United States v. Caramadre

A federal grand jury returned an indictment charging defendant-appellant Joseph Caramadre with masterminding one of the most avaricious frauds in Rhode Island history. Caramadre went to trial, but things did not go well for him and, after four days, he entered into a plea agreement with the government. The district court accepted his changed plea.

Some months later (but before sentencing), Caramadre experienced a change of heart. Represented by new counsel, he sought to retract his guilty plea. Following a multi-day evidentiary hearing, the district court denied his motion. Sentencing ensued.

Caramadre’s appeals, taken collectively, advance an infinity of arguments, characterized by clangorous sound and unrestrained fury. But fiery rhetoric alone is not enough to breathe life into moribund arguments and, after close scrutiny, we conclude that Caramadre’s appeals are without merit. Accordingly, we affirm the judgment below.

I. BACKGROUND

We sketch the origin and travel of the case, assuming the reader’s familiarity with a number of other judicial opinions. See, e.g., W. Reserve Life Assur. Co. of Ohio v. ADM Assocs., LLC, 737 F.3d 135 (1st Cir.2013); United States v. Caramadre, No. 11-186, 2014 WL 409336 (D.R.I. Feb. 3, 2014); United States v. Caramadre, No. 11-186, 2013 WL 7138109 (D.R.I. Nov. 26, 2013); United States v. Caramadre, No. 11-186, 2013 WL 7138106 (D.R.I. Nov. 6, 2013); United States v. Caramadre, 957 F.Supp.2d 160 (D.R.I.2013); W. Reserve Life Assur. Co. of Ohio v. ADM Assocs., LLC, 116 A.3d 794 (R.I.2015).

Under the government’s theory of the case, Caramadre — a lawyer and accountant — and his codefendant, Raymour Ra-dhakrishnan, engaged for well over a *365 decade in a scheme to defraud various financial institutions. Caramadre and Ra-dhakrishnan implemented the scheme by fraudulently obtaining the identifying information of terminally ill individuals through material misrepresentations and omissions. They then invested in variable annuities and corporate bonds with death-benefit features, using the identities of these unwitting individuals as measuring lives. When a terminally ill individual died, Caramadre and Radhak-rishnan cashed in the annuities and bonds and captured the profits. 1

Based on the scope of the fraud alleged in the sixty-six-count indictment and the large number of anticipated government witnesses, the trial was expected to last over three months. On November 19, 2012 — four days into the trial — Caramadre and Radhakrishnan entered into plea agreements and admitted their guilt to two counts: one count of wire fraud and one count of conspiracy to commit wire fraud, mail fraud, and identity theft. The district court accepted their pleas, and the government later dismissed the remaining counts. 2

Nearly two months passed. Carama-dre’s attorneys then moved to withdraw from their representation of him, and his new counsel informed the district court that Caramadre intended to seek leave to retract his guilty plea. Caramadre filed such a motion on February 28, 2013. The government objected, and the district court held a protracted evidentiary hearing. The court denied the motion from the bench at the conclusion of the hearing and followed up with a fuller exposition in a written rescript issued on August 1, 2013. See Caramadre, 957 F.Supp.2d at 186.

On December 16, 2013, the district court sentenced Caramadre to a six-year term of immurement. The court had previously referred the question of restitution to a magistrate judge. Prior to the imposition of the prison sentence, the magistrate judge conducted an evidentiary hearing and recommended restitution of approximately $46,000,000. See Caramadre, 2013 WL 7138109 at *2; Caramadre, 2013 WL 7138106 at *19. Over Caramadre’s protest, the district court adopted the magistrate judge’s recommendation. See Cara-madre, 2014 WL 409336, at *1.

Caramadre timely appealed and, on September 8, 2014, he tendered his opening brief to this court. The brief referred to statements allegedly made by the district court at an unrecorded and untranscribed chambers conference held on January 15, 2013. Because those statements were not part of the record, we struck his brief and ordered him to refile it without reference to anything supposedly said at the conference. Caramadre complied.

But that was not the end of the matter: Caramadre moved in the district court for a statement of what had transpired at the January 15 conference. See Fed. R.App. P. 10(c). On January 5, 2015, the district court rejected Caramadre’s version of what had occurred and substituted its own recollection. See United States v. Caramadre, No. 11-186 (D.R.I. Jan. 5, 2015) (unpublished order). Caramadre again appealed, sparking a new round of appellate briefing.

*366 Caramadre’s appeals raise a golconda of issues. We discuss here only those claims of error that possess a patina of plausibility. The rest are either patently meritless, insufficiently developed, or both. Consequently, we reject them out of hand.

II. PLEA-WITHDRAWAL MOTION

Caramadre offers several arguments in support of his assertion that the district court erred in denying his motion to withdraw his guilty plea. These include claims that the court employed the wrong legal standard in deciding the motion, that the court abused its discretion in balancing the relevant factors, and that the court “exhibited bias and prejudged the motion.” We find none of these claims persuasive.

A. Legal Standard.

The logical starting point is Caramadre’s claim that the .district court used an “erroneous” legal standard when ruling on the motion to withdraw. This claim presents a pure question of law and, thus, engenders de novo review. 3 See United States v. Gates, 709 F.3d 58, 69 (1st Cir.2013).

