted that, because the language of the Guidelines and ACCA residual clauses are the same, “we apply the ACCA standard to determine whether an offense qualifies as a crime of violence under section 4B1.2.” In re Sealed Case, 548 F.3d at 1089. That textual linkage did not change when the residual clause was held unconstitutional. The “grave uncertainty about how to estimate the risk posed by a crime” and the “uncertainty about how much risk it takes for a crime to qualify as a violent felony,” Johnson, 135 S.Ct. at 2557-2558, brood just as heavily over the Guidelines’ application as they did over the statute.
Furthermore, constitutional challenges may be brought against the Guidelines even though they are only advisory. See Peugh v. United States, — U.S. -, 133 S.Ct. 2072, 2082, 186 L.Ed.2d 84 (2013). That is because the Guidelines “impose a series of requirements on sentencing courts that cabin the exercise of * * * discretion.” Id. at 2084. “Common sense indicates that in general, this system will steer district courts to more within-Guidelines sentences.” Id.
So too with a constitutional vagueness challenge. Where the Guidelines “exert controlling influence on the sentence that the court will impose,” Peugh, 133 S.Ct. at 2085, an unconstitutionally vague Guidelines provision that has the effect of doubling or tripling a defendant’s sentence is constitutionally troublesome in its own right. Indeed, multiple courts of appeals have ruled that the Guidelines’ residual clause, like ACCA’s residual clause, is unconstitutionally vague, as the government has repeatedly conceded. See United States v. Pawlak, 822 F.3d 902, 911 (6th Cir. 2016) (holding so explicitly); United States v. Madrid, 805 F.3d 1204, 1210 (10th Cir. 2015) (same); United States v. Townsend, 638 Fed.Appx. 172, 177-178 (3d Cir. 2015) (same); see also United States v. Fields, 823 F.3d 20, 33 (1st Cir. 2016) (noting the government’s concession); United States v. Martinez, 821 F.3d 984, 988 (8th Cir. 2016) (same); United States v. Maldonado, 636 Fed.Appx. 807, 810 (2d Cir. 2016) (same); Ramirez v. United States, 799 F.3d 845, 856 (7th Cir. 2015) (“proceeding] on the assumption that the Supreme Court’s reasoning applies to section 4B1.2 as well”). But see United States v. Matchett, 802 F.3d 1185, 1195 (11th Cir. 2015) (holding Peugh inapplicable to due process challenges). Given the breadth of authority from both the Supreme Court and other circuits, combined with the government’s concession in this case and others, the error was plain. See In re Sealed Case, 573 F.3d 844, 851 (D.C. Cir. 2009) (error can be plain despite a circuit split on the issue).
Finally, the district court’s plain error unquestionably affected Sheffield’s substantial rights. See Olano, 507 U.S. at 734, 113 S.Ct. 1770. Application of the career-offender enhancement dramatically increased Sheffield’s sentencing exposure. As the district court specifically mentioned, “[a]bsent this career offender adjustment,” Sheffield would have faced “a sentencing range of 92 to 115 months.” Sentencing Tr. 12. In other words, because of the enhancement, Sheffield was sentenced to double the top of the Guidelines range he otherwise would have faced.
Nor can the increase be sustained on the alternative ground of relying on the “elements clause.” That clause requires a conviction to be based on a crime that “has as an element the use, attempted use, or threatened use of physical force against the person of another.” U.S.S.G. § 4B1.2(a). Importantly, “[i]n determining *314whether [a] crime is a violent felony, we consider the offense generically, that is to say, we examine it in terms of how the law defines the offense and not in terms of how an individual offender might have committed it on a particular occasion.” Begay v. United States, 553 U.S. 137, 141, 128 S.Ct. 1581, 170 L.Ed.2d 490 (2008). '“The prior conviction qualifies as an ACCA predicate only if the statute’s elements are the same as, or narrower than, those of the generic offense.” Descamps, 133 S.Ct. at 2281.
Under that test, D.C.’s attempted robbery statute is not categorically a crime that “has as an element the use, attempted use, or threatened use of physical force against the person of another,” § 4B1.2. At the time Sheffield was convicted, D.C.’s “[a]ttempt to commit robbery” provision read: “Whoever attempts to commit robbery, as defined in § 22-2801, by an overt act, shall be imprisoned for not more than 3 years or be fined not more than $500, or both.” D.C. Code § 22-2802. The robbery statute, in turn, defines the offense as: “by force or violence, whether against resistance or by sudden or stealthy seizure or snatching, or by putting in fear, * * * tak[ing] from the person or immediate actual possession of another anything of value[J” Id. § 22-2801.
As the “stealthy seizure” clause indicates, D.C.’s robbery statute includes “offenses that fail to qualify as crimes of violence under section 4B1.2.” In re Sealed Case, 548 F.3d at 1089. By the same token, the attempted robbery statute also “[is] not categorically [a] crime[] of violence,” as the government recognizes. Gov’t Supp. Br. 8; see also Mathis v. United States, — U.S. -, 136 S.Ct. 2243, 2248, 195 L.Ed.2d 604 (2016) (“[I]f the crime of conviction covers any more conduct than the generic offense, then it is not an ACCA [offense] — even if the defendant’s actual conduct (i.e., the facts of the crime) fits within the generic offense’s boundaries.”).
There is just one more wrinkle in this inquiry. The Supreme Court has allowed for a “modified categorical approach” to identifying crimes of violence for offenses that do not categorically satisfy the elements clause in those rare instances when that statutory offense is “divisible.” A statute is divisible if it “list[s] potential offense elements in the alternative, [and thus] renders opaque which element played a part in the defendant’s conviction.” Descamps, 133 S.Ct. at 2283. Under that modified categorical approach to divisible statutes, a court can review certain types of documents to determine whether the particular crime that the defendant committed necessarily included an element of violence, within the meaning of ACCA or Guidelines § 4B1.1. See Descamps, 133 S.Ct. at 2283-2284.
