*327 A federal jury convicted Roberto Trinidad del Carpio Frescas of engaging in wire fraud and then laundering the proceeds. He cheated Mexican "investors" out of at least $5 million in a multi-year transnational Ponzi scheme. Del Carpio nonetheless brings a variety of challenges to his convictions, restitution order, and sentence. We affirm the convictions and restitution order in full. But the district court's Guidelines calculation was off by a single point. So under current Supreme Court precedent and the facts of this case, we have no choice but to vacate the sentence and remand for the limited purpose of resentencing.

I.

El Paso police first learned of del Carpio's Ponzi scheme in 2011. Their first witness was Luz Elva Martinez Rivera. In her thirty years working as a school teacher in Mexico, Martinez Rivera had saved $165,000. Believing del Carpio's promise that he would pay her 15% interest, she drove from her home in Chihuahua, Mexico, to El Paso, Texas, and deposited every penny into del Carpio's bank account. She lost everything.

After speaking with Martinez Rivera, Detective Nichole Ramm spoke with more than 100 other victims. Most were from Chihuahua, Mexico. All had similar stories to tell: Del Carpio held himself out as a stock broker, solicited their investments, promised them big returns, and took their money. When they asked for status updates, del Carpio often responded evasively. Eventually he stopped responding entirely.

The government charged del Carpio with twenty-five counts of wire fraud and ten counts of money laundering. 18 U.S.C. §§ 1343 , 1957(a). Specifically, del Carpio "caused [certain writings or signals] to be transmitted by means of wire communication in foreign and interstate commerce" to further a scheme to defraud others of their money. Then he transferred those ill-gotten gains to himself and his family for personal use. An eleven-day trial included testimony from three investigators, three bank employees, ten of the victims named in the indictment, and seven other victims impacted by the scheme. The jury convicted del Carpio on thirty-four counts.

Prior to the sentencing hearing, the probation office prepared a Pre-Sentence Report ("PSR") under the 2015 Sentencing Guidelines. The PSR grouped all thirty-four counts together under Chapter 3, Part D of those Guidelines. It then identified money laundering as the relevant offense guideline for the group.

Next, the PSR identified the base offense level for money laundering. The money laundering guideline required a base offense level equal to "[t]he offense level for the underlying offense for which the laundered funds were derived"-in this case, wire fraud. U.S.S.G. § 2S1.1(a). Based on the wire fraud provisions, the PSR assigned a base offense level of 7. Then the PSR identified the relevant specific offense characteristics under Chapter Two and two adjustments under Chapter Three of the Guidelines:

*328 • 18 points under § 2B1.1(b)(1)(J) because del Carpio caused more than $3.5 million in economic loss;

• 6 points under § 2B1.1(b)(2)(C) because del Carpio caused hardship to more than 25 people;

• 2 points under § 2B1.1(b)(10)(B) because del Carpio committed much of his scheme from outside the United States;

• 1 point under § 2S1.1(b)(2)(A) because del Carpio was convicted of money laundering under 18 U.S.C. § 1957 ;

• 2 points under § 3B1.3 because del Carpio abused his victims' trust; and

• 4 points under § 3B1.1(a) because del Carpio organized or led a scheme "that involved five or more participants or was otherwise extensive."

Adding those together, the PSR calculated an offense level of 40. Del Carpio fell in criminal history category I. So the Guidelines yielded a range of 292 to 365 months in prison.

Del Carpio contested one of the specific offense characteristics under Chapter Two-namely, that his offense caused hardship to more than 25 people. See U.S.S.G. § 2B1.1(b)(2)(C). And he contested both of the Chapter Three enhancements-the abuse-of-trust enhancement and the leadership enhancement. See id. §§ 3B1.3, 3B1.1(a). After an evidentiary hearing, the court concluded that both enhancements applied in full. But it modified the specific offense characteristics under Chapter Two. It concluded del Carpio caused hardship to at least 5 people, but perhaps not 25. So it applied 4 points under § 2B1.1(b)(2)(B) rather than 6 points under § 2B1.1(b)(2)(C). The court also granted a 2-point reduction because del Carpio had assisted investigators. Combining this new offense level of 36 with a criminal history category of I yielded a Guidelines range of 188 to 235 months.

In his allocution at sentencing, del Carpio suggested he ran a legitimate business that just turned south. "I am a man fearful of God," he said. The court rebuked him: "What did the conversation with God sound like when you took that poor school teacher's life savings that she worked all her life to save?" The court sentenced del Carpio to concurrent sentences of 235 months for the wire fraud counts and 120 months for the money laundering counts. Two weeks later, the court began its restitution hearing. A month after that, the court ordered del Carpio to pay back $5,402,661.

Del Carpio appealed the district court's judgment and sentence, as well as its restitution order.

II.

We affirm del Carpio's convictions because sufficient evidence supports them. We also affirm the district court's restitution order.

A.

Del Carpio challenges the evidentiary sufficiency of his convictions on a handful of wire fraud and money laundering counts. In a sufficiency challenge, the question is not "whether [this court] believes that the evidence at the trial established guilt beyond a reasonable doubt." Jackson v. Virginia , 443 U.S. 307 , 318-19, 99 S.Ct. 2781 , 61 L.Ed.2d 560 (1979). Rather, the familiar test is "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." Id. at 319 , 99 S.Ct. 2781 .

*329 1.

We start with wire fraud. In a wire fraud prosecution, the government must prove that (1) a scheme to defraud exists, (2) the defendant used wire communications in interstate or foreign commerce to further that scheme, and (3) the defendant had specific intent to defraud. See 18 U.S.C. § 1343 ; United States v. Harris , 821 F.3d 589 , 598 (5th Cir. 2016). Del Carpio challenges the sufficiency of the government's wire fraud evidence on Counts 13, 23, and 24. We affirm the sufficiency of each in turn.

In Count 13, the government charged del Carpio with fraudulently inducing Rodrigo Muñiz Vallina to wire him $100,000 on July 5, 2011. Del Carpio challenges the jury's guilty verdict on this count based only on wire fraud's second element-that the government failed to show this money moved in interstate or foreign commerce. Relatedly, he challenges the district court's decision permitting the government to introduce a summary chart ("Exhibit 42") that purported to list the transaction details for each wire fraud count.

Del Carpio complains that Exhibit 42 created the false appearance that the $100,000 moved from Mexico to New York. The chart lists Intercam Casa de Bolsa (in Mexico) as the "Origin Bank" and Bank of America (in New York) as the "Destination Bank." Jurors relying on that chart, then, would conclude the money moved in foreign commerce. In reality, del Carpio contends, the money only moved from one New York bank to another New York bank.

The district court did not abuse its discretion in admitting Exhibit 42. For starters, the court, the prosecutor, and defense counsel repeatedly reminded the jury that the chart was not evidence. Plus, the chart was generally consistent with the testimony at trial. Roxanne Hollingsworth, a Bank of America employee, testified that the July 5th transaction was requested by an originator in Chihuahua, Mexico, before passing through an originating bank in New York (Standard Chartered Bank, Ltd.), and landing at a receiving bank in New York (Bank of America). Based on these facts, Hollingsworth agreed that "this wire transfer also [was] interstate." Del Carpio never objected to her conclusion. He did not even ask Hollingsworth about it on cross.