It is common ground that a defendant has no absolute right to withdraw a guilty plea. See United States v. Ramos-Mejía, 721 F.3d 12, 14 (1st Cir.2013); Gates, 709 F.3d at 68. When a defendant moves to withdraw a guilty plea after the court has accepted it but before the court has sentenced him, he may do so only if he “can show a fair and just reason for requesting the withdrawal.” Fed.R.Crim.P. 11(d)(2)(B); see Gates, 709 F.3d at 68; United States v. Marrero-Rivera, 124 F.3d 342, 347 (1st Cir.1997). The burden rests with the defendant to make this showing. See Marrero-Rivera, 124 F.3d at 347.

Critical to the plea-withdrawal inquiry is whether the original guilty plea was knowing, intelligent, and voluntary. See United States v. Aker, 181 F.3d 167, 170 (1st Cir.1999) (citing Fed.R.Crim.P. 11). Other factors, however, may weigh in the balance. The court may consider, for example, “the plausibility and weight of the reason given for the withdrawal, the timing of the request, whether the defendant is now colorably asserting legal innocence, and whether the original plea was pursuant to a plea agreement.” Id. If these factors, taken together, tilt in favor of allowing withdrawal, the court must then weigh the prejudice that the government would suffer if the plea were to be vacated. See Gates, 709 F.3d at 69; United States v. Doyle, 981 F.2d 591, 594 (1st Cir.1992).

In the case at hand, the district court expressly acknowledged that the “fair and just reason” standard controlled its inquiry. Caramadre, 957 F.Supp.2d at 166. It proceeded to identify and evaluate all of the relevant factors. See id. at 166, 181-86. Caramadre nonetheless persists in his claim of error, hanging his hopes on two sentences in the district court’s lengthy rescript: “As the above discussion makes pellucid, Caramadre entered into a knowing, intelligent, and voluntary plea, and thus no just reason exists for allowing him to withdraw it. Still, for completeness sake, the Court will briefly address the additional factors enumerated by the First Circuit.” Id. at 181. Caramadre urges that these sentences demonstrate that the court conflated the “generous” fair and just reason for permitting withdrawal of a guilty plea with the “stricter” standard for holding a plea invalid.

*367 This is nonsense on steroids. Rule 11 considerations are a paramount concern in a plea-withdrawal inquiry. See United States v. Santiago Miranda, 654 F.3d 130, 136 (1st Cir.2011); United States v. Richardson, 225 F.3d 46, 51 (1st Cir.2000) (quoting United States v. Cotal-Crespo, 47 F.3d 1, 3 (1st Cir.1995)). Thus, the court below appropriately focused, at the outset of its inquiry, on whether Caramadre’s plea was knowing, intelligent, and voluntary.

Here, moreover, Caramadre’s plea-withdrawal motion — which alleged that his plea had been involuntary and that he was not competent to have tendered it — invited this very focus. Caramadre cannot now fault the district court for accepting this invitation and beginning its analysis with the very factors that he himself had stressed.

In any event, the district court did not simply examine Rule 11 considerations and stop there. Although the court stated that it would address the other factors “briefly,” Caramadre, 957 F.Supp.2d at 181, this was nothing more than self-deprecating litotes. What followed was a thorough analysis of the other factors. See id. at 181-86.

The short of it is that Caramadre’s contention that the district court premised its decision entirely on the validity of his plea (and, thus, used an erroneous legal standard) turns a blind eye to a generous portion of the district court’s reasoning. Reading the district court’s rescript as a whole, Caramadre’s claim is fanciful. We summarily reject it. 4

B. Abuse of Discretion.

In the absence of legal error, we review decisions denying plea-withdrawal motions solely for abuse of discretion. See United States v. Merritt, 755 F.3d 6, 9 (1st Cir.2014). Within this rubric, findings of fact are reviewed for clear error. See Gates, 709 F.3d at 69. The defendant bears the devoir of persuasion. See Merritt, 755 F.3d at 9.

Caramadre’s primary argument is that the district court abused its discretion in balancing the factors relevant to whether he should be allowed to withdraw his plea. In his view, the district court did not appreciate that a “perfect storm” of events “overbore his will and induced him to enter a guilty plea” that was involuntary. This argument has several subsets, which we discuss below.

1. The Rule 11 Colloquy. The most heated among these sub-arguments is Car-amadre’s claim that the ehange-of-plea colloquy was too scanty with respect to the district court’s inquiry into his medications and history of depression. At the change-of-plea hearing, the district court asked Caramadre if he was being treated for mental illness. He responded that he had been treated for depression “both lately and for the last 20 years.” One of Cara-madre’s lawyers then proffered a list of Caramadre’s current medications. The court reviewed this list and asked Carama-dre to confirm that he fully understood all *368 the proceedings and that his medications did not impede his understanding. Cara-madre and his counsel confirmed both points.

Before us, Caramadre complains that the court failed to probe deeply enough into the effects of his medications. Relat-edly, he suggests that his counsel should not have vouched for his clarity of mind without consulting his physicians.

We start with first principles. Where, as here, a defendant confirms during a change-of-plea colloquy that he is taking medication, the district court has a duty to inquire into the effects of the medication and the defendant’s capacity to plead guilty. See United States v. Savinon-Acosta, 232 F.3d 265, 268 (1st Cir.2000). The dispositive feature of this inquiry is whether the medication is in fact causing such an impairment. See id. A district court often may satisfy this basic obligation when it queries a defendant about whether the medication he is taking has impaired his ability to understand the proceedings. See United States v. Morrisette, 429 F.3d 318, 322 (1st Cir.2005); Cody v. United States, 249 F.3d 47, 53 (1st Cir.2001); see also United States v. Román-Orench, — Fed.Appx. -, - (1st Cir.2015) [No. 13-2082, slip op. at 4]. But context is crucial, and in some situations the court’s obligation does not end there. Thus, the “better practice” is for a district court “to identify which drugs, a defendant is taking, how recently they have been taken and in what quantity, and (so far as possible) the purpose and consequences.” Savino n-Acosta, 232 F.3d at 268.

