RECOMMENDED FOR FULL-TEXT PUBLICATION Pursuant to Sixth Circuit I.O.P. 32.1(b) File Name: 18a0141p.06 UNITED STATES COURT OF APPEALS FOR THE SIXTH CIRCUIT UNITED STATES OF AMERICA, Plaintiff-Appellee, ┐ │ │ > v. MALIK F. FARRAD, Defendant-Appellant. Nos. 16-5102/6730 │ │ │ │ │ ┘ Appeal from the United States District Court for the Eastern District of Tennessee at Knoxville. No. 3:14-cr-00110-1—Thomas A. Varlan, Chief District Judge. Argued: May 3, 2018 Decided and Filed: July 17, 2018 Before: MOORE, THAPAR, and BUSH, Circuit Judges. _________________ COUNSEL ARGUED: Michael M. Losavio, Louisville, Kentucky, for Appellant. Luke A. McLaurin, UNITED STATES ATTORNEY’S OFFICE, Knoxville, Tennessee, for Appellee. ON BRIEF: Michael M. Losavio, Louisville, Kentucky, for Appellant. Luke A. McLaurin, UNITED STATES ATTORNEY’S OFFICE, Knoxville, Tennessee, for Appellee. _________________ OPINION _________________ KAREN NELSON MOORE, Circuit Judge. Although the phrase “going undercover” may still connote high-stakes masquerading, twenty-first century undercover investigations can begin and end in cyberspace. This case stems from an undercover investigation on Facebook. Nos. 16-5102/6730 United States v. Farrad Page 2 The subject of the investigation, Defendant-Appellant Malik Farrad, was a felon prohibited by federal law from possessing a firearm. In June 2015, after a two-day trial, a jury found Farrad guilty of breaking that prohibition. No physical evidence was presented, no witness claimed to have seen Farrad with a gun, and Farrad himself never made any statements suggesting that he owned a gun; instead, the Government relied primarily on photographs obtained from what was evidently Farrad’s Facebook account. To help prove its case, however, the Government called two police officers: Officer Garrison, who testified that criminals are particularly likely to upload photos of criminal deeds soon after committing those deeds, and Officer Hinkle, who testified at length about the similarities between the photos and a real gun, as well as the dissimilarities between the photos and the closest fake gun of which he was aware. Farrad now challenges the sufficiency of the evidence, the admission of the photos into evidence, and the testimony that Officers Garrison and Hinkle were allowed to offer, as well as the district court’s denial of his motion for a new trial, his sentencing as an armed career criminal, and the district court’s failure to suppress the photos on Fourth Amendment grounds. For the reasons that follow, we AFFIRM. I. BACKGROUND A. Investigation, Indictment, and Pre-Trial Motions After serving time in prison for a previous felony, Farrad was released from federal custody in January 2013. See, e.g., R. 5-1 (Warrant Application at 3) (Page ID #9). Farrad came to the attention of local law enforcement sometime after June 10 of that same year, when “[v]arious confidential informants and concerned citizens” evidently “reported observing Farrad to be in possession of one or more firearms while in Johnson City, Tennessee.” Id. at 4 (Page ID #10). Some time later, a Johnson City police officer named Thomas Garrison, using an undercover account, sent Farrad “a friend request on Facebook.” R. 71 (Trial Tr. Vol. I at 81, 90–91) (Page ID #679, 688–89). After Farrad “accept[ed] the friend request,” Garrison was able to see more of Farrad’s photos. See id. at 81 (Page ID #679). “One [photo] in particular” “caught [his] interest”: a photo that showed what appeared to be three handguns “sitting on a closed toilet lid in a bathroom.” Id. at 81–82 (Page ID #679–80); see also Appellant’s App’x at Nos. 16-5102/6730 United States v. Farrad Page 3 6. The photo had been uploaded on October 7, 2013. R. 71 (Trial Tr. Vol. I at 83) (Page ID #681); Appellant’s App’x at 7. Garrison brought the photo to the attention of Johnson City police officer and FBI task force officer Matthew Gryder, who applied on October 25, 2013, for a warrant to search Facebook’s records for “information associated with the Facebook user ID MALIK.FARRAD.5.” R. 5-1 (Warrant Application at 4, 10–11) (Page ID #10, 16–17). A federal magistrate judge granted the warrant application. R. 5-4 (Search and Seizure Warrant at 1) (Page ID #25). The warrant mandated execution “on or before November 6, 2013,” id., and the return executed by federal law enforcement indicates that the warrant was “served electronically” on Facebook on November 1, 2013, id. at 2 (Page ID #26). The resulting data yielded a series of additional photos that are central to this case: some show a person who looks like Farrad holding what appears to be a gun, see, e.g., Appellant’s App’x at 11–14, while others show a closer-up version of a hand holding what appears to be a gun, see, e.g., id. at 16–26.1 While none of the photos shows a calendar, date, or one-of-a-kind distinguishing feature, the person in the photos has relatively distinctive tattoos, and some of the photos show, as backdrop, the décor of the room in which they were taken. See id. at 6, 11–12, 19–20, 22–23, 25–26. Facebook records revealed that the photos had been uploaded on October 11, 2013. See id. at 12, 15, 18, 21, 24, 26. In September 2014, a federal grand jury charged Farrad with having, “on or about October 11, 2013, . . . knowingly possess[ed] . . . a firearm, namely, a Springfield, Model XD, .45 caliber, semiautomatic pistol.” R. 3 (Indictment) (Page ID #3). On March 26, 2015, Farrad filed a pro se motion seeking an evidentiary hearing, dismissal of the indictment against him, and suppression of the Facebook photos on Fourth Amendment grounds. R. 22 (Pro Se Mot.) (Page ID #85–91). The magistrate judge assigned to Farrad’s case denied that motion on April 9, 2015, on the grounds that Farrad already had appointed counsel and the local rules prohibited a 1Another photo appeared to show guns, money, and marijuana on top of a stove. R. 28-2 (Gov’t Tr. Br., Ex. 2) (Page ID #115). This photo was never offered at trial. Appellee’s Br. at 13 n.5. Nos. 16-5102/6730 United States v. Farrad Page 4 represented party from “act[ing] in his or her own behalf” without “an order of substitution.” R. 24 (Order at 1) (Page ID #93) (quoting E.D. Tenn. Local Rule 83.4(c)). Farrad’s trial counsel did not renew Farrad’s motion. The parties did, however, litigate the admission of the photos on evidentiary grounds. The Government argued that the Facebook photos qualified as business records under Federal Rule of Evidence 803(6) and that they were, as such, self-authenticating under Federal Rule of Evidence 902(11). R. 26 (Gov’t’s Mot. in Limine at 1) (Page ID #100). In support of its assertion, the Government introduced a certification by a Facebook-authorized records custodian, who attested that the records provided by Facebook—including “search results for basic subscriber information, IP logs, messages, photos, [and] other content and records for malik.farrad.5”—“were made and kept by the automated systems of Facebook in the course of regularly conducted activity as a regular practice of Facebook” and “made at or near the time the information was transmitted by the Facebook user.” R. 26-1 (Facebook Certification) (Page ID #105). In addition to disputing admissibility under Federal Rules of Evidence 401, 402, 403, 404, 405, and 406, R. 30 (Def.’s Response to Gov’t’s Tr. Br. at 3) (Page ID #119), Farrad’s trial counsel argued that the photos, despite the custodian’s affidavit having been “done correctly under the federal rules,” were “hearsay within hearsay” and did not “authenticate who took the pictures, when the pictures were taken, by whom, at what time,” R. 60 (Pretrial Conf. Tr. at 11) (Page ID #456). All that the custodian could attest to, trial counsel emphasized, was “that at some point these pictures were uploaded to what [was] allegedly [Farrad’s] Facebook account”; the custodian could not “testify as to . . . who took [the photos], when they were taken, where they were taken.” Id. at 12 (Page ID #457). On June 15, 2015, the district court concluded that it had “found no indication of a lack of trustworthiness” and that the photos qualified as business records under Rules 803(6) and 902(11). See R. 73 (Pretrial Hr’g Tr. at 33) (Page ID #872). It also determined that the photos were relevant. See id. at 35–37 (Page ID #874–76). B. Trial Trial began the next day. At trial, the jury first heard from Garrison, who not only detailed his discovery of the precipitating toilet-seat photo, R. 71 (Trial Tr. Vol. I at 83) (Page ID #681), but also, in his capacity as an experienced user of social media in the service of police Nos. 16-5102/6730 United States v. Farrad Page 5 investigations, see id. at 80 (Page ID #678), discussed broader trends in how people who have committed crimes behave on social-media platforms. After Garrison conceded that users “can upload . . . pictures to Facebook that were taken at different times,” id. at 84 (Page ID #682), the following exchange occurred: GOVERNMENT: Okay. Now, in your training and experience and drawing upon the hundreds of cases you said you’ve been involved in using social media, when you come across people involved in criminal conduct who have uploaded photographs to their social media account, how quickly do they do that, relative to when that photograph is actually taken? DEFENSE COUNSEL: Object as speculation, Your Honor. THE COURT: Your response. GOVERNMENT: Your Honor, Mr. Garrison has testified about his training in social media investigations, hundreds of cases he said. He can certainly testify as to what his experience has been when people upload photographs in the past, other investigations he’s been involved in. THE COURT: objection. With those parameters in mind, the Court will overrule the GARRISON: Generally, in my experience, it’s been more of a—you know, like I say, it can be instantaneous. But it is more of a present-type of thing. Id. Although Garrison admitted that he could not “think of a specific instance” in which he had “talked to a target about specific things on Facebook” or recall “specific instances of training” on the subject, id. at 85 (Page ID #683), the Government asked him to explain why “people choose to post criminal conduct that they’re involved in, to social media,” id. at 86 (Page ID 684). This exchange followed: GARRISON: I would say that it mirrors the same reason that—that people in general post things on Facebook. But criminals specifically, they like to brag about their—their activities, they’re proud of it, and just like anyone, they want to let their friends know what they’re doing, let their friends know, you know, where they’re at, what’s going on. GOVERNMENT: In terms of your training and experience, is it more likely or less likely in terms of all of the cases you’ve seen before, that someone uploads those photographs immediately as opposed to sitting and waiting to upload them weeks or months or years later? Nos. 16-5102/6730 United States v. Farrad Page 6 GARRISON: In my experience, I would consider that more likely. GOVERNMENT: Now, in the context of social media applications that someone uses on their cell phone, say, does the use of those applications on someone’s cell phone make it more or less likely in your training and experience that those photographs are going to be uploaded at the time they’re actually created? Id. Defense counsel again objected on speculation and relevancy grounds, and the district court again allowed the questioning to continue “within the limited circumstances of [Garrison’s] experience.” Id. at 86–87 (Page ID #684–85). Garrison answered that “the apps makes it easier and more likely for someone to immediately upload a photograph taken with their cell phone.” Id. This line of questioning then concluded with the following exchange: GOVERNMENT: In contrast, then, how many times relative to instances in which people have uploaded immediately, contrasted that [sic] to the number of times you’ve seen photographs depicting criminal conduct that have been held back for extended periods of time, weeks, months or years? GARRISON: I would—I would consider that more rare. I can’t—I can’t think of a specific number or a—or an instance just off the top of my head. Id. at 87–88 (Page ID #685–86). Garrison also explained via direct examination that while digital photographs generally contain “metadata” that preserves, for example, “the time and date that the photograph was taken,” “Facebook actually strips that metadata as the photograph is uploaded to Facebook,” rendering the date of creation unknown. Id. at 89 (Page ID #687).2 The jury also heard about and saw the photos obtained from Farrad’s Facebook account, chiefly through testimony by Gryder. See id. at 94–109 (Page ID #692–707). Gryder identified Farrad and, while introducing the photos, noted that each photo had come from a Facebook account identified as “Malik.Farrad.5,” registered to a “Malik