On appeal, however, del Carpio insists Hollingsworth's testimony shows the money moved only intra state-between two New York banks. That is irrelevant even if true: All that needs to move across national or state lines is a "writing[ ], sign[ ], signal[ ], picture[ ], or sound[ ]" that furthers the fraud. 18 U.S.C. § 1343 . Therefore, a customer in Mexico who sends a digital request-by email, phone, or some other conceivable means-to his bank in New York asking it to transfer money to another New York bank transmits an inter national "writing," "signal," or "sound" that facilitates an intrastate transfer. See United States v. Arledge , 553 F.3d 881 , 892 (5th Cir. 2008) (facsimile); United States v. Johnson , 700 F.2d 163 , 176-77 (5th Cir. 1983) (phone call); United States v. Foster , 878 F.3d 1297 , 1304-05 (11th Cir. 2018) (email). The evidence was sufficient. 1

*330 In Counts 23 and 24, the government charged del Carpio with fraudulently inducing Miguel Luevano Gutierrez and Augustine Jiminez Leyva to wire him $13,970 and $2,988, respectively. Del Carpio challenges the jury's guilty verdicts on those counts based only on wire fraud's third element-that the government showed no specific intent to defraud. Del Carpio insists that because neither victim testified, we do not know why they sent money to him.

That does not cut it. Detective Ramm testified that she spoke with 110 victims, that "all of these people gave [her] reports" and produced documents showing the "investments" they made with del Carpio, and that two of the victims she spoke with were Luevano Gutierrez and Jiminez Leyva. "In general, the [victims'] allegation was theft ... and that they had invested money that they did not believe had been invested."

This evidence may not have been detailed. But "the relevant question is 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." Jackson , 443 U.S. at 319 , 99 S.Ct. 2781 . A jury could infer Luevano Gutierrez and Jiminez Leyva complained to the El Paso police because they were drawn to del Carpio by the same lure as his other victims-the false promise of bountiful returns. The evidence on Counts 23 and 24 was sufficient.

2.

It is also a federal crime to launder dirty money. See 18 U.S.C. § 1957 (a). In a money laundering prosecution, the government must prove that (1) the defendant knowingly engaged in a monetary transaction, (2) the transaction involved criminally derived property worth more than $10,000, and (3) the defendant knew the property was derived from criminal activity. United States v. Rodriguez , 278 F.3d 486 , 490 (5th Cir. 2002). Del Carpio challenges the sufficiency of the government's money laundering evidence in Counts 27 and 32. We affirm as to both.

In Counts 27 and 32, the government charged del Carpio with laundering some of the fraudulent proceeds by transferring money from his Wells Fargo account on March 18, 2011 ($20,000) and March 25, 2011 ($21,859.84). Del Carpio challenges the jury's guilty verdicts on these counts based only on money laundering's second element-that he transferred criminally derived funds. That is so, he says, because his Wells Fargo account had at least some clean money in it. In particular, del Carpio says the account contained clean funds totaling $57,800 just before the two March withdrawals totaling $41,859.84.

But the government contends del Carpio's Wells Fargo account contained no clean funds at all. It points to Secret Service Agent Brian Cummings's testimony. El Paso police asked for his help on "interacting with the banks, issuing subpoenas, analyzing the cash flow and getting the story of what happened with the money." At trial, Agent Cummings testified that:

All money going into any of the Bank of America or Wells Fargo accounts we've looked at, which usually routed through Chase, with some exceptions, it was deposited directly, but it always came from investor/victims . And we made sure to verify through cash flow analysis that any time we saw a large transaction *331 or large purchase like that, we wanted to know where the source of the money was.

(Emphasis added.)

Del Carpio responds that this testimony is ambiguous. Was Cummings referring to all deposits to del Carpio's Wells Fargo account, or just a subset of the deposits "we've looked at"? It is a fair question. But it is precisely the kind of question the trier of fact is best situated to answer. And del Carpio's counsel never cross-examined Cummings on the point. (The cross-examination barely exceeds a single page in the trial transcript.) Based on the evidence at trial, a reasonable juror could find del Carpio withdrew dirty money.

B.

Del Carpio also challenges the district court's restitution order directing him to pay $5,402,661 under the Mandatory Victim Restitution Act. See 18 U.S.C. §§ 3663A, 3664. We review this order through a deferential abuse-of-discretion lens, and the defendant bears the burden of showing he is entitled to a reduction based on returns he made to victims. United States v. De Leon , 728 F.3d 500 , 506-07 (5th Cir. 2013). A defendant can offset the restitution amount only if "there is ... evidence that [not] doing so would result in double recovery to the victim." United States v. Taylor , 582 F.3d 558 , 567 (5th Cir. 2009) (per curiam).

Mike Petron, a certified public accountant and a certified fraud examiner, testified at the restitution hearing. He initially determined that del Carpio caused around $6.5 million in losses. Petron reviewed both the government's and del Carpio's submissions and concluded del Carpio had already returned a little over $1 million of that money. Petron therefore offset that amount against the losses del Carpio caused. But Petron refused to credit an additional $254,297 "where there was no third-party documentation" supporting del Carpio's request for an offset. So Petron settled on a total restitution amount of $5,402,661. The district court agreed.

Del Carpio claims the court should have offset an additional $181,138.05, leaving $5,221,522.95 in restitution. First, he says the government provided records from JP Morgan showing that he paid $158,788.05 to various victims. He cites to pages in "Government Exhibit 3." But as far as we can tell, that document is nowhere in the record on appeal. And in all events, it was undisputed in the district court that no third-party documentation supported offsetting these amounts. Below, del Carpio simply dumped a bunch of documents on the government's expert and said, in effect, figure it out. He cannot prove a double recovery for the first time on appeal by creating tables in his brief that he never showed to Petron and that are unsupported by record citations. "Judges are not like pigs, hunting for truffles buried in the record." Albrechtsen v. Bd. of Regents of Univ. of Wisconsin Sys. , 309 F.3d 433 , 436 (7th Cir. 2002) (Easterbrook, J.) (quotation omitted).

Second, del Carpio says his claim is supported by "investor-victim testimony." But in the district court, he conceded that his victims "may not have remembered [the] amounts." Third, del Carpio says certain emails show he paid $22,350 to Martinez Rivera. But the district court did not abuse its discretion by demanding bank documentation rather than email screenshots. We affirm the restitution order in full.

III.

Finally, del Carpio brings two challenges to the district court's sentence. First, he says the court failed to calculate *332 the offense levels for wire fraud and money laundering separately. Second, he says the court erred in the way it applied the abuse-of-trust and leadership enhancements. He concedes he never raised either objection below. We need not consider the first argument because, under current Supreme Court precedent, del Carpio must be resentenced under the second.