Here, the district court inquired into what medications Caramadre was taking and Caramadre’s ability to understand the proceedings. The court also elicited from Caramadre an assurance that his medications were not preventing him from participating fully in the change-of-plea colloquy. In addition, the court had some other assurances. For one thing, Caramadre’s behavior during the change-of-plea colloquy corroborated his statements to the court. For another thing, Caramadre’s lawyer vouched for his client’s ability to understand the proceedings. A district court may reasonably rely on the assurances of the defendant and his counsel to help to ascertain the defendant’s mental clarity. 5 See id. at 269. Finally, the court’s duty to delve into the specifics of a defendant’s medications is relaxed to some degree where, as here, there are no “other identifiable red flags in [the defendant’s] performance at the hearing.” United States v. Kenney, 756 F.3d 36, 47 (1st Cir.), cert. denied, — U.S. -, 135 S.Ct. 770, 190 L.Ed.2d 640 (2014).

To be sure, Caramadre had an impressive list of medications, along with a history of depression and anxiety. Given these facts, we think that the district court’s handling of this issue was marginal at best. A deeper dive into the effects of the medications and Caramadre’s psychiatric history may well have been warranted. But *369 our standard of review is deferential, see Morrisette, 429 F.3d at 322, and in all events, two other sets of considerations impel us to find that any error was harmless.

First, Caramadre has never made an explicit claim that either his medication regime or his history of depression and anxiety actually impaired his ability to understand the change-of-plea colloquy. Though he vigorously assails the manner in which the district court conducted that colloquy, his assignments of procedural error are untethered to any actual consequences. As such, they cannot ground his claim that the district court abused its discretion in denying his plea-withdrawal motion. See Savinon-Acosta, 232 F.3d at 268 (explaining that “merely technical failures to comply with Rule 11 are often found harmless”); United States v. Pellerito, 878 F.2d 1535, 1542 (1st Cir.1989) (explaining that “[t]here must be some evidence that the medication affected [the defendant’s] rationality”).

Second, the lengthy evidentiary hearing that the district court conducted on Cara-madre’s plea-withdrawal motion yielded fully supportable findings that refuted his claim that either his medications or his mental health history tainted his plea. As discussed in greater detail infra, the doctors who submitted affidavits regarding Caramadre’s mental state in the period leading up to his guilty plea failed to cast any plausible doubt on his rationality. Furthermore, Caramadre’s former attorneys testified extensively about his overall lucidity and clarity of mind.

That completes this phase of our inquiry. Viewing the record as a whole, we can discern no reversible error in the district court’s Rule 11 colloquy.

2. Caramadre’s Stated Reasons. Caramadre next advances a slew of arguments underpinning his claim that the district court improvidently rejected his stated reasons for seeking to withdraw his plea. We briefly address the least frivolous of these arguments — that he was not competent at the time of the plea hearing, that his counsel provided ineffective assistance, and that he believed that he would be dissembling if he entered a guilty plea — and otherwise rely on the district court’s cogent analysis. See Caramadre, 957 F.Supp.2d at 181-86.

Caramadre argues that his mental state was too fragile to permit him to enter a valid plea. He attributes his instability both to his depressed mental state and to his wife’s emotional breakdown on the second day of trial. He claims that the confluence of these conditions catapulted him into a “downward spiral,” rendering him incompetent to enter a guilty plea.

To succeed on such a claim, Caramadre must show more than a mere “sensitivity] to external considerations.” Pellerito, 878 F.2d at 1541. Rather, he must show that his decision to change his plea occurred under so much duress that it could no longer be considered a product of free will. See id.

In an attempt to carry this burden, Car-amadre submitted affidavits from two of his doctors, an affidavit from his wife’s doctor, and an affidavit from a psychiatric consultant. The district court reviewed these submissions and found them wanting. See Caramadre, 957 F.Supp.2d at 169-71. After careful consideration, we conclude that this finding was well within the encincture of the court’s discretion.

The affidavits of Caramadre’s doctors were of little force. While they purported to describe his mental state during the four days of trial, neither doctor had evaluated Caramadre (or even spoken to him) during that period. By the same token, the doctor who cared for Mrs. Caramadre *370 ventured no opinion regarding Carama-dre’s mental health.

The affidavit of Caramadre’s retained expert was more to the point: that physician stated that it was “reasonable to conclude” that Caramadre was not competent to plead. But even this witness did not opine that Caramadre in fact lacked the capacity to plead.

We think it is significant that the district court, in refusing to find Caramadre incompetent to plead based on these four affidavits, did not view them in a vacuum. The court appropriately considered, for example, Caramadre’s own participation in negotiating the terms of the plea agreement, see United States v. Ramos, 810 F.2d 308, 313 (1st Cir.1987); the conclusions of Caramadre’s principal lawyer about his client’s mental clarity, see Savinon-Acosta, 232 F.3d at 269; and the court’s own observations of Caramadre over a prolonged period, see United States v. Buckley, 847 F.2d 991, 998-1000 (1st Cir.1988); see also Román-Orench, — Fed.Appx. at —— [slip op. at 5].