Farrad” with the email address “AllaFarrad@gmail.com,” and associated with Knoxville, Tennessee. Id. at 96–98 (Page ID #694–96). As Gryder explained, aside from the toilet-seat photo, all of the Facebook photos had been uploaded on October 11, 2013. Id. at 99–104 (Page ID #697–702). Gryder further explained that he had visited Farrad’s residence as part of his investigation and that he had 2The parties did stipulate, however, that the photographs at issue “were taken and created after May 29th, 2012, and after [Farrad] was convicted of a crime punishable by imprisonment for more than one year.” R. 71 (Trial Tr. Vol. I at 109–10) (Page ID #707–08). Nos. 16-5102/6730 United States v. Farrad Page 7 learned that Farrad had occupied two different units in the same complex, first residing in Apartment 15 and then moving to Apartment 9. Id. at 105 (Page ID #703). Gryder then identified Apartment 15 as the backdrop of several of the photos, noting an apparent match with the “mirror in the bathroom, . . . the color of the door, and the color of the walls,” as well as with a distinctive paint job on “the edge of [a] door frame.” Id. at 107 (Page ID #705). The jury later learned from the property manager of the apartment complex that Farrad had lived in Apartment 15 from February 6, 2013, until October 15, 2013. Id. at 119–20 (Page ID #717–18). The jury also heard from Morristown police officer Kenneth Hinkle, who served as his department’s armorer, had been a gunsmith “for over 30 years,” “started apprenticing” when he was thirteen years old, and had “handled or worked on” “[t]housands” of firearms over the years. Id. at 124–25 (Page ID #722–23). With reference to the Facebook photos, a real Springfield XD .45 caliber handgun, and photos of a real Springfield XD .45 caliber handgun, Hinkle pointed out various commonalities that led him to conclude that the item in the Facebook photos was a real Springfield XD .45 caliber handgun.3 See id. at 126–46 (Page ID #724–44). These commonalities included, for example, distinctive symbols and markings on the gun’s slide, id. at 131–32 (Page ID #729–30), a common lever above the trigger guard and ridges across that lever, id. at 137 (Page ID #735), cocking serrations along the rear left side of the slide, id. at 139–40 (Page ID #737–38), and the shape and position of the firing pin on the rear of the slide, id. at 143–44 (Page ID #741–42). Compare Appellant’s App’x at 16–17, 26, with id. at 34, 36, 38. Hinkle then discussed the possibility of “any replicas, toys, fakes, airsoft[]” versions of the Springfield XD .45 caliber handgun. R. 71 (Trial Tr. Vol. I at 146) (Page ID #744). Hinkle testified that he had “not been able to locate, during any of [his] research, a toy, airsoft, BB firearm of the XD, XD series pistol in any caliber.” Id. The Government then asked if he knew whether Springfield had “ever provided the licenses to manufacture an imitation of this firearm to another company.” Id. Hinkle responded: “No, sir, they have not.” Id. 3Real Springfield XD .45 caliber handguns are “manufactured in . . . Croatia,” Hinkle explained, R. 71 (Trial Tr. Vol. I at 127) (Page ID #725), thus putatively satisfying the interstate-commerce element of 18 U.S.C. § 922(g)(1). Nos. 16-5102/6730 United States v. Farrad Page 8 Hinkle proceeded to compare the Facebook photos against real and photographic versions of what he testified was the nearest non-gun comparator to the Springfield XD .45 caliber handgun: an airsoft replica of a Springfield XD(m) .40 caliber handgun. See id. at 147–50 (Page ID #745–48). He noted, for example, distinctions between the muzzles, including the presence of an orange tip, id. at 151–52 (Page ID #749–50), and between the cocking serrations, id. at 154–55 (Page ID #752–53). He compared the shape of the rear of the replica XD(m) .40 to the shape of the rear of the real XD .45 and the shape of the rear of the gun in the Facebook photos, noting that the former was angled along the side whereas the latter two were boxier. See id. at 155–58 (Page ID #753–56). Compare Appellant’s App’x at 16–17 (Facebook photos, boxy), with id. at 36–38 (real XD .45, boxy), with id. at 42–44 (replica XD(m) .40, trapezoidal). At the end of Hinkle’s direct examination, Hinkle concluded that the two were “totally different firearms” and suggested again that the image in the photos must be “an actual firearm” because “[t]hey never made an airsoft.” Id. at 159 (Page ID #757). Although Farrad’s attorney attempted to get Hinkle to admit on cross examination that he could not tell from photographs whether something was in fact a “real firearm,” Hinkle insisted that he could tell from the photographs alone.4 Id. at 169 (Page ID #767). Following Hinkle’s testimony, the Government rested, and Farrad moved without argument for a directed verdict. Id. at 186 (Page ID #784). Following the Government’s equally cursory opposition and a brief recitation of the standard, the district court denied Farrad’s motion. Id. at 187 (Page ID #785). Farrad did not testify, nor did the defense call any other witnesses. See R. 72 (Trial Tr. Vol. II at 7) (Page ID #795). Closing arguments occurred on June 17, 2015. The Government began its closing argument by noting that the offense at issue was “charged to have happened on or about October the 11th, 2013, . . . because that was the date the vast majority of the photos . . . were uploaded to Facebook by the defendant.” Id. at 9 (Page ID #797). The Government argued that while it could not prove exactly when the photos were uploaded because “Facebook strips all of that data out of the photographs that it receives when people upload them,” Garrison’s testimony “based 4Hinkle did admit that he could not see every single requisite marking or a serial number in the photos. R. 71 (Trial Tr. Vol. I at 176–77) (Page ID #774–75). Nos. 16-5102/6730 United States v. Farrad Page 9 upon his training and experience” was that “when people are involved in criminal conduct and then they document that conduct in photos or videos, they’re uploading it right away.” Id. The Government also suggested that Garrison’s testimony was “in line with what we do on a daily basis.” Id. at 10 (Page ID #798). Farrad’s trial counsel argued, meanwhile, that the Government’s case was built on “[p]ure speculation,” and that people project false images of themselves all the time, for all sorts of reasons—for example, employing “a sign out front that says this house is protected by a security system” when in fact no such security system exists, or a sticker that says “[t]his house protected by Smith & Wesson” when in fact no firearm exists. Id. at 14–15 (Page ID #802–03). Defense counsel also emphasized that there was no physical evidence connecting Farrad to a real firearm, argued that the lack of physical evidence belied the Government’s argument, and questioned whether the photos were clear enough or Hinkle believable enough to conclude beyond a reasonable doubt that the photos in fact showed a real firearm. Id. at 17–21 (Page ID #805–09). The district court then instructed the jury, noting among other directives that because “[t]he indictment charges the crime happened on or about October 11, 2013, . . . the government does not have to prove that the crime happened on that exact date, but the government must prove that the crime happened reasonably close to that date.” Id. at 33 (Page ID #821). Farrad summarily renewed his motion for a judgment of acquittal after the jury had been excused for deliberations, which motion the district court again summarily denied. Id. at 44–45 (Page ID #832–33). The jury found Farrad guilty. R. 40 (Verdict Form) (Page ID #159). C. Sentencing and Motion for New Trial Prior to sentencing, the probation department determined that Farrad qualified as an armed career criminal under 18 U.S.C. § 924(e) (“the ACCA”) based on a Tennessee conviction for simple robbery and federal convictions for eight counts of distribution and intent to distribute crack cocaine. See R. 53 (Revised Presentence Investigation Report (“PSR”) at 6–9) (Page ID #411–14). Farrad objected, arguing (among other issues) that the record did not establish that his eight drug-trafficking convictions were committed on different occasions, that the application of Nos. 16-5102/6730 United States v. Farrad Page 10 the ACCA to him at sentencing violated his Fifth Amendment right to due process and Sixth Amendment right to a jury trial, and that his eight drug-trafficking convictions did not categorically qualify as serious drug offenses under the ACCA. R. 47 (Def.’s Objections to PSR) (Page ID #311–28); R. 48 (Def.’s Supp. Objections to PSR) (Page ID #515–24). On January 14, 2016, the district court rejected each of these contentions. R. 65 (Sentencing Tr. at 53–65) (Page ID #560–72). After applying the sentencing factors to Farrad’s case, the district court pronounced a sentence of 188 months of imprisonment. Id. at 73 (Page ID #580); see also R. 56 (Judgment at 1) (Page ID #423). One week later, Farrad moved pro se for a new trial, arguing in part that the Government had presented “false and misleading perjured testimony to the jury” in the form of Hinkle’s claims regarding the existence of fake-gun versions of the Springfield XD .45. R. 59 (Pro Se Mot. for New Trial at 6–12) (Page ID #439–40). The district court appointed counsel to assist Farrad, R. 70 (Mem. & Order) (Page ID #597–98), and counsel filed a reply that—potentially contrary to Hinkle’s testimony—pointed out two websites offering to sell “inert replicas or simulated versions of the Springfield [XD] .45” produced by a company called Ring Manufacturing, R. 88 (Def.’s Reply re Mot. for New Trial at 1–2) (Page ID #932–33); see also R. 88-1 (Ex. 1, Def.’s Reply re Mot. for New Trial) (Page ID #936–37) (website printouts). On November 15, 2016, the district court denied Farrad’s motion, stating that it could not “find that Hinkle’s testimony was false and that the government knew it was false,” given both that Farrad had “provide[d] no evidence as to the date Ring Manufacturing began producing the firearm replicas of the Springfield .45 caliber Model XD” and that no evidence had been presented suggesting knowledge of any falsity. R. 89 (Mem. Op. & Order at 11) (Page ID #948). The district court also concluded that any falsity was nevertheless immaterial, given that the replicas were “painted bright blue for quick recognition as inauthentic” and therefore “would be easily distinguishable from an actual firearm and easily distinguishable from the firearms in the photographs.” Id. at 11–12 (Page ID #948–49). Farrad filed timely notices of appeal following both (1) the entry of final judgment after sentencing, R. 58 (First Notice of Appeal) (Page ID #433), and (2) the denial of his motion for Nos. 16-5102/6730 United States v. Farrad Page 11 new trial, R. 90 (Second Notice of Appeal) (Page ID #951). Those appeals have since been consolidated and are now before this court. II. DISCUSSION Farrad raises seven arguments on appeal: (1) that there was insufficient evidence introduced at trial to support his conviction; (2) that the Facebook photos should not have been admitted into evidence; (3) that Officers Hinkle and Garrison should not have been permitted to testify as experts; (4) that the district court should have granted Farrad’s motion for a new trial; (5) that Farrad did not in fact qualify as an armed career criminal under the ACCA; (6) that finding him to be an armed career criminal at sentencing violated his Fifth and Sixth Amendment rights; and (7) that the district court should have excluded the Facebook photos on Fourth Amendment grounds.5 We consider each in turn. A. Sufficiency of the Evidence “We review a challenge to the sufficiency of the evidence de novo, considering ‘whether, after viewing the evidence in the light most favorable to the prosecution, any rational trier of fact could have found the essential elements of the crime beyond a reasonable doubt.’” United States v. Tocco, 200 F.3d 401, 424 (6th Cir. 2000) (quoting Jackson v. Virginia, 443 U.S. 307, 319 (1979)). “A defendant making such a challenge bears a very heavy burden,” id., especially given that “[c]ircumstantial evidence alone is sufficient to sustain a conviction,” United States v. Spearman, 186 F.3d 743, 746 (6th Cir. 1999) (citation omitted), cert. denied, 528 U.S. 1033 (1999), and a “jury may draw any reasonable inferences from direct, as well as circumstantial, proof,” Tocco, 200 F.3d at 424. Accordingly, we “will reverse a judgment for insufficiency of evidence only if this judgment is not supported by substantial and competent evidence upon the 5Farrad also briefly argues, as part of this final claim, that if the problem is that Farrad’s trial counsel never renewed the motion that the magistrate judge denied, then Farrad has a viable claim for ineffective assistance of counsel. Appellant’s