Given del Carpio's failure to object, our review is for plain error. Federal Rule of Criminal Procedure 52(b) provides that "[a] plain error that affects substantial rights may be considered even though it was not brought to the court's attention." This language requires a four-part showing: The defendant must show that (1) the district court committed an error, (2) the error is plain, (3) the error affects his substantial rights, and (4) failure to correct the error would seriously affect the fairness, integrity, or public reputation of judicial proceedings. See United States v. Sanchez-Hernandez , 931 F.3d 408 , 410-11, 2019 WL 3333731 , at *2 (5th Cir. July 25, 2019).

As explained above, the district court sentenced del Carpio under the money laundering guideline, § 2S1.1(a). The note to that guideline says the "application of any Chapter Three adjustment[s] shall be determined based on the offense covered by this guideline ( i.e. , the laundering of criminally derived funds) and not on the underlying offense from which the laundered funds were derived." U.S.S.G. § 2S1.1 application note 2(C); see United States v. Salgado , 745 F.3d 1135 , 1138-39 (11th Cir. 2014) (collecting cases). Here the district court imposed two adjustments under Chapter Three-the abuse-of-trust enhancement (2 points) and the leadership enhancement (4 points). But it based both on del Carpio's wire fraud conduct, not his money laundering conduct. That violated the application note.

The government argues the district court based both enhancements on the money laundering conduct. It is possible that one set of conduct could be relevant for assessing a defendant's leadership in both a money laundering scheme and in the underlying crime that produced the dirty money. In that case, Application Note 2(C) surely does not put relevant conduct out of bounds simply because it also applies to the underlying criminal offense. Cf. United States v. Lopez , 743 F. App'x 489 , 494-95 (3d Cir. 2018).

But that is not what happened here. For both enhancements, the court fixated on the wire fraud conduct, not the money laundering conduct. Regarding the abuse-of-trust enhancement, the district court emphasized how del Carpio held himself out as "a very sophisticated investor and broker and trader" to his wire fraud victims. Similarly, when discussing the leadership enhancement, the court noted the scheme was "extensive" because "the scheme was, I'm going to make as many of these people believe this is real and then get them to go out and convince other people to come invest ." (Emphasis added.) Focusing on the wire fraud is understandable, given the egregiousness of del Carpio's wire fraud crimes and his failure to object below. It was error nonetheless.

The error increased the total offense level by a single point, 2 which increased *333 the Guidelines range. Under current Supreme Court precedent, and on the facts of this case, that means the error affected del Carpio's substantial rights. See Molina-Martinez v. United States , --- U.S. ----, 136 S. Ct. 1338 , 1345, 194 L.Ed.2d 444 (2016). And it means we must "exercise o[ur] discretion" to remand. Rosales-Mireles v. United States , --- U.S. ----, 138 S. Ct. 1897 , 1909, 201 L.Ed.2d 376 (2018).

* * *

We AFFIRM del Carpio's conviction and the district court's restitution order. We VACATE del Carpio's sentence and REMAND to allow the district court to resentence him in accordance with this opinion. Nothing in this opinion precludes the district court from exercising its discretion to depart from the Guidelines and choose any sentence permitted by 18 U.S.C. § 3553 .

ANDREW S. OLDHAM, Circuit Judge, concurring:

Today's result might surprise the uninitiated: Based on a one-point offense-level miscalculation in the advisory Guidelines, the United States must restart its criminal-justice machinery so it can fix a mistake that's supposedly so "plain" it cannot be ignored but also so subtle that del Carpio ignored it below. This result is particularly surprising because, not so long ago, the Supreme Court told us that "[m]eeting all four prongs of [plain-error review] is difficult, as it should be." Puckett v. United States , 556 U.S. 129 , 135, 129 S.Ct. 1423 , 173 L.Ed.2d 266 (2009). But this case illustrates it's no longer that difficult. So I agree current Supreme Court precedent requires that del Carpio be resentenced. I write separately to explain how we got here.

I.

Failure to raise an error ordinarily insulates it from appellate review. That's a bedrock principle of American law. As the Supreme Court put it:

No procedural principle is more familiar to this Court than that a constitutional right may be forfeited in criminal as well as civil cases by the failure to make timely assertion of the right before a tribunal having jurisdiction to determine it.

Yakus v. United States , 321 U.S. 414 , 444, 64 S.Ct. 660 , 88 L.Ed. 834 (1944). The Court has offered numerous justifications for this raise-it-or-lose-it rule in criminal cases. Principally it aligns incentives: The raise-it-or-lose-it rule motivates the defendant-who might otherwise be tempted to hold objections in reserve for appeal-to bring any errors to the attention of the court as early as possible. See Puckett , 556 U.S. at 134 , 129 S.Ct. 1423 ; Henderson v. United States , 568 U.S. 266 , 286, 133 S.Ct. 1121 , 185 L.Ed.2d 85 (2013) (Scalia, J., dissenting). It allows the person best situated to avoid the error-the trial judge-to "focus on it." Henry v. Mississippi , 379 U.S. 443 , 462, 85 S.Ct. 564 , 13 L.Ed.2d 408 (1965) (Harlan, J., dissenting); see also Puckett , 556 U.S. at 134 , 129 S.Ct. 1423 . In the process, it "enables the record to be made ... when the recollections of witnesses are freshest, not years later" after an appeal. Wainwright v. Sykes , 433 U.S. 72 , 88, 97 S.Ct. 2497 , 53 L.Ed.2d 594 (1977). And it conserves judicial resources by fixing errors before they necessitate retrials. See Henry , 379 U.S. at 448 , 85 S.Ct. 564 (majority op.); id. at 462-63 , 85 S.Ct. 564 (Harlan, J., dissenting).

That's not to say the Court has always recognized the rule's virtues. For example, during the ancien regime of Fay v. Noia , 372 U.S. 391 , 83 S.Ct. 822 , 9 L.Ed.2d 837 (1963), the Court regularly ignored violations *334 of the raise-it-or-lose-it rule. See, e.g. , Henry , 379 U.S. at 448-49 , 85 S.Ct. 564 . At least as to prisoners in state custody, the Court held procedural rules (like raise-it-or-lose-it) could be enforced only where the prisoner "understandingly," "knowingly," and "deliberately bypassed" the chance to object. Noia , 372 U.S. at 438-39 , 83 S.Ct. 822 . That led to predictable results. The absence of an enforceable raise-it-or-lose-it rule incentivized a defendant to "sandbag[ ] the court [by] remaining silent about his objection and belatedly raising the error only if the case [did] not conclude in his favor." Puckett , 556 U.S. at 134 , 129 S.Ct. 1423 . It vitiated otherwise-final convictions. See, e.g. , Lefkowitz v. Newsome , 420 U.S. 283 , 293, 95 S.Ct. 886 , 43 L.Ed.2d 196 (1975). And it created myriad anomalies. See, e.g. , Davis v. United States , 411 U.S. 233 , 238-39, 93 S.Ct. 1577 , 36 L.Ed.2d 216 (1973) (rejecting Noia 's deliberate-bypass rule as applied to a federal prisoner).