Balancing the tepid evidence contained in the affidavits against the court’s first-hand knowledge of what had transpired, we descry no abuse of discretion in its determination that Caramadre’s assertion of incompetence was not a fair and just reason for withdrawing his plea. When all is said and done, Caramadre is simply complaining that the district court weighed his proffered evidence less heavily than he would have liked. That is not enough: a district court does not abuse its discretion when it evaluates a body of evidence, chooses between two inferences which, though conflicting, are both rational, and offers plausible reasons for its choice. See Pellerito, 878 F.2d at 1538.

The case law supplies a final check. When the results of the evidentiary hearing are considered, Caramadre’s case is not materially different from the mine-run of analogous cases. See, e.g., Santiago Miranda, 654 F.3d at 137-39 (upholding plea where defendant argued involuntariness based on prescription drug abuse, lack of sleep, and familial pressure); United States v. Sousa, 468 F.3d 42, 46 (1st Cir.2006) (upholding plea despite defendant’s argument that “distressing news” about wife’s terminal illness impaired his capacity to plead); Aker, 181 F.3d at 170-71 (upholding plea notwithstanding defendant’s claim of depression over wife’s death and inability to sleep); Pellerito, 878 F.2d at 1541-42 (upholding plea where defendant claimed “an agitated emotional state” along with abuse of anti-anxiety drugs). Normally, such situations are fact-specific and, thus, are apt to be grist for the district court’s mill. See Merritt, 755 F.3d at 9 (noting that “a district court’s close relationship to the plea process affords it a superior coign of vantage”); Pellerito, 878 F.2d at 1538 (noting that, when “[confronted with an attempt at plea retraction, the trial judge must make an idiocratic, particularistic, fact-bound assessment”). Based on the teachings of the case law, we will not second-guess the trier’s informed determination of the voluntariness of the defendant’s plea without good reason. See United States v. Austin, 948 F.2d 783, 786 (1st Cir.1991). In this instance, we discern no good reason: the district court reviewed all of the evidence and supportably found that neither Caramadre’s history of depression and anxiety nor his wife’s breakdown comprised a fair and just reason allowing him to retract his plea. See Caramadre, 957 F.Supp.2d at 168-74. Caramadre has offered nothing that would give us a principled basis to second-guess this finding.

3. Ineffective Assistance of Counsel. Caramadre goes on to contend that he should have been allowed to withdraw his *371 guilty plea because his first set of attorneys provided ineffective assistance to him. Once he perceived his attorneys’ ineptitude at trial, his thesis runs, he “suddenly]” came to the realization that he had no choice but to plead guilty and throw himself upon the mercy of the court. . This contention is hopeless.

The court below accurately rehearsed the standard for assessing an ineffective assistance of counsel claim in the context of a plea-withdrawal motion. See Caramadre, 957 F.Supp.2d at 174-75. In fine, the challenger must demonstrate that counsel’s performance fell below an objective threshold of reasonable care and that this deficient performance prejudiced him. See Turner v. United States, 699 F.3d 578, 584 (1st Cir.2012); see generally Strickland v. Washington, 466 U.S. 668, 687, 104 5.Ct. 2052, 80 L.Ed.2d 674 (1984). In the plea-withdrawal context, the prejudice element requires a showing of “a reasonable probability that, but for counsel’s errors, he would not have pleaded guilty and would have insisted on going to trial.” Moreno-Espada v. United States, 666 F.3d 60, 64 (1st Cir.2012) (quoting United States v. Colón-Torres, 382 F.3d 76, 86 (1st Cir.2004)); see United States v. Isom, 85 F.3d 831, 837 (1st Cir.1996).

Although the district court correctly explained that “[cjounsel’s alleged ineffectiveness is only relevant to the extent it affected Caramadre’s decision to plead guilty,” Caramadre, 957 F.Supp.2d at 174 n. 9, its analysis focused on the trial performance of Caramadre’s lawyers (specifically, their purported failure to investigate witnesses and cross-examine them adequately). This focus was misplaced: when a defendant pleads guilty and later tries to withdraw his plea, the ineffective assistance of counsel inquiry must focus on his lawyer’s preparation, advice, and overall performance in counseling the defendant about whether to plead guilty. See Austin, 948 F.2d at 786-87; United States v. DeSimone, 736 F.Supp.2d 477, 486 (D.R.I.2010).

This is not to suggest that trial performance is wholly irrelevant to the ineffective assistance of counsel inquiry in the guilty plea context. A lawyer’s trial performance may be so deficient that it compels a defendant to plead under duress. But such trial performance is relevant to the ineffective assistance inquiry only to the extent that it affects the knowing and voluntary nature of a defendant’s decision to plead guilty.

Given this legal landscape, the district court’s focus on the trial performance of Caramadre’s lawyers was misplaced. Car-amadre did not tie his counsel’s trial performance to the voluntariness of his guilty plea and, thus, the meat of his argument— that he would not have pleaded guilty had his counsel performed better at trial — is inapposite. By his framing of this issue, Caramadre attempted to shoehorn a claim of ineffective assistance at trial into a plea-withdrawal inquiry. That attempt necessarily fails. 6 See Isom, 85 F.3d at 837; Austin, 948 F.2d at 786.

Caramadre’s assignment of error collapses when we reorient the ineffective assistance of counsel inquiry along the proper axis. The record does not support a claim that, but for his attorneys’ poor advice about the desirability of a plea, Caramadre “would not have pleaded guilty and would have insisted on going to trial.” *372 Moreno-Espada, 666 F.3d at 64 (quoting Colón-Torres, 382 F.3d at 86). Nor does Caramadre explain why the advice he was given was deficient. So, too, he wholly neglects to explain why, given better advice, he would have wanted the trial to continue.

