PUBLISHED
UNITED STATES COURT OF APPEALS FOR THE FOURTH CIRCUIT
No. 24-4231
UNITED STATES OF AMERICA, Plaintiff – Appellee,
v. ROBERT SMITH, JR., Defendant – Appellant.
Appeal from the United States District Court for the Eastern District of Virginia, at Alexandria. Claude M. Hilton, Senior District Judge. (1:23−cr−00090−CMH−1)
Argued: September 10, 2025 Decided: October 15, 2025
Before WILKINSON, THACKER, and HEYTENS, Circuit Judges.
Affirmed by published opinion. Judge Wilkinson wrote the opinion, in which Judge Thacker and Judge Heytens joined.
ARGUED: Frances H. Pratt, OFFICE OF THE FEDERAL PUBLIC DEFENDER, Alexandria, Virginia, for Appellant. Nicholas Allen Durham, OFFICE OF THE UNITED STATES ATTORNEY, Alexandria, Virginia, for Appellee. ON BRIEF: Geremy C. Kamens, Federal Public Defender, Valencia D. Roberts, Assistant Federal Public Defender, OFFICE OF THE FEDERAL PUBLIC DEFENDER, Alexandria, Virginia, for Appellant. Jessica D. Aber, United States Attorney, OFFICE OF THE UNITED STATES ATTORNEY, Richmond, Virginia, for Appellee.
2
WILKINSON, Circuit Judge:
The hearing for Robert Smith, Jr.’s guilty plea brushed over several parts of Rule 11 in the Federal Rules of Criminal Procedure. So he appeals his conviction, challenging both the validity of his plea and the district court’s denial of his motion to withdraw it.
While we recognize the oversights below and reiterate the value in district courts meticulously following Rule 11, we must also respect the finality of guilty pleas. Smith has not shown a reasonable probability that he would not have pleaded guilty had there been no Rule 11 error, and the district court properly exercised its discretion in denying his motion to withdraw the plea.
I.
As recited in his written plea agreement, Smith bought fifty-nine firearms over the span of sixteen months. J.A. 61. For every transaction, he filled out a firearms transaction record as prescribed by the Bureau of Alcohol, Tobacco, Firearms and Explosives. J.A. 61; see also 27 C.F.R. § 478.124(a). In each, Smith entered one of two addresses in Virginia as his home. J.A. 61. The problem? He did not live in either of them. J.A. 62. In fact, one did not even exist. J.A. 62.
A grand jury returned a twenty-four-count indictment against Smith, reflecting the twenty-four forms he had completed, for lying to a federal firearms licensee. J.A. 25–27; see also 18 U.S.C. § 924(a)(1)(A). In the lead-up to trial, the government assembled a host of compelling facts indicating Smith committed these crimes. It was prepared to show, for instance, not only that one of his two listed addresses did not exist, but also that the
3
residents of the other address did not know Smith, and that GPS evidence consistently linked his cell phone to the same home in Maryland. J.A. 17–18, 20–21.
The day before trial, the parties negotiated a plea deal: Smith would admit guilt to one count and waive his right to appeal. J.A. 50, 52–53. The government, in exchange, would drop the other counts and recommend a three-level sentence reduction for acceptance of responsibility under Sentencing Guideline § 3E1.1. J.A. 52–53.
At the plea hearing, the district court touched on many provisions in Rule 11. For instance, Smith confirmed knowing and waiving his constitutional rights to plead not guilty, to a jury trial, and to cross-examine witnesses. J.A. 43–44; see also Fed. R. Crim. P. 11(b)(1)(B)–(C), (E)–(F). He also stated he understood the Sentencing Guidelines might affect his sentence and that, by pleading guilty, he could receive up to five years of imprisonment, three years of supervised release, $250,000 in fines, and a $100 special assessment. J.A. 43; see also Fed. R. Crim. P. 11(b)(1)(H), (J), (L)–(M). And, after mentioning his brief stint in college, Smith confirmed reviewing his written plea agreement, which contained thorough descriptions of the sentencing process and the waiver of his right to appeal. J.A. 42, 51–53; see also Fed. R. Crim. P. 11(b)(1)(M)–(N).
