Federal law makes it a crime to possess body armor after having been convicted of a “crime of violence” as defined in 18 U.S.C. § 16 (“section 16”). See 18 U.S.C. § 931. Appellant David Fish possessed body armor after having been convicted of several crimes, and the district court ruled that at least one of those several offenses qualified as a crime of violence under section 16. Following that ruling, Fish entered a conditional plea of guilty, reserving the right to challenge on this timely appeal the determination that he had previously been convicted of a crime of violence.
In defense of the district court’s determination, the government points to four crimes under Massachusetts law for which Fish had previously been convicted: breaking and entering in the daytime, see Mass. Gen. Laws ch. 266, §§ 17-18, breaking and entering at night, see id. § 16, assault and battery with a dangerous weapon, see id. ch. 265, § 15A(b), and possession of a burglarious instrument, see id. ch. 266, § 49. Notwithstanding their aptly-styled titles, we find that none of those crimes, as defined under Massachusetts law, qualifies as a crime of violence under section 16. We therefore reverse Fish’s conviction.
I. Facts
The following facts are taken from the prosecution’s presentation at Fish’s plea colloquy. At the colloquy, Fish admitted to all facts necessary to support his guilty plea. The facts are uncontested on appeal.
On June 18, 2009, Fish, who was working as an auto mechanic at a repair shop in Pittsfield, Massachusetts, reported to the Pittsfield Police Department that someone had broken into a police vehicle that had been left at the shop for repair. After examining the vehicle, police discovered that several bulletproof vests were missing. In early July, the department learned through a cooperating witness that Fish was offering to sell six bulletproof vests. An undercover officer contacted Fish through the cooperating witness and purchased five vests, which the department afterwards identified as five of the vests that had been stolen from its vehicle. The vests, manufactured outside Massachusetts, had traveled interstate for sale in the Commonwealth.
A federal grand jury eventually returned an indictment charging Fish with a single count of possessing body armor in violation of 18 U.S.C. § 931(a), the federal body armor statute, which prohibits any person who “has been convicted of a felony that is ... a crime of violence (as defined *4in [section 16])” from possessing body armor that has been “sold or offered for sale[ ] in interstate or foreign commerce.” See also 18 U.S.C. § 921(a)(35). Fish moved to dismiss the indictment, claiming that none of his prior convictions qualified as a “crime of violence” under section 16, and that the body armor statute was unconstitutional. In response, the government argued that Fish’s convictions for assault and battery with a dangerous weapon and breaking and entering all qualified as crimes of violence under section 16, and that the body armor statute was constitutional.
The district court denied the motion to dismiss, finding that “[a]t a minimum, the convictions for breaking and entering satisfy the requirements of section 16].” The court also rejected Fish’s challenge to the constitutionality of the body armor statute. Fish entered a conditional guilty plea under Rule 11(a)(2) of the Federal Rules of Criminal Procedure, preserving his right to challenge on appeal the district court’s denial of his motion to dismiss the indictment. On June 7, 2012, the district court entered a final judgment, sentencing Fish to forty-eight months’ probation, with ten months to be served in a community corrections facility. This appeal followed.
II. Standard of Review
Whether a prior conviction is a qualifying offense under section 16 is a question of law that we review de novo. See Aguiar v. Gonzales, 438 F.3d 86, 88 (1st Cir.2006).
III. Analysis
The difficulty posed by this and similar cases arises from the fact that there is no master list of offenses that qualify as crimes of violence. Rather, section 16 sets forth two qualitative definitions of the term “crime of violence,” leaving it to the courts to measure each crime against these definitions, which read as follows:
(a) an offense that has as an element the use, attempted use, or threatened use of physical force against the person or property of another, or
(b) any other offense that is a felony and that, by its nature, involves a substantial risk that physical force against the person or property of another may be used in the course of committing the offense.
18 U.S.C. § 16.
The candidates for satisfying these definitions are legion and varied. Each state defines its own crimes, generally without reference to (and often, we presume, without knowledge of) the section 16 definitions. Similar-sounding crimes may have different elements from state to state. E.g., Sykes v. United States, — U.S. -, 131 S.Ct. 2267, 2295, 180 L.Ed.2d 60 (2011) (Kagan, J., dissenting). The elements of each crime may be defined by statute, e.g., Mass. Gen. Laws ch. 266, § 16, or by case law, e.g., Commonwealth v. Burno, 396 Mass. 622, 625, 487 N.E.2d 1366 (1986) (discussing the elements of Mass. Gen. Laws ch. 265, § 15A).
Compounding the difficulty of working with section 16’s two qualitative definitions is the fact that Congress has also adopted an entirely separate, but quite similar, definition of the term “violent felony” as used in the Armed Career Criminal Act, 18 U.S.C. § 924(e) (“ACCA”). ACCA defines “violent felony” as follows:
[T]he term “violent felony” means any crime punishable by imprisonment for a term exceeding one year ... that
(i) has as an element the use, attempted use, or threatened use of physical force against the person of another; or
(ii) is burglary, arson, or extortion, involves use of explosives, or otherwise *5involves conduct that presents a serious potential risk of physical injury to another.
The partially overlapping, two-pronged definitions of the terms “crime of violence” and “violent felony” have given rise to multiple lines of precedent, each offering both the advantages and the limitations of cross-over application by analogy. E.g., United States v. Leahy, 473 F.3d 401, 412 (1st Cir.2007) (noting that although we and the Supreme Court have treated the “risk of physical injury” provisions as reaching conduct beyond the scope of section 16(b), neither our decisions nor the Supreme Court’s “in any way suggest! ] that the reverse is true”). Adding further insight, but perhaps further confusion as well, the United States Sentencing Guidelines define the term “crime of violence” using language that is almost, but not quite, the same as the language that ACCA uses to define the term “violent felony.” See U.S.S.G. § 4B1.2 (defining “crime of violence” under the career offender guideline); compare United States v. Willings, 588 F.3d 56, 58 n. 2 (1st Cir.2009) (“[T]he terms ‘crime of violence’ under the career offender guideline and ‘violent felony’ under the ACCA are nearly identical in meaning, so that decisions construing one term inform the construction of the other.”) with United States v. Giggey, 551 F.3d 27, 36 (1st Cir.2008) (en banc) (pointing out differences).