That ends this aspect of the matter. Caramadre has not offered any support for the proposition that his attorneys were deficient in advising him about his guilty plea. Nor did he make any developed argument to this effect before the district court. Any such argument is, therefore, doubly waived. See United States v. Zannino, 895 F.2d 1, 17 (1st Cir.1990).

4. The Truth, the Whole Truth, and Nothing but the Truth. Caramadre has one last shot in his sling. Alford pleas aside, 7 a defendant who wishes to plead guilty to a criminal charge must admit that he committed the acts upon which the charge is predicated. See United States v. Broce, 488 U.S. 563, 570, 109 S.Ct. 757, 102 L.Ed.2d 927 (1989). That admission must be truthful; mere lip service is not enough. Here, Caramadre admitted his guilt when he changed his plea. In his plea withdrawal motion, however, he reversed course and claimed that he had lied with the knowledge and encouragement of his lawyers. The district court rejected this claim, see Caramadre, 957 F.Supp.2d at 185, and so do we.

Caramadre attempts to bolster his version of events by describing two communications that he had with his attorneys. For one thing, prior to the change-of-plea hearing, Caramadre sent an e-mail to one of his former lawyers inquiring about the possibility of an Alford plea, which “would eliminate m[y] needing to lie.” For another thing, Caramadre says that he told one of his former attorneys on the night before he changed his plea that he would “be lying” if he admitted guilt.

Confessing to the commission of a felony does not always come easily (particularly for a person who, like Caramadre, had been holding himself out as a pillar of the community and living a life of high-profile respectability). It is, therefore, not uncommon for persons accused of reprehensible crimes to waffle even when discussing the extent of their involvement with their counsel. That may well be what happened here: in conversations with his counsel, Caramadre equivocated from time to time about his guilt.

The cheese became binding, however, when the change-of-plea hearing began and Caramadre faced the district court. That is, literally and figuratively, the moment of truth — and in this instance, Caramadre unhesitatingly agreed under oath with the prosecutor’s version of the relevant events and unambiguously admitted his guilt. When Caramadre sang a much different song during the plea-withdrawal hearing, the district court determined that he was prevaricating then and that he had told the truth at the change-of-plea hearing. This determination was nothing more or less than a credibility call and, as such, is deserving of considerable deference. See, e.g., United States v. Patrone, 948 F.2d 813, 816 (1st Cir.1991); *373 United States v. Green, 887 F.2d 25, 28 (1st Cir.1989).

In addition, the court’s determination was consistent with the testimony of Cara-madre’s former attorneys. They vouchsafed that, based on their investigation and Caramadre’s admissions to them over the course of their extended representation of him, they were convinced that he was factually guilty and that his admissions of guilt at the change-of-plea hearing were genuine. To cinch the matter, Carama-dre’s former attorneys “emphasized [to Caramadre] the importance and necessity of telling the truth and not lying to the [cjourt” at the change-of-plea hearing. Caramadre, 957 F.Supp.2d at 185. The district court credited the former attorneys’ testimony, see id., and the record contains no compelling reason for rejecting that assessment.

In an effort to blunt the force of this reasoning, Caramadre posits that his case is analogous to United States v. DeSimone, 736 F.Supp.2d 477 (D.R.I.2010). There, the defendant did not agree with the recitation of the facts contained in his plea agreement and asked his attorney whether he had to lie in order to plead guilty. See id. at 479-80. The attorney “left [the] [defendant with the impression that lying to the [c]ourt was necessary to get his plea accepted.” Id. at 486. The defendant proceeded with his plea but later sought to retract it. The district court allowed him to do so, concluding that “a fair and just reason” existed for withdrawing the plea. Id.

The court below distinguished DeSi-mone on a number of grounds. 8 See Caramadre, 957 F.Supp.2d at 185. We agree that the two cases are not fair congeners. Unlike in DeSimone, Caramadre did not take issue with the prosecution’s version of the facts when he changed his plea; and more importantly, Caramadre’s former lawyers testified that they had instructed him not to lie. The district court not only found this testimony credible but also accepted the lawyers’ testimony that they would not have allowed Caramadre to plead guilty if they thought that doing so would require him to prevaricate. See Caramadre, 957 F.Supp.2d at 185.

C. Bias.

Caramadre insists that the district court’s refusal to permit him to withdraw his guilty plea was infected by judicial bias. He concentrates his fire principally on the “text and tenor” of the district court’s rescript. We begin our discussion with the government’s contention that Car-amadre’s bias claim has been waived and then proceed to address the components of that claim.

1. Waiver and Standard of Review.

Almost seven months elapsed between the denial of Caramadre’s plea-withdrawal motion and his sentencing. During this period, he never sought to have the district court recuse itself. The government’s argument that this inaction constituted a waiver of the bias claim has some support in the case law. See, e.g., United States v. DiPina, 230 F.3d 477, 486 (1st Cir.2000) (finding waiver when defendant had nei *374 ther moved for recusal nor otherwise raised claim of judicial bias in district court). But other cases indicate that plain error review may be appropriate when a party raises a bias-based recusal argument for the first time on appeal. See, e.g., United States v. Reynolds, 646 F.3d 63, 74 (1st Cir.2011); United States v. Cruz-Mercado, 360 F.3d 30, 36 (1st Cir.2004).

We are sensitive to a judge’s unflagging duty to be impartial. Given the importance of impartiality, we think that the better rule is that a claim of judicial bias, raised for the first time on appeal, should be reviewed for plain error. Consequently, we reject the government’s waiver argument and hold instead that Caramadre’s bias-based recusal claim engenders plain error review.