Eventually, the Court relented. It overruled Noia 's deliberate-bypass standard and again enforced the raise-it-or-lose-it rule. See Sykes , 433 U.S. at 88 , 97 S.Ct. 2497 ("The reasons for our rejection of [ Noia ] are several."). That restored the incentives for defendants and trial judges alike to focus on errors at the earliest possible time. See id. at 89 , 97 S.Ct. 2497 . It alleviated anomalies by ensuring state and federal prisoners faced the same raise-it-or-lose-it rules. See id. at 84 , 97 S.Ct. 2497 . And in the process, it "ma[de] a major contribution to finality in criminal litigation." Id. at 88 , 97 S.Ct. 2497 .

II.

The plain-error doctrine is (or at least was) a narrow exception to the raise-it-or-lose-it rule. Plain error originated in federal common law. Then it was codified in Federal Rule of Criminal Procedure 52. Then it largely reverted to federal common law. With these changes, the plain-error doctrine has waxed and waned in its penalization of criminal defendants who fail to object at trial. But at least when it comes to sentencing appeals, today's federal common law of plain error is eerily reminiscent of Noia .

A.

The concept of plain error originated in Wiborg v. United States , 163 U.S. 632 , 16 S.Ct. 1127 , 41 L.Ed. 289 (1896). Captain Wiborg, First Mate Petersen, and Second Mate Johansen used a Dutch steamer named The Horsa to aid Cuban revolutionaries in their bid for independence from Spain. Wiborg piloted the ship from the port at Philadelphia to an area just outside the jurisdiction of the United States; there he took aboard men and arms for a military expedition to Cuba. The government charged Wiborg and his two mates with violating a neutrality law enacted by Congress. The jury convicted them. The Supreme Court noted none of the officers requested a directed verdict at the close of the government's case. Nonetheless, the Court held, federal courts have inherent power to correct certain "plain errors": "[I]f a plain error was committed in a matter so absolutely vital to defendants, we feel ourselves at liberty to correct it." Id. at 658 , 16 S.Ct. 1127 . The Court did not cite anything for that proposition. But the Court used that plain-error power to hold "the jury should have been instructed to acquit" Petersen and Johansen because they merely followed "the captain's orders." Id. at 659 , 16 S.Ct. 1127 .

The Court returned to the plain-error doctrine in United States v. Atkinson , 297 U.S. 157 , 56 S.Ct. 391 , 80 L.Ed. 555 (1936). Atkinson was a civil case. The civil jury determined plaintiff's hearing loss constituted a "total disability" under a Veterans'

*335 Administration regulation. The Court noted the government forfeited any argument to the contrary, which required affirmance under the raise-it-or-lose-it rule:

The verdict of a jury will not ordinarily be set aside for error not brought to the attention of the trial court. This practice is founded upon considerations of fairness to the court and to parties and of the public interest in bringing litigation to an end after fair opportunity has been afforded to present all issues of law and fact.

Id. at 159 , 56 S.Ct. 391 . Although Atkinson was a civil case, the Court noted the exception Wiborg created for criminal cases: "In exceptional circumstances, especially in criminal cases, appellate courts, in the public interest, may, of their own motion, notice errors to which no exception has been taken, if the errors are obvious, or if they otherwise seriously affect the fairness, integrity, or public reputation of judicial proceedings." Id. at 160 , 56 S.Ct. 391 . 3 In the next sentence, the Court held it irrelevant: "But no such case is presented here." Ibid.

While Atkinson 's mention of "the fairness, integrity, or public reputation of judicial proceedings" was dicta, it nonetheless stuck. In 1940, Congress empowered the Supreme Court to create rules governing practice and procedure in federal criminal cases. See Act of June 29, 1940, 76 Pub. L. No. 445, 54 Stat. 688 . The Supreme Court responded in 1944 by adopting Federal Rule of Criminal Procedure 52(b) : "Plain errors or defects affecting substantial rights may be noticed although they were not brought to the attention of the court." 4 The Advisory Committee's note says "[t]his rule is a restatement of existing law." FED . R. CRIM . P. 52 advisory committee's note to 1944 adoption. And the Supreme Court later held this "restatement of existing law" included both Wiborg and the "fairness, integrity, public reputation" dicta from Atkinson . See United States v. Young , 470 U.S. 1 , 15 & n.12, 105 S.Ct. 1038 , 84 L.Ed.2d 1 (1985).

From the very beginning, some have questioned this approach. Justice Frankfurter, for example, dissented from the 1944 promulgation of the Criminal Rules. See Order, Rules of Criminal Procedure, 323 U.S. 821 (1944); id. at 821 (Mem. of Frankfurter, J.). From 1789 until after the Civil War, Frankfurter noted, the justices rode circuit. That gave them "intimate, *336 first-hand experience with the duties and demands of trial courts." Id. at 822 . But for the 50 years preceding promulgation of the Criminal Rules, the justices had been "necessarily removed from direct, day-by-day contact with trials in the district courts." Ibid. That made Frankfurter question the wisdom of the rules in toto . Ibid. And as if to predict the common-law evolution of Rule 52, Frankfurter further objected: "Such a code can hardly escape provisions in which lurk serious questions for future adjudication by this Court." Ibid.

Plain error was one of those lurking questions. Take for example Young . The government charged Young with mail fraud. During closing argument, the defense accused the prosecution of "reprehensible" conduct and said Young was "the only one in this whole affair that has acted with honor and with integrity." Young , 470 U.S. at 4-5 , 105 S.Ct. 1038 . In its rebuttal, the prosecutor described Young's fraud and said: "I don't know whether you call it honor and integrity, I don't call it that"; "I call it fraud." Id. at 5 , 105 S.Ct. 1038 . Young did not object. But the Tenth Circuit held it was plain error anyway.

In considering the plain-error question, the Supreme Court started with the drafting history of Rule 52(b). See id. at 15 n.12, 105 S.Ct. 1038 . Relying on a "preliminary draft" of Rule 52(b), the Court noted it was intended "to enable the courts of appeals to review prejudicial errors so that any miscarriage of justice may be thwarted." Ibid. (citation omitted); see also ibid. (dismissing the Rule's text as "misleading"). Such reliance on drafting history seems foreign in today's text-first (if not text-only) world. See SCALIA & GARNER, READING LAW 383 (2012) (describing the 1970s and 1980s as the "heyday of legislative history"). Just this Term, in fact, the Supreme Court pointedly criticized a 1974 D.C. Circuit decision for using drafting history to engraft unenacted words onto a rule. See Food Mktg. Inst. v. Argus Leader Media , --- U.S. ----, 139 S. Ct. 2356 , 2364, 204 L.Ed.2d 742 (2019) ("We cannot approve such a casual disregard of the rules of statutory interpretation.").