But D.C.’s attempted robbery statute is not divisible — that is, it does not set out “multiple, alternative versions of the crime” that include both violent and nonviolent elements. Descamps, 133 S.Ct. at 2284. To be sure, this court has held that D.C.’s robbery statute is divisible because robbery may be accomplished “against resistance” or “by putting in fear,” or alternatively, “by sudden or stealthy seizure or snatching.” In re Sealed Case, 548 F.3d at 1089 (citing D.C. Code § 22-2801).5 That means that a person can be convicted of committing either the violent version of robbery or the nonviolent, stealthy version of robbery. And if a defendant is convicted of the violent version, the defendant has *315committed a crime that has as an element the use of physical force against the person of another for purposes of Sentencing Guideline § 4B1.1.
The same cannot be said, though, of D.C.’s attempted robbery statute. That statute does not include those same alternative versions of the crime. “The elements of attempted robbery are that (1) the defendant committed an act which was reasonably adapted to the commission of the offense of robbery, (2) at the time the act was committed, the defendant acted with the specific intent to commit the offense of robbery, and (3) the act went beyond mere preparation, and carried the project forward to within dangerous proximity of the criminal end to be sought.” Robinson v. United States, 608 A.2d 115, 116 (D.C. 1992).
Thus, to convict a defendant of attempted robbery, D.C. law requires a jury to find beyond a reasonable doubt only that the defendant committed an act in furtherance of and with the specific intent to commit generic robbery, not any specific type of robbery (whether violent or stealthy). Nothing in the statutory text or case law requires a jury, in convicting a defendant of attempted robbery, to first find that the defendant committed one of multiple alternative elements, one of which is a crime of violence under the elements clause. Quite the contrary, attempted robbery is a loosely defined crime, with an expansive overt act requirement that is not tied to any specific type of robbery — violent or otherwise. Indeed, in Jones v. United States, 386 A.2d 308 (D.C. 1978), the D.C. Court of Appeals upheld a conviction for attempted robbery under D.C. law where the defendant had carefully planned a bank robbery, had “conducted a dry run,” was armed, “was proceeding toward the bank according to plan and was no further than four blocks away, turning back only when he heard police sirens and concluded that something had gone wrong.” Id. at 312-313. That conviction required no finding as an element of how the ultimate robbery might have been committed in terms of violence or stealth.
Because attempted robbery is an indivisible crime, we may not consult documents — such as the indictment or plea colloquy — to determine whether Sheffield’s offense qualifies as a crime of violence. See Mathis, 136 S.Ct. at 2254 (The modified categorical approach “is not to be repur-posed as a technique for discovering whether a defendant’s prior conviction, even though for a too-broad crime, rested on facts (or otherwise said, involved means) that also could have satisfied the elements of a generic offense.”). D.C.’s attempted robbery statute simply does not qualify as a crime of violence as a categorical matter. See Descamps, 133 S.Ct. at 2281-2282.
In sum, the district court’s plain error under the residual clause affected Sheffield’s substantial rights because his sentence cannot be saved under the elements clause. The unlawfulness of his sentence necessarily affects the fundamental fairness and integrity of his conviction. Sheffield is entitled to a resentencing without the career-offender enhancement.6
VII
We affirm Sheffield’s conviction, but vacate his sentence and remand for resen-tencing consistent with this decision.
So ordered.
. The Eleventh Circuit appears to have had conflicting precedent on the issue. Compare United States v. Gonzalez, 71 F.3d 819, 827 n.18 (11th Cir. 1996) ("[SJince the government declined to press this standing issue before the district court, we conclude that this issue has been waived.”), with United States v. Braithwaite, 709 F.2d 1450, 1453-1454 (11th Cir. 1983) (addressing issue despite the government's failure to "challenge [the defendant's] standing to assert a violation of his fourth amendment rights at the district court level”).
. One other circuit has continued to hold that the government need not obey ordinary argument-preservation rules when it comes to Fourth Amendment standing, but has done so precisely because it is the defendant's burden to establish the invasion of a protected expectation of privacy. See United States v. Paopao, 469 F.3d 760, 764 (9th Cir. 2006).
. In a supplemental filing with this Court, the government offered United States v. McCarson, 527 F.3d 170 (D.C. Cir. 2008), as an example of a decision upholding the admission of a decade-old conviction for firearm possession under Rule 404(b). Gov’t 7/5/16 Letter at 2. That opinion did not address the staleness of the conviction. See Cooper Industries, Inc. v. Aviall Services, Inc., 543 U.S. 157, 170, 125 S.Ct. 577, 160 L.Ed.2d 548 (2004) (" 'Questions which merely lurk in the record, neither brought to the attention of the court nor ruled upon, are not to be considered as having been so decided as to constitute precedents.’ ”) (quoting Webster v. Fall, 266 U.S. 507, 511, 45 S.Ct. 148, 69 L.Ed. 411 (1925)).
. We accordingly need not decide whether Sheffield's conviction would in fact have qualified as a crime of violence at the time of sentencing.
. This court's determination that D.C.'s robbery statute is divisible preceded the Supreme Court's recent decision in Mathis v. United States, - U.S. -, 136 S.Ct. 2243, 195 L.Ed.2d 604 (2016), which cast additional light on when the modified categorical approach applies.
. Last month, the Supreme Court granted certiorari in Beckles v. United States, No. 15-8544, to decide whether Johnson applies to enhancements under Sentencing Guidelines § 4B1.2. The government, however, has conceded that Sheffield is entitled to a new sentencing and has not requested that his case be held pending Bedeles. We agree that disposition is appropriate.