Plain error review requires a four-part showing: “(1) that an error occurred (2) which was clear or obvious and which not only (3) affected the defendant’s substantial rights, but also (4) seriously impaired the fairness, integrity, or public reputation of judicial proceedings.” United States v. Duarte, 246 F.3d 56, 60 (1st Cir.2001).

2. Merits. Judges have a duty to sit unless some compelling reason for recusal exists. See United States v. Snyder, 235 F.3d 42, 46 (1st Cir.2000). Not every hint of bias is disqualifying: after all, a judge is expected to make judgments, a process which entails forming opinions about the credibility of witnesses and the intrinsic merit (or lack of merit) of cases that he hears. See Liteky v. United States, 510 U.S. 540, 550-51, 114 S.Ct. 1147, 127 L.Ed.2d 474 (1994). In order for us to find disqualifying bias and overrule a judge’s decision (explicit or implicit) that no sound basis for his recusal exists, an appellant must show that the judge’s actions were “so extreme as to display [a] clear inability to render fair judgment.” Id. at 551, 114 S.Ct. 1147.

To support his bias-based recusal claim, Caramadre relies on a string of strongly worded statements excerpted from the district court’s rescript denying his motion to withdraw. Specifically, he points to the following:

• the district court’s characterization of his plea-withdrawal motion as “entirely meritless, bordering on frivolous”;

• the district court’s conclusion that none of the evidence presented by Caramadre — including his medical affidavits — “even remotely supported]” his claim of incompetence;

• the district court’s dismissive treatment of Caramadre’s ineffective assistance of counsel claim;

• the district court’s intimation that Car-amadre was suggesting that his former attorneys deliberately undermined his defense in order to pressure him into pleading guilty;

• the district court’s statement that Car-amadre’s plea-withdrawal motion was “an incredibly cynical and disturbing effort to manipulate the court and the criminal justice system”; and

• the district court’s suggestion that the actions of Caramadre’s new counsel might subject him to disciplinary review.

These statements, taken collectively, show that the district court did not think much of Caramadre’s plea-withdrawal motion. Admittedly, the court couched its findings and conclusions in blunt language. But trial judges are not required either to mince words or to sugar-coat their views. See Logue v. Dore, 103 F.3d 1040, 1045 (1st Cir.1997). Blunt language, without more, does not translate into a showing of judicial bias. See United *375 States v. Rodríguez-Rivera, 473 F.3d 21, 27-28 (1st Cir.2007).

The Supreme Court has taught that “remarks during the course of trial that are critical or disapproving of, or even hostile to, counsel, the parties, or their cases” are usually insufficient to prove bias. Liteky, 510 U.S. at 555, 114 S.Ct. 1147. The same is true of a court’s “expressions of impatience, dissatisfaction, annoyance, and even anger.” Id. at 555-56, 114 S.Ct. 1147. The case at hand falls within these general rules, not within the long-odds exceptions to them. Though the court below employed forceful rhetoric, its comments, without exception, are supported by a reasonable view of the record evidence. The court did not cross the Liteky line.

This conclusion is borne out by contrasting the district court’s rhetoric with words and conduct that have been held insufficient to require recusal. See, e.g., United States v. Ofray-Campos, 534 F.3d 1, 32-34 (1st Cir.2008) (finding no judicial bias though judge, inter alia, interrupted counsel during opening and closing statements, told counsel to “shut up” during a sidebar conference, and made demeaning remarks about counsel’s performance); Rodríguez-Rivera, 473 F.3d at 26-29 (finding no judicial bias though judge, inter alia, reprimanded counsel in open court, commented unfavorably on counsel’s objections, and “made a series of unpredictable and adverse rulings” against the defendant); DiPina, 230 F.3d at 486 (finding no judicial bias though judge characterized defendant’s legal arguments as “worthless” and remarked on his “criminal conduct”). By comparison, the rhetoric challenged here easily passes muster.

We summarize succinctly. While Cara-madre has directed a barrage of epithets at the district court, he has fallen far short of showing that the court was biased against him. Put another way, the district court’s unflattering assessment of Carama-dre’s litigation strategy and substantive claims does not sink to the level of disqualifying bias. On this record, we simply cannot find that the district court’s words displayed an inability to render a fair judgment. Cf. Christian Recorder-Proverbs (Mar. 22, 1862) (“Sticks and stones will break my bones, but words will never harm me.”). We conclude, therefore, that there was no error, plain or otherwise, in the district court’s failure to recuse itself sua sponte.

3. Rule 10(c). This brings to the fore Caramadre’s appeal of the district court’s disposition of his Rule 10(c) motion. Cara-madre asseverates that statements made by the district court during an unrecorded and untranscribed chambers conference held on January 15, 2013, show that the court pre-judged his plea-withdrawal motion and exhibited bias against him.

We set the stage. Caramadre’s version of what transpired at the chambers conference is contained in an affidavit of successor counsel, appended to his Rule 10(c) motion. The district court discarded this account and substituted its own summary of what was said and done.

Caramadre challenges the court’s substituted version. Relying on his lawyer’s affidavit, he attributes certain statements to the court. A representative sampling follows:

• that the first week of trial had been a “complete, unmitigated disaster” for Caramadre;

• that the government had made a “compelling, overwhelming presentation of evidence of guilt” during the four days of trial;

• that had the trial continued, it would have been from Caramadre’s point of *376 view, “a train wreck for the next three months”; and

• that Caramadre had changed his plea because “he was getting killed at trial.”