A third and greater complexity arises from the fact that many crimes are defined in a manner broad enough to cover both conduct that clearly meets one or both of the section 16 definitions and conduct that clearly does not. For example, in Massachusetts, the broad definition of simple assault and battery encompasses both a devastating beating and a tap on the shoulder. See generally United States v. Holloway, 630 F.3d 252 (1st Cir.2011) (discussing the Massachusetts simple assault and battery statute).
The Supreme Court has grappled repeatedly with this third complexity, establishing and then refining a set of rules to be employed in classifying a defendant’s prior offenses of conviction. These rules derive in great part from the need to honor the requirements of the Sixth Amendment’s right to jury trial. Their principal purpose is to ensure that before we send a person to jail for doing “X,” either the person must admit to “X” or a jury (or jury-waived court) must convict the person of doing “X” following a fair trial. See Shepard v. United States, 544 U.S. 13, 24, 125 S.Ct. 1254, 161 L.Ed.2d 205 (2005).
The first set of rules to be applied forms what is known as the “categorical” approach. Aguiar v. Gonzales, 438 F.3d 86, 88 (1st Cir.2006). The categorical approach requires an assessment of “the elements of the statute of conviction, not ... the facts of each defendant’s conduct.” Taylor v. United States, 495 U.S. 575, 601, 110 S.Ct. 2143, 109 L.Ed.2d 607 (1990). In other words, without regard to the specific facts of each defendant’s offense, we compare the elements of the crime for which the defendant was previously convicted with Congress’s definition of the type of crime that may serve as a predicate offense. Under this approach, we “look[] only to the statutory definition of the state crime and the fact of conviction to determine whether the conduct criminalized by the statute, including the most innocent conduct, qualifies as a crime of violence.” Karimi v. Holder, 715 F.3d 561, 567 (4th Cir.2013) (internal quotation marks omitted); see also Aguiar, 438 F.3d at 89. For example, if a state defines the elements of burglary so as not to require unlawful entry, such that its statute encompasses both shoplifting and a classic midnight *6break-in of a bank, then under the categorical approach a conviction under that law is not considered to be a conviction for so-called “generic” burglary. Descamps v. United States, — U.S. -, 133 S.Ct. 2276, 2281, 2293, 186 L.Ed.2d 438 (2013) (defining the “generic” version of a crime as “the offense as commonly understood”); Shepard, 544 U.S. at 16-18, 125 S.Ct. 1254.1
A second set of rules recognizes an exception to the categorical approach. If an offense’s elements are overbroad — if, that is, they encompass conduct that does not require all the elements necessary to render the offense a predicate — we are sometimes authorized to apply the “modified” categorical approach. Under that approach, we first determine whether the prior conviction took place under a “divisible” statute. Descamps, 133 S.Ct. at 2281-82. A statute is divisible if it sets forth one or more elements of a particular offense in the alternative. Id. (“[A divisible] statute sets out one or more elements of the offense in the alternative — for example, stating that burglary involves entry into a building or an automobile.”). When confronted with such a statute, we are permitted to “consult a limited class of documents, such as indictments and jury instructions, to determine which alternative formed the basis of the defendant’s prior conviction.” Id.; see also Shepard, 544 U.S. at 17, 26, 125 S.Ct. 1254. We then analyze the prior conviction not in relation to all the statute’s elements, but instead in relation only to the narrower subset of elements that actually gave rise to the conviction. E.g., Descamps, 133 S.Ct. at 2281-82.
Third, in assessing whether the elements of the candidate proposed as a predicate crime are overbroad, we need not consider fanciful, hypothetical scenarios. See Gonzales v. Duenas-Alvarez, 549 U.S. 183, 193, 127 S.Ct. 815, 166 L.Ed.2d 683 (2007) (“[T]o find that a state statute creates a crime outside the generic definition of a listed crime in a federal statute requires more than the application of legal imagination to a state statute’s language. It requires a realistic probability, not a theoretical possibility, that the State would apply its statute.... ”); see also James v. United States, 550 U.S. 192, 208, 127 S.Ct. 1586, 167 L.Ed.2d 532 (2007) (“We do not view [the categorical approach] as requiring that every conceivable factual offense covered by a statute must necessarily present a serious potential risk of injury before the offense can be deemed a violent felony.”). But the two approaches remain stringent: they are governed by the basic principle that a state’s definition of a crime is overbroad if its elements allow for a conviction without satisfying the elements Congress has provided to define the required predicate offense. See generally Descamps, 133 S.Ct. at 2283-85 (“[I]f the statute sweeps more broadly than the generic crime, a conviction under that law cannot count as an ACCA predicate, even if the defendant actually committed the offense in its generic form.”); Taylor, 495 U.S. at 599, 110 S.Ct. 2143 (“If the state statute is narrower than the generic view ... the conviction necessarily implies that the defendant has been found guilty of all the elements of generic burglary. And if the defendant was convicted of burglary in a State where the generic definition has been adopted, with minor variations in terminology, then the trial court need find *7only that the state statute corresponds in substance to the generic meaning of burglary”). With these rules in mind, we turn now to analyzing whether any of the four crimes to which the government points qualifies as a crime of violence under section 16.
A. Daytime and Nighttime Breaking and Entering
Because the district court based its judgment on Fish’s prior convictions for “B & E Daytime Felony” under an unidentified statute and for breaking and entering a building in the nighttime with the intent to commit a felony, see Mass. Gen. Laws ch. 266, § 16, we begin our analysis by considering the applicability of section 16 to these offenses. The government’s brief on appeal argues that even though the records of the former conviction state only that Fish was convicted of a “B & E Daytime,” one could conclude from them that Fish had been convicted under a statute, Mass. Gen. Laws ch. 266, § 17, which requires as an element that a person lawfully in the structure broken into have been put in fear. Before oral argument, however, the government submitted a Rule 28(j) letter in which it withdrew that interpretation of the records of conviction. Then, at oral argument, the government said it had “trouble making sense of’ the records of conviction as they related to the statute, ultimately conceding that we should analyze the least culpable conduct that qualifies as daytime B & E, see Aguiar v. Gonzales, 438 F.3d 86, 88 (1st Cir.2006). Because that conduct overlaps in all material respects with nighttime B & E, we analyze the two offenses together.
