A jury convicted Kurt Mallory, Margaret McKnight, and Susan Pioch on various charges arising from a scheme to pilfer Martin Fewlas's sizeable estate. The Defendants appeal their convictions, and Mallory and Pioch further challenge their sentences. We AFFIRM in part, REVERSE in part, VACATE Mallory's and Pioch's sentences in part, and REMAND for further proceedings consistent with this opinion.

I.

Martin Fewlas looked the part of a bar-dweller. Most days, he frequented various neighborhood establishments, drinking beer. At one such locale, he befriended Margaret McKnight. McKnight's mother worked as a bartender, and McKnight often came by after school. Eventually, McKnight became a bartender too. She developed a close friendship with Fewlas-so much so that Fewlas invited McKnight to rent the upstairs apartment in his duplex.

For seventeen years, McKnight lived in this upstairs apartment with her boyfriend, Kurt Mallory. Kurt's estranged father, Gary, later joined the couple during the summers in an attempt to reconcile with his son. Over the years, Margaret fed and cleaned Fewlas when he came home from the bars. Perhaps Fewlas saw McKnight as the daughter he never had. But, as any family knows, sometimes familiarity breeds contempt, and Fewlas and *589 McKnight did not always get along. In particular, Fewlas disliked and distrusted Kurt. And at one point, Fewlas wanted to evict the couple because they had fallen behind on rent.

Eventually, Fewlas passed away. But as it turns out, there was more to him than met the eye. Fewlas had amassed a fortune of over $2.2 million. And in his will, he left it all to McKnight.

McKnight immediately went on a spending spree. She withdrew over $600,000 and made numerous purchases, including a new duplex and a used car dealership. To make these purchases, McKnight took out the money in 171 different transactions-all in amounts less than $10,000. Unsurprisingly, this suspicious conduct got the Internal Revenue Service's attention. And when the IRS delved into the case, they started to suspect that Fewlas had not left his estate to McKnight after all.

The plot soon unraveled. The IRS interviewed Gary Mallory, and he confessed that he had forged Fewlas's signature on the will. According to Gary, attorney Susan Pioch had prepared the fake will, and Kurt asked him to sign Fewlas's name, offering him a cut in return. Gary agreed. But when Gary asked for more money, his co-conspirators spurned his request. So Gary, aggrieved that no honor exists among thieves, tried contacting the probate court to alert them about the fraud and stop McKnight from getting Fewlas's money. For whatever reason, the court brushed off his phone call. When the IRS later got involved, however, Gary cooperated and pled guilty to his involvement in the scheme. His confession set off a chain reaction that resulted in a fifty-nine-count indictment and subsequent convictions across the board.

Kurt Mallory, McKnight, and Pioch now appeal.

II.

Kurt Mallory first claims the district court erred by allowing the government to play the videotaped deposition of Gary Mallory. About six months after the defendants were indicted, the government deposed Gary. At the time, Gary was seventy-six years old and in poor health. He was suffering from a number of debilitating conditions, including lumbar disc disease, diabetes, and recurring bladder cancer. The pain in his back-where two vertebrae were "disintegrating"-was severe. R. 307, Pg. ID 3773-74. So the government, concerned that Gary might not live until trial, sought to preserve his testimony by deposing him. Kurt objected to the deposition, claiming an insufficient time to prepare. But the district court let it go forward anyway. Since Gary could not travel, the deposition took place at a federal courthouse in Arizona, where Gary resided.

After the deposition, Gary's health further deteriorated. He was admitted to the hospital twice-first for hematuria (blood in his urine), and second when he experienced "[a]cute renal failure" (serious kidney malfunction). R. 102-7, Pg. ID 679. He was also diagnosed with dementia. These and his other "chronic" conditions rendered him "homebound" and "unable to safely travel." R. 171-1, Pg. ID 1312; R. 102-5, Pg. ID 668. So when the time for trial approached, the government sought to play the videotape of Gary's deposition in lieu of having him testify at trial. Kurt opposed the government's request, arguing that Gary should have to testify by live video feed from the federal courthouse in Arizona. In light of Gary's condition, however, the district court granted the government's request to play the videotape of the deposition.