Br. at 69. But as the Government points out, Appellee’s Br. at 55–56, “[a]s a general rule, a defendant may not raise ineffective assistance of counsel claims for the first time on direct appeal, since there has not been an opportunity to develop and include in the record evidence bearing on the merits of the allegations.” United States v. Martinez, 430 F.3d 317, 338 (6th Cir. 2005) (quoting United States v. Wunder, 919 F.2d 34, 37 (6th Cir. 1990)). Particularly given the scant briefing devoted to this question, we hew to that general rule here. See, e.g., id. Farrad may, of course, bring this claim in a 28 U.S.C. § 2255 motion for postconviction relief. Nos. 16-5102/6730 United States v. Farrad Page 12 record as a whole,” regardless, of “whether the evidence is direct or wholly circumstantial.” United States v. Stone, 748 F.2d 361, 363 (6th Cir. 1984). “To obtain a conviction pursuant to § 922(g)(1), the government must prove beyond a reasonable doubt: (1) that the defendant has a prior conviction for a crime punishable by imprisonment for a term exceeding one year; (2) that the defendant thereafter knowingly possessed the firearm and ammunition specified in the indictment; and (3) that the possession was in or affecting interstate commerce.” United States v. Schreane, 331 F.3d 548, 560 (6th Cir. 2003) (quoting United States v. Daniel, 134 F.3d 1259, 1263 (6th Cir. 1998)). In addition, “[w]hen ‘on or about’ language is used in an indictment, proof of the exact date of an offense is not required as long as a date reasonably near that named in the indictment is established.” United States v. Ford, 872 F.2d 1231, 1236 (6th Cir. 1989); see also United States v. Harris, 293 F.3d 970, 975 (6th Cir. 2002) (“Under 18 U.S.C. § 922(g), the government must prove beyond a reasonable doubt that (1) the defendant had a previous felony conviction and (2) on or about the night in question, the defendant possessed a firearm.”). “Under § 922(g), either actual or constructive possession is sufficient.” Harris, 293 F.3d at 975. As the foregoing helps make clear, there are a few potential theories on which Farrad’s trial and appellate counsel could have challenged his indictment and conviction, respectively. First, could any rational juror have concluded beyond a reasonable doubt that the photos were taken on or about October 11? (Call this the “date” theory.) Second, could any rational juror have concluded beyond a reasonable doubt that the item in the photos was not a fake gun? (Call this the “replica” theory.) Third, could any rational juror have concluded beyond a reasonable doubt that the photos were not altered to suggest the appearance of a real gun? (Call this the “Photoshop” theory.) Fourth, could any rational juror have concluded that Farrad was in fact the person in the photos? (Call this the “lookalike” theory.)6 We discuss each in turn. 6Farrad also gestures at another possible theory: that there is “no evidence of who made or uploaded the picture.” Appellant’s Br. at 40. Where the charge is one of possession, however, this question goes more to authenticity (and thus admissibility) than it does to sufficiency of the evidence, and we therefore discuss it below with reference to Farrad’s evidentiary claims. With regard to sufficiency, meanwhile, this question is subsumed by the Photoshop theory; aside from potential alteration (which itself goes more to admissibility), we are aware of no reason why it would matter who took the photo, so long as a rational juror could conclude that Farrad was actually in the photo holding a real gun. Nos. 16-5102/6730 United States v. Farrad Page 13 1. Date Theory As noted above, and as the jury was instructed here, “[w]hen ‘on or about’ language is used in an indictment, proof of the exact date of an offense is not required as long as a date reasonably near that named in the indictment is established.” Ford, 872 F.2d at 1236; R. 72 (Trial Tr. Vol. II at 33) (Page ID #821); see also R. 3 (Indictment) (Page ID #3). In other words, sufficient proof of “a date reasonably near” October 11, 2013, was required to support Farrad’s conviction. See, e.g., Harris, 293 F.3d at 975; United States v. M/G Transp. Servs., Inc., 173 F.3d 584, 588–89 (6th Cir. 1999); see also United States v. Grubbs, 506 F.3d 434, 439 (6th Cir. 2007) (“[I]t is not enough that the defendant possessed a firearm at some unidentified point in the past; the evidence must prove that the defendant possessed the same handgun ‘identified in the indictment.’” (quoting United States v. Arnold, 486 F.3d 177, 183 (6th Cir. 2007) (en banc))). Here, Farrad was charged with having knowingly possessed a Springfield XD .45 “on or about October 11, 2013.” R. 3 (Indictment) (Page ID #3). The jury heard that nearly all of the Facebook photos in evidence had been uploaded on that day (with the other having been uploaded four days earlier), R. 71 (Trial Tr. Vol. I at 83, 99–104) (Page ID #681, 697–702), and the jury also heard Garrison’s testimony that, in his training and experience, “criminals specifically . . . like to brag about their . . . activities,” and thus tend to upload inculpatory materials near in time to commission of criminal acts, id. at 86 (Page ID 684). The jury also heard, however, that Facebook users “can upload . . . pictures to Facebook that were taken” at any given time, id. at 83 (Page ID #681), and that Facebook “strips [a photograph’s] metadata as the photograph is uploaded to Facebook,” making it impossible to tell when an uploaded photo was actually taken, id. at 89 (Page ID #687). And the jury likewise heard that while the backdrop of some of the Facebook photos matched Farrad’s apartment, Farrad had lived in that apartment as early as February 6, 2013. Id. at 119–20 (Page ID #717–18). In other words, there was a potential gap of as long as eight months between when Farrad could have taken the photo and the date alleged in the indictment. Regardless of the merit of the date theory, however—a question on which we express no opinion—the undisputable problem here for Farrad is that his trial counsel never focused on this Nos. 16-5102/6730 United States v. Farrad Page 14 theory, see R. 71 (Trial Tr. Vol. I at 17–21) (Page ID #802–11),7 or filed any motions relating to it. And his appellate counsel, moreover, eschewed this argument in his briefing and then disclaimed it at oral argument.8 Oral Arg. at 9:05–9:22, 11:56–12:14, 29:22–29:37, 31:39– 32:17. Accordingly, while the date issue is relevant to our opinion for reasons discussed below, see section II.C.2 infra, it is not an argument that could support relief here. 2. Replica Theory Farrad’s trial counsel and appellate counsel both focused, instead, primarily on the replica theory. See R. 71 (Trial Tr. Vol. I at 17–21) (Page ID #802–04); Appellant’s Br. at 40–41. But there are two big problems with the replica theory—one general and legal, the other specific and factual—in Farrad’s case. Even if Farrad could overcome the general and legal problem—an issue on which we express no ultimate opinion—the facts of his case doom the argument. The legal problem is this: Farrad is not the first person in our circuit to argue that what appears in images to be a gun is really a sophisticated replica, though he may be the first to argue it in this particular context. If a defendant is convicted of armed robbery, however, and claims the gun that witnesses saw was merely a convincing fake, reasonable jurors may infer—at least when corroborated by images or the testimony of a witness with law-enforcement experience— that the gun was real. See, e.g., United States v. Cobb, 397 F. App’x 128, 131–32 (6th Cir. 2010) (upholding conviction based on witness testimony and law-enforcement analysis of video); United States v. Conner, 306 F. App’x 978, 981–82 (6th Cir. 2009) (upholding conviction based on testimony of three eyewitnesses alongside video and photographs from surveillance camera); United States v. Crowe, 291 F.3d 884, 887 (6th Cir. 2002) (upholding conviction based on 7Trial counsel did make one statement in closing argument that can theoretically be understood as having invoked the date theory. See R. 72 (Trial Tr. Vol. II at 17) (Page ID #805) (“The government’s own witness, first witness again, when they take the photos, they’re going to upload them really quick. Okay. If that’s the case, where are they? If they were taken in that apartment in Tennessee, where are they?”). But we cannot say that this glancing reference qualified as having raised the issue. 8Appellate counsel also suggested at oral argument that trial counsel had “stipulated that the photos were made at or about the time set out in the indictment,” id. at 9:05–9:22; see also id. at 31:47–31:59. As noted, the actual stipulation of the parties was that the “photographs seized from Facebook . . . were taken and created after May 29, 2012”—“after the defendant was convicted of a crime punishable by imprisonment for more than one year.” See R. 71 (Trial Tr. Vol. I at 109–10) (Page ID #707–08); see also Oral Arg. at 11:56–12:14 (Government reporting same). Nos. 16-5102/6730 United States v. Farrad Page 15 testimony of eyewitness law-enforcement agent with “extensive experience and training in handling firearms”). “Indeed, ‘[t]he mere possibility that the object seen by witnesses may have been a sophisticated toy or other facsimile does not necessarily create a reasonable doubt, nor is the government required to disprove that theoretical possibility.’” Crowe, 291 F.3d at 887 (quoting United States v. Jones, 16 F.3d 487, 491 (2d Cir. 1994)). If Farrad is like the bank robber who later protests that what his victims saw was not real, his claim fails. There are reasons to believe that Farrad might not be sufficiently comparable to the bank robber in this analogy. After all, in the armed-robber cases, there is a clear (if often unstated) reason why any reasonable juror could credit witness testimony and image-based evidence to conclude beyond a reasonable doubt that a robber carried a real gun: robberies are violent and dangerous undertakings, and the fact that someone undertook to rob with what appears to be a weapon makes it somewhat more likely that they undertook to rob with an actual weapon. Cf. United States v. Medved, 905 F.2d 935, 940 (6th Cir. 1990) (discussing the dangers that arise in an armed robbery regardless of whether a weapon is or merely appears real). For much the same reason, people who show up to book-club meetings usually bring real, rather than fake, copies of the assigned book; they could bring a fake copy with blank pages inside, but the point is generally to bring a real one. Facebook photos are, at least arguably, different. As Farrad’s trial counsel noted in closing argument, R. 72 (Trial Tr. Vol. II at 14–15) (Page ID #802–03), people advertise false images of themselves all the time—putting signs on their doors suggesting that they have a fancy alarm system, a fierce guard dog, or a high-powered firearm that they do not in fact have. The Facebook poseur is in this sense not like the bank robber: his incentive is not necessarily to show a real gun, which is costlier, harder to procure, and, needless to say, more likely to subject its owner to criminal liability. The simple fact that a person is taking photos and posting them to Facebook does not necessarily allow the kind of behavior-based inference that armed robbery allows. In any event, however, the replica theory falters on the facts, because this is not a case in which a jury was asked to draw a conclusion from a few grainy or ill-lit photos, or asked to guess at an object out of focus or in the distance. Rather, the jury in Farrad’s case saw seven different Nos. 16-5102/6730 United States v. Farrad Page 16 photos, all from different angles, some remarkably close-up and of seemingly high resolution. See Appellant’s App’x at 5–26. And it heard exceedingly meticulous testimony from Hinkle— pointing to, among other things, markings and serrations on the slide, a lever and ridges on the trigger guard, and the shape and position of the firing pin—that both (a) convincingly likened the item in the close-up photographs to a real Springfield XD .45, R. 71 (Trial Tr. Vol. I at 126–46) (Page ID #724–44), and (b) discredited the closest non-gun comparator that Hinkle could find, id. at 146–58 (Page ID #744–56). While images in a different case might be more suspicious, the combination of Hinkle’s meticulous testimony and the plethora of photographic evidence here provided grounds for a rational juror to conclude that the item in the photos was in fact a real Springfield XD .45. 