Federal Rule of Appellate Procedure 10(c) provides: -

If the transcript of a hearing or trial is unavailable, the appellant may prepare a statement of the evidence or proceedings from the best available means, including the appellant’s recollection. The statement must be served on the appellee, who may serve objections or proposed amendments within 14 days after being served. The statement and any objections or proposed amendments must then be submitted to the district court for settlement and approval. As settled and approved, the statement must be included by the district clerk in the record on appeal.

As a threshold matter, the government suggests that, because Caramadre appealed the district court’s Rule 10(c) ruling separately, we lack jurisdiction over that appeal. We do not agree. The district court’s Rule 10(c) order was a final order, filed after Caramadre’s conviction and sentence had already been appealed. It was, therefore, appealable under 28 U.S.C. § 1291. This makes perfect sense: the core purpose of Rule 10(c) would be frustrated if a district court’s version of events was inoculated against judicial review. See Bergerco, U.S.A. v. Shipping Corp. of India, Ltd., 896 F.2d 1210, 1214-15 (9th Cir.1990); cf. United States v. Mori, 444 F.2d 240, 246 (5th Cir.1971) (explaining that a district court may consider a motion to correct the record under Federal Rule of Appellate Procedure 10(e) “even after appeal has been taken”).

We review a district court’s disposition of a Rule 10(c) motion for abuse of discretion. Cf. United States v. Pagán-Ferrer, 736 F.3d 573, 582 (1st Cir.2013), cert. denied, sub nom. Vidal-Maldonado v. United States, — U.S. -, 134 S.Ct. 2839, 189 L.Ed.2d 810 (2014) (reviewing denial of Rule 10(e) motion for abuse of discretion). The movant (here, Carama-dre) must establish that “the trial court’s account is patently unreasonable or deliberately false,” Rogan v. Menino, 175 F.3d 75, 80 (1st Cir.1999), and that this account prejudiced the presentation of his claims on appeal, see In re Cambridge Literary Props., Ltd., 271 F.3d 348, 349 (1st Cir.2001).

Caramadre upbraids the district court for relying on its own recollection of the chambers conference rather than accepting the version of events proffered by Car-amadre’s counsel. Relatedly, he submits that the district court improvidently expanded the record by adding explanations for why it made certain statements.

Caramadre’s insistence that the district court had no right to set forth its own version of events reads Rule 10(c) in too grudging a manner. In terms, the rule provides that once an appellant has “prepare[d] a statement of the evidence or proceedings from the best available means,” the statement must “be submitted to the district court for settlement and approval.” Fed. R.App. P. 10(c). The phrase “settlemeht and approval” is generous in its scope, and nothing prohibits a court from drawing on its own memory of events in the “settlement and approval” process.- Indeed, it would be folly for a judge to close his eyes to case-related matters within his personal knowledge. We hold, therefore, that a district court may rely on its own recollection of relevant events in settling and approving a proposed Rule 10(c) statement. See United States v. Kenney, 911 F.2d 315, 317-18 (9th Cir.1990); see also United States v. *377 Brown, 202 F.3d 691, 696-97, 697 n. 8 (4th Cir.2000) (noting approvingly that in weighing parties’ competing versions of what occurred at a hearing, district court necessarily relied on its own recollection).

Caramadre also argues that the district court was obliged to adopt his version of the facts because his counsel’s notes were “contemporaneously recorded” and the government never challenged their accuracy. But a rule to this effect would reduce the district court’s role to that of a rubber stamp, and we do not think that the law imposes so counterintuitive a requirement. A case in point is United States v. Keskey, 863 F.2d 474, 478 (7th Cir.1988), in which the Seventh Circuit rejected a similar argument. Simply put, the district court was not obliged to elevate the lawyer’s notes over its own recollection.

Caramadre’s contention that the district court had no authority to elaborate on what was said is likewise unavailing. Common sense suggests that, in the Rule 10(c) settlement and approval process, a district court must have the power to contextualize what was said. And though the district court’s Rule 10(c) statement goes beyond mere eontextualization, that overreach makes no difference here: even were we to accept lock, stock, and barrel the version of events limned in Carama-dre’s Rule 10(c) statement, Caramadre’s claim of judicial bias would fail. The carefully culled statements reflect nothing more than the district court’s decidedly negative evaluation of Caramadre’s attempt to withdraw his plea. Those statements are insufficient to demonstrate that the district court harbored a disqualifying bias against Caramadre. See supra Part 11(C)(2).

III. SENTENCING

Caramadre attempts to challenge his sentence on two grounds. He asserts both that the district court engaged in vindictive sentencing and that its order for $46,000,000 in restitution is insupportable. The government submits that these claims are barred by the waiver-of-appeal provision contained in Caramadre’s plea agreement, 9 and we agree. We explain briefly.

Our case law makes pellucid that “[a] defendant who waives his right to appeal and thereafter attempts to avoid the effect of the waiver must confront the waiver head-on.” United States v. Miliano, 480 F.3d 605, 608 (1st Cir.2007). Such waivers are “presumptively valid,” subject to three “stringent criteria.” United States v. Teeter, 257 F.3d 14, 23, 25 (1st Cir.2001). First, the plea agreement must clearly “elucidat[e] the waiver and delineat[e] its scope.” Id. at 24. Second, the court’s inquiries at the change-of-plea colloquy must “suffice[ ] to ensure that the defendant freely and intelligently agreed to waive [his] right to appeal.” Id. at 24. Third, pretermitting the right to appeal must not result in a “miscarriage of justice.” Id. at 25.