As foreign as Young might seem by today's textualist standards, the Court still easily reversed the finding of plain error. The Court noted it would be an "extravagant protection" of the defendant's rights "to use the plain-error doctrine to consider trial court errors not meriting appellate review absent timely objection." Young , 470 U.S. at 16 , 105 S.Ct. 1038 (quotation omitted). The Court further worried that correcting a forfeited error "would skew the Rule's careful balancing of our need to encourage all trial participants to seek a fair and accurate trial the first time around against our insistence that obvious injustice be promptly redressed." Id. at 15 , 105 S.Ct. 1038 (quotation omitted). The Court then held the prosecutor's comments did not rise to the level of "obvious injustice." Ibid. Thus, whatever its faults as a matter of textualism, Young understood the plain-error standard as a restriction on the power of appellate courts to correct forfeited errors.

The Court reaffirmed that view in United States v. Olano , 507 U.S. 725 , 113 S.Ct. 1770 , 123 L.Ed.2d 508 (1993). The question presented there was whether the presence of alternate jurors during jury deliberations constituted reversible error. Olano did not object at trial. And the government on appeal conceded the presence of the alternates constituted (1) "error" that was (2) "plain." But the Court identified two additional hurdles in Rule 52(b). The third hurdle, the Court said, is the plain error must "affect substantial rights." Id. at 734 , 113 S.Ct. 1770 . That is, it must be prejudicial. Ibid. Fourth and finally, the Court *337 noted, the plain error must be the kind that warrants the appellate court's discretionary intervention. See id. at 735 , 113 S.Ct. 1770 . After all, the Court emphasized, " Rule 52(b) is permissive, not mandatory." Ibid. Therefore, "[i]f the forfeited error is 'plain' and 'affects substantial rights,' the court of appeals has authority to order correction, but is not required to do so." Ibid. (alteration omitted). The Court then held the error was plain but nonetheless did not affect Olano's substantial rights.

The Court again enforced the raise-it-or-lose-it rule in Johnson v. United States , 520 U.S. 461 , 117 S.Ct. 1544 , 137 L.Ed.2d 718 (1997). The government indicted the defendant for perjury. The trial judge determined the materiality of the statement as a question of law; Johnson did not object. See id. at 464 , 117 S.Ct. 1544 . But as the Supreme Court later held in a different case, the materiality of a false statement is an issue of fact for the jury. See United States v. Gaudin , 515 U.S. 506 , 522-23, 115 S.Ct. 2310 , 132 L.Ed.2d 444 (1995). The question presented in Johnson was whether the defendant's failure to object at trial insulated the error from appellate review.

The Supreme Court held yes. As in Young , the Court "cautioned against any unwarranted expansion of Rule 52(b)." Johnson , 520 U.S. at 466 , 117 S.Ct. 1544 . And the Court emphasized that it "ha[s] no authority to make" "an exception" to the rule. Ibid. The Court assumed the unobjected-to error was so serious that it affected Johnson's substantial rights. Id. at 469 , 117 S.Ct. 1544 . Even so, the Court held, it could not satisfy the dicta imported from Atkinson -that is, the error would not "seriously affect the fairness, integrity or public reputation of [the] judicial proceedings." Id. at 469-70 , 117 S.Ct. 1544 (alteration omitted) (quoting Atkinson , 297 U.S. at 160 , 56 S.Ct. 391 ). Johnson powerfully illustrates the raise-it-or-lose-it rule: If the defendant had raised the materiality objection at trial, she would have received the benefit of Gaudin on appeal; having not raised it, however, she lost it.

So too in United States v. Cotton , 535 U.S. 625 , 122 S.Ct. 1781 , 152 L.Ed.2d 860 (2002). The government's indictment omitted a fact that would increase the statutory maximum for the defendant's crime (drug quantity). The defendant forfeited any objection to that omission. While the case was pending on appeal, the Supreme Court announced Apprendi v. New Jersey , 530 U.S. 466 , 120 S.Ct. 2348 , 147 L.Ed.2d 435 (2000). And Apprendi held that "any fact that increases the penalty for a crime beyond the prescribed statutory maximum must be submitted to a jury, and proved beyond a reasonable doubt.'' Id. at 490 , 120 S.Ct. 2348 . The question presented in Cotton was whether plain-error review allowed the Court to overlook the defendant's forfeiture. Again, no. As in Johnson , the Court assumed Cotton could meet prong three of the plain-error standard. And it held that Cotton's claim nonetheless failed at prong four. In doing so, the Court yet again emphasized "the longstanding rule 'that a constitutional right may be forfeited in criminal as well as civil cases by the failure to make timely assertion of the right.' " Cotton , 535 U.S. at 634 , 122 S.Ct. 1781 (quoting Yakus , 321 U.S. at 444 , 64 S.Ct. 660 ).

This long line of cases-from Wiborg to Atkinson to Young to Olano to Johnson to Cotton -illustrates how hard it is to overcome a forfeited error. That is why, of course, it's the raise-it-or- lose -it rule. Even rights as fundamental as those protected by Apprendi are lost if they're not preserved. And even in Wiborg , the two defendants who overcame the raise-it-or-lose-it rule did so because they were in effect *338 innocent of the charges. See 163 U.S. at 659 , 16 S.Ct. 1127 . That is obviously an extraordinarily high bar. Cf. Herrera v. Collins , 506 U.S. 390 , 417, 113 S.Ct. 853 , 122 L.Ed.2d 203 (1993) (holding "the threshold showing for such an assumed right [of actual innocence] would necessarily be extraordinarily high"). All of this explains why the Court once said "[m]eeting all four prongs [of plain error] is difficult, as it should be." Puckett , 556 U.S. at 135 , 129 S.Ct. 1423 (quotation omitted).

B.

Something changed after Puckett . Instead of requiring actual innocence, now we are satisfied by one-point math errors in the advisory Guidelines range. Instead of considering it an unwarranted "extravagance" to reach forfeited errors, now we reach them as a matter of course. Instead of saying a forfeited error "may" be noticed under certain conditions, now we say errors "must" be noticed in ordinary cases. Instead of viewing plain error as a restriction on our appellate discretion , now it expands our appellate obligations . And instead of making defendants run the gauntlet of a "difficult" four-part test, now plain error accommodates almost all but those who intentionally relinquish their rights.

It is the Supreme Court's prerogative to tell us how to apply these rules. And we'll do our best to follow them. It is noteworthy, however, just how much plain error has changed in the 10 years since Puckett . When we walk through the four prongs of that doctrine, it looks more and more like Noia .

1.

Prong one of the plain error standard is whether an "error" occurred. See Olano , 507 U.S. at 732-33 , 113 S.Ct. 1770 . The background appellate principle is that there is no "error" to correct if the defendant failed to raise it below. That's the raise-it-or-lose-it rule from Yakus . And it forms the backdrop for Federal Rule of Criminal Procedure 51, which explains how to preserve an error:

A party may preserve a claim of error by informing the court-when the court ruling or order is made or sought-of the action the party wishes the court to take, or the party's objection to the court's action and the grounds for that objection. If a party does not have an opportunity to object to a ruling or order, the absence of an objection does not later prejudice that party.