On appeal, Kurt contends that the playing of Gary's deposition violated his right *590 under the Sixth Amendment's Confrontation Clause. See U.S. Const. amend. VI ("In all criminal prosecutions, the accused shall enjoy the right ... to be confronted with the witnesses against him ...."). Specifically, Kurt argues (1) that the district court erred in deeming Gary "unavailable" to testify, and (2) that he did not have a meaningful opportunity to cross-examine Gary at his deposition. We address each argument in turn.

Unavailability. Kurt maintains that Gary was not "unavailable" to testify at trial. In order for Gary's deposition to be played at trial without running afoul of the Confrontation Clause, the government needed to show that he was unavailable. Crawford v. Washington , 541 U.S. 36 , 68, 124 S.Ct. 1354 , 158 L.Ed.2d 177 (2004) ; United States v. Quinn , 901 F.2d 522 , 526 (6th Cir. 1990). Illness of an elderly witness can render him unavailable. United States v. Campbell , 845 F.2d 1374 , 1377-78 (6th Cir. 1988) ; accord United States v. Porter , 886 F.3d 562 , 567 (6th Cir. 2018) ; see Fed. R. Evid. 804(a)(4) ("A declarant is considered to be unavailable as a witness if the declarant ... cannot be present or testify at the trial or hearing because of death or a then-existing infirmity, physical illness, or mental illness[.]"). To establish unavailability on account of illness, the government must show that the witness's "specific symptoms" and "the duration and the severity of the illness" preclude the witness from testifying. Porter , 886 F.3d at 567 (quoting Burns v. Clusen , 798 F.2d 931 , 937-38 (7th Cir. 1986) ); Stoner v. Sowders , 997 F.2d 209 , 213 (6th Cir. 1993). We review the district court's finding of unavailability for abuse of discretion. Porter , 886 F.3d at 567 .

Kurt argues that the district court abused its discretion in two principal respects. First, he maintains that the government provided stale medical records that failed to establish that Gary's condition was "significantly different" than it was at the time of his deposition. But if anything, Gary's condition worsened after his deposition, particularly in light of his hospitalizations and dementia diagnosis. And while the government did not provide records detailing Gary's condition for several months leading up to trial, the records it did provide characterized Gary's condition as "chronic" and gave no indication that his condition would improve. R. 171-1, Pg. ID 1312. Our circuit does not require the government to disprove the possibility of "miraculous rejuvenation" in order to establish an elderly witness's unavailability on account of illness. Campbell , 845 F.2d at 1378 . And while Campbell involved a shorter records-gap, the same principle holds true here. See id. ; United States v. McGowan , 590 F.3d 446 , 454 (7th Cir. 2009) (affirming a finding of unavailability based on "the consistency of the reports regarding [the witness's] deteriorating health over time, including a report of her condition a few months before trial"); accord United States v. Bond , 362 F. App'x 18 , 22 (11th Cir. 2010) (per curiam). The district court did not abuse its discretion by declining to require the government to provide more medical records confirming what everyone already knew.

Second, Kurt maintains that the district court should have required Gary to testify by live video feed from the Arizona courthouse. Kurt points out that, in the months leading up to trial, Gary attended his own sentencing hearing at that very courthouse. But Gary's presence at his sentencing hearing did not compel the district court to mandate his live testimony at Kurt's trial. As the district court noted, Gary was so ill that he could not leave his home-not even to see his doctor. So whatever the circumstances that permitted Gary to attend his sentencing hearing, the *591 district court was not obligated to ignore Gary's homebound status. This is especially true given that Gary's sentencing hearing was fourteen minutes-hardly the pressure-filled hours Gary would likely endure in giving trial testimony. McGowan , 590 F.3d at 454 (noting the "rigors of testifying at trial"). Thus, the district court did not abuse its discretion in deeming Gary unavailable.