Caramadre’s plea agreement stated in pertinent part that: “Defendant hereby *378 waives [his] right to appeal the convictions and sentences imposed by the Court, if the sentences imposed by the Court are at or below the government’s maximum recommended sentence.” This language is direct and to the point; it clearly elucidates the waiver. What is more, the district court took pains at the ehange-of-plea hearing to ensure that Caramadre understood the effect of the waiver. Nor does Caramadre argue that his term of immurement exceeded the boundaries adumbrated in the plea agreement (which limited any prison sentence to a maximum of ten years). Withal, Caramadre tries to skirt the waiver in three different ways. None of his arguments is convincing.

Caramadre’s first sortie is stillborn. He argues that the plea agreement as a whole is invalid because he should have been allowed to withdraw his plea. We already have explained why the premise of this argument is wrong, see supra Part II, so we say no more about it.

Caramadre’s most loudly bruited claim implicates the scope of the appeal waiver. 'He posits that the plea agreement did not foreclose him from appealing the restitution order. In support, he notes the lack of any explicit reference to restitution in the waiver-of-appeal provision; and he points to language elsewhere in the plea agreement stating that the amount of restitution would be determined in the future. Thus, Caramadre says, the restitution order does not fall within the scope of the waiver-of-appeal provision.

This claim, though forcefully presented, runs headlong into our decision in United States v. Okoye, 731 F.3d 46 (1st Cir.2013). There, the plea agreement included a waiver-of-appeal provision similar to Cara-madre’s: neither provision made any explicit mention of restitution. See id. at 48. We nonetheless concluded that the waiver provision applied to a restitution order imposed as part of the defendant’s sentence. See id. at 49-50. We explained that the plea agreement as a whole “unambiguously established that [the defendant’s] sentence would include ‘restitution in the amount of loss’ ” and, thus, the appeal waiver extended to the restitution award. Id. at 49.

Okoye and this case are on all fours. Caramadre resists this obvious congruence, though, trying to distinguish Okoye on the ground that the plea agreement there contemplated a specific amount of restitution whereas the plea agreement here stated that the amount of restitution was yet to be determined. This is a distinction without a difference. That Cara-madre’s plea agreement did not specify a specific restitutionary amount has no bearing at all on whether restitution should properly be considered part of Carama-dre’s “sentence.”

At the expense of carting coal to Newcastle, we add that the waiver-of-appeal provision applies even more clearly here than in Okoye. There, the waiver provision stated that the defendant “[would] not file a direct appeal nor collaterally challenge any prison sentence of 27 months or less.” Id. at 48 (emphasis in original). The use of the modifying adjective “prison” gave rise to a colorable argument that the portion of the sentence to which the waiver applied did not include restitution. The defendant made this argument, and the Okoye court debunked it. See id. at 49-50. This argument is not available to Caramadre; the waiver-of-appeal provision contains no comparable modifier.

That restitution is a part of Caramadre’s sentence scarcely can be doubted. See 18 U.S.C. § 3663A(a)(l); United States v. Salas-Fernández, 620 F.3d 45, 47 & n. 2 (1st Cir.2010). Here, moreover, Caramadre’s plea agreement affirms that the government was “free to recommend *379 any combination of supervised release, fines, and restitution which it deems appropriate.” The clear implication of this statement is that restitution would be part of Caramadre’s sentence. 10

In a last-ditch effort to elude the grasp of the appeal waiver, Caramadre invokes the miscarriage of justice exception. See Teeter, 257 F.3d at 25-26. This assignment of error need not detain us.

The miscarriage of justice exception is to be applied “sparingly and without undue generosity.” Id. at 26. It is not intended to redress “mere ‘garden-variety’ claims of error.” United States v. Rivera-López, 7 36 F.3d 633, 635 (1st Cir.2013) (quoting Teeter, 257 F.3d at 26). Caramadre’s claim of vindictive sentencing is reminiscent of the bias claims that we already have rejected, see supra Part 11(C), and he has made no showing that this claim comes within the narrow confines of the miscarriage of justice exception.

. In the first place, the plea agreement capped Caramadre’s exposure with respect to incarceration at ten years. This was considerably below the top of his guideline sentencing range. Even so, the district court sentenced him to only a six-year prison term. Surely, that was not a miscarriage of justice.

His plaint that the restitution amount is similarly excessive is unimpressive. That amount was calculated by the magistrate judge after a three-day evidentiary hearing and confirmed by the district court. To cinch the matter, the $46,000,000 total, though large, has ample footing in the record.

To be sure, Caramadre has left no doubt but that he considers his sentence “unjust.” But a defendant’s dissatisfaction with his sentence, no matter how profound, cannot constitute a basis for circumventing a waiver-of-appeal provision to which he agreed. See United States v. Edelen, 539 F.3d 83, 86-87 (1st Cir.2008). To allow Caramadre to frustrate his appeal waiver in the circumstances of this case would cheat the government of one of the salient benefits of the bargain that it struck with Caramadre.

IV. CONCLUSION

To recapitulate, Caramadre — ably represented by experienced counsel — elected to plead guilty to serious charges. When he thereafter had a change of heart and sought to retract his guilty plea, the district court gave him every opportunity to demonstrate a fair and just reason for doing so. The district court’s determination that Caramadre failed in this effort was neither infected by legal error nor constituted an abuse of discretion. The sentencing determinations that followed are insulated from review because Cara-madre, appropriately warned, waived his right to appeal his sentence as part of the plea agreement that he negotiated and signed.

We need go no further. For the reasons elucidated above, Caramadre’s appeals are futile. He has reaped what he has sown.

Affirmed.