FED . R. CRIM . P. 51(b). Obviously, if a defendant did "have an opportunity to object" but did not "inform[ ] the court," then he has not "preserve[d] a claim of error."

You might reasonably wonder why we have Rule 51 if Rule 52 allows review of unpreserved errors. After all, a defendant can ignore an error under the former with impunity as long as he can get review of it under the latter. One answer is that plain-error review under Rule 52 applies only to forfeitures, not waivers: "Deviation from a legal rule is 'error' unless the rule has been waived ." Olano , 507 U.S. at 732-33 , 113 S.Ct. 1770 (emphasis added). Intentionally relinquishing a right (waiver) "extinguish[es] an 'error' under Rule 52(b)." Id. at 733 , 113 S.Ct. 1770 . Merely "fail[ing] to make the timely assertion of a right" (forfeiture) does not. Ibid.

In this case, the probation office and the district court gave del Carpio and his attorneys from the Federal Public Defender's Office an opportunity to review the Pre-Sentence Report ("PSR") before the sentencing hearing. Del Carpio and his attorneys discussed the report. And they offered ten written objections to it. These included all manner of considered objections to the Guidelines and their application *339 to del Carpio's crimes. Then they agreed everything else in the PSR was correct. Does that mean del Carpio forfeited objections eleven through infinity? Or does it mean he waived them by agreeing to the correctness of the unobjected-to portions of the PSR?

In other areas where the raise-it-or-lose-it rule applies, it wouldn't matter. Take, for example, the man who stands trial for murder in state court. There he fails to object to the admission of his confession. It's irrelevant whether he forfeited the objection in ignorance or waived it on purpose. Because he's a state prisoner, his claim is defaulted and hence unreviewable in all events. See Sykes , 433 U.S. at 88-91 , 97 S.Ct. 2497 . If this state prisoner wants to overcome that default, he must run the difficult gauntlet of showing cause and prejudice. See id. at 90 , 97 S.Ct. 2497 . Ever since " Sykes limited [ Noia ] to its facts," the Supreme Court's decisions "have been unanimous in applying the cause and prejudice standard" to state prisoners who violate the raise-it-or-lose-it rule. Coleman v. Thompson , 501 U.S. 722 , 747, 111 S.Ct. 2546 , 115 L.Ed.2d 640 (1991).

We must apply a different rule to del Carpio because he's a federal prisoner. The Federal Public Defender's Office recently advised us in another case that it would be "a sea change in the law" to enforce the raise-it-or-lose-it rule for PSR-based errors. Oral Argument at 33:38, United States v. Sanchez-Hernandez , 931 F.3d 408 , 2019 WL 3333731 (5th Cir. July 25, 2019), http://www.ca5.uscourts.gov/OralArgRecordings/18/18-40211_4-3-2019.mp3. That is so, we are told, even where the defendant affirmatively agreed that the PSR was correct. See Sanchez-Hernandez , 931 F.3d at 411 n. 2, 2019 WL 3333731 , at *2 n.2 (noting the defendant filed a petition for panel rehearing not to change the outcome of his appeal but to remove a footnote suggesting the word "Yes" can ever constitute a waiver). And the government implicitly agrees by its failure to contest the issue in its briefs here and elsewhere. That suggests-when it comes to federal prisoners-we can overlook a PSR "error" at prong one only if the defendant "understandingly," "knowingly," and "deliberately bypassed" the chance to object. Noia , 372 U.S. at 438-39 , 83 S.Ct. 822 .

I can imagine a few explanations for this anomaly, but I'm not sure how persuasive they are. First, you might say it makes sense to apply the strict cause-and-prejudice rule on collateral review (Sykes) and a more forgiving no-cause rule on direct review (del Carpio). After all, one reason our habeas rules are so strict is because the state prisoner already had his direct appeal in the state system. See, e.g. , Langley v. Prince , 926 F.3d 145 , 155-56 (5th Cir. 2019) (en banc). Fair enough. But it means the state prisoner never gets review of the forfeited error: Florida's courts refused to consider whether Sykes's confession was inadmissible because he failed to raise it (and hence lost it) at trial; then the federal courts enforced that default on collateral review without considering the merits. See Sykes , 433 U.S. at 91 , 97 S.Ct. 2497 . We apply that approach even when it means no court (state or federal) ever will review an "alleged constitutional error [that] impaired the truthfinding function of the [state] trial." Coleman , 501 U.S. at 747 , 111 S.Ct. 2546 (citing Engle v. Isaac , 456 U.S. 107 , 102 S.Ct. 1558 , 71 L.Ed.2d 783 (1982) ). But we cannot apply that approach to a non-constitutional math error in del Carpio's Guidelines calculation. That seems odd.

Second, you might say federal courts always require knowing and deliberate waivers before ignoring important errors. Again, no. " 'Waiver' is a vague term used *340 for a great variety of purposes, good and bad, in the law." Green v. United States , 355 U.S. 184 , 191, 78 S.Ct. 221 , 2 L.Ed.2d 199 (1957). But "[a]lmost without exception, the requirement of a knowing and intelligent waiver has been applied only to those rights which the Constitution guarantees to a criminal defendant in order to preserve a fair trial." Schneckloth v. Bustamonte , 412 U.S. 218 , 237, 93 S.Ct. 2041 , 36 L.Ed.2d 854 (1973) (emphasis added). For example, the knowing-and-deliberate waiver doctrine originates in a case where the State denied a criminal defendant his right to counsel at trial. See Johnson v. Zerbst , 304 U.S. 458 , 464, 58 S.Ct. 1019 , 82 L.Ed. 1461 (1938). But the Supreme Court has said "[t]here is a vast difference between those rights that protect a fair criminal trial and" other constitutional rights-like the Fourth Amendment right to be free from unreasonable searches. Bustamonte , 412 U.S. at 241 , 93 S.Ct. 2041 ; see also id. at 246 , 93 S.Ct. 2041 (holding the State need not prove knowing-and-deliberate consent to search); cf. United States v. Shepperson , 739 F.3d 176 , 179-80 (4th Cir. 2014) (refusing to apply Johnson v. Zerbst to waiver of statutory counsel rights). And there is a vaster difference still between constitutional trial rights and the non-constitutional, non-statutory, non-trial right to a correct Guidelines calculation. So why are we applying the hallowed Johnson v. Zerbst standard to this error in the first place?