3. Photoshop Theory Farrad argues only briefly on appeal that the photos could have been manipulated, see Appellant’s Br. at 40, and we can likewise deal with this theory quickly. In the abstract, the Photoshop theory raises some of the same concerns as the replica theory, given that a person seeking to project a (false) image of himself might well manipulate an image to make it look as if they were holding a firearm. (Similarly, a person wishing ill on another might manipulate an image to inculpate them.) But these particular concerns generally cash out as a question of admissibility (an issue discussed in more detail below): whether “the item is what the proponent claims it is.” FED. R. EVID. 901(a); see, e.g., Griffin v. Bell, 694 F.3d 817, 826–27 (7th Cir. 2012) (affirming exclusion of video of altercation in question from trial where party offering it was unable to authenticate it); People v. Price, 80 N.E.3d 1005, 1009–10 (N.Y. 2017) (vacating conviction and ordering new trial where prosecution failed to authenticate photo of defendant holding gun ostensibly used in robbery). See generally Paul W. Grimm et al., Authentication of Social Media Evidence, 36 AM. J. TRIAL ADVOC. 433 (2013); Elizabeth A. Flanagan, Note, #Guilty? Sublet v. State and the Authentication of Social Media Evidence in Criminal Proceedings, 61 VILL. L. REV. 287 (2016). Moreover, while the various questions raised by potentially altered social-media images “could conceivably be quite interesting,” United States v. Thomas, 701 F. App’x 414, 419 (6th Cir. 2017), this case does not present those questions. The jury heard evidence that the photos in question came from a Facebook account registered to a Nos. 16-5102/6730 United States v. Farrad Page 17 Malik Farrad from Knoxville, Tennessee, and saw photos that appeared to show Farrad in his own apartment—largely ruling out the possibility that this was the work of a forger. See R. 71 (Trial Tr. Vol. I at 96–98) (Page ID #694–96); Appellant’s App’x at 11–14. Farrad’s trial counsel, meanwhile, never cross-examined a witness on potential alteration of the images, presented any evidence to that effect, or argued that theory to the jury. This theory is, accordingly, not a winning one for Farrad on appeal. 4. Lookalike Theory Similar defects apply to the lookalike theory. Farrad adverts to this argument only briefly as well, asserting that “the pictures did not clearly disclose who, if anyone, was holding a gun.” Appellant’s Br. at 41. But again, the jury heard that personal details of the Facebook account that posted the photos matched Farrad’s personal details, R. 71 (Trial Tr. Vol. I at 96–98) (Page ID #694–96); saw photos that looked like Farrad, Appellant’s App’x at 11–14; heard and saw evidence suggesting that the photos were taken in an apartment belonging to Farrad, R. 71 (Trial Tr. Vol. I at 105) (Page ID #703); Appellant’s App’x at 19–20, 22–23, 25–28; and heard a defense theory that all but admitted that Farrad was the person in the photos, R. 72 (Trial Tr. Vol. II at 15) (Page ID #803) (“Maybe it’s a case of where Mr. Farrad needed to let the world know that he could protect himself, let them think that he can.”). The jury could easily infer as much. *** With Farrad’s trial and appellate counsel having eschewed the date theory, Farrad’s sufficiency-of-the-evidence challenge must rise and fall with the replica, Photoshop, and lookalike theories. But the number and detail of the photos, coupled with Hinkle’s meticulous testimony and other corroborating details (such as the appearance of Farrad’s apartment), render each uniquely ill-suited to Farrad’s specific case. We therefore reject Farrad’s sufficiency-ofthe-evidence challenge on the record and arguments before us. Nos. 16-5102/6730 United States v. Farrad Page 18 B. Admissibility of the Photos Farrad also argues that the district court erred in admitting the Facebook photos into evidence in the first place. We review evidentiary challenges for an abuse of discretion. 9 United States v. Morales, 687 F.3d 697, 701–02 (6th Cir. 2012). Abuses of discretion in evidentiary rulings, however, merit reversal only if the error is not harmless—“that is, only if the erroneous evidentiary ruling affected the outcome of the trial.” United States v. Marrero, 651 F.3d 453, 471 (6th Cir. 2011). Farrad’s evidentiary challenge raises a number of issues, but his most central and colorable challenge is to the district court’s authentication of the Facebook images. 10 See Appellant’s Br. at 31–33. “Like other evidence, photographs must be authenticated prior to being admitted into evidence.” Hartley v. St. Paul Fire & Marine Ins. Co., 118 F. App’x 914, 921 (6th Cir. 2004). “To satisfy [this requirement], the proponent must produce evidence sufficient to support a finding that the item is what the proponent claims it is.” FED. R. EVID. 901(a); see also United States v. Jones, 107 F.3d 1147, 1150 n.1 (6th Cir. 1997) (“The [authentication] rule requires only that the court admit evidence if sufficient proof has been introduced so that a reasonable juror could find in favor of authenticity or identification. The rest is up to the jury.” (quoting 5 JACK B. WEINSTEIN ET AL., WEINSTEIN’S EVIDENCE ¶ 901(a), at 901–19 (1996))). This task can be accomplished in a number of ways—with testimony from someone with knowledge of the evidence offered, for example, or by pointing to distinctive characteristics that establish authenticity. See FED. R. EVID. 901(b)(1), (4). Some items, however, “are selfauthenticating; they require no extrinsic evidence of authenticity in order to be admitted.” FED. 9This standard converges with de novo review of legal questions and clear-error review of factual determinations, “because it is an abuse of discretion to make errors of law or clear errors of factual determination.” See United States v. McDaniel, 398 F.3d 540, 544 (6th Cir. 2005). 10Farrad also briefly raises a challenge based on Federal Rules of Evidence 403 and 404 to a photo or photos showing guns, drugs, and U.S. currency. See Appellant’s Br. at 37–38; see also R. 28-2 (Ex. 2, Gov’t’s Trial Br.) (Page ID #115). This challenge lacks merit, however, because, as the Government notes, the Government never introduced any such photos despite the district judge’s having allowed such photos in. Appellee’s Br. at 27 n.9; accord R. 71 (Trial Tr. Vol. I at 76–184) (Page ID #674–782) (Government’s case in chief); Appellant’s App’x at 5–26. Any error, in other words, was clearly harmless. See, e.g., Marrero, 651 F.3d at 471. Nos. 16-5102/6730 United States v. Farrad Page 19 R. EVID. 902. This category of self-authenticating evidence includes “certified domestic records of a regularly conducted activity”—that is, a business “record that meets the requirements of Rule 803(6)(A)–(C),” so long as properly certified by a “custodian or other qualified person” and so long as the evidence is subject to challenge by the opposing party. FED. R. EVID. 902(11). The relevant portion of Rule 803 encompasses: A record of an act, event, condition, opinion, or diagnosis if: (A) the record was made at or near the time by—or from information transmitted by—someone with knowledge; (B) the record was kept in the course of a regularly conducted activity of a business, organization, occupation, or calling, whether or not for profit; [and] (C) making the record was a regular practice of that activity[.] FED. R. EVID. 803(6)(A)–(C). The final two prongs in Rule 803(6), in turn, complete the set of prerequisites for Rule 803(6)’s exception to the rule against hearsay: the first three conditions must have been demonstrated by a qualifying certification from an authorized source, FED. R. EVID. 803(6)(D), and the opponent must not have “show[n] that the source of information or the method or circumstances of preparation indicate a lack of trustworthiness,” FED. R. EVID. 803(6)(E). Here, it is undisputed that the Government secured the requisite certification from Facebook to qualify the photos as self-authenticating business records. See R. 26-1 (Facebook Certification) (Page ID #105); R. 60 (Pretrial Conf. Tr. at 11) (Page ID #456). Instead, the principal question is whether the images qualified as self-authenticating business records at all.11 11Farrad also briefly suggests that there may have been a Confrontation Clause problem with authentication by remote affidavit. See Appellant’s Br. at 31 (citing Melendez-Diaz v. Massachusetts, 557 U.S. 305 (2009)). Farrad did not raise this issue below, however, and it is made in such cursory fashion here as to be forfeited altogether. See, e.g., Puckett v. Lexington-Fayette Urban Cty. Gov’t, 833 F.3d 590, 610–11 (6th Cir. 2016). Even if we did not deem it forfeited, it is unlikely that it would have been a winning argument on plain-error review in light of the Supreme Court’s discussion of the “narrowly circumscribed” exception at common law that allowed a clerk to present a “certificate authenticating an official record.” Melendez-Diaz, 557 U.S. at 322; see also id. at 311 n.1 (“Contrary to the dissent’s suggestion, we do not hold, and it is not the case, that anyone whose testimony may be relevant in establishing the chain of custody, authenticity of the sample, or accuracy of the testing device, must appear in person as part of the prosecution’s case.” (citation omitted)). Other circuits, meanwhile, that have reached this question in the years following Melendez-Diaz have rejected Farrad’s argument. See United States v. Mallory, 461 F. App’x 352, 356–57 (4th Cir. 2012); United States v. Yeley-Davis, 632 F.3d 673, 680–81 (10th Cir. 2011); see Nos. 16-5102/6730 United States v. Farrad Page 20 Farrad argues, as he did to the district court, that they did not so qualify, given that Facebook could not authenticate “who took the pictures, when the pictures were taken, by whom or at what time, what they actually showed.” Appellant’s Br. at 30; accord R. 60 (Pretrial Conf. Tr. at 11) (Page ID #456). The district court disagreed. R. 73 (Pretrial Hr’g Tr. at 33) (Page ID #872). 12 If Farrad is right, the question becomes whether the evidence would have been admissible for some other reason, or whether the error was in any event harmless to the outcome. See, e.g., United States v. Henderson, 626 F.3d 326, 334 (6th Cir. 2010). No opinion in this circuit has spoken directly to the proper standard for assessing the admissibility of photographs taken from an online social-media platform like Facebook. After reviewing analogous cases from inside this circuit and persuasive authority from other circuits, we conclude that the district court was correct to admit the photos, but that it should have done so under a more traditional standard: regular authentication under Rule 901. One small issue is worth addressing at the outset: whether, as Farrad has argued throughout, the “pictures of guns were all out-of-court ‘statements’ that Farrad illegally possessed a firearm” and thus hearsay. Appellant’s Br. at 34; see also R. 60 (Pretrial Conf. Tr. at 11) (Page ID #456). They were not. Farrad is correct, of course, that hearsay is any out of court statement “offer[ed] in evidence to prove the truth of the matter asserted,” FED. R. EVID. 801(c), and that hearsay encompasses “nonverbal conduct, if the person intended it as an assertion,” FED. R. EVID. 801(a). A paradigmatic case of such nonverbal assertive conduct is “the act of pointing to identify a suspect in a lineup.” FED. R. EVID. 801 advisory committee’s note to 1972 proposed rules. And as Farrad points out, Appellant’s Br. at 34, in United States v. Martinez, 588 F.3d 301 (6th Cir. 2009), we ruled that a video in which a doctor demonstrated the correct way to perform nerve-block injections qualified as hearsay, given that it was offered to show that the defendant had performed such injections incorrectly. Id. at 311. But demonstrating a particular procedure for the purpose of showing whether another has erred is more intrinsically and specifically assertive than simply appearing (or having one’s possessions appear) in a photo, even if it may also United States v. Anekwu, 695 F.3d 967, 973–74 (9th Cir. 2012) (arriving at the same conclusion, but on plainerror review). 