SELYA, Circuit Judge.
In cases in which defendants are accused of trafficking in narcotics, drug quantity is often both an-element of the offense and a critical integer in the sentencing calculus. These appeals illustrate that duality and, at the same time, serve to explicate the shifting standards of proof
that pertain. Because the court below applied these standards appropriately, we affirm both of the challenged convictions and two of the three challenged sentences. With respect to the third sentence, though, the government concedes that the district court relied on too weak a foundation in classifying the defendant as a career offender and we are not persuaded by the government’s attempt to brand the error harmless. We therefore remand that sentence for further proceedings consistent with this opinion.
I. BACKGROUND
We start with an overview of the case, drawing relevant facts from the trial transcripts, line sheets of recorded telephone calls introduced into evidence, and (where appropriate) undisputed portions of the defendants’ presentence investigation reports.
This case has its roots in an investigation into drug dealing in and around Worcester, Massachusetts, commenced by-the Drug Enforcement Administration (DEA) and local police officers. The probe initially focused on defendants James Dun-ston and Sergio Hernandez. Its scope later expanded to include defendant Anthony Wooldridge.
The investigation was launched with a number of controlled buys: between January and June — all dates are in 2012 unless otherwise indicated — an undercover DEA agent bought nearly forty-four grams of crack cocaine from Dunston and Hernandez in sixteen separate transactions. Intelligence gained during these sorties furnished the basis for properly authorized wiretaps on both Hernandez’s telephone and the telephone of yet another cocon-spirator (Richard Cruz). Between June and September, the agents intercepted close to 30,000 calls and text messages. The wiretapped intercepts indicated that Dunston, Hernandez, and Wooldridge were regularly dealing crack cocaine in Worcester and its environs.
■ The wiretaps revealed, inter alia, that the defendants acquired powder cocaine on no fewer than seven occasions in the summer months and attempted at least one further acquisition. Shortly after receiving the powder, the defendants promptly converted it to crack. They frequently discussed crack conversion techniques, described the results of particular conversions, and boasted about them ability to convert powder to crack without losing any appreciable drug weight.
It is said that all good things come to an end and, in July, Wooldridge was arrested during a traffic stop after police officers conducted a pat-frisk and found ninety-three grams of crack cocaine on his person. Dunston and Hernandez were arrested in September. All three were charged with conspiring to possess with intent to distribute both crack cocaine and powder cocaine. See 21 U.S.C. §§ 841(a)(1), 846. Additionally, Wooldridge was charged with possession of crack cocaine with intent to distribute, see M § 841(a)(1), and Hernandez was charged with being a felon in possession of a firearm and ammunition, see 18 U.S.C. § 922(g)(1).
After some preliminary skirmishing, all three defendants waived indictment and pleaded guilty to superseding informations charging them with, as relevant here, conspiring to possess with intent to distribute crack cocaine. The informations specified that the charged conspiracy “involved 280 grams or more of a mixture and substance containing a detectable amount of cocaine base” — a quantity sufficient to trigger a ten-year mandatory minimum sentence. See 21 U.S.C. § 841(b)(l)(A)(in).
Each defendant pleaded guilty to the underlying conspiracy offense, reserving,
however, the right to contest at a bench trial whether the amount of crack cocaine reasonably foreseeable or attributable to him was 280 grams or more (thus exposing him to the mandatory minimum sentence). The district court thereafter held an eight-day bench trial, at which it heard, inter alia, recordings of intercepted calls as well as testimony from á DEA agent (Timothy Boyle), who interpreted the slang and jargon that permeated in the recordings. At the conclusion of all the evidence, the court found “beyond a reasonable doubt that over 280 grams of cocaine base [wa]s attributable and reasonably foreseeable to all defendants.”
The district court ordered the probation department to prepare individual presen-tence investigation reports. At the disposition hearings, the court adopted the reports without change and classified all three defendants as career offenders under the sentencing guidelines. It found Wooldridge responsible for at least 840 grams of crack cocaine and sentenced him to a 182-month term of immurement. With respect to Dunston and Hernandez, the court found each of them responsible for at least 2.8 kilograms (2,800 grams) of crack cocaine; sentenced Dunston to 144 months’ imprisonment; and sentenced Hernandez to 162 months’ imprisonment. These timely appeals followed.