Which brings us, third and finally, back to where we started: Olano . The Court began its opinion in that case with a discussion of the "basic principles" that govern plain error. 507 U.S. at 737 , 113 S.Ct. 1770 . That included a discussion of prong one even though, as noted above, Olano turned exclusively on prong three. See Part II.A, supra . And as one of the "basic principles" that governs prong one, the Court quoted Johnson v. Zerbst for the proposition that a "waiver is the intentional relinquishment or abandonment of a known right." Olano , 507 U.S. at 733 , 113 S.Ct. 1770 (citation omitted). Then it distinguished a knowing-and-intentional waiver from an unintentional forfeiture. See ibid. But in the very next sentence, the Court emphasized the knowing-and-intentional-waiver rule does not apply to all errors; rather, it "depend[s] on the right at stake." Ibid. And at no point did Olano purport to overrule Bustamonte 's longstanding refusal to extend the knowing-and-intentional-waiver rule beyond constitutional trial rights. Yet somehow-perhaps through the dint of repetition over many decades-we've lost that nuance. And now we routinely apply Johnson v. Zerbst (by way of Olano ) to all sorts of non-constitutional, non-statutory, and non-trial errors like the Guidelines miscalculation here.

All of this raises more questions than it answers. But one thing is clear: The line between waiver and forfeiture does little to constrain appellate review of PSR errors today. I suppose we still must find a mistake of law. But that would also be true if the defendant had objected and plain error did not apply. So this first hurdle is not much of a hurdle at all.

2.

If there was an error, Rule 52(b) instructs us to ask whether it was "plain." Here too we find an obstacle doing very little obstructing. You might have thought "plain" error refers only to an error that is "manifest," "unmistakable," "obvious," or "easily ... recognizable." WEBSTER'S NEW INTERNATIONAL DICTIONARY 1878 (2d ed. 1941); 7 OXFORD ENGLISH DICTIONARY 936 (1st ed. 1933). And given the sophistication of federal sentencing proceedings, you might think manifest, obvious, and easily *341 recognizable errors would be few and far between. Federal taxpayers devote significant resources to these proceedings. For example, the sentencing and restitution hearings in this case spanned more than six days, with numerous attorneys, forensic experts, fact witnesses, and experienced personnel from the Probation Office, the U.S. Attorney's Office, and the Federal Public Defender's Office. What are the odds that all these folks missed an unmissably obvious error?

Under our current plain-error doctrine, those odds are much higher than you might think. The first reason is a function of timing. According to the Supreme Court, we must judge the plainness of an error "at the time of appellate consideration." Johnson , 520 U.S. at 467-68 , 117 S.Ct. 1544 . That means a district judge's decision that was plainly correct at the time it was made can be plainly wrong at the time of appeal. Ibid. Perhaps that result makes sense when an objection would have been useless in light of then-binding precedent. But the time-of-review rule applies even where binding precedent did not provide an answer at the time of trial. Henderson , 568 U.S. at 269 , 133 S.Ct. 1121 . That, of course, is when an objection is "eminently useful." Id. at 284 , 133 S.Ct. 1121 (Scalia, J., dissenting); see United States v. Mouling , 557 F.3d 658 , 664 (D.C. Cir. 2009).

The second reason is a function of scope. A "plain" error sounds like an error that might give rise to a claim for malpractice against a lawyer who missed it. But we sometimes treat an error as "plain" even when it's obvious to no one.

Take the "plain" error in this case. Here is my highly condensed summary of it. Ordinarily, offenses covered by § 2A3.5; §§ 2B1.1, 2B1.4, 2B1.5, 2B4.1, 2B5.1, 2B5.3, 2B6.1 ; §§ 2C1.1, 2C1.2, 2C1.8 ; §§ 2D1.1, 2D1.2, 2D1.5, 2D1.11, 2D1.13 ; §§ 2E4.1, 2E5.1 ; §§ 2G2.2, 2G3.1 ; § 2K2.1 ; §§ 2L1.1, 2L2.1 ; § 2N3.1 ; § 2Q2.1 ; § 2R1.1 ; §§ 2S1.1, 2S1.3 ; §§ 2T1.1, 2T1.4, 2T1.6, 2T1.7, 2T1.9, 2T2.1, and 2T3.1 are grouped under § 3D1.2(d). But if you read application note 6 to Guideline § 2S1.1 (which is specifically referenced in § 3D1.2(d)), then you'd know these wire fraud and money laundering counts should be grouped under § 3D1.2(c), not § 3D1.2(d). Now we should turn to § 3D1.3(a), which says for offenses grouped under § 3D1.2(c), we use the offense level for the most serious count. Here you might think that's Count 26: Money Laundering. But obviously, you don't want to forget about application note 2 in the commentary following § 3D1.3. It says the offense levels should be calculated separately for wire fraud and money laundering, and when you do that, the relevant offense level comes from the wire fraud counts (1-24). The wire fraud base offense level is 7 under § 2B1.1(a)(1). We'd add 18 points for the amount of loss under § 2B1.1(b)(1)(J). Then we'd add 4 points because del Carpio caused substantial financial hardship to five or more victims under § 2B1.1(b)(2)(B). Then we'd add 2 points because a substantial part of the scheme was committed outside of the United States under § 2B1.1(b)(10)(B). Then we'd add Chapter Three enhancements for extensive leadership (4 points under § 3B1.1(a)) and abuse of trust (2 points under § 3B1.3)). Then we'd subtract two points for substantial assistance under § 5K1.1. That brings us to a grand total offense level of 35. Which is plainly different than the 36-point offense level calculated by the district court. The district court thought money laundering would be the higher offense level because it took the base offense level from § 2B1.1 (31 points) and added 1 point as required by § 2S1.1(b)(2)(A) for offenses involving a conviction under 18 U.S.C. § 1957 . Then it added 6 points for the Chapter Three enhancements *342 under §§ 3B1.1(a) and 3B1.3. You might think the district court was correct to do that as required by § 1B1.5(c). But in cases where § 2S1.1(a)(1) applies, we need to apply the Chapter Three adjustments based only on the money laundering conduct as required by application note 2 subsection (C) to the commentary following § 2S1.1. Which means, ironically, we don't apply the Chapter Three enhancements to the money laundering guideline when calculating the money laundering offense level.

Does this strike anyone as plain and obvious? I think it's amazing the district court managed to err by only a single point.

Del Carpio's changing litigation positions powerfully illustrate how not plain and not obvious a plain error can be. Recall everyone (including del Carpio's attorneys) missed the supposedly obvious mistake described above. Then in his opening appellate brief, del Carpio had to devote ten pages, five tables, two footnotes, and more than fifty citations to the Guidelines and cases interpreting the Guidelines to explain this supposedly "plain" one-point error. It's both so subtle that everyone missed it and so obvious that we now wonder how anyone missed it. It's like Fermat's Last Theorem before and after Andrew Wiles proved it. 5 But query whether it's still fair to say "that plain-error review is not a grading system for trial judges." Henderson , 568 U.S. at 278 , 133 S.Ct. 1121 .

3.

The third prong requires a defendant to prove the plain error "affects [his] substantial rights." FED. R. CRIM . P. 52(b). That means he "must show a reasonable probability that, but for the error, the outcome of the proceeding would have been different." Molina-Martinez v. United States , --- U.S. ----, 136 S. Ct. 1338 , 1343, 194 L.Ed.2d 444 (2016) (quotation omitted). This third hurdle is particularly low in Guidelines cases. The Supreme Court recently predicted a defendant will normally be able to clear the third hurdle by pointing to an erroneous Guidelines range "to show a reasonable probability that the district court would have imposed a different sentence under the correct range." Id. at 1349 .