12The district court also, as noted above, determined that the photos were relevant. R. 73 (Pretrial Hr’g Tr. at 35–37) (Page ID #874–76). Nos. 16-5102/6730 United States v. Farrad Page 21 be true in some philosophical sense that posing for a photograph does itself encode the general sentiment: here I am. In any case, as the Government notes, Appellee’s Br. at 32, even if the photos were statements, they would have (so long as authenticated) qualified as statements of a party opponent and thus were not hearsay all the same. See FED. R. EVID. 801(d)(2)(A). The question, then, is the central one: the authentication of the photos. Although it did not discuss the business-records exception, the closest analogue from our precedents is United States v. Thomas, 701 F. App’x 414 (6th Cir. 2017)—a case that itself illustrates that the business-records exception was not necessary to admit the photos here. The defendant in that case, Jabron Thomas, was charged with armed bank robbery, and he was identified at trial in part through photos obtained from Facebook and another social-media platform, Instagram. Id. at 418. He argued that the photos “could not be authenticated and were thus inadmissible in part because [the investigator who obtained them] admitted that he did not know who created the Facebook page or whether the Facebook page itself was authentic.” Id. at 419. We turned that argument aside, however, noting that there was simply enough evidence presented “to support a finding that the [photograph] [was] what the proponent claim[ed] it to be.” Id. (quoting FED. R. EVID. 901(a)). The photos “appeared to show Thomas with distinctive tattoos” and clothing, which was enough given that “the government was not seeking to authenticate Jabron Thomas’s [or any] Facebook page or Instagram page (nor any of the factual information contained therein, such as Thomas’s workplace)[,] and the government was not even necessarily presenting the photographs as ‘pictures of Jabron Thomas’—the jury was free to consider the photographs as identifying Thomas or not.” Id. Given that there was sufficient evidence that the photos were what they were claimed to be, in short, we ruled that there was no abuse of discretion in admitting them. Id. at 419–20. That rationale is all that is needed to resolve the ultimate question here. As already mentioned above, not only did the details of the account match Farrad, R. 71 (Trial Tr. Vol. I at 96–98) (Page ID #694–96), but more importantly, the photos appeared to show Farrad, his tattoos, and (perhaps most probatively) distinctive features of Farrad’s apartment, as confirmed by police investigation, R. 71 (Trial Tr. Vol. I at 105) (Page ID #703); Appellant’s App’x at 6, 11–14, 19–20, 22–23, 25–28. The Government was at most seeking to introduce these photos as Nos. 16-5102/6730 United States v. Farrad Page 22 evidence uploaded by a particular Facebook account that tended to show Farrad in possession of a real gun; the photos were not, however, offered as definitive and irrebuttable proof. See, e.g., R. 60 (Pretrial Conf. Tr. at 16) (Page ID #461); see also Oral Arg. at 21:28–21:38, 22:10–22:24. No specific evidence was shown to suggest that the photographs were not “accurate representation[s] of the scene depicted.” United States v. Hobbs, 403 F.2d 977, 979 (6th Cir. 1968). In short, while there were still questions about the photos that merited probing, those questions were not so glaring as to prevent the photos from clearing the relatively lower hurdle of authentication.13 See Thomas, 701 F. App’x at 418; Jones, 107 F.3d at 1150 n.1. The district court was correct to admit them. Whether the photos should have been admitted as self-authenticating under the businessrecords exception, on the other hand, is a different question. As the Third Circuit recently observed in United States v. Browne, 834 F.3d 403 (3d Cir. 2016), cert. denied, 137 S. Ct. 695 (2017), a case involving Facebook “chats” admitted against a defendant, “Rule 803(6) is designed to capture records that are likely accurate and reliable in content, as demonstrated by the trustworthiness of the underlying sources of information and the process by which and purposes for which that information is recorded.” Id. at 410 (citing FED. R. EVID. 803 advisory committee’s note to 1972 proposed rules; United States v. Gurr, 471 F.3d 144, 152 (D.C. Cir. 2006); E.C. Ernst, Inc. v. Koppers Co., 626 F.2d 324, 330–31 (3d Cir. 1980)). To the extent that the chats in Browne were records, however, Facebook had no oversight or particular interest in ensuring that they were trustworthy. As the Third Circuit reasoned: 13This case is thus different, for example, from the Seventh Circuit decision that Farrad cites, Griffin v. Bell, 694 F.3d 817 (7th Cir. 2012). In that § 1983 appeal, in dicta, Judge Rovner explained that even if Griffin’s challenge to the exclusion of a video (and photographs distilled from it) that he claimed depicted a fight between him and a police officer had not been waived, there was still no abuse of discretion in the district court’s having declined to admit the evidence. Id. at 819–20, 826–27. But that was in part because, in addition to the video’s creator being unavailable and other open questions about the video, “the video portrayed only a small part of the incident” (and the photos “even less”), which naturally made the video a potentially unhelpful (or even misleading) account of how the very much disputed events at issue had unfolded. See id. Here, the district court knew more about the photos at issue and their authenticity as accurate depictions of the relevant scene, and there was significantly less reason to worry that they presented a one-sided or confusing view of a complex and fast-moving event. Nos. 16-5102/6730 United States v. Farrad Page 23 Here, Facebook does not purport to verify or rely on the substantive contents of the communications in the course of its business. At most, the records custodian employed by the social media platform can attest to the accuracy of only certain aspects of the communications exchanged over that platform, that is, confirmation that the depicted communications took place between certain Facebook accounts, on particular dates, or at particular times. This is no more sufficient to confirm the accuracy or reliability of the contents of the Facebook chats than a postal receipt would be to attest to the accuracy or reliability of the contents of the enclosed mailed letter. Id. at 410–11 (citing United States v. Jackson, 208 F.3d 633, 637–38 (7th Cir. 2000)); accord United States v. Hillis, 656 F. App’x 222, 229 (6th Cir. 2016) (“Nonetheless, if records contain information obtained from a customer, the information still falls within the business-records exception if the business’s standard practice is to verify the information provided by the customer.” (citing cases)), cert. denied, 137 S. Ct. 1359 (2017). While allowing that a different question would be presented “[i]f the Government . . . had sought to authenticate only the timestamps on the Facebook chats, the fact that the chats took place between particular Facebook accounts, and similarly technical information,” the Third Circuit concluded that, because “the Government’s interest [lay] in establishing the admissibility of the chat logs in full, . . . the Facebook records [were] not business records under Rule 803(6) and thus [could not] be authenticated by way of Rule 902(11).” Browne, 834 F.3d at 411. The Third Circuit’s approach accords with our approach in Thomas, see 701 F. App’x at 419 (“[W]e see no reason to depart from the ordinary rule that photographs, including socialmedia photographs, are authenticated by ‘evidence sufficient to support a finding that the [photograph] is what the proponent claims it is.”), and it also accords with the approaches taken by the Second Circuit, United States v. Vayner, 769 F.3d 125, 131–33 (2d Cir. 2014), and Fifth Circuit, United States v. Barnes, 803 F.3d 209, 217 (5th Cir. 2015).14 It also fits with common sense: it is not at all clear, nor has the Government been able to explain adequately, why our rules of evidence would treat electronic photos that police stumble across on Facebook one way 14The only circuit that is potentially in conflict is the Fourth Circuit, but even that is not clearly the case. The Fourth Circuit, after all, affirmed on a similar question (1) after noting specific extrinsic evidence that seemed to authenticate the social-media evidence at issue, and (2) after explaining that the district court had determined that the prosecution had satisfied both Rule 902(11) and Rule 901(a). See United States v. Hassan, 742 F.3d 104, 133– 34 (4th Cir. 2014). Nos. 16-5102/6730 United States v. Farrad Page 24 and physical photos that police stumble across lying on a sidewalk a different way. 15 We therefore conclude that while it was an error for the district court to deem the photographs selfauthenticating business records, that error was harmless because admission was proper under Rule 901(a) regardless. C. Expert Testimony from Officers Hinkle and Garrison Farrad also argues that Officers Hinkle and Garrison “were permitted to improperly speculate and opine on matters relating to social media and give expansive testimony as to complete inferences from the photograph to fabricate Farrad’s guilt, all without foundation as required by and in violation of” Federal Rules of Evidence 701, 702, and 703.16 Appellant’s Br. at 35. The Government counters that “the record demonstrates that both Officers Garrison and Hinkle were qualified to offer the expert testimony that they provided.” Appellee’s Br. at 34. We conclude that while Hinkle’s testimony easily passes muster, Garrison’s fails. Nevertheless, given that neither Farrad’s trial counsel nor his appellate counsel argued the date issue (as discussed above), we conclude that the error was harmless. “This Court reviews a district court’s decision to admit proposed expert testimony for abuse of discretion.”17 United States v. LaVictor, 848 F.3d 428, 440 (6th Cir.), cert. denied, 137 S. Ct. 2231 (2017). “When a defendant fails to object to expert testimony at trial,” however, “this court reviews for plain error the district court’s admission of that testimony.” 18 United 15The Government attempted at oral argument to ground this distinction in an assertedly higher likelihood of contemporaneity ascribable to photos found on Facebook. See Oral Arg. 23:25–25:24. But as discussed above with reference to the replica theory, upload dates are not nearly as telling as the Government suggests. Moreover, their providing a particular cutoff time before which a photo must have been taken does not suitably distinguish electronic photos from physical photos, which also, by nature, must have been taken prior to whatever date they are discovered. 16There does not appear to have been a Rule 703 problem in this case, and for reasons discussed below, the crucial testimony here was not lay testimony under Rule 701. Accordingly, we focus on Rule 702, which is the most colorable part of Farrad’s claim. 17As noted above, “[a] district court abuses its discretion if it bases its ruling on an erroneous view of the law or a clearly erroneous assessment of the evidence.” LaVictor, 848 F.3d at 440 (quoting Best v. Lowe’s Home Cntrs., Inc., 563 F.3d 171, 176 (6th Cir. 2009)). 18“A ‘plain error’ is an error that is clear or obvious, and if it affects substantial rights, it may be noticed by an appellate court.” United States v. Oliver, 397 F.3d 369, 375–76 (6th Cir. 2005) (quoting United States v. Barajas–Nunez, 91 F.3d 826, 830 (6th Cir. 1996)). “We ‘should correct a plain forfeited error affecting substantial Nos. 16-5102/6730 United States v. Farrad Page 25 States v. Rodgers, 85 F. App’x 483, 487 (6th Cir. 2004). Here, as the Government notes, Appellee’s Br. at 34, Farrad’s trial counsel objected only to Garrison’s testimony. R. 71 (Trial Tr. Vol. I at 84–87) (Page ID #682–85). Accordingly, we review the district court’s allowance of Garrison’s testimony for an abuse of discretion and the district court’s allowance of Hinkle’s testimony for plain error. “A witness who is qualified as an expert by knowledge, skill, experience, training, or education may testify in the form of an opinion or otherwise if”: (a) the expert’s scientific, technical, or other specialized knowledge will help the trier of fact to understand the evidence or to determine a fact in issue; (b) the testimony is based on sufficient facts or data; (c) the testimony is the product of reliable principles and methods; and (d) the expert has reliably applied the principles and methods to the facts of the case. FED. R. EVID. 702. Accordingly, in addition to assessing a witness’s qualifications, “the trial judge must ensure that any and all scientific testimony or evidence admitted is not only relevant, but reliable.”19 Daubert v. Merrell Dow Pharm., Inc., 509 U.S. 579, 589 (1993); see also, e.g., LaVictor, 848 F.3d at 441 (“For expert testimony to be admissible, the court must find the expert to be: (1) qualified; (2) her testimony to be relevant; and (3) her testimony to be reliable.”). rights if the error seriously affect[s] the fairness, integrity or public reputation of judicial proceedings.’” Id. at 376 (quoting United States v. Olano, 507 U.S. 725, 736 (1993) (alteration in original) (internal quotation marks omitted)). 