II. ANALYSIS
The defendants’ appeals have been consolidated, and we turn first to the claims of error that implicate their convictions. From that point, we proceed to their claims of sentencing error.
A. Lay Opinion Testimony.
Dunston contends that the district court should not have allowed Agent Boyle to testify about the meaning of slang terms and jargon used in the course of the wiretapped conversations. He submits that although Boyle may have been qualified to give some lay opinion testimony, the government failed to erect an adequate foundation for his interpretations of particular slang terms. To illustrate his point, Dun-ston notes that Boyle was allowed to testify that the word “tweezy” referred to crack cocaine and that the phrase “step up a yard” referred to turning powder into crack. In Dunston’s view, Boyle’s familiarity with the defendants’ conversations did not qualify him to give an informed opinion about the meaning of these and other specific phrases used within those conversations. We do not agree.
All three defendants raised this issue below, but only Dunston pursues it on appeal. Objections to the admission of evidence are reviewed for abuse of discretion. See United States v. Valdivia, 680 F.3d 33, 50 (1st Cir. 2012). The parties agree that Agent Boyle’s testimony should be considered lay'opinion testimony. Hence, Federal Rule of Evidence 701 controls. See id. Rule 701 permits the admission of lay opinion testimony “rationally based on the witness’s perception” that would help the fact-finder “determinfe] a fact in issue.”
Application of Rule 701 in the drug-trafficking context is not novel: “we have long held that government witnesses with experience in drug investigations may explain the drug trade and translate coded language” for factfinders through lay opinion testimony. United States v. Rosado-Pérez, 605 F.3d 48, 56 (1st Cir. 2010); accord United States v. Hoffman, 832 F.2d 1299, 1310 (1st Cir. 1987) (holding that “interpretation of codes and jargon used in the drug trade can be supplied through one experienced in the field”). Such testimony is especially useful where, as here, it can afford the factfinder the benefit of specialized knowledge. See United States
v. Albertelli, 687 F.3d 439, 446 (1st Cir. 2012).
Of course, such interpretive testimony must be anchored in the witness’s personal experience in the field, see Hoffman, 832 F.2d at 1310, and his experience-based understanding of the meaning of the terms used, United States v. Prange, 771 F.3d 17, 28 & n.3 (1st Cir. 2014). Put another way, “an interpretation of a phrase or reference ought to be explicable”' — a standard that typically requires the witness to point to similar statements surrounding similar events. Albertelli, 687 F.3d at 450.
In the case at hand, the government erected a sturdy foundation for Boyle’s testimony. The record reflects that Boyle had a twenty-four-year career in law enforcement, with significant experience in undercover drug investigations. He had received specialized training in narcotics enforcement, had participated in over fifteen wiretap investigations, and had supervised more than twenty other such investigations. As part of his duties, he had reviewed audio and videotape from undercover crack cocaine purchases “hundreds” of times and had taken part at least once in an undercover operation in which he observed powder being converted to crack.
In this particular probe, Boyle reviewed nearly all of the 30,000 calls and texts collected during the wiretaps. He not only drew on his extensive experience to inform his understanding of specific slang terms but also took into account the context in which those terms were used. Boyle explained that he often listened to “several calls leading up to” the use of a given bit of slang as well as “conversations that would take place after that” to ensure his understanding.
Confronted with objections, the district court prudently allowed defense counsel to conduct a voir dire and question Boyle about his credentials, his experience, and his knowledge. This additional safeguard, coupled with Boyle’s cross-examination by all three defense attorneys, mitigated any risk of unfair prejudice from his testimony. See United States v. Henry, 848 F.3d 1, 13-14 (1st Cir. 2017); Albertelli, 687 F.3d at 447.