That won't always be true. Id. at 1346-47 ; see Griffith v. United States , 871 F.3d 1321 , 1338 (11th Cir. 2017) (noting Molina-Martinez 's "prediction is not, however, a presumption"). And we have a duty to conduct a case-specific inquiry. In every case, we have to ask: "What was driving this judge's decision to impose this sentence for this defendant? In answering that question, we apply no presumptions or categorical rules." Sanchez-Hernandez , 931 F.3d at 411 , 2019 WL 3333731 , at *3.

Unfortunately, however, courts seem to "misunderstand [the Supreme Court's] predictions as veiled directives." Molina-Martinez , 136 S. Ct. at 1349 , 1351 n.4 (Alito, J., concurring in part and concurring in the judgment). Our Court and others routinely conclude criminal defendants have cleared the third hurdle whenever they show a Guidelines calculation error. See, e.g. , United States v. Islas-Saucedo , 903 F.3d 512 , 520-21 (5th Cir. 2018) (calculation *343 error affected substantial rights where incorrect Guidelines range was 37-46 months, correct range was 24-30, and district court sentenced defendant to 42 months); United States v. Corbett , 921 F.3d 1032 , 1040-41 (11th Cir. 2019) (calculation error affected substantial rights where incorrect Guidelines range was 30-37 months, correct range was 24-30 months, and district court sentenced defendant to 12 months); United States v. Hurlburt , 835 F.3d 715 , 725-26 (7th Cir. 2016) (en banc) (finding error affected substantial rights after "presum[ing] the improperly calculated [G]uideline[s] range influenced the judge's choice of sentence").

That powerfully illustrates the common-law changes to Rule 52. Recall for example Olano . In that case, the trial court plainly violated the defendant's rights by allowing alternates into the jury room during deliberations. That contravened the express protections of Federal Rule of Criminal Procedure 24(c). See 507 U.S. at 737 , 113 S.Ct. 1770 . And the Court noted the potential for prejudice because alternates could either participate in the deliberations (verbally or non-verbally) or could chill the jurors' deliberations. See id. at 739 , 113 S.Ct. 1770 . Still, in the absence of actual proof of prejudice, the Court refused to find the plain error affected Olano's substantial rights. By contrast, current Supreme Court precedent requires us to find the one-point Guidelines error here affected del Carpio's substantial rights-even though the Guidelines are purely advisory, and even though the district court can resentence del Carpio to the same or even a higher sentence on remand. In Guidelines cases at least, the third "hurdle" is not much of a hurdle at all.

4.

The fourth and final hurdle says we "may" exercise discretion to correct an error where it "seriously affects the fairness, integrity[,] or public reputation of judicial proceedings." Olano , 507 U.S. at 732 , 113 S.Ct. 1770 (quotation omitted). It's worth remembering this hurdle is not in Rule 52(b) 's text:

Plain Error. A plain error that affects substantial rights may be considered even though it was not brought to the court's attention.

FED. R. CRIM . P. 52(b). All the rule says is that we "may ... consider[ ]" an error when the first three requirements are met. The rule takes care to enumerate those three requirements and no others. See SCALIA & GARNER , supra , at 107-11. What's left is discretion. See Henderson , 568 U.S. at 283 n.1, 133 S.Ct. 1121 (Scalia, J., dissenting) (noting hurdle 4 pertains "to the word 'may' ").

As noted in Part II.A, supra , the Supreme Court has read a lot into "may." That one word includes the "fairness, integrity, or public reputation of judicial proceedings" dicta from Atkinson . It also includes the Advisory Committee's drafting notes and the "miscarriage of justice" exception from Young . But these precedents at least read Rule 52(b) as a limitation on appellate review. Atkinson and Young said appellate courts "should not exercise [ Rule 52(b) ] discretion unless the error seriously affects the fairness, integrity[,] or public reputation of judicial proceedings." Olano , 507 U.S. at 732 , 113 S.Ct. 1770 ; see also Young , 470 U.S. at 15 , 105 S.Ct. 1038 ("The Rule authorizes the Courts of Appeals to correct only ... those errors that seriously affect the fairness, integrity[,] or public reputation of judicial proceedings." (quotation omitted)). That is, satisfaction of the Atkinson dicta and the Young drafting history authorized but did not compel a court to grant relief. At bottom, " Rule 52(b) is permissive, not mandatory."

*344 Olano , 507 U.S. at 735 , 113 S.Ct. 1770 ; see also id. at 737 , 113 S.Ct. 1770 .

The Supreme Court's decision in Rosales-Mireles took a different path. There the Court concluded we "should" correct an error whenever it may affect the judiciary's reputation. 138 S. Ct. at 1906 . And it further concluded a Guidelines calculation error will affect the judiciary's reputation "[i]n the ordinary case." Id. at 1911 . That means, "[i]n the ordinary case," everything permitted is also required. Ibid. It also means, in this case, we must correct the Guidelines error. Ante , at 12, 105 S.Ct. 1038 . But it's not obvious how much discretion we have in the matter.

* * *

So here's how the hurdles appear to the defendant on the starting line of a Guidelines appeal. Hurdles 1 and 2 are almost illusory. We assume defendants clear hurdle 3 when the Probation Office makes a one-point math error that would've eluded anyone except Andrew Wiles. And "may" means "should" at hurdle 4. At some point, the hurdlers are running a dash. And even the Federal Public Defender's Office says this is an "odd position." Oral Argument at 6:21, United States v. Del Carpio Frescas , No. 17-50245, http://www.ca5.uscourts.gov/OralArgRecordings/17/17-50245_4-30-2019.mp3.

The oddity of it is not what worries me, however. What worries me is that we've seen this movie before. In the mine-run sentencing case, "[a] defendant could simply relax and wait to see if the sentence later struck him as satisfactory," United States v. Vonn , 535 U.S. 55 , 73, 122 S.Ct. 1043 , 152 L.Ed.2d 90 (2002), "belatedly raising [an] error only if the case does not conclude in his favor," Puckett , 556 U.S. at 134 , 129 S.Ct. 1423 . That's exactly what happened when Noia required federal courts to ignore the raise-it-or-lose-it rule in the "ordinary" case. And that movie did not end well. See Sykes , 433 U.S. at 88 , 97 S.Ct. 2497 .

A plain error doctrine with no hurdles will encourage defense counsel to focus on anything but the Guidelines in the district court. In this case, del Carpio's sentencing and restitution hearings spanned six days and included more than two dozen witnesses. The record is over 2,500 pages long. But somehow, despite all that effort, no one noticed the Guidelines miscalculation that's supposedly so "plain" today. Given the current plain-error doctrine, del Carpio's trial counsel was quite right to focus on anything except proving Fermat's Last Theorem in the Guidelines calculation. After all, the latter can always be done on appeal. What once was "extravagant," Young , 470 U.S. at 16 , 105 S.Ct. 1038 , is now de rigueur .