Meaningful Opportunity . Kurt also argues that he did not have a meaningful opportunity to cross-examine Gary at his deposition. The Confrontation Clause mandates such an opportunity, but not one that is "effective in whatever way, and to whatever extent, the defense may wish." United States v. Owens , 484 U.S. 554 , 559, 108 S.Ct. 838 , 98 L.Ed.2d 951 (1988) (quoting Kentucky v. Stincer , 482 U.S. 730 , 739, 107 S.Ct. 2658 , 96 L.Ed.2d 631 (1987) ). And while cross-examination before trial may not be as searching as cross-examination at trial, prior testimony is nevertheless admissible so long as the defendant was not "significantly limited" in his cross-examination. California v. Green , 399 U.S. 149 , 166, 90 S.Ct. 1930 , 26 L.Ed.2d 489 (1970) ; see also Crawford , 541 U.S. at 68 , 124 S.Ct. 1354 (confirming admissibility of prior testimony subject to the declarant's unavailability and "a prior opportunity for cross-examination"). Since Kurt raised his meaningful-opportunity argument in a motion for a new trial under Federal Rule of Criminal Procedure 33, we review the district court's decision to deny him a new trial for abuse of discretion. 1 United States v. Munoz , 605 F.3d 359 , 366 (6th Cir. 2010) ; see United States v. Stewart , 5 F. App'x 402 , 407 (6th Cir. 2001) (per curiam).

Kurt cites four reasons why he could not meaningfully cross-examine Gary at the deposition. First, he maintains that the two-month period that counsel had to prepare for the deposition was insufficient in light of voluminous discovery. But he fails to identify what he wishes counsel could have reviewed but did not. This indicates that counsel's preparation, even if hurried, was not so rushed as to significantly limit his ability to cross-examine Gary, let alone compel a new trial. See Green , 399 U.S. at 166 , 90 S.Ct. 1930 ; United States v. Casamento , 887 F.2d 1141 , 1173 (2d Cir. 1989) (holding that eleven days' notice of declarant's deposition was sufficient, even though "defense counsel understandably would have preferred more time to prepare for [the] deposition").

Second, Mallory asserts that he received discovery after the deposition that he could have used to impeach Gary's testimony. The only discovery he points to, however, is forensic testing of the forged will. And while Kurt claims he was not privy to the results of the forensic tests before the deposition, he does not explain what additional cross-examination he would have performed with the results of those tests. See Mechler v. Procunier , 754 F.2d 1294 , 1299 (5th Cir. 1985) (rejecting Confrontation Clause claim premised on autopsy and *592 ballistics report unavailable before preliminary hearing since "counsel ha[d] not demonstrated how additional questioning on the subject would have aided" cross-examination conducted at the hearing). Moreover, since Kurt did not mention this issue in his motion for a new trial, the district court did not have an opportunity to consider it.

Third, Kurt notes that he did not know of Gary's dementia diagnosis at the time of the deposition, and he would have liked to cross-examine Gary with that information. Yet Gary's medical records from before his deposition indicate that he suffered from "weakness and episodes of confusion." R. 102-2, Pg. ID 640. In fact, Pioch's counsel asked Gary at his deposition whether he had been tested for Alzheimer's, a line of questioning Kurt's counsel evidently chose not to pursue. See United States v. Richardson , 781 F.3d 237 , 245 (5th Cir. 2015) (denying Confrontation Clause claim where proposed additional questioning did not present "a 'new and significantly material line of cross examination that was not at least touched upon ' " in prior testimony (quoting Mancusi v. Stubbs , 408 U.S. 204 , 215, 92 S.Ct. 2308 , 33 L.Ed.2d 293 (1972) ) ). More tellingly, the court offered Kurt the opportunity to impeach Gary's deposition testimony at trial-if not with a medical expert, then at least with the mere fact of his dementia. But for whatever reason, Kurt did not do so, even though Gary, being absent, could not have responded. So here again, Kurt's actions belie an unmet need to confront Gary about his dementia. And if Kurt felt it unnecessary to present the jury with Gary's dementia at his first trial, the "interest of justice" does not demand a second. Fed. R. Crim. P. 33(a).

Fourth, Kurt claims that he learned of conflicting stories about the will signing from Gary's daughters only after the deposition, and thus did not have the information to impeach Gary's account of what went on. But Kurt interviewed the daughters before the deposition and even asked Gary about the "several different versions" of the will signing that he had told one of his daughters. R. 307, Pg. ID 3828-29. This indicates that Kurt had a meaningful opportunity to cross-examine Gary about the conflicting stories.

Kurt thus fails to identify a basis for concluding that his opportunity to confront Gary was not meaningful-much less one that would compel the district court to grant a new trial. Therefore, his Confrontation Clause claim falls short.

III.