19Although Daubert mostly “limited” its “discussion . . . to the scientific context because that [was] the nature of the expertise offered” in that case, Daubert, 509 U.S. at 590 n.8, the Court did make some statements that are understood as having broader applicability. We have made clear, for example, that despite Daubert’s focus on scientific experts, the above-quoted “language relative to the ‘gatekeeper’ function of federal judges is applicable to all expert testimony offered under Rule 702.” Berry v. City of Detroit, 25 F.3d 1342, 1350 (6th Cir. 1994). It may also be fair to rely on other generalized statements from Daubert—for instance, that “the word ‘knowledge’ connotes more than subjective belief or unsupported speculation.” Daubert, 509 U.S. at 590. Similarly, it seems reasonable to rely on its broader observation that while it is impermissible to require “certainty,” even for scientific testimony, “[p]roposed testimony must be supported by appropriate validation—i.e., ‘good grounds,’ based on what is known.” Id. Nos. 16-5102/6730 United States v. Farrad Page 26 1. Officer Hinkle Hinkle easily passes this test, particularly on plain-error review. As Hinkle testified, he served as his police department’s armorer and had worked as a gunsmith “for over 30 years.” R. 71 (Trial Tr. Vol. I at 124) (Page ID #722). He “started apprenticing” when he was thirteen years old, and at the time of his testimony he had “handled or worked on” “[t]housands” of firearms. Id. at 124–25 (Page ID #722–23). He had been through—and led—numerous trainings and had testified in court as a firearms expert “[o]ver a dozen” times. Id. at 125 (Page ID #723). His testimony, moreover, reads as cogent, meticulously informed, and relevant to the question at hand. See id. at 127–58 (Page ID #725–56). In short, because Hinkle was qualified, his testimony was relevant, and his testimony was reliable, the district court did not err, let alone plainly err, in allowing his testimony. See, e.g., LaVictor, 848 F.3d at 441. 2. Officer Garrison Abuse of Discretion. — Garrison’s testimony is a different story. First of all, it is important to be clear that Garrison was in fact offering expert testimony. See, e.g., Berry v. City of Detroit, 25 F.3d 1342, 1349–50 (6th Cir. 1994). Compare FED. R. EVID. 701, with FED. R. EVID. 702. Though Garrison likely offered some lay testimony about social-media usage— detailing, for example, the basics of when one can upload a picture onto Facebook and how becoming someone’s “friend” allows a viewer to see more of their content, R. 71 (Trial Tr. Vol. I at 81, 83–84) (Page ID #679, 681–82)—the testimony of Garrison’s that is at issue here was distinct from that. As we have explained, “the advisory committee notes to [Rule 702] indicate that a wide array of expert testimony is contemplated by the Rule”: The rule is broadly phrased. The fields of knowledge which may be drawn upon are not limited merely to the “scientific” and “technical” but extend to all “specialized” knowledge. Similarly, the expert is viewed, not in a narrow sense, but as a person qualified by “knowledge, skill, experience, training, or education.” Thus within the scope of the rule are not only experts in the strictest sense of the word, e.g., physicians, physicists, and architects, but also the large group sometimes called “skilled” witnesses, such as bankers or landowners testifying to land values. Nos. 16-5102/6730 United States v. Farrad Page 27 Jones, 107 F.3d at 1159 (quoting FED. R. EVID. 702 advisory committee’s note to 1972 proposed rules). Accordingly, a witness testifies as an expert when his “expertise is largely a product of his ‘knowledge, skill, experience, training, [and] education.’” Id. (quoting FED. R. EVID. 702). That is clearly what happened in the testimony at issue here. Garrison was asked to testify based on his “training and experience,” “drawing upon the hundreds of cases [he] said [he had] been involved in using social media.” R. 71 (Trial Tr. Vol. I at 84) (Page ID #682). Moreover, he was not asked to testify simply about general social-media habits, but about a purportedly niche area of social-media activity: how criminals behave on social media. See id. (“[W]hen you come across people involved in criminal conduct who have uploaded photographs to their social media account, how quickly do they do that, relative to when that photograph is actually taken?”); see also id. (“And, again, in your training and experience, typically, why do people upload photographs of themselves involved in criminal activity to social media, such as Facebook?”); id. at 86 (Page ID #684) (“Mr. Garrison, in your training and experience why do people choose to post criminal conduct that they’re involved in, to social media?); id. (“[C]riminals specifically, they like to brag about their—their activities, they’re proud of it, and just like anyone, they want to let their friends know what they’re doing, let their friends know, you know, where they’re at, what’s going on.”); id. at 87–88 (“In contrast, then, how many times relative to instances in which people have uploaded immediately, contrasted that to the number of times you’ve seen photographs depicting criminal conduct that have been held back for extended periods of time, weeks, months or years?”). The body of knowledge that Garrison was drawing on, in other words, was specifically presented by the Government as falling “beyond the ken of the average juror,” United States v. Rios, 830 F.3d 403, 413 (6th Cir. 2016) (quoting United States v. Amuso, 21 F.3d 1251, 1263 (2d Cir.), cert. denied, 513 U.S. 932 (1994)), cert. denied, 137 S. Ct. 1120 (2017). The next question is whether Garrison’s expert testimony was scientific or nonscientific. In Berry, a case about a purported expert with police experience who sought to testify about the effect of departmental disciplinary failures, we drew a comparison between two witnesses called to testify about bumblebees to illustrate the difference. 25 F.3d at 1348–50. An aeronautical engineer, for example, could testify as a scientific expert, because “flight principles have some Nos. 16-5102/6730 United States v. Farrad Page 28 universality,” and thus “the expert could apply general principles to the case of the bumblebee.” Id. “On the other hand . . . a beekeeper with no scientific training at all would [also] be an acceptable expert witness if a proper foundation were laid for his conclusions”; although he might “not know any more about flight principles than the jurors, . . . he has seen a lot more bumblebees than they have.” Id. Garrison, like the purported expert in Berry, was presented as the latter type: he did not claim to have done any scientific research on how people who commit crimes use social media, but he had supposedly seen a lot more criminals using social media. See R. 71 (Trial Tr. Vol. I at 84–87) (Page ID #682–85). The ultimate set of questions is whether the district court abused its discretion in finding Garrison’s nonscientific testimony, over Farrad’s repeated objections, R. 71 (Trial Tr. Vol. I at 84–88) (Page ID #682–85), to be qualified, relevant, and reliable. LaVictor, 848 F.3d at 441. Relevance is easy for the Government: one of the necessary questions for the jury was whether Farrad had possessed a firearm “on or about” the dates that he uploaded the Facebook photos, see, e.g., R. 3 (Indictment) (Page ID #3), and to the extent Garrison could offer evidence based on his training and experience that it was indeed “more likely” that a criminal’s Facebook upload follows hot on the heels of an actus reus, see R. 71 (Trial Tr. Vol. I at 86) (Page ID #684), then it was ostensibly more likely that Farrad had done what was charged in the indictment. The stumbling block for the Government is whether Garrison was sufficiently qualified and, relatedly, whether his testimony was sufficiently reliable. Following our framing from Berry, because Garrison lacked “formal training that would allow [him] to testify from a scientific standpoint,” in order for his “testimony to be admissible, a foundation would have to have been laid based upon [Garrison’s] firsthand familiarity with” criminal social-media habits. 25 F.3d at 1350. Yet even assuming for the sake of argument that criminal social-media behavior is a real “field,” and not “so broad as to be devoid of meaning” (“like declaring an attorney an expert in the ‘law’”), see id. at 1352, Garrison was repeatedly unable to establish the requisite familiarity. The Government largely relies on Garrison’s initial testimony about his training and experience. See Appellee’s Br. at 34–36. If Garrison’s initial testimony were all that we had to go on, such a foundation might well have been established. Garrison testified, after all, that he Nos. 16-5102/6730 United States v. Farrad Page 29 had used social media, as part of his job, “basically on a daily basis for several years,” stating that he and his colleagues would “usually . . . run somebody on Facebook or whatever,” such that the number of cases that included some social-media investigation “could potentially be in the hundreds.” R. 71 (Trial Tr. Vol. I at 80) (Page ID #678). He added that, “in attending different narcotics investigations, training, and in-services, there’s been specific blocks on conducting investigations through and with social media.” Id. Though vague, that account suggests a decent basis of knowledge. We have, however, “counseled against putting some general seal of approval on an expert after he has been qualified but before any questions have been posed to him.” Berry, 25 F.3d at 1351. “The issue with regard to expert testimony is not the qualifications of a witness in the abstract, but whether those qualifications provide a foundation for a witness to answer a specific question.” Id. And the problem for the Government is that the wheels of Garrison’s expertise came off the wagon upon basic probing. Consider, for example, this exchange on direct examination, which followed a prior objection along the same lines by Farrad: GOVERNMENT: And, again, in your training and experience, typically, why do people upload photographs of themselves involved in criminal activity to social media, such as Facebook? DEFENSE COUNSEL: Same objection, Your Honor, speculation. GOVERNMENT: Again, Your Honor, the same explanation. Mr. Garrison has established himself as someone who’s been involved in numerous investigations involving social media. THE COURT: But I’m not sure he said how he is qualified to answer that additional question. You might need to lay additional foundation for him to answer that. GOVERNMENT: Mr. Garrison, have you had an opportunity to discuss with targets of other investigations who have used social media about the reasons why they did that? DEFENSE COUNSEL: Object to leading, Your Honor. THE COURT: I’ll overrule that objection. Go ahead. WITNESS: I cannot—I can’t think of a specific instance where I’ve talked to a target about specific things on Facebook. Nos. 16-5102/6730 United States v. Farrad Page 30 GOVERNMENT: Has the reasonings [sic] for why people upload these photographs to social media, was that a part of your training that you went through, the bases for why people do that? WITNESS: I know that has been discussed before. I can’t remember specific instances of training but, you know, we’ve discussed it with other investigators or other law enforcement officers. Id. at 84–85 (Page ID #682–83). Farrad’s counsel persisted in his objections. See, e.g., id. at 86–87 (Page ID #684–85). Nevertheless, the trial court allowed Garrison to continue testifying as an expert in criminal social-media behavior. And yet Garrison’s inability to recall even one specific instance of training or experience relating to the question at hand persisted. See id. at 87–88 (“Q. In contrast, then, how many times relative to instances in which people have uploaded immediately, contrasted that to the number of times you’ve seen photographs depicting criminal conduct that have been held back for extended periods of time, weeks, months or years? A. I would—I would consider that more rare. I can’t—I can’t think of a specific number or a—or an instance just off the top of my head.”). Garrison not only failed to offer methodology or data, see Berry, 25 F.3d at 1352, but he also failed to offer even an anecdote. Allowing this testimony to continue was an abuse of discretion. Because Garrison could not remember even a single apposite experience or training that spoke to the immediacy of social-media posting by criminals, it is clear that Garrison was not qualified to testify as an expert about the immediacy of social-media posting by criminals (again assuming that criminal social-media behavior is even a cognizable field), see, e.g., LaVictor, 848 F.3d at 441, and therefore should not have been held up to the jury as such an expert. Garrison’s base of knowledge did not “provide a foundation for [him] to answer [the Government’s] specific question[s],” Berry, 25 F.3d at 1351, and indeed his base of knowledge, as Farrad observes, Appellant’s Br. at 36, appears to have been impermissibly close to “subjective belief or unsupported speculation,” Daubert, 509 U.S. at 590. And for essentially the same reason, it is clear that Garrison’s testimony was not reliable. See, e.g., LaVictor, 848 F.3d at 