At the same time, the colloquy fell short in other respects. The district court did not explicitly ask Smith if he understood the appeal waiver. It did not specifically describe how the Sentencing Guidelines come into play and how they are ultimately nonbinding. And it did not ask numerous questions that are often deployed to supplement Rule 11, such as whether Smith was then under the influence of drugs or alcohol. Even so, Smith’s retained attorney raised no objections, and the district court accepted his plea. J.A. 46.
4
Roughly three months later, Smith disclosed that he had terminated his attorney for ineffective assistance. J.A. 64. According to Smith, his counsel had not communicated with him at all since the plea hearing and had filed a sentencing memorandum without his knowledge. J.A. 64. The attorney had also purportedly given Smith a statement of facts to sign that differed in several ways from the one formally filed with his plea agreement. J.A. 72. And one week before the Rule 11 hearing, his plea counsel was recommended for disbarment. See In re Johnson, Report and Recommendation of the Board on Professional Responsibility, No. 20-BD-020, at 4 (D.C. Ct. App. July 18, 2023). Worse still, his plea counsel has since been disbarred and twice suspended.
The district court accordingly appointed a public defender. J.A. 79. After another three months, Smith moved to withdraw his plea. J.A. 99. Reciting the above findings, he argued that his guilty plea was not knowing and voluntary because of his plea counsel’s repeated ineffectiveness. J.A. 102–09.
The district court denied the motion, however, and sentenced Smith to fourteen months’ imprisonment with two years of supervised release. J.A. 295–98; 357–58. Smith appeals, raising two challenges: First, citing the deficiencies in his Rule 11 colloquy, he again argues his plea was not knowing and voluntary. Second, given both these deficiencies and his plea counsel’s purportedly ineffective assistance, Smith contends the district court abused its discretion in denying the motion to withdraw his plea.
II.
At the outset, we recognize that the plea colloquy was not what it should have been.
Though the district court asked Smith if he knew the Sentencing Guidelines would affect
5
his sentence, J.A. 43, it did not specifically discuss, among other things, its authority to depart from a Guidelines-based sentence. This violates Rule 11(b)(1)(M). Likewise, the district court failed to address the appellate waiver in Smith’s plea agreement. This violates Rule 11(b)(1)(N). The district court also did not ask Smith if he was then being treated for mental illness or addiction, or if he was under the influence or narcotics or alcohol. These omissions run counter to the recommended process for evaluating whether a guilty plea is knowing and voluntary. See Fed. Jud. Ctr., Benchbook for U.S. District Court Judges § 2.01(D)(6)–(7), at 64–65 (6th ed. 2013). And the district court never asked whether all plea offers had been conveyed to Smith, nor did it go over any part of the plea agreement with him. These oversights too ignore best practice. Id. § 2.01(E), at 65–66.
Time, time, and time again we have reminded district courts to hew closely to Rule 11. See, e.g., United States v. Lockhart, 947 F.3d 187, 193–94 (4th Cir. 2020) (en banc); United States v. Sanya, 774 F.3d 812, 817 (4th Cir. 2014). After all, “[t]hat a guilty plea is a grave and solemn act to be accepted only with care and discernment has long been recognized.” Brady v. United States, 397 U.S. 742, 748 (1970). The slate of constitutional rights signed away in a plea agreement, as well as the resulting risk of criminal punishment, “demand[] the utmost solicitude of which courts are capable in canvassing the matter with the accused to make sure he has a full understanding of what the plea connotes and of its consequence.” Boykin v. Alabama, 395 U.S. 238, 243–44 (1969).
To that end, it is not too much to ask that district courts conduct their plea colloquies with precision. True enough, Rule 11’s obligations and disclosures are perhaps redundant to the experienced trial judge. But the same cannot be said for the defendant opting to plead
6
guilty—the party that Rule 11 is designed to protect. See Halliday v. United States, 