Next, we turn to McKnight's and Pioch's objections to the court's admission of testimony by a government handwriting expert. The expert, Larry Olson, testified that the signature on the forged will was "probably" not Fewlas's, but instead a "simulation" performed by someone else. R. 308, Pg. ID 4052. McKnight and Pioch concede that Olson was qualified to offer this opinion, but dispute its reliability.

District courts are the "gatekeep[ers]" of expert testimony. Daubert v. Merrell Dow Pharm., Inc. , 509 U.S. 579 , 597, 113 S.Ct. 2786 , 125 L.Ed.2d 469 (1993). Federal Rule of Evidence 702 tasks them with determining whether an expert's "scientific, technical, or other specialized knowledge" will assist the trier of fact in doing its job, and whether the expert's testimony "is the product of reliable principles and methods." Fed. R. Evid. 702(a), (c). To perform this inquiry, district courts often consider whether an expert's theory has been tested, is the subject of peer review and publication, has a permissible error rate, follows established standards, and receives "general acceptance" within a "relevant scientific community."

*593 Daubert , 509 U.S. at 593-94 , 113 S.Ct. 2786 . These so-called Daubert factors, while perhaps most apt in evaluating a purely scientific discipline, can also apply in evaluating non-scientific fields that are "technical" or "specialized" in nature. Kumho Tire Co., Ltd. v. Carmichael , 526 U.S. 137 , 149-53, 119 S.Ct. 1167 , 143 L.Ed.2d 238 (1999). But the factors are not mandatory in every case. Id. at 150-51 , 119 S.Ct. 1167 . The questions of what factors to apply and what conclusion to draw about an expert's reliability are entrusted to the district court's discretion. Id. at 152-53 , 119 S.Ct. 1167 .

The reliability of expert handwriting analysis has come before our court before. In United States v. Jones , our court upheld the admissibility of such testimony. 107 F.3d 1147 , 1161 (6th Cir. 1997). In so holding, Jones explained that handwriting analysis is not a science per se . Id. at 1157 . Handwriting analysts "do not concentrate on 'proposing and refining theoretical explanations about the world,' " as scientists do. Id. (quoting Daubert , 509 U.S. at 590 , 113 S.Ct. 2786 ). Instead, handwriting analysts "use their knowledge and experience to answer the extremely practical question of whether a signature is genuine or forged." Id. Handwriting analysts see things in handwriting that laypeople do not-both because of analysts' training in the minutiae of loops, swoops, and dotted "i"s, and because of the volume of handwriting they inspect-and therefore assist the trier of fact by bringing their training and experience to bear. See id. at 1159-61. Thus, while handwriting analysis may not boast the "empirical" support underpinning scientific disciplines, it is nevertheless "technical" or "specialized" knowledge that, subject to thorough gatekeeping, is a proper area of expertise. Id. at 1157-61 ; see Fed. R. Evid. 702(a).

Our court decided Jones without the benefit of Kumho Tire . In Kumho Tire , the Supreme Court clarified that the Daubert factors may also be useful in scrutinizing non-scientific expertise. 526 U.S. at 149 , 119 S.Ct. 1167 . In doing so, however, the Court instructed that application of the Daubert factors should be flexible and that one or more factors may or may not apply depending on the type of expertise at hand. Id. at 150-51 , 119 S.Ct. 1167 . In fact, the Court referenced handwriting analysis as an area where strict Daubert -type analysis might be less appropriate, indicating that "the relevant reliability concerns may focus upon personal knowledge or experience." Id. at 150 , 119 S.Ct. 1167 . Since Jones predated Kumho Tire , it did not apply the Daubert factors in evaluating the handwriting analysis at issue. Jones , 107 F.3d at 1158-59 . Still, Jones 's focus on handwriting analysts' experience-based expertise is consistent with Kumho Tire , even though Daubert -type inquiries may also be appropriate in evaluating such testimony.