441. While courts over the past half-century have understandably come to “rely on . . . the notion that trained, experienced officers develop rarefied and reliable insight into crime,” see generally Nos. 16-5102/6730 United States v. Farrad Page 31 Anna Lvovsky, The Judicial Presumption of Police Expertise, 130 HARV. L. REV. 1995, 1999 (2017), there are limits to such reliance, see, e.g., Berry, 25 F.3d at 1348–54. Garrison’s testimony breaches those limits. Harmless Error. — Farrad “is not entitled to a new trial,” however, “unless it is more probable than not that the error materially affected the verdict,” LaVictor, 848 F.3d at 448 (citation and internal quotation marks omitted)—that is, unless the error was not harmless. In theory, he has a colorable argument. As already discussed at length, the prosecution was required to prove beyond a reasonable doubt that Farrad possessed a gun “on or about” October 11, 2013—the date that most of the photos were uploaded. See, e.g., R. 3 (Indictment) (Page ID #3); R. 71 (Trial Tr. Vol. I at 99–104) (Page ID #697–702). In the absence of metadata or any other evidence tying the charge against Farrad to that date, id. at 89 (Page ID #687), Garrison’s testimony was potentially necessary, see R. 72 (Trial Tr. Vol. II at 9) (Page ID #797), and indeed the Government’s only fallback to it was a speculative (and, by nature, nonevidentiary) statement at closing argument that immediate uploading was “in line with what we do on a daily basis,” id. at 10 (Page ID #798). The problem is, first, that it is harder to conclude that Garrison’s testimony materially affected the verdict when Farrad’s trial counsel did not argue the date theory. R. 71 (Trial Tr. Vol. I at 17–21) (Page ID #802–11); cf. Rose v. Clark, 478 U.S. 570, 582 n.11 (1986) (“Harmless-error analysis addresses [the] question: what is to be done about a trial error that, in theory, may have altered the basis on which the jury decided the case, but in practice clearly had no effect on the outcome?”). And it is, in any event, impossible for us to deem the error harmful when Farrad’s appellate counsel also declined to argue the date theory. Oral Arg. at 9:05–9:22, 11:56–12:14, 29:22–29:37, 31:39–32:17. Thus, while Farrad can, like any other petitioner, seek post-conviction relief for ineffective assistance of trial and appellate counsel under § 2255 (an issue, again, on which we express no ultimate opinion), we must deem the trial court’s error harmless. Nos. 16-5102/6730 United States v. Farrad Page 32 D. Motion for New Trial Farrad also appeals the district court’s denial of his motion for a new trial. “We review a district court’s denial of a motion for a new trial for abuse of discretion.” United States v. Callahan, 801 F.3d 606, 616 (6th Cir. 2015). “To prevail on a claim that the government presented perjured testimony, [Farrad] must show ‘(1) that the statements were actually false; (2) the statements made were material; and (3) [the] prosecution knew they were false.’” United States v. Pierce, 62 F.3d 818, 834 (6th Cir. 1995) (quoting United States v. Farley, 2 F.3d 645, 655 (6th Cir. 1993)); accord Akrawi v. Booker, 572 F.3d 252, 265 (6th Cir. 2009). We can leave aside the first two of these prongs, because the fatal prong for Farrad is the third. Here, Farrad’s challenge is to Hinkle’s testimony (1) that he had “not been able to locate, during any of [his] research, a toy, airsoft, BB firearm of the XD, XD series pistol in any caliber,” and (2) when asked whether Springfield had “ever provided the licenses to manufacture an imitation of this firearm to another company,” his response that “they ha[d] not.” R. 71 (Trial Tr. Vol. I at 146) (Page ID #744); see Appellant’s Br. at 43–44. But as the district court correctly noted, R. 89 (Mem. Op. & Order at 11) (Page ID #948), Farrad failed to provide, and still has not provided, any reason to conclude that either Hinkle or the prosecution knew that this was false testimony. Of course, in referencing the blue replicas, Farrad has pointed to evidence that could conceivably render Hinkle’s testimony incorrect. R. 88-1 (Ex. 1, Def.’s Reply re Mot. for New Trial) (Page ID #936–37). But aside from Hinkle’s general knowledgeability (which, while impressive, is not tantamount to omniscience), the record offers no reason to conclude that the replicas even existed at the time of Hinkle’s testimony, let alone that Hinkle in turn knew about them and chose not to disclose their existence during his testimony. Simply proving that a witness was mistaken, meanwhile, does not entitle a defendant to a new trial. Pierce, 62 F.3d at 834. Farrad may, again, seek post-conviction relief under § 2255 based on any potential failure by trial counsel to discover the replicas and use them to impeach Hinkle’s testimony at trial, but the district court did not abuse its discretion in denying Farrad’s motion for a new trial.20 20Farrad also gestures briefly at an additional asserted ground for a new trial, based on purported police misconduct in gaining access to Farrad’s Facebook photos and the district court’s asserted unwillingness to let him cross-examine Garrison on this issue at trial. Appellant’s Br. at 49–50. But Farrad has offered no evidence Nos. 16-5102/6730 United States v. Farrad Page 33 E. Farrad’s ACCA Predicate Offenses Farrad also challenges the district court’s determination that he qualified as an armed career criminal under the ACCA, 18 U.S.C. § 924(e)(1). That provision applies, in relevant part, to any “person who violates section 922(g) . . . and has three previous convictions . . . for a violent felony or a serious drug offense, or both, committed on occasions different from one another,” and it provides for a fifteen-year mandatory minimum. Id. Ordinarily, “[a] district court’s interpretation and application of the ACCA is a question of law, reviewed de novo.” United States v. Stafford, 721 F.3d 380, 395–96 (6th Cir. 2013). When a party fails to object at sentencing, however, review is for plain error only. See, e.g., United States v. Southers, 866 F.3d 364, 366 (6th Cir. 2017). Farrad challenges both the designation of his Tennessee conviction for simple robbery and the designation of eight federal drug-trafficking convictions as ACCA predicates. Appellant’s Br. at 52–60. But we need not focus on the Tennessee robbery conviction, because if even three of the federal drug-trafficking convictions qualify, the existence of an additional predicate is purely academic.21 See, e.g., United States v. Phillips, 752 F.3d 1047, 1049 (6th Cir. 2014). Farrad attacks the eight federal drug-trafficking convictions on three grounds: (1) that they included “aiding and abetting modifiers” and therefore potentially encompassed a broader swath of wrongdoing than the ACCA, Appellant’s Br. at 54–55; (2) that they should not have been counted as separate convictions, Appellant’s Br. at 54–58; and (3) that they cannot be suggesting police misconduct, Farrad’s trial counsel laid no foundation for this line of inquiry at trial and then abandoned it almost immediately after beginning it, see R. 71 (Trial Tr. Vol. I at 92–92) (Page ID #690–91), and, as discussed more below, Farrad never renewed his motion to suppress after it was dismissed on procedural grounds, R. 24 (Order at 1) (Page ID #93). In short, there was no abuse of discretion here either. 21For completeness, we note that Farrad also failed to raise to the district court the objections to the Tennessee robbery conviction that he raises here, see R. 47 (Def.’s Objections to PSR) (Page ID #311–28); R. 48 (Def.’s Supp. Objections to PSR) (Page ID #515–24), so he accordingly qualifies only for plain-error review. Southers, 866 F.3d at 366. To the extent that Farrad seeks to argue that Tennessee robbery is not a violent felony under the ACCA, Appellant’s Br. at 53, his claim is doomed by binding circuit precedent regardless. See Southers, 866 F.3d at 366–69. To the extent he challenges the conviction based on the fact that he was only seventeen years old when he was arrested for that crime, Appellant’s Br. at 52–53, that argument also fails under binding circuit precedent given that Farrad was evidently convicted and sentenced as an adult. See United States v. Mitchell, 743 F.3d 1054, 1067 (6th Cir. 2014); R. 53 (Revised PSR at 7) (Page ID #412). There was, in other words, no plain error on either of these grounds. Nos. 16-5102/6730 United States v. Farrad Page 34 counted against him because he was found guilty “in absentia,” Appellant’s Br. at 58–60. None of these arguments is compelling. With regard to the first, Farrad’s characterization of his federal convictions is misleading. The eight convictions—on Counts 12, 14, 16–20, and 22—were for violations of 21 U.S.C. § 841(a)(1), (b)(1)(B)–(C), which are valid ACCA predicates carrying a maximum punishment of more than ten years of imprisonment. See 18 U.S.C. § 924(e)(2)(A)(i); R. 51-2 (Federal Drug-Trafficking Indictment at 8–11) (Page ID #369–72). Furthermore, if we look to the relevant charging document, see Shepard v. United States, 544 U.S. 13, 26 (2005); see also Mathis v. United States, 136 S. Ct. 2243, 2249 (2016), it is clear that Farrad was not charged with or convicted of a crime that categorically falls outside of the ACCA’s sweep: although each of the relevant counts in the indictment makes reference to aiding and abetting, the aiding and abetting at issue is alleged to have been done by others, while Farrad is charged with having violated the statute, plain and simple. R. 51-2 (Federal Drug-Trafficking Indictment at 8–11) (Page ID #369–72). Farrad’s first challenge fails. Second, Farrad’s argument that his drug-trafficking convictions should not be counted separately is also unavailing. “In this circuit, two offenses were committed on different occasions under the ACCA if 1) it is possible to discern when the first offense ended and the subsequent point at which the second offense began; 2) the offender could have withdrawn from crime after the first offense ended and not committed the second offense; or 3) the offenses were committed at different residences or business locations.” United States v. Pham, 872 F.3d 799, 802 (6th Cir. 2017), cert. denied, 138 S. Ct. 1018 (2018). This analysis necessarily entails the use of Shepard documents. Id. “The government prevails if it meets even one of the tests.” Id. It is clear from looking at the indictment that the Government meets two of these tests. Counts 12, 14, 16–20, and 22 charge that the relevant conduct occurred “on or about” different and specific dates, ranging from November 3, 1998, to January 7, 1999. R. 51-2 (Federal DrugTrafficking Indictment at 8–11) (Page ID #369–72). The jury convicted Farrad on each of these separate counts. R. 51-3 (Federal Drug-Trafficking Verdict Form) (Page ID #374–78). It is clearly “possible to discern when [each] offense ended and the subsequent point at which the [next] offense began,” just as it is clearly possible that Farrad “could have withdrawn from crime Nos. 16-5102/6730 United States v. Farrad Page 35 after [each] offense ended and not committed the [next] offense,” despite the fact that each was in the context of an ongoing conspiracy. Pham, 872 F.3d at 802; see also United States v. Brady, 988 F.2d 664, 669 (6th Cir. 1993) (en banc); United States v. Roach, 958 F.2d 679, 684 (6th Cir. 1992). This challenge thus fails. Third and finally, Farrad argues that his federal drug-trafficking convictions cannot count as ACCA predicates because he was convicted in absentia. This argument, Farrad concedes, Appellant’s Br. at 58–59, is to be reviewed only for plain error in light of Farrad’s failure to raise it in the district court. The problem for Farrad is that there is no evidence that Farrad was in fact impermissibly convicted in absentia. Federal Rule of Criminal Procedure 43(c) provides: A defendant who was initially present at trial, or who had pleaded guilty or nolo contendere, waives the right to be present under the following circumstances: (A) when the defendant is voluntarily absent after the trial has begun, regardless of whether the court informed the defendant of an obligation to remain during trial; (B) in a noncapital case, when the defendant is voluntarily absent during sentencing; or (C) when the court warns the defendant that it will remove the defendant from the courtroom for disruptive behavior, but the defendant persists in conduct that justifies removal from the courtroom. FED. R. CRIM. P. 43(c). Here, the only indication that Farrad was convicted in absentia is the PSR’s statement to that effect. R. 53 (Revised PSR at 9) (Page ID #414) (“Defendant Farrad was found guilty, in absentia, by jury trial of distribution of approximately one kilogram of ‘crack’ cocaine.”); see also Reply Br. at 24–25. But “the PSR itself is not evidence” in the classic, conclusive sense, United States v. Traylor, 