Where malefactors try to mask their criminal activities by using codes, a law enforcement officer who is equipped by knowledge, experience, and training to break those codes can help to inform the factfinder’s understanding. So it is here: the government provided the district court with ample reason to conclude that Boyle was knowledgeable about the idiom of the drug trade and, in particular, the vernacular of this group of miscreants. On this record, we hold, without serious question, that the admission of Boyle’s lay opinion testimony was comfortably within the ambit of the district court’s discretion.
B. Sufficiency of the Evidence.
The next leg of our journey takes us to Dunston’s and Wooldridge’s claims that the district court lacked sufficient evidence to find them guilty of the charged offense. These claims share a common focus: they are aimed at the district court’s finding, at trial, that at least 280 grams of crack cocaine was “attributable and reasonably foreseeable” by the defendant making the claim. Such a defendant-specific finding is an element of the offense of conviction: it means that a defendant cannot, simply by reason of his membership in a conspiracy that traffics in large amounts of drugs, automatically be “saddled with the full weight of the conspiracy’s wrongdoing.” United States v. Sepulveda, 15 F.3d 1161, 1197 (1st Cir. 1993). He can only be held responsible for drugs that he “personally handled or anticipated
handling,” as well as “drugs involved in additional acts that were reasonably foreseeable by him and were committed in furtherance of the conspiracy.” Id.
Here, the district court held a bench trial on the drug-weight issue — a trial that was necessary because any fact that increases the mandatory minimum sentence for an offense must, absent a plea or a stipulation, be submitted to a factfin-der to be determined beyond a reasonable doubt. See Alleyne v. United States, — U.S. -, 133 S.Ct. 2151, 2155, 186 L.Ed.2d 314 (2013) (citing Apprendi v. New Jersey, 530 U.S. 466, 483 n.10, 490, 120 S.Ct. 2348, 147 L.Ed.2d 435 (2000)); United States v. Pizarro, 772 F.3d 284, 292-93 (1st Cir. 2014). At the end of the trial, the court found that each of the defendants was responsible for at least 280 grams of crack cocaine.
Hernandez does not appeal the drug-quantity determination as it relates to him. Dunston and Wooldridge do appeal and, as to each of them, we must “exam-in[e] the facts and inferences in the light most favorable to the verdict.” United States v. O’Donnell, 840 F.3d 15, 18 (1st Cir. 2016) (citation omitted); cf. United States v. Colon-Solis, 354 F.3d 101, 103 (1st Cir. 2004) (requiring “defendant-specific determination of drug quantity” for this purpose). The ultimate question is whether, after viewing the evidence in the light most hospitable to the government, a rational trier of fact could have found the drug-weight elements beyond a reasonable doubt. See United States v. Grace, 367 F.3d 29, 34 (1st Cir. 2004). We must answer this question separately as to each of the two appealing defendants, see Colon-Solis, 354 F.3d at 103, exercising de novo review. See O’Donnell, 840 F.3d at 18.
With this preface, we turn to the substance of the sufficiency claims. We begin with Dunston and then move to Wool-dridge.
1. Dunston. Dunston contends that the court’s finding that he was responsible for at least 280 grams of crack cocaine was plucked out of thin air. Pointing out that the authorities never seized any of the drugs to which he alluded in the wiretapped conversations, he insists that it remains unclear how much powder was converted to crack.
Dunston is whistling past the graveyard. The government presented evidence linking Dunston to at least eight separate transactions. We need examine only two of these occurrences to defuse Dunston’s attack.
On June 26, the government intercepted a telephone call in which Dunston and Hernandez discussed their plans to cook 200 grams of powder cocaine recently purchased from Wooldridge. The two men split the contraband, and Dunston told Hernandez that he was going to “step up a yard right now and see what it is.” This meant, Agent Boyle testified, that Dunston intended to convert 100 grams of powder into crack. Later that day, Dunston and Hernandez spoke again. When Hernandez asked, “how did that come out,” Dunston replied that the result was “[a]ll yellow” and that he “[g]ot back, like, 2 extra grams. It was, like, 102 when it dried all the way out, bone dry.” Boyle testified that the yellow tinge of Dunston’s product was emblematic of crack cocaine and that, since converting powder to crack requires water, Dunston’s reference to drying the product was also consistent with making crack.