Here, the district court faithfully applied Daubert , Jones , and Kumho Tire in deeming Olson's handwriting analysis admissible. The court conducted thorough voir dire to ascertain Olson's experience and methodology. Olson testified to his thirty-one years' experience as an ink chemist and forensic document examiner at the IRS National Forensic Laboratory, during which he has performed countless handwriting analyses and testified in court on multiple occasions. See Jones , 107 F.3d at 1160 . He explained that his laboratory is accredited by an international organization that polices "general standards [practiced] throughout the discipline." R. 307, Pg. ID 3692; see Daubert , 509 U.S. at 594 , 113 S.Ct. 2786 . In addition, Olson walked through the principles and basic approach he used in performing his analysis. To perform the analysis, Olson studied approximately *594 ninety-one known examples of Fewlas's signature. From those samples, he discerned various unique characteristics, many of which he then found lacking in the signature on the forged will. As Olson explained, this approach embodies two precepts-no two people write exactly alike, and no one person writes exactly the same every time-which he represented as having been tested in various studies and experiments. See United States v. Prime , 431 F.3d 1147 , 1153 (9th Cir. 2005) (affirming admission of handwriting expert citing one of the same studies). Those studies and experiments, according to Olson, further establish that his mode of analysis is highly accurate. See id. (observing that admitted handwriting expert cited low error rate); accord United States v. Mooney , 315 F.3d 54 , 62 (1st Cir. 2002). Moreover, Olson testified that his laboratory requires document examiners to review each other's work, and that in this case, another document examiner not only reviewed his work but independently verified his opinion. See Prime , 431 F.3d at 1153 (highlighting similar review and verification); accord United States v. Crisp , 324 F.3d 261 , 271 (4th Cir. 2003) ; Mooney , 315 F.3d at 62 . Based on this testimony, the district court did not abuse its discretion in deeming Olson's testimony reliable. Jones , 107 F.3d at 1161 ; see Prime , 431 F.3d at 1153-54 (affirming admission of handwriting expert offering similar testimony); accord, e.g. , Crisp , 324 F.3d at 270-71 ; Mooney , 315 F.3d at 62-63 .

McKnight and Pioch make two primary arguments otherwise. First, they take issue with the district court's characterization of handwriting analysis as a "science," pointing to certain studies and district court decisions that have questioned its scientific underpinnings. See, e.g. , United States v. Saelee , 162 F.Supp.2d 1097 , 1101 (D. Alaska 2001). But this argument misses the mark. Handwriting analysis, of course, is not a science- Jones makes that much clear. 107 F.3d at 1157 . The district court's loose language in describing handwriting analysis as a science, however, was more of an afterthought to otherwise thorough gatekeeping. The court's voir dire demonstrates that, rather than viewing handwriting analysis as a science, it sought to ascertain whether Olson's experience-based expertise was reliable. Plus, the district court immediately conditioned its "science" misnomer by noting that handwriting analysis does not involve "beakers and computers and yardsticks." R. 307, Pg. ID 3765. This statement reflects the district court's proper understanding of the true nature of handwriting expertise.

Moreover, while McKnight and Pioch are correct to point out that some have called handwriting analysis's scientific underpinnings into question, that debate does not mean the district court abused its discretion here. For one thing, McKnight and Pioch did little to impugn the studies and experiments validating handwriting analysis on which Olson relied, offering no argument about testing, peer review, or error rate after voir dire . For another, focusing exclusively on the scientific shortcomings of handwriting analysis-a nonscientific area of expertise-ignores its foundations in "personal knowledge [and] experience." See Kumho Tire , 526 U.S. at 151-52 , 119 S.Ct. 1167 ; Jones , 107 F.3d at 1158-61 . And in the end, even if some have called handwriting analysis into question, an expertise need not be perfect to be admissible. Mooney , 315 F.3d at 63 (noting that Daubert does not require that expert testimony be "unassailable"). As Jones teaches, expert testimony that satisfies Rule 702 remains subject to cross-examination on reliability grounds. 107 F.3d at 1161 . McKnight and Pioch were free to expose handwriting analysis's purported failings to the jury by attacking the weight of *595 Olson's testimony, but elected not to do so. Thus, McKnight's and Pioch's challenges to the "science" of handwriting analysis are unavailing.

Second, McKnight and Pioch argue that Olson failed to properly account for Fewlas's alcohol use and other factors that might impact his handwriting. But in both voir dire and before the jury, Olson testified that he considered whether these factors might have impacted Fewlas's writing and concluded that any impact did not change his opinion. McKnight and Pioch provide no support for the notion that this aspect of Olson's opinion, even if flawed, rendered his testimony so unreliable as to be inadmissible. Rather, their argument bears on the proper weight to give his testimony. See id. at 1161 . And McKnight and Pioch each cross-examined him on this point. McKnight and Pioch are obviously disappointed with the weight the jury gave Olson's testimony, but that does not render his testimony inadmissible.