511 F. App’x 449, 452 (6th Cir. 2013); cf. United States v. Roark, 403 F. App’x 1, 4 n.2 (6th Cir. 2010) (arguing that “[t]he information in the PSR is evidence, at least in the sense that it may be considered at sentencing”), and in any event, as Federal Rule of Criminal Procedure 43 makes clear, it is not necessarily impermissible for a defendant to be found guilty in absentia, so long as that defendant “was initially present at trial,” FED. R. CRIM. P. 43(c). Moreover, as the Government notes, Farrad has twice appealed his prior convictions, and neither appeal noted any concern regarding an impermissible exclusion of Farrad from his own trial. See United States v. Farrad, 76 F. App’x 42 (6th Cir. 2003); United Nos. 16-5102/6730 United States v. Farrad Page 36 States v. Farrad, 41 F. App’x 707 (6th Cir. 2002). In short, the district court did not plainly err on this ground. F. Fifth and Sixth Amendment Challenge to ACCA Sentencing Farrad’s challenge to his sentence also includes a broader constitutional attack on his having been deemed an armed career criminal under 18 U.S.C. § 924(e) without an indictment, information, or jury finding on the question. Appellant’s Br. at 61–62 (alleging violations of the Fifth and Sixth Amendments). But, as Farrad concedes, Appellant’s Br. at 12, 62, his argument is foreclosed by binding precedent. United States v. Nagy, 760 F.3d 485, 487–88 (6th Cir. 2014) (noting that Almendarez-Torres v. United States, 523 U.S. 224 (1998), “which held that prior convictions that enhance a defendant’s sentence are not elements of a crime that must be submitted to a jury,” Nagy, 760 F.3d at 487, “is still good law and will remain so until the Supreme Court explicitly overrules it,” id. at 488 (citation omitted)); see also United States v. Pritchett, 749 F.3d 417, 434 (6th Cir. 2014); United States v. Mack, 729 F.3d 594, 609 (6th Cir. 2013). Farrad understandably “makes this argument in order to preserve it for Supreme Court review,” Appellant’s Br. at 12, but we are not empowered to grant him relief on it. See Salmi v. Sec’y of Health & Human Servs., 774 F.2d 685, 689 (6th Cir. 1985). G. Fourth Amendment Suppression Farrad’s final claim is that the Government’s acquisition of Farrad’s Facebook photos, and the introduction of those photos against him at trial, violated the Fourth Amendment. 22 Appellant’s Br. at 63–69. The parties disagree about the standard of review: Farrad argues that we should use the normal standard—de novo for law, clear error for facts, see, e.g., United States v. Pacheco, 841 F.3d 384, 389 (6th Cir. 2016)—while the Government argues that plain error applies. Compare Appellant’s Br. at 63, with Appellee’s Br. at 51–52. This disagreement stems from the procedural history. After Farrad moved for suppression pro se, R. 22 (Pro Se Mot. at 3– 4) (Page ID #87–88), the magistrate judge denied Farrad’s motion because he was already represented by counsel, R. 24 (Order at 1) (Page ID #93), and Farrad’s trial counsel did not 22As noted above, Farrad also discusses a potential ineffective-assistance-of-counsel claim here. We decline to review that claim at this stage, given that it is underdeveloped and Farrad is still on direct appeal. See, e.g., Martinez, 430 F.3d at 338. Nos. 16-5102/6730 United States v. Farrad Page 37 renew the motion. Farrad argues that because the magistrate judge nevertheless had discretion to consider Farrad’s pro se motion, see E.D. Tenn. Local Rule 83.4(c), the district court’s ruling amounted to a denial of his motion on the merits. Appellant’s Br. at 63–64. The Government argues against such an inference and adds that plain-error review should apply regardless because Farrad’s argument on appeal is different from the argument in his pro se motion. Appellee’s Br. at 52 n.12. The Government is correct: plain-error review applies. Its second rationale is true only in large part; Farrad makes several arguments against the warrant, and all but one went unmade below. Compare R. 22 (Pro Se Mot. at 3–4) (Page ID #87–88) (arguing that evidence relied on to secure search warrant was fruit of illegal search and seizure), with Appellant’s Br. at 64–69 (arguing (1) that “affidavit was facially insufficient,” (2) that “execution was facially flawed,” and (3) that affidavit was “fruit of the poisonous tree”). Having not been raised in Farrad’s pro se motion, Farrad’s first two arguments clearly qualify only for plain-error review. United States v. Lopez-Medina, 461 F.3d 724, 739 (6th Cir. 2006). But more foundationally, the magistrate judge’s very brief order denying Farrad’s pro se motion was clearly based on the procedural problem of having a defendant with counsel filing pro se motions, and nothing in that order discussed the merits or suggested that Farrad’s trial counsel could not immediately file a comparable motion. See R. 24 (Order at 1–2) (Page ID #93–94). Although the magistrate judge’s order does not use the words “without prejudice,” any fair reading of it suggests as much. Cf. Robert N. Clemens Tr. v. Morgan Stanley DW, Inc., 485 F.3d 840, 845 (6th Cir. 2007) (presuming, based on different circumstances, that “the district court must have intended to dismiss the Plaintiffs’ claims with prejudice” (emphasis added)). In short, if the district court was to review this motion on the merits, it was for Farrad’s trial counsel to renew the motion. Because trial counsel did not, review here is entirely for plain error. The district court did not, in turn, plainly err in failing to order suppression of its own accord. First, with regard to Farrad’s argument that the warrant application was facially deficient, Appellant’s Br. at 64–65, Farrad has not provided sufficient grounds to conclude that the issuing judge erred in granting the application, see Illinois v. Gates, 462 U.S. 213, 236 (1983) (“A magistrate’s determination of probable cause should be paid great deference by reviewing Nos. 16-5102/6730 United States v. Farrad Page 38 courts.” (citation and internal quotation marks omitted)), let alone that the executing officer’s reliance on that decision was objectively unreasonable, see United States v. Leon, 468 U.S. 897, 923–24 (1984). The warrant application established not only that Farrad had been convicted of a felony and that various individuals had reported seeing him armed, but also that law enforcement had come across a photograph on what appeared to be Farrad’s Facebook page showing what appeared to be three handguns. R. 5-1 (Warrant Application at 3–4) (Page ID #9–10). That was a colorable basis for the district court to assume good-faith reliance in the absence of a renewed motion. And while Farrad is correct that the application was inartfully worded with regard to the location of what was to be searched, compare R. 5-1 (Warrant Application at 1) (Page ID #7) (providing, in response to prompt for location of “person or property” to be searched, both that reviewer should “See Attachment A” and that it was “located in the Eastern District of Tennessee”), with id. at 11 (Page ID #17) (Attachment A, providing that “[t]his warrant applies to information . . . that is stored at premises owned, maintained, controlled, or operated by Facebook, a company headquartered in Menlo Park, California”), for the district judge sua sponte to have faulted the affidavit on that ground would have been the apotheosis of “interpreting [an affidavit] in a hypertechnical, rather than a commonsense, manner”—exactly what the Supreme Court has directed courts not to do. See Gates, 462 U.S. at 236 (citation omitted). There was certainly no plain error. There was also no plain error based on improper execution. Farrad seems to aim this challenge in part at the fact that the warrant provides for execution “on or before November 6, 2013,” R. 5-4 (Search and Seizure Warrant at 1) (Page ID #25), whereas Facebook’s certification is dated April 17, 2015, R. 26-1 (Facebook Certification) (Page ID #105). See Appellant’s Br. at 66. But this argument appears to conflate the date of a private third party’s evidentiary certification with the date of the Government’s warrant execution. Indeed, the return form certified by the executing officer makes clear that the warrant was “served electronically” on Facebook on November 1, 2013, regardless of whether Facebook responded right away. R. 5-4 (Search and Seizure Warrant at 2) (Page ID #26). While a search made by a private entity “act[ing] at the direction of law enforcement agents . . . must comport with the Fourth Amendment,” United States v. Boumelhem, 339 F.3d 414, 425 (6th Cir. 2003), Farrad has Nos. 16-5102/6730 United States v. Farrad Page 39 pointed to no authority or rationale to suggest that a date of execution similarly binds a third party’s certification of its records for evidentiary purposes.23 This argument lacks merit. Farrad’s improper-execution challenge seems to encompass the warrant’s execution outside of the issuing district as well. See Appellant’s Br. at 66–69. But Farrad again points to no authority to suggest that a federal magistrate judge in the Eastern District of Tennessee could not permissibly issue a warrant to search electronic records bearing on a suspected in-district crime that were located in another district, and indeed federal law appears to allow exactly that. See 18 U.S.C. § 2703(a) (“A governmental entity may require the disclosure by a provider of electronic communication service of the contents of a wire or electronic communication . . . pursuant to a warrant issued using the procedures described in the Federal Rules of Criminal Procedure . . . by a court of competent jurisdiction.”); id. § 2711(3)(A)(i) (defining a “court of competent jurisdiction” as “any district court of the United States (including a magistrate judge of such a court) or any United States court of appeals that . . . has jurisdiction over the offense being investigated”). There was, accordingly, no plain error here either.24 Finally, Farrad argues that “[t]he search warrant was obtained with illegally obtained evidence and anything received as a result of that search warrant would be ‘fruit of the poisonous tree.’” Appellant’s Br. at 67. But in addition to the fact (noted above) that Farrad has pointed to no record evidence to support his assertions regarding an illegal search, Farrad’s claim also fails because there is no mention in the affidavit of the ostensibly illegal search to which Farrad alludes. See R. 5-1 (Warrant Application at 3–4) (Page ID #9–10). Thus, while the district court (if presented with the issue) could have conceivably “look[ed] beyond the four corners of the warrant affidavit to information that was known to the officer and revealed to the issuing 23This logic holds even if we assume that what Farrad means to challenge is not the evidentiary certification, see Appellant’s Br. at 66 (“The certification of Facebook was dated April 17, 2015, outside the scope of the November, 2013 deadline . . . .”), but rather Facebook’s production of the records themselves, which appears to have occurred sometime in 2014. See R. 26 (Gov’t’s Mot. in Limine at 1) (Page ID #100); Appellant’s App’x at 4. In either case, the warrant was still executed on time. 24Farrad also very briefly seeks to raise an issue akin to an issue addressed in Carpenter v. United States, 138 S. Ct. 2206 (2018). See Appellant’s Br. at 68–69. But his argumentation is too perfunctory for us to consider it on appeal. Puckett, 833 F.3d at 610–11. In any event, we cannot see how it could have been plain error for the district court to assume sufficient compliance with the traditional requirements of probable cause in this case in light of Carpenter, given that Carpenter concerns the permissibility of obtaining records under 18 U.S.C. § 2703 with less than probable cause. See Carpenter, 138 S. Ct. at 2212. Nos. 16-5102/6730 United States v. Farrad Page 40 magistrate” to assess good-faith reliance, United States v. Frazier, 423 F.3d 526, 536 (6th Cir. 2005), and while Farrad could conceivably have had grounds to seek a hearing under Franks v. Delaware, 438 U.S. 154, 155–56 (1978), because no evidence was ever offered that the warrant or its issuance were based on any impermissible police activity, there was no plain error in the district court’s failure to inquire into or order suppression on its own. III. CONCLUSION The bottom line in this case—that Farrad has been sentenced to serve 188 months in prison because the Government found Facebook photos of him with what appears to be a gun— may well raise a lay reader’s hackles. There are likewise aspects of Farrad’s trial and conviction—the date issue, Officer Garrison’s testimony—that are at least debatably troubling from a legal perspective. Nevertheless, we are not empowered to grant relief based on arguments not made or where errors were harmless. With that observation, and for the foregoing reasons, we AFFIRM.