When judges sit as factfinders, they are not obliged to put their common sense into cold storage. Given the wiretaps and Boyle’s testimony, the district court surely had a rational basis for finding, beyond any reasonable doubt, that Dunston was responsible for at least 200 grams of crack
stemming from the June 26 events. For one thing, the court supportably could have found Dunston responsible for the 100 grams that he himself cooked. For another thing, the court supportably could have found that the remaining 100 grams, retained by Hernandez, was to be converted from powder to crack as part of the conspiracy and that the conversion was reasonably foreseeable to Dunston.
Dunston protests. He says that there is no evidence that this latter 100 grams of powder was actually converted to crack and that this part of the district court’s drug-quantity calculation was faulty because it implicitly assumed that 100 grams of powder cocaine, when converted, would yield 100 grams of crack cocaine. He insists that, even in optimal conditions, 100 grams of powder cannot be converted into more than 89.2 grams of crack.
This attack misfires. The court had before it evidence that, on both June 26 and June 29 (discussed infra) Dunston himself had performed conversion operations and had gotten better than a one-to-one yield. Given that evidence, the court was entitled to apply a one-to-one conversion ratio in its appraisal of the conspiracy’s output.
The court also heard the particulars of a transaction that occurred on June 29. On that date, the government intercepted a call between Hernandez and a third party, during which Hernandez bought 200 grams of powder cocaine. Just over two hours later, Dunston called Hernandez to inform him that he planned to “do the whole thing together.” He subsequently reported that the result was “206 wet,” and yellow in color, but quickly turning white. Given this evidence and Boyle’s explanatory testimony, the district court could have attributed, beyond any reasonable doubt, an additional 200 grams of crack to Dunston. After all, proof beyond a reasonable doubt does not require proof to an absolute certainty. See United States v. Lasseque, 806 F.3d 618, 623 (1st Cir. 2015), cert. denied, — U.S. -, 136 S.Ct. 1472, 194 L.Ed.2d 567 (2016).
Dunston counters that the government offered no direct or tangible evidence that this -batch of cocaine ever existed. But it is old hat that proof of a defendant’s guilt can be based, in whole or in part, on circumstantial evidence. See United States v. George, 841 F.3d 55, 63 (1st Cir. 2016); United States v. Williams, 717 F.3d 35, 39-40 (1st Cir. 2013). In this instance, the district court, qua factfinder, drew a series of reasonable inferences from the evidence, and its conclusion that Dunston was responsible for this additional 200 grams of crack is unimpugnable.
For these reasons, Dunston’s conviction must stand.
2. Wooldridge. The record also provides ample support for the district court’s decision to attribute at least 280
grams of crack cocaine to Wooldridge. Wooldridge was arrested with ninety-three grams of crack on his person. In addition, there was cogent evidence that Wooldridge sold his codefendants the 200 grams of powder cocaine mentioned in the June 26 conversation. As explained above, 200 grams of powder could be found, beyond a reasonable doubt, to equal 200 grams of crack. Because the production of crack was a reasonably foreseeable act in furtherance of the conspiracy, the district court did not err either in finding that Dunston and Hernandez converted their shares of the powder sold by Wooldridge into crack or in holding Wooldridge responsible for a corresponding amount of crack. See Se-pulveda, 15 F.3d at 1197.
In an effort to stave off this conclusion, Wooldridge insists that he was not yet part of the charged conspiracy when he sold drugs to his codefendants at the end of June. Building on that foundation; he posits that any crack cocaine resulting from that sale cannot be attributed to him. This is revisionist history, and we reject it. Wooldridge pleaded guilty to being a member of the conspiracy beginning in August of 2011 — -long before the sale occurred. He is bound by his plea. See United States v. Laracuent, 778 F.3d 347, 351 (1st Cir.), cert. denied, — U.S. -, 135 S.Ct. 2875, 192 L.Ed.2d 908 (2015).