Accordingly, the district court did not abuse its discretion in admitting Olson's handwriting analysis in this case.

IV.

The Defendants raise two additional challenges to their convictions. Neither has merit.

Exhibit . Pioch and Kurt Mallory call into question one of the government's exhibits at trial: an unexecuted power of attorney that Pioch allegedly fabricated. In their view, the government failed to authenticate the document as Pioch's work. To authenticate a document, the proponent must provide evidence sufficient for the factfinder to conclude that the document is what it is claimed to be. Fed. R. Evid. 901(a) ; United States v. Farrad , 895 F.3d 859 , 876 (6th Cir. 2018). That evidence can come from witnesses with personal knowledge about the document or from "distinctive characteristics of the item, taken together with all the circumstances." Fed. R. Evid. 901(b)(1), (4). Here, the government provided sufficient evidence to authenticate the document as Pioch's creation. First, the government obtained the document from Pioch's proposed trial exhibits. See United States v. Crosgrove , 637 F.3d 646 , 658 (6th Cir. 2011) (relying in part on evidence of where a document originated in deeming the document properly authenticated). Second, Pioch testified that her secretary "could have" prepared the document. R. 312, Pg. ID 4876; see Crosgrove , 637 F.3d at 658 . And third, the document bore Pioch's signature line. See Jones , 107 F.3d at 1150 (observing that an individual's signature on a letter, when combined with other distinctive evidence of authenticity, could demonstrate authorship). Together, this foundational evidence was sufficient to clear authentication's "relatively low[ ] hurdle," Farrad , 895 F.3d at 878 , leaving to the jury whether Pioch in fact fabricated the document.

Kurt further objects that the exhibit inflicted unfair prejudice that substantially outweighed its probative value, thus violating Federal Rule of Evidence 403. But the prejudice Kurt claims is simply the possibility that the jury believed the exhibit was in fact Pioch's work. That "prejudice" is not what Rule 403 envisions. Old Chief v. United States , 519 U.S. 172 , 180, 117 S.Ct. 644 , 136 L.Ed.2d 574 (1997) (explaining that unfair prejudice under Rule 403"speaks to the capacity of some concededly relevant evidence to lure the factfinder into declaring guilt on a ground different from proof specific to the offense charged" (emphasis added) ); United States v. Bruce , 437 F. App'x 357 , 363 (6th Cir. 2011) (rejecting Rule 403 argument that jury may have failed to appreciate *596 "discrepancies" in weighing properly authenticated exhibits). Rather, Kurt's complaint of prejudice merely highlights the exhibit's probative value in undermining Pioch's credibility. His objection under Rule 403 therefore fails.

Reid Technique . McKnight argues that a government investigator made improper use of the so-called Reid Technique, an interrogation method used to elicit a confession. But despite the investigator's purported use of the technique, McKnight never confessed, and she does not bring an involuntary-confession claim. Instead, she complains of "false investigation," charging that the investigator unfairly assumed her guilt and marshaled evidence to support his assumption. McKnight Opening Br. 13. But McKnight fails to identify any constitutional provision that bars an investigator from assuming guilt and trying to prove it. See Garner v. Harrod , 656 F. App'x 755 , 760 (6th Cir. 2016) ("[T]here is no constitutionally protected right to the manner in which a criminal investigation is conducted."). And while McKnight alludes to fabrication of evidence, she fails to specify any evidence the investigator fabricated. See id. at 760-61 . Absent that sort of misconduct, it is the Constitution's various safeguards, including trial by an impartial jury, the requirement of proof beyond reasonable doubt, and confrontation of witnesses, that guard against unfounded charges of guilt. See U.S. Const. amend. VI ; In re Winship , 397 U.S. 358 , 363-64, 90 S.Ct. 1068 , 25 L.Ed.2d 368 (1970). McKnight received these guarantees and thoroughly cross-examined the government's investigator. The Constitution requires nothing more.

V.