Both daytime and nighttime B & E may be committed by breaking into a “building, ship, vessel or vehicle.” Id. at § 16; see also id. at § 18 (“... building, ship or motor vehicle or vessel ... ”). Presumably because the breaking need not involve the use of force, e.g., Commonwealth v. Burke, 392 Mass. 688, 688-90, 467 N.E.2d 846 (1984), but instead may involve simply walking through an unlocked door, see Commonwealth v. Tilley, 355 Mass. 507, 508, 246 N.E.2d 176 (1969), the government does not argue that either of Fish’s B & E convictions qualifies as a crime of violence under section 16(a), which is limited to felonies having “as an element,” the “use, attempted use, or threatened use of physical force.” We therefore limit our analysis to section 16(b), which applies to all felonies that, by their nature, “involve! ] a substantial risk that physical force against the person or property of another may be used.”
Though the applicability of section 16(b) to the two Massachusetts B & E offenses is a question of first impression in this circuit, our analysis does not take place on a blank slate. In United States v. Brown, 631 F.3d 573 (1st Cir.2011), we analyzed the nighttime B & E statute and held that, even as narrowed under the modified approach to include only “night-time burglary of a building,” nighttime B & E did not qualify as a “crime of violence” under the residual clause of the career offender provision of the sentencing guidelines, U.S.S.G. § 4B1.2.2 A year later, we held in United States v. Farrell, 672 F.3d 27, 37 (1st Cir.2012), that in light of Brown, a district court had committed plain error by holding that Massachusetts’ section 18, the daytime B & E statute, was a “violent felony” under the Armed Career Criminal Act, 18 U.S.C. § 924(e).
*8We based our holding in Brown almost entirely on the breadth of the “building” element under Massachusetts law. Acknowledging that the term “includes not just stores and office buildings but an array of structures — detached garages and storage facilities, for example — that may invite theft of property but would only rarely expose individuals to violence,” we found the “threat of violence” in “so broadly defined a universe” to be “fairly speculative.” 631 F.3d at 579. Then, in Farrell, when we considered the “building” element alongside the possibility of “ship” and “vessel” break-ins, we found that the Brown rationale “applie[d] with even more force.” 672 F.3d at 35. We noted that “happening upon a person is far less likely to take place while breaking and entering a vessel than it is while burglarizing a building.” Id. at 37.
The government concedes the correctness of Brown and Famll, but asks us to limit those holdings on the ground that, unlike section 16(b), neither ACCA nor the guidelines provision takes account of the risk of the use of force against property. This is a fair point. The problem, though, is that the Massachusetts offense plainly does not require any conduct that involves or substantially risks the use of force against property. Rather, it reaches such non-forceful acts as walking through an unlocked door without permission. See Tilley, 355 Mass, at 508, 246 N.E.2d 176 (“In this Commonwealth the opening of a closed but unlocked door or window is a breaking.” (internal quotation marks omitted)). And since we are limited to our common sense — the government has given us nothing else on which to rely — we must view it as entirely plausible that the offense frequently involves such conduct (which is presumably why police frequently remind property owners to lock doors and windows).
This conclusion likewise dooms the government’s final argument, that we should write off as not the “ordinary case” any application of the Massachusetts statutes to conduct that does not pose the relevant risks. Without an empirical foundation for its proposed application of the “ordinary case” approach, the government directs our attention to the Supreme Court’s suggestion in Leocal v. Ashcroft, 543 U.S. 1, 10, 125 S.Ct. 377, 160 L.Ed.2d 271 (2004), that generic burglary is the prototypical section 16(b) offense. At oral argument, the government pressed the analogy to Leocal, implying that Leocal’s discussion had turned on the risk of violence to property. But that opinion, though it discussed section 16(b) in great depth, did no such thing. Rather, the Supreme Court suggested that burglary of a building is a section 16 offense because it “involves a substantial risk that the burglar will use force against a victim,” not because it raises any concern about harm to property. Id. And since we already held in Brown and Farrell that the breaking and entering statutes at issue here are broader than generic burglary and do not present a requisite risk of the type with which Leo-cal was in fact concerned — that is, the risk of harm to persons — we fail to see how Leocal supports the government’s position.
Having twice determined that the Massachusetts breaking and entering statutes, applying as they do to nonviolent entries of rarely-occupied structures through unlocked doors or windows, do not necessarily involve conduct that would pose a risk of physical injury or of the use of force, we now hold that Fish’s prior convictions for daytime B & E and nighttime B & E are not categorically crimes of violence under section 16(b).
B. Assault and Battery with a Dangerous Weapon
The next offense to which the government points is the Massachusetts crime *9of assault and battery with a dangerous weapon (“ABDW”), Mass. Gen. Laws ch. 265, § 15A(b). The name of this offense marks it as a strong candidate for classification as a crime of violence. Indeed, convictions for ABDW often arise from the intentional use of dangerous force against another, causing serious injury. E.g., Commonwealth v. Vick, 454 Mass. 418, 910 N.E.2d 339 (2009) (shooting with intent to murder and causing serious bodily injury).
The government, with good reason, nevertheless declines to argue that ABDW qualifies under section 16(a). As we have noted, section 16(a) requires that a predicate offense have “as an element the use, attempted use, or threatened use of physical force.” The Supreme Court recently held, in the context of ACCA’s force clause, 18 U.S.C. § 924(e)(2)(B)(i), that “the phrase ‘physical force’ means violent force,” see Johnson v. United States, 559 U.S. 133, 140, 130 S.Ct. 1265, 176 L.Ed.2d 1 (2010), and we see no reason to think the same would not apply to the same phrase in section 16(a). And since ABDW may be accomplished by a mere “touching, however slight,” see United States v. Hart, 674 F.3d 33, 42 (1st Cir.2012), it does not have “as an element the use” of physical force.3 As a result, it is overbroad.
The government therefore focuses its argument on section 16(b), which contains no requirement that violent force be employed. Section 16(b) does, however, require a “substantial risk” that physical force “may be used” in the course of committing an offense. In theory, it might be possible to construe the reference to the “use[ 1” of force so broadly as to encompass offenses involving strict liability, negligence, or recklessness, so long as some adequate level of violent impact were involved. Just such a construction was urged on the Supreme Court in Leocal v. Ashcroft, 543 U.S. 1, 9, 125 S.Ct. 377, 160 L.Ed.2d 271 (2004), a section 16(b) case involving a Florida conviction for driving under the influence and causing bodily injury. The Court, however, rejected the government’s argument that “the ‘use’ of force does not incorporate any mens rea component.” Id. Rather, it reasoned, “ ‘use’ requires active employment,” because “[wjhile one may, in theory, actively employ something in an accidental manner, it is much less natural to say that a person actively employs physical force against another person by accident.” Id. at 9-10, 125 S.Ct. 377 (emphasis in original). Although the Supreme Court explicitly limited its reasoning to negligence-or-less crimes, LeocaVs rationale would seem to apply equally to crimes encompassing reckless conduct wherein force is brought to bear accidentally, rather than being actively employed. It is therefore not surprising that our sister circuits have concluded, with striking uniformity, that section 16(b) does not *10reach recklessness offenses.4 On the force of Leocal’s logic, we hold the same.
So the key question is whether Massachusetts ABDW allows convictions based on mere recklessness. The answer is clearly yes, as long as the recklessness causes non-trivial bodily harm. E.g., Commonwealth v. Burno, 396 Mass. 622; 487 N.E.2d 1366 (1986). Indeed, “[i]n Massachusetts, conduct that underlies a conviction for operating under the influence and causing serious bodily injury may also be charged as ABDW.” Hart, 674 F.3d at 43 n. 8 (1st Cir.2012). The government does not challenge the accuracy of this description of Massachusetts ABDW. Instead, the government argues that, in fact, Massachusetts ABDW is typically applied to conduct involving the active employment of force against another, so we should simply ignore, as not the “ordinary case,” convictions involving mere recklessness.
In support of this position, the government relies on United States v. Hart, 674 F.3d 33, 41-44 (1st Cir.2012), in which we determined that Massachusetts ABDW qualifies as a “violent felony” under ACCA’s residual clause. In analyzing ACCA’s applicability to the ABDW offense, we first found that ABDW posed a “serious risk of injury, comparable to the degree of risk posed by [ACCA’s] enumerated offenses.” Id. Clearly ABDW does, in all of its applications (and thus in the “ordinary case,” see James v. United States, 550 U.S. 192, 208, 127 S.Ct. 1586, 167 L.Ed.2d 532 (2007)), pose such a risk-even in its reckless form, which expressly requires injury that is “more than transient.” Burno, 396 Mass, at 627, 487 N.E.2d 1366; see also United States v. Glover, 558 F.3d 71, 81 (1st Cir.2009) (concluding that because ABDW requires as an element that a defendant have effected a touching with a dangerous weapon, the offense “ineluctably poses a serious potential risk of physical injury”). Equally clearly, and contrary to the dissent’s suggestion that section 16(b) “does not differ from the ACCA’s residual clause in any relevant respects,” see Dissenting Op. at 20, this is not the risk that must be assessed in a section 16(b) analysis. See Leocal, 543 U.S. at 10 n. 7, 125 S.Ct. 377 *11(holding that section 16(b) “plainly does not encompass all offenses which create a ‘substantial risk’ that injury will result from a person’s conduct”, because “[t]he ‘substantial risk’ in § 16(b) relates to the use of force, not to the possible effect of a person’s conduct”); Aguiar, 438 F.3d at 88.
Having determined that ABDW posed a sufficient risk of injury to qualify under ACCA’s residual clause, we proceeded, pursuant to the Supreme Court’s analysis in Begay v. United States, 553 U.S. 137, 142, 128 S.Ct. 1581, 170 L.Ed.2d 490 (2008), to inquire into whether ABDW was “roughly similar in kind to the [offenses enumerated in ACCA’s residual clause].” Hart, 674 F.3d at 43-44; see also Begay, 553 U.S. at 143, 128 S.Ct. 1581 (“[T]o give effect to every clause and word [of 18 U.S.C. § 924(e)(2)(B)(ii) ], we should read the [example crimes in section 924(e)(2)(B)(ii) ] as limiting the crimes that clause (ii) covers to crimes that are roughly similar, in kind as well as in degree of risk posed, to the examples themselves.” (internal citation and quotation marks omitted)). In order to satisfy Begay’s test for “rough[ ]” similarity to burglary, arson, extortion, and crimes involving the use of explosives — crimes that are listed in ACCA, but not in section 16(b) — an offense must “typically involve purposeful, violent and aggressive conduct.” Hart, 674 F.3d at 43-14 (quoting Begay, 553 U.S. at 144-45, 128 S.Ct. 1581) (internal quotation marks omitted). Over protest from the defendant to the effect that ABDW is occasionally applied to reckless conduct — and, in particular, to reckless driving causing injury — we found that such a fact pattern did not “represent the vast majority of ABDW convictions,” 674 F.3d at 44, and could therefore not defeat the conclusion that ABDW was “typically” purposeful, violent, and aggressive. Id.
We need not question Hart’s holding as to ABDW’s similarity to ACCA’s listed offenses. But that holding, based as it is on an inquiry into whether ABDW is “typically purposeful, violent, and aggressive,” cannot establish that ABDW satisfies section 16(b). To the extent that the “typically purposeful, violent, and aggressive” test requires that an offense involve purposefulness at all,5 the test looks only to the “usual circumstances of the crime.” See 674 F.3d at 44 (“ ‘Adjectives like “purposeful” and “aggressive” denote qualities that are ineluctably manifested in degree and appear in different combinations.’ ” (quoting United States v. Williams, 529 F.3d 1, 7 n. 7 (1st Cir.2008))). Section 16(b), by contrast, requires that an offense, in every realistically probable application, involve a substantial risk that physical force will be brought to bear in a manner such that it can be said to have been “used.” See Leocal, 543 U.S. at 8-12, 125 S.Ct. 377; Gonzales v. Duenas-Alvarez, 549 U.S. 183, 193, 127 S.Ct. 815, 166 L.Ed.2d 683 (2007) (“[To find that a state statute is overbroad] requires a realistic probability, not a theoretical possibility, that the State would apply its statute to conduct that falls outside the generic definition of a crime. To show that realistic probability, an offender, of course, may show that the statute was so *12applied ... ”); see also James v. United States, 550 U.S. 192, 208, 127 S.Ct. 1586, 167 L.Ed.2d 532 (2007) (citing Duenas-Alvarez with approval and noting that “[o]ne can always hypothesize unusual cases in which even a prototypically violent crime might not present a genuine risk of injury — for example, an attempted murder where the gun, unbeknownst to the shooter, had no bullets.”).
Finding no comfort in Hart’s holding, the government points out that our opinion in Hart employed language that can be read to go beyond what Begay required. Specifically, Hart’s analysis of the “purposeful, violent, and aggressive” test contains the following references to the “ordinary case”:
It is true that an ABDW conviction may rest on a recklessness theory, and it is not insignificant that reckless ABDW may be committed with a seemingly innocent object used in a dangerous fashion, as in the case of reckless, vehicular ABDW. But this fact pattern does not represent the vast majority of ABDW convictions, and our analysis under the residual clause is explicitly, and necessarily, limited to the “ordinary case.” James v. United States, 550 U.S. 192, 208, 127 S.Ct. 1586, 167 L.Ed.2d 532 (2007).
... In considering the ‘ordinary case [ ]’ of ABDW, James, 550 U.S. at 208, 127 S.Ct. 1586, we must conclude that a composite of purposeful, violent, and aggressive conduct is the norm. See Begay, 553 U.S. at 144-45, 128 S.Ct. 1581.
674 F.3d at 43-44 (some citations omitted). The government argues that this language, in combination with Hart’s citations to James, should be read as license to use the “ordinary case” approach to ignore reckless ABDW in determining whether Massachusetts ABDW satisfies the section 16(b) test. For the following reasons, we disagree.
As an initial matter, the license the government would draw from this language rests on dictum. As we have explained, Begay’s test for similarity to ACCA’s enumerated offenses was never intended to operate as a rigorous comparison between the conduct necessarily underlying a prior conviction and the conduct described in a recidivist statute. See Sykes v. United States, — U.S. -, 131 S.Ct. 2267, 2275, 180 L.Ed.2d 60 (2011) (“The phrase ‘purposeful, violent, and aggressive’ has no precise textual link to the residual clause.”). Rather, after first employing the categorical approach to define the elements of an offense without reference to the actual facts of a defendant’s conduct, Begay trains its focus on whether that offense is, in addition to meeting ACCA’s textual requirement, “roughly similar” to the offenses listed in ACCA, so as to avoid the absurd application of ACCA to crimes that “though dangerous, are not typically committed by those whom one normally labels ‘armed career criminals.’ ” Begay, 553 U.S. at 146, 128 S.Ct. 1581. Because our observation in Hart that reckless, vehicular ABDW “does not represent the vast majority of ABDW convictions” was enough to satisfy this peimissive standard, it was unnecessary to further inquire into whether the “ordinary case” of ABDW involves a risk of the “use” of physical force as required by Leocal. Any conclusion we drew as to that question would, as dictum, therefore not bind us here. See Kosereis v. Rhode Island, 331 F.3d 207, 213 (1st Cir.2003) (“Dicta, of course, is not binding on future panels.”); see also Diaz-Rodriguez v. Pep Boys Corp., 410 F.3d 56 (1st Cir.2005) (“[Although a newly constituted panel ordinarily may not disregard the decision of a previous panel, principles of stare decisis do not preclude us from dis*13claiming dicta in a prior decision.”); Pierre N. Leval, Judging Under the Constitution: Dicta About Dicta, 81 N.Y.U. L.Rev. 1249, 1263 (2006) (“Among the most common manifestations of disguised dictum occurs where the court ventures beyond the issue in controversy to declare the solution to a further problem — one that will arise in another case, or in a later phase of the same case.”).
The government concedes that Hart is, “to be sure, ... not dispositive” of this case, and we take that concession at face value. The dissent, however, overshoots the government, proposing that we transform the Begay test — an inquiry designed to narroiv ACCA’s application even when a crime, in all its actual applications, poses the risk that ACCA’s text requires — into one that broadens section 16(b)’s application. For this simple reason, we cannot accede to the dissent’s suggestion that Be-gay’s focus on the “usual circumstances” of an offense now binds us to conclude that everything outside those “usual circumstances” is, in James’s terms, “hypothesize[d].” See James, 550 U.S. at 208, 127 5.Ct. 1586.6 Neither James nor Begay suggests such an approach, and to adopt it would be to abandon section 16(b)’s requirements in favor of an ill-fitting and less demanding test designed to aecommo-date the text, purpose, and legislative history of a materially different statute. Such a result simply cannot be what Be-gay, which mentioned the “risk” inquiry only to demonstrate that it was not the inquiry at issue, had in mind.7
Our analysis under section 16(b) is therefore governed by James, Duenas-Al-varez, and Leocal, not by Begay. And in defining the “ordinary case” as it applies to the “risk” inquiry, James explains that sentencing courts may disregard only “hy-pothetize[d]” factual scenarios. 550 U.S. at 208, 127 S.Ct. 1586. Duenas-Alvarez, which James cites in the course of its explanation of the “ordinary case,” likewise permits exclusion only of applications that exist solely in “legal imagination.” 549 U.S. at 193, 127 S.Ct. 815. Unlike Begay’s “roughly similar” test, the analysis described in James and Duenas-Alvarez grants us no license to ignore actual cases on the ground that they are not “typical” or do not represent the “majority” of convictions. Thus, though we do not read our opinion in Hart as having gone out of its way, in cryptic dictum, to violate James and Duenas-Alvarez, we conclude that we would be bound to follow those two Supreme Court cases over any dictum the government might find to the contrary in *14Hart’s application of the less-demanding “typically purposeful, violent, and aggressive” test. See generally United States v. Dancy, 640 F.3d 455, 470 (1st Cir.2011).8
In so concluding, we acknowledge that at least one court has, in an analogous situation, relied on James to find license under the “ordinary case” approach to look only to what it imagined might be the typical case of conviction, in the process ignoring a state statute’s overbreadth even in the face of actual applications of the statute to conduct that failed to meet the textual requirements of the federal statute at issue. See, e.g., United States v. Mayer, 560 F.3d 948, 960-63 (9th Cir.2009).9 Such a freewheeling interpretation of James would seem to conflict not only with James and Duenas-Alvarez, but also with the Supreme Court’s recent decision in Descamps v. United States, — U.S. -, 133 S.Ct. 2276, 2285-86, 186 L.Ed.2d 438 (2013), which again reaffirmed that the only way a facially overbroad statute can qualify as an ACCA predicate is by application of the modified categorical approach. Though it is theoretically possible to read Descamps as having no application to the theory the dissent proposes, we think it unlikely that the Supreme Court took and decided the Descamps case, in which it yet again clarified the ornate rules that govern the categorical and modified categorical approaches, all in the service of a procedure that ends with the excision of real applications of broad offenses based on non-empirical determinations that they do not present the ordinary case.
We are guided here not merely by the thrust of Descamps, but by its language, as well. Descamps contains myriad warnings to the effect that “[wjhether the statute of conviction has an overbroad or *15missing element, the problem is the same: Because of the mismatch in elements, a person convicted under that statute is never convicted of the generic crime.” Id. at 2292. In this case, the dissent can avoid the application of that principle only by suggesting that we not consider whether the statute is overbroad until we have already whitewashed its overbroad, actual applications.
To adopt that approach would ensnare us into deciding how big a “minority” of actual convictions for unqualifying offenses under an overly broad definition we may permissibly ignore. One option, in theory, would be to find empirical tools for confidently gauging whether actual convictions met whatever definition of minority we might invent. See Mayer, 560 F.3d at 952 (Kozinski, C.J., dissenting from denial of rehearing en banc) (“Don’t even think about how a court is supposed to figure out whether a statute is applied in a certain way ‘most of the time.’ (A statistical analysis of the state reporter? A survey? Expert evidence? Google? Gut instinct?)”). The only alternative would be to wipe out the categorical approach and directly reject Descamps. The first option is impossible, the second foreclosed.
In view of the unavoidable complexity of the foregoing, we also consider a simple hypothetical. Imagine that Massachusetts defined the current elements of ABDW solely by statute, rather than in its case law. Keeping the elements the same, the statute would, in substance, read as follows:
Assault and Battery with a Dangerous Weapon is:
(1) The intentional and unjustified touching of another by use of a dangerous weapon,
or,
(2) The intentional commission of a wanton or reckless act causing more than transient or trifling injury to another.
See Hart, 674 F.3d at 42, 43 n. 7.
We do not understand the dissent or the government to go so far as to argue, counter to the law of ten circuits, that a conviction under part (2) of our hypothetical statute would serve as a predicate offense under section 16(b). See, e.g., Leocal, 543 U.S. at 9-10, 125 S.Ct. 377. And if a defendant’s conviction were simply for ABDW (as in the present case), with no indication as to whether the charge was under a particular subdivision, one would have to assume that the conviction might have taken place under part (2). Aguiar 438 F.3d at 89 (“[Ojnly the minimum criminal conduct necessary to sustain a conviction under a given statute is relevant.” (internal citation omitted)). So the question arises: given such a statute, and an actual conviction not specified as to whether it arose under part (1) or part (2), would the possibility of a conviction under part (2) be ignored as outside of the “ordinary case”? Clearly, the answer must be no.
If that is the case, then why would one reach a different result here? True, our hypothetical is easy because the elements are plainly defined by statute. But because the provenance of a crime’s elements tells one nothing about how the crime is committed, we see no reason why that fact should be decisive.
This hypothetical also serves to illustrate our reading of Hart. Absent any Shepard-wppmved documents telling us which provision of the hypothetical statute had given rise to a conviction, our analysis of the statute under ACCA would replicate Hart’s holding that ABDW is a violent felony under ACCA’s residual clause. Thus, we would first ask whether, in all but imagined, hypothetical circumstances, the statute involved a “serious potential *16risk of physical injury.” We would have to conclude that it did: section (2) makes injury an explicit textual requirement, and although section (1) does not explicitly require injury, it plainly requires conduct that creates a serious potential risk thereof. Under ACCA (unlike section 16(b)), we would then apply Begay’s similarity test to see whether ABDW should nevertheless be disqualified. Because that inquiry, unlike the “ordinary case” analysis, is satisfied so long as the “typical” violation of the statute involves purposefulness, we would, just as in Hart, find Begay satisfied. Cf. United States v. Johnson, 616 F.3d 85, 91 n. 4 (2d Cir.2010) (finding Begay’s “roughly similar” test satisfied even on the assumption that an “overwhelming majority” of convictions under a statute, but not all of them, “involve[d] violent and aggressive behavior”).
To summarize our analysis of ABDW: the elements of Massachusetts ABDW are satisfied when the intentional commission of a reckless act causes more than trifling injury; convictions for ABDW for such reckless conduct are not merely hypothetical possibilities, but instead actually occur; we agree with ten Circuits that reckless conduct bereft of an intent to employ force against another falls short of the mens rea required under section 16(b) as interpreted in Leocal; no Shepard-approved documents tell us that Fish’s ABDW conviction was not such an offense; therefore, his ABDW conviction is not a crime of violence under section 16(b). And in response to our learned colleague’s considered dissent, we agree with both Fish and the government that Hart does not dictate a contrary result. To the extent that Hart can be read as using the “ordinary case” notion of James to erase from our consideration of ABDW its actual applications to reckless conduct, we find such a construction of James to be unnecessary to Hart’s actual holding that Massachusetts ABDW survives examination under Begay’s similarity test. The similarity test requires only that an offense “typically” involve a purposeful use of force.10
Finally, the very complexity of the government’s attempt to prove that every person convicted of ABDW in Massachusetts is, per se, a violent offender, without any adjudication or admission necessitating the conclusion, should itself give us pause. If someone with Fish’s record had asked whether he could lawfully buy body armor, no one (other than five Supreme Court Justices) could have confidently answered the question. In such a case, we cannot simply combine intricate statutory interpretations with judicial hunches about the conduct underlying prior convictions in order to imprison as a violent felon one whose conduct no jury has necessarily found to satisfy the elements that make an offense a crime of violence as defined by Congress. See Leocal, 543 U.S. at 11 n. 8, 125 S.Ct. 377 (noting that because “ § 16 is a criminal statute”, “the rule of lenity ap*17plies”); cf. Alleyne v. United States, — U.S. -, 133 S.Ct. 2151, 2156, 186 L.Ed.2d 314 (2013) (“The Sixth Amendment ..., in conjunction with the Due Process Clause, requires that each element of a crime be proved to the jury beyond a reasonable doubt.”). We therefore hold that because ABDW, as defined by Massachusetts law, does not in form or application require a risk of the use of force, it is not a crime of violence as defined in section 16(b).
C. Burglarious Tools
The government points us last to Fish’s prior conviction under the Massachusetts statute prohibiting the making, possession, and use of burglarious instruments. That statute reads as follows:
Whoever makes or mends, or begins to make or mend, or knowingly has in his possession, an engine, machine, tool or implement adapted and designed for cutting through, forcing or breaking open a building, room, vault, safe or other depository, in order to steal therefrom money or other property, or to commit any other crime, knowing the same to be adapted and designed for the purpose aforesaid, with intent to use or employ or allow the same to be used or employed for such purpose, or whoever knowingly has in his possession a master key designed to fit more than one motor vehicle, with intent to use or employ the same to steal a motor vehicle or other property therefrom, shall be punished by imprisonment in the state prison for not more than ten years or by a fine of not more than one thousand dollars and imprisonment in jail for not more than two and one half year's.
Mass. Gen. Laws ch. 266, § 49. Fish argues that because the government never raised the burglarious instruments conviction until this appeal, we should not consider the offense. While we note the peculiarity of placing an appellate court in the position of finding facts to satisfy an element of an offense, we need not address Fish’s contention: We ultimately conclude, as perhaps the government did when it determined not to argue the issue in the district court, that the burglarious tools statute is overbroad, as well.
The problem for the government is that Massachusetts courts have made clear that a “tool or implement ... designed for ... breaking open a building, room, vault, safe or other depository” as described in the first clause of section 49 can be a master key, so long as the master key is not one for an automobile. Commonwealth v. Til-ley, 306 Mass. 412, 417, 28 N.E.2d 245 (1940) (“Keys expressly made to fit a particular lock for the purpose of wrongfully gaining access to a depository in which goods were kept, in order to steal them, are tools and implements of the kind and character described in the statute.”).11 Given the possibility that a defendant might be convicted of making or possessing a master key without any attempt to use it — a crime that strikes us as posing a relatively low risk of the ultimate use of physical force against persons or property — we cannot conclude that the burglarious tools offense, even as limited under the modified approach, categorically constitutes a crime of violence.
IV. Conclusion
It is no secret that the statutes Congress chose to enact in its understandable *18effort to focus on violent conduct are imperfect. See, e.g., Descamps v. United States, 133 S.Ct. 2276, 2293-94 (Kennedy, J., concurring) (“If Congress wishes to pursue its policy in a proper and efficient way without mandating uniformity among the States with respect to their criminal statutes for scores of serious offenses, and without requiring the amendment of any number of federal criminal statutes as well, Congress should act at once.”); Derby v. United States, — U.S. -, 131 S.Ct. 2858, 180 L.Ed.2d 904 (2011) (Scalia, J., dissenting from denial of certiorari and so criticizing ACCA’s residual provision); Sykes v. United States, — U.S. -, 131 S.Ct. 2267, 2295, 180 L.Ed.2d 60 (2011) (Kagan, J., joined by Ginsburg, J., dissenting and lamenting the Supreme Court’s difficulties in crafting a workable approach); Chambers v. United States, 555 U.S. 122, 131-32, 129 S.Ct. 687, 172 L.Ed.2d 484 (2009) (Alito, J., joined by Thomas, J., concurring in the judgment) (“[Ojnly Congress can rescue the federal courts from the mire into which ACCA’s draftsmanship and Taylor's ‘categorical approach’ have pushed us.”). As has been pointed out elsewhere, see, e.g., Sykes, 131 S.Ct. at 2284 (Scalia, J., dissenting), the great variation between the different states’ criminal statutes has flummoxed the federal courts. Though the duty here undertaken seems a better fit for Congress or an administrative agency, we have for now no choice but to do our best to give effect to Congress’s expressed intent.
As a result, our holding may appear odd to the reasonably discerning citizen, particularly from afar. Convictions under statutes with names connoting violence are sometimes deemed not to be crimes of violence, even if it is likely that most such convictions arise from violent conduct. This apparent anomaly arises largely because many states have stretched these violence-connoting rubrics to encompass conduct that Congress does not define as a crime of violence. Driving under the influence and accidentally causing serious injury thus gets grouped together with pistol-whipping a bank teller, and prosecutors and courts are left to choose between two unpalatable options: either we may deem non-violent individuals who likely are in fact violent, or we may falsely assume that every person convicted under an over-broad statute is in fact a violent criminal. Since the constitution prohibits us from charting the latter course, we will take the former unless Congress changes the law or the Supreme Court instructs otherwise.
Fish’s conviction is reversed, and the case is remanded for dismissal. So ordered.
. In the context of statutes other than section 16, courts are occasionally tasked with defining an offense by the full range of conduct it proscribes, inquiring not into whether that conduct is overbroad, but instead into whether it "typically” involves certain characteristics. See Begay v. United States, 553 U.S. 137, 128 S.Ct. 1581, 170 L.Ed.2d 490 (2008).
. The guidelines provision covers any offense that "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.” Id.
. As we explained in Hart,
Massachusetts ABDW may be committed (1) intentionally or (2) wantonly or recklessly. The former theory requires the intentional and unjustified use of force upon the person of another, however slight. The latter calls for the intentional commission of a wanton or reckless act (something more than gross negligence) causing physical or bodily injury to another. In the case of reckless or wanton ABDW, the victim's injury must be more than transient or trifling and severe enough to interfere with health or comfort.
674 F.3d at 43 n. 7 (citations and internal quotation marks omitted). Both theories of ABDW require that the offense involve the employment of a dangerous weapon, but the definition of "dangerous weapon” includes both items that are dangerous "per se” and otherwise innocuous items that, as used, are "capable of producing serious bodily harm.” Id. at 42-43 (citation and internal quotation marks omitted).
. See Jobson v. Ashcroft, 326 F.3d 367, 373 (2d Cir.2003); Tran v. Gonzales, 414 F.3d 464, 469-70 (3d Cir.2005) ("[U]se of force is an intentional act.”); Bejarano-Urrutia v. Gonzales, 413 F.3d 444, 447 (4th Cir.2005) ("[T]he conclusion of the Leocal Court that 'in no ordinary or natural sense can it be said that a person risks have to use physical force against another person in the course of operating a vehicle while intoxicated and causing injury' strongly indicates that the result in Leocal would have been the same even had a violation of the statute there at issue required recklessness rather than mere negligence.”); United States v. Chapa-Garza, 243 F.3d 921 (5th Cir.2001) (felony driving while intoxicated does not qualify under section 16 because it does not necessarily involve intentional use of force or recklessness as to the possibility of intentional use of force); United States v. Portela, 469 F.3d 496 (6th Cir.2006) ("[A] crime requiring only recklessness does not qualify as a 'crime of violence’ under 18 U.S.C. § 16.”); Jimenez-Gonzalez v. Mukasey, 548 F.3d 557, 560 (7th Cir.2008) ("Today we join our sister circuits and hold that reckless crimes are not crimes of violence under Section 16(b)”); United States v. Torres-Villalobos, 487 F.3d 607, 615 (8th Cir.2007) (reckless manslaughter not a "crime of violence” after Leocal)', Fernandez-Ruiz v. Gonzales, 466 F.3d 1121, 1129-30 (9th Cir.2006) (cited in Covarrubias Teposte v. Holder, 632 F.3d 1049 (9th Cir.2011) (intentionally discharging firearm with reckless disregard as to whether it will hit an inhabited dwelling is not a crime of violence)); United States v. Zuniga-Soto, 527 F.3d 1110, 1124 (10th Cir.2008) (“[Rjecklessness falls into the category of accidental conduct that the Leocal Court described as failing to satisfy the use of physical force requirement under either of § 16’s definitions of ‘crime of violence.’ ”); United States v. Palomino Garcia, 606 F.3d 1317, 1335-36 (11th Cir.2010).
. Though the phrase “purposeful, violent, and aggressive” would seem, on its face, to require purposefulness, violence, and aggression, it is by now well-established that the test may be satisfied by any offense that “contemplates purposefulness, but not necessarily conduct that is deliberately violent or aggressive as a matter of course.” Hart, 674 F.3d at 44 n. 9; see also United States v. Williams, 529 F.3d 1, 7 n. 7 (1st Cir.2008) (noting that even ACCA’s example crimes "satisfy [the 'purposeful, violent, and aggressive’] requirements only in some measure” and that drug trafficking crimes, which "involve purposeful conduct but are only sometimes violent or aggressive,” may satisfy Begay).
. The dissent suggests that "the examples of 'unusual' cases that James gave are not so farfetched.” See Dissenting Op. at 22. But James gave no examples of merely "unusual” cases. Rather, James made clear that the examples it provided were of "hypothesizefd] unusual” cases, see 550 U.S. at 208, 127 S.Ct. 1586 (emphasis added), provided in order to demonstrate only that "ACCA does not require metaphysical certainty” that a defendant’s underlying conduct would have met a federal recidivist statute's requirements. Id. (citing Duenas-Alvarez, 549 U.S. at 193, 127 S.Ct. 815). Notwithstanding our dissenting colleague's objection to the examples provided in James, that case’s warning against relying on imagined, hypothetical scenarios has no application here, because the defendant points to cases in which the ABDW statute has in fact been applied to conduct falling outside section 16(b)'s bounds.
. Moreover, we are unable to reconcile the dissent's suggestion that section 16(b) “does not differ from the ACCA's residual clause in any relevant respects,” Dissenting Op. at 20, with the Supreme Court's suggestion in Leo-cal that the two statutes are meaningfully distinct. See 543 U.S. at 10 n. 7, 125 S.Ct. 377; see generally Hohn v. United States, 524 U.S. 236, 252-53, 118 S.Ct 1969, 141 L.Ed.2d 242 (1998) (Supreme Court decisions "remain binding precedent” until that Court "see[s] fit to reconsider them”).
. In agreeing with the government that Hart is not dispositive of the case before us, we do not, as the dissent suggests, "apply[] the ordinary case rule differently to Section 16(b) than the ACCA.” Dissenting Op. at 21. To the contrary, we acknowledge that the ordinary case rule allows courts to disregard imagined, hypothetical scenarios when matching an offense to the two statutes' "risk” requirements. But what does not apply to section 16(b) (particularly to broaden it) is the assessment, under Begay, of what an offense "typically” involves. That assessment, which permits a court to look only to the usual circumstances under which an offense is committed, applies only to ACCA.
. Though our dissenting colleague also claims support in Delgado-Hernandez v. Holder, 697 F.3d 1125, 1129 (9th Cir.2012), and United States v. Johnson, 616 F.3d 85 (2d Cir.2010), see Dissenting Op. at 22-23 & nn. 15-16, neither opinion even feints toward an analysis different from the one we employ. In Delgado-Hemandez, the Ninth Circuit held that California's kidnapping statute was a crime of violence only after scouring reported cases to ensure that only by "adopting] a Pollyannaish outlook at the margins of the statute” could it "imagine” a scenario in which the offense did not involve at least "a substantial risk of force.” Delgado-Hernandez, 697 F.3d at 1129. And in Johnson, the Second Circuit applied James precisely as we understand it, concluding that Connecticut’s prison rioting statute applied (both in theory and in fact) only to conduct involving the requisite risk. See 616 F.3d at 94 ("Every violation of prison rules creates a risk that fellow inmates will join in the disturbance, oppose it with force, or simply use its occurrence to engage in other acts of violence.”). If the language our colleague quotes from Johnson seems inconsistent with that understanding of James, see Dissenting Op. at 23 n. 16, that is perhaps because the language is plucked not from the section of Johnson entitled “Similar in Degree of Risk Posed,” but instead from a separate section of the opinion-one entitled "Similar 'In Kind.' " See 616 F.3d at 89-93. The latter section, which makes not a single reference to the "ordinary case,” demonstrates little more than that like us, the Second Circuit understands the Begay inquiry to permit a court to look only to the usual circumstances of an offense. See supra note 8. Neither case contains any indication whatsoever that the same applies to either ACCA’s or section 16(b)’s "risk” requirement.
. Our dissenting colleague, proposing that we should "treat Begay and James interchangeably," points to two cases that he suggests have so held. See Dissenting Op. at 20 (citing United States v. Dismuke, 593 F.3d 582 (7th Cir.2010); United States v. Stinson, 592 F.3d 460, 466 (3d Cir.2010)). But neither of those cases supports our colleague’s conclusion that the James rule permits us to disregard actual applications of a statute to conduct that fails to meet ACCA's "risk” requirement. Rather, in Dismuke, the defendant conceded that the "risk” requirement was satisfied, thus taking it off the table completely. See 593 F.3d at 591 n. 3. And in Stinson, the Third Circuit concluded that although the language of Pennsylvania's resisting arrest statute suggested the possibility of overbroad application, the statute had never been so applied. See 592 F.3d at 466 ("[W]e have found no decision under Pennsylvania law that affirmed a conviction for resisting arrest based on a defendant’s inaction or simply 'lying down’ or 'going limp.' ”).
. Though the offense of possession of an automobile master key may no longer be charged under the first clause, see Commonwealth v. Collardo, 13 Mass.App.Ct. 1013, 1013-14, 433 N.E.2d 487 (Mass.App.Ct.1982), the government provides us no reason to conclude that the possession of a non-automobile master or duplicate key could not be charged under the first clause.