That ends this aspect of the matter. We conclude that the evidence was sufficient to allow the district court to find beyond a ■reasonable doubt that at least 280 grams of crack were attributable to Wooldridge. Consequently, his conviction must stand.
C. Dunston’s Sentence.
Dunston’s claim of sentencing error is easily dispatched. He does not challenge the district court’s guideline calculations but, rather, maintains that his 144-month sentence offends the parsimony principle, see United States v. Sepúlveda-Hernández, 817 F.3d 30, 34 (1st Cir. 2016) (citing 18 U.S.C. § 3553(a)), because it is greater than necessary to achieve the purposes of sentencing. Under our case law, we treat such a claim as one of substantive unreasonableness. See id.
Dunston did not raise this claim in the district court, and the standard of review is unsettled. See United States v. Ruiz-Huertas, 792 F.3d 223, 228 (1st Cir.) (noting uncertainty with respect to standard of review regarding unpreserved claims of substantive unreasonableness), cert. denied, — U.S. -, 136 S.Ct. 258, 193 L.Ed.2d 191 (2015). For simplicity’s sake, we assume — favorably to Dunston — that review is for abuse of discretion.
Dunston was subject to the career offender guideline. See USSG § 4B1.1(a). The court below set his guideline sentencing range (GSR) at 262-327 months and sentenced him appreciably below the bottom of that range. A defendant who challenges a below-the-range sentence as substantively unreasonable must carry a heavy burden. See United States v. Montero-Montero, 817 F.3d 35, 37 (1st Cir. 2016). The greater the drop from the bottom of the range, the heavier the lift. See id.
There is more than one reasonable sentence in any given case, and the relevant inquiry is whether the sentence imposed “resides within the expansive universe of reasonable sentences.” United States v. King, 741 F.3d 305, 308 (1st Cir. 2014). In mounting this inquiry, we concentrate on whether the sentence reflects “a plausible ... rationale and a defensible result.” United States v. Martin, 520 F.3d 87, 96 (1st Cir. 2008).
The court below supplied a plausible rationale for Dunston’s sentence. It explained that a significant period of incar
ceration was warranted to protect the public from future crimes, provide deterrence, and punish the defendant. Given the scope and duration of the conspiracy, the force of this explanation cannot be gainsaid. The court added, though, that “incarceration at or near the advisory guideline range seems to be more than what is necessary to meet the goals of sentencing.” For that reason, it executed a sharp downward variance.
Despite the court’s seeming leniency, Dunston assails the result, arguing that his ultimate sentence was skewed by a “mechanical application” of the career offender guideline. This argument is constructed out of whole cloth: it overlooks the undeniable fact that the sentencing court’s downward variance was anything but mechanical, yielding a sentence almost ten years below the nadir of the guideline range. The challenged sentence, which took into account the nature of Dunston’s crime, his history, and his personal characteristics, was wholly defensible and well within the court’s discretion. See King, 741 F.3d at 309.
D. Wooldridge’s Sentence.
Wooldridge challenges the amount of crack cocaine attributed to him for sentencing purposes. This drug quantity (at least 840 grams) is separate from the lesser drug quantity (280 grams) needed to establish Wooldridge’s guilt. See supra Part IIB. We explain briefly why two distinct findings are necessary.
At trial, the government had the burden of proving the drug quantity charged as an element of the offense (280 grams or more) beyond a reasonable doubt. See Alleyne, 133 S.Ct. at 2155. At sentencing, however, “a defendant-specific determination of drug quantity [i]s a benchmark for individualized sentencing under the guidelines.” Colon-Solis, 354 F.3d at 103. The sentencing court is charged with making such a defendant-specific determination. See United States v. Correy, 773 F.3d 276, 279-80, 280 n.4 (1st Cir. 2014). All drugs “attributable to[ ] or reasonably foreseeable by” a defendant may be included in that defendant’s individualized total. United States v. Cintrón-Echautegui, 604 F.3d 1, 5 (1st Cir. 2010). The government bears the burden of proving drug quantity at sentencing by a preponderance of the evidence. See id. at 6.
The sentencing guidelines direct a sentencing court to consider relevant conduct, see USSG § lB1.3(a), when assessing the drug quantity for which a defendant is to be held accountable at sentencing. Where, as here, a defendant has been convicted as a coconspirator, his relevant conduct includes not only his own acts and omissions but also the reasonably foreseeable acts and omissions of other coconspirators in furtherance of the conspiracy. See id. § lB1.3(a)(l)(B).
A sentencing determination of defendant-specific drug quantity does not require mathematical precision: it need only be a “reasonable approximation of the weight of the controlled substance(s) for which the defendant should be held responsible.” United States v. Demers, 842 F.3d 8, 12 (1st Cir. 2016). As we have said, calculating drug quantities “is not an exact science,” and the sentencing court “need not be precise to the point of pedantry.” United States v. Platte, 577 F.3d 387, 392 (1st Cir. 2009).
Such drug-quantity determinations are quintessentially factual in nature, and we review them for clear error. See Cintrón-Echautegui, 604 F.3d at 6. That standard dictates that a finding will stand unless a reviewing court, after assessing the whole of the record, is firmly convinced that a mistake has been made. See id. “[W]here there is more than one plausible
view of the circumstances, the sentencing court’s choice among supportable alternatives cannot be clearly erroneous.” United States v. Ruiz, 905 F.2d 499, 508 (1st Cir. 1990).
Against this backdrop, Wool-dridge argues that the district court committed clear error in finding him accountable for at least 840 grams of crack cocaine. That finding, combined with other uncontroversial guideline calculations and Wool-dridge’s designation as a career offender, see USSG § 4Bl.l(a), resulted in a GSR of 262-327 months. The district court proceeded to impose a below-the-range sentence of 132 months.
Although the district court varied downward from the guideline range, any error in calculating that range would likely be material on appeal. See Molina-Martinez v. United States, — U.S. -, 136 S.Ct. 1338, 1346, 194 L.Ed.2d 444 (2016) (“In most cases a defendant who has shown that the district court mistakenly deemed applicable an incorrect, higher Guidelines range has demonstrated a reasonable probability of a different outcome.”). Here, however, the district court’s drug-quantity determination (and, thus, its configuration of the GSR) is not clearly erroneous.
The district court identified three transactions that linked Wooldridge to well over 840 grams of crack cocaine. First, Wool-dridge sold powder cocaine to Hernandez and Dunston at the end of June and, as discussed above, see supra Part IIB, that sale formed a solid basis for a finding attributing approximately 200 grams of crack cocaine to ■ him. Second, the court supportably tied Wooldridge to a purchase of 300 grams of powder cocaine that occurred on July 31.
The record includes discussions among all three defendants about the purchase and about their plans to convert the powder into crack. That evidence, coupled with the court’s permissible use of a one-to-one conversion ratio, was sufficient to ground attribution of 300 grams of crack cocaine to Wooldridge.
Third, and finally, the court reasonably attributed an additional 497 grams of crack cocaine to Wooldridge with respect to events occurring on August 31. At that time, federal agents seized a package sent by Cruz (in Puerto Rico) to an acquaintance of Hernandez (in Massachusetts). The package was found to contain 497 grams of powder cocaine. Although the evidence indicated that Hernandez was the prime mover in arranging for the shipment, both Wooldridge and Dunston encouraged his efforts. Thus, the 497 grams of powder cocaine, destined for conversion into crack cocaine, could fairly be attributed to Wooldridge for sentencing purposes. See Cintrón-Echautegui, 604 F.3d at 5; USSG § 1B1.3(a)(1)(B).
The fact that the package was intercepted before its contents reached the defendants is of no moment. The record makes manifest that the defendants, in furtherance of the conspiracy, sought the shipment of powder cocaine as a prelude to conversion and would in all probability have converted it to crack cocaine if given the chance. Indeed, the government introduced evidence showing that, on no fewer than seven separate occasions between June and August, the defendants procured powder cocaine and converted it into crack cocaine within a matter of days. When drug traffickers have forged a well-defined pattern and practice for their operations, a sentencing court need not turn a blind eye to that pattern and practice.
Wooldridge tries to distance himself from this drug weight, asseverating that no part of the August 31 shipment should be attributed to him because his coconspirators had “cut[ ] [him] out of the picture” by then. But the evidence of exclusion is meager, and the district court was not obliged to resolve this issue in Wooldridge’s favor. Just as withdrawal from a conspiracy requires more than an empty claim of disaffection, see United States v. Piper, 298 F.3d 47, 53 (1st Cir. 2002) (explaining that, in order to withdraw, a conspirator must take affirmative action to disavow or defeat the conspiracy), so too does a claim of exclusion. There is simply no credible evidence that either Wooldridge or his partners in crime took any such action here. To the contrary, the record reflects that Wooldridge contacted Hernandez at the end of July in search of cocaine and exhibited an abiding intent to remain involved in the conspiracy. Given the absence of any competent evidence of withdrawal, the sentencing court did not clearly err in finding the coconspirators’ acts at the end of August to be a part of Wooldridge’s relevant conduct.
E. Hernandez’s Sentence.
Hernandez’s claim of sentencing error similarly embodies a challenge to the district court’s drug-quantity determination. The court found him responsible for at least 2.8 kilograms of crack cocaine. Combined with other uncontroversial guideline calculations and a career offender enhancement, this drug quantity resulted in a GSR of 292-365 months. The district court imposed a below-the-range sentence of 162 months.
The record is replete with circumstantial evidence that Hernandez headed up the conspiracy. He was involved in every transaction that we have discussed: the events of June 26 (200 grams), the events of June 29 (200 grams), the events of July 31 (300 grams), and the events of August 31 (497 grams). In consequence of this involvement, 1,197 grams of crack cocaine were fairly attributable to him.
Furthermore, the record contains proof sufficient to support the attribution to Hernandez, using a one-to-one conversion ratio, of an additional 1,000 grams of crack cocaine:
• 200 grams of powder cocaine purchased and converted on July 7;
• 100 grams of powder cocaine converted on July 15;
• 500 grams of powder cocaine purchased on August 5 and converted the following day; and
• 200 grams of powder cocaine purchased on August 28 and at least partially converted the next day.
When these quantities are combined with the nearly forty-four grams of crack sold by Hernandez to undercover officers at the start of the probe, Hernandez can properly be held responsible for a minimum of 2.2 kilograms of crack cocaine.
To boost that total to 2.8 kilograms, the district court reasonably relied on two additional events. First, the evidence showed that, on July 29, Hernandez arranged the payment of $19,000 to Cruz for powder cocaine previously supplied. The court sup-portably concluded that this payment represented the purchase of roughly 500 grams of powder cocaine. See Sepulveda, 15 F.3d at 1201 (holding that when cash- is likely dedicated to the purchase of contraband, “a sentencing court may convert the cash into equivalent amounts of narcotics” for sentencing purposes). After applying a one-to-one conversion ratio, this transaction yielded 500 grams of crack cocaine attributable to Hernandez.
The court reasonably added another 100 grams, based on evidence that Hernandez
was actively dealing crack cocaine between January and June (when the wiretaps were instituted). While this finding represented an estimate, it was by all odds a conservative estimate — especially since Hernandez, pleaded guilty to conspiring to deal crack as early as August of 2011. We have consistently upheld a sentencing court’s use of reasonable estimates in assessing drug quantity, see, e.g., United States v. Bernier, 660 F.3d 543, 546 (1st Cir. 2011); Platte, 577 F.3d at 392; United States v. Ventura, 353 F.3d 84, 88 (1st Cir. 2003), and that praxis has particular appeal when — as in this case — the sentencing court, in fashioning its estimate, has taken a conservative approach, cf. United States v. Sklar, 920 F.2d 107, 113 (1st Cir. 1990) (admonishing sentencing courts, in this context, to “err on the side of caution” (citation omitted)).
This gets the grease from the goose. While we could analyze these multiple transactions in more granular detail, doing so would serve no useful purpose. It suffices to say that the record discloses ample evidence to ground a finding that Hernandez was responsible for at least 2.8 kilograms of crack cocaine. As a result, we discern no clear error in the district court’s drug-quantity determination.
Hernandez has one last string to his bow. He takes issue with the district court’s designation of him as a career offender under the sentencing guidelines. Specifically, he asserts that the court improperly relied, over objection, on a portion of his presentence investigation report, which cited only criminal offender record information (CORI) to lay out his criminal history.
The government bears the burden of establishing the existence of at least two predicate offenses to trigger a career offender enhancement under the sentencing guidelines. See United States v. Bryant, 571 F.3d 147, 153 (1st Cir. 2009) (citing USSG § 4B1.1(a)). The government may carry this burden in divers ways, such as by furnishing “a certified copy of the conviction or an equivalent proffer” or by pointing to official court documents of the type that engender a presumption of reliability. Id at 153, 155 (quoting United States v. McKenzie, 539 F.3d 15, 19 (1st Cir. 2008)). But anecdotal CORI descriptions, if objected to, are not enough. See id. at 154-55; cf. United States v. Brown, 510 F.3d 57, 74 (1st Cir. 2007) (observing that “when a defendant challenges a conviction laid out in the [presentence investigation report], more is required”).
At Hernandez’s sentencing, the government argued for a career offender enhancement noting that the CORI material described in the presentence investigation report identified three potential predicate offenses. Hernandez objected. The district court stayed the proceedings to give the government the opportunity to make a more persuasive evidentiary proffer; but when the government produced nothing more, the court overruled Hernandez’s objection and sentenced him as a career offender.
On appeal, Hernandez insists that the record before the district court was insufficient to show the required predicate offenses. The government concedes this point and instead proffers a series of docket sheets, never made available to the district court, inviting us to take judicial notice of them. These docket sheets, it says, will confirm the existence of the required predicate offenses (albeit belatedly) and render any error harmless.
We decline the government’s invitation. As a general matter, we do hot consider evidentiary materials that were not proffered in the district court. See United States v. Farrell, 672 F.3d 27, 30-
31 (1st Cir. 2012); United States v. Kobrosky, 711 F.2d 449, 457 (1st Cir. 1983). Although we have occasionally taken judicial notice of state court documents identified for the first time on appeal, see, e.g., Farrell, 672 F.3d at 31; United States v. Mercado, 412 F.3d 243, 247-48 (1st Cir. 2005), we cannot do so here. The proffered documents bear no hallmarks of authenticity and, in all events, are subject to interpretation. A career offender designation can expose a defendant to a significantly higher sentence, and it remains the district court’s duty to ensure that documents offered to prove the existence of predicate offenses are “sufficiently reliable.” Bryant, 571 F.3d at 154.
Although we conclude that the district court erred, we think it would be premature to vacate Hernandez’s sentence. Instead, we remand so that the district court may hold a hearing and afford the parties an opportunity to present evidence anent Hernandez’s prior convictions. Should the district court find that the evidence presented is sufficiently reliable and establishes the existence of the requisite number of predicate offenses, it should, within sixty days from the date hereof, report its findings and conclusions to this court. If, however, the court finds to the contrary, it should, within sixty days from the date hereof, report its findings and conclusions to this court and proceed to vacate Hernandez’s sentence and resen-tence him without reference to the career offender guideline.
III. CONCLUSION
We need go no further. For the reasons elucidated above, we affirm the judgment in appeal number 15-1812, thus affirming Dunston’s conviction and sentence. We likewise affirm the judgment in appeal number 15-2000, thus affirming Wool-dridge’s conviction and sentence. With respect to appeal number 15-1999 (Hernandez), we remand for further proceedings consistent with this opinion and for the time being retain appellate jurisdiction over this appeal.
So ordered.