After the jury handed down its verdict, Kurt Mallory sought a new trial under Federal Rule of Criminal Procedure 33, arguing in part that the verdict was against the manifest weight of the evidence. The district court rejected his request. Kurt argues that the district court applied the incorrect legal standard in denying his motion.

Rule 33 permits a new trial if a verdict is against the "manifest weight" of the evidence. United States v. Hughes , 505 F.3d 578 , 592 (6th Cir. 2007). Such a motion calls on the trial judge to take on the role of a thirteenth juror, weighing evidence and making credibility determinations firsthand to ensure there is not a miscarriage of justice. See id. at 593 (citing United States v. Lutz , 154 F.3d 581 , 589 (6th Cir. 1998) ). This differs from a motion under Rule 29, which challenges the sufficiency of the evidence. Rule 29 asks whether " any rational trier of fact could have found the essential elements of the crime beyond a reasonable doubt," Jackson v. Virginia , 443 U.S. 307 , 319, 99 S.Ct. 2781 , 61 L.Ed.2d 560 (1979) -not whether the trial judge himself believes the manifest weight of the evidence supports the verdict, see Hughes , 505 F.3d at 592-93 . So while Rule 29 requires the court to view the evidence in a light most favorable to the prosecution, id. at 592 , Rule 33 does not.

Here, Kurt claims the district court conflated these rules. And he is right. The court repeatedly characterized its task as evaluating the sufficiency of the evidence, rather than weighing the evidence for itself. See, e.g. , R. 369, Pg. ID 6053 (characterizing issue as "insufficiency of the evidence"); id. , Pg. ID 6054 (criticizing "evidence-parsing arguments more suitable to closing to a jury than in new trial motions"); id. , Pg. ID 6059 (reaching a conclusion as to "Sufficiency of the Evidence"). And the court "[v]iewed [the evidence] most favorably in support of the *597 guilty verdict" and found it "clearly sufficient to support" Kurt's conviction-thus applying Rule 29 and not Rule 33. Id. , Pg. ID 6059. Confusingly, the court also made statements implicating Rule 33. It referenced the manifest-weight standard at one point and thoroughly analyzed Pioch's credibility. Id. , Pg. ID 6054, 6056-59. But these moments of clarity do not cure the court's application of the wrong legal standard. When read as a whole, the court's order demonstrates that it asked the wrong question in denying Kurt's motion.

In the end, the manifest weight of the evidence may support the verdict. But as an appellate court, this is not for us to say. The judge that saw the witnesses and sat with the evidence at trial must make that call. United States v. Ashworth , 836 F.2d 260 , 266 (6th Cir. 1988) (articulating "the appropriate standard for district courts to follow" in evaluating Rule 33 motions and highlighting that "[t]he court of appeals ... does not sit as a 'thirteenth juror' to judge the credibility of witnesses[ or] reweigh the evidence" (emphases omitted) ). We therefore remand for the district court to apply the correct legal standard in considering Kurt's motion.

VI.

At sentencing, the district court enhanced Kurt Mallory's and Pioch's sentencing ranges after concluding that they caused financial hardship to the putative beneficiary of Fewlas's estate. See U.S. Sentencing Guidelines Manual § 2B1.1(b)(2)(A)(iii) (U.S. Sentencing Comm'n 2016). But the Sentencing Guidelines did not contain the financial-hardship enhancement at the time of Kurt's and Pioch's misconduct. See U.S. Sentencing Guidelines Manual § 2B1.1(b)(2)(A) (U.S. Sentencing Comm'n 2014). Thus, to apply it in their case violates the Ex Post Facto Clause. U.S. Const. art. I, § 9, cl. 3 ; Peugh v. United States , 569 U.S. 530 , 544, 550, 133 S.Ct. 2072 , 186 L.Ed.2d 84 (2013). The government agrees and asks that we vacate those parts of their sentences affected by the error. 2 We vacate their sentences accordingly and remand for resentencing.

* * *

We AFFIRM McKnight's and Pioch's convictions, as well as Kurt Mallory's conviction to the extent expressed in this opinion, REVERSE the district court's denial of Kurt's motion for a new trial, VACATE Kurt's and Pioch's sentences to the extent impacted by the financial-hardship enhancement, and REMAND for further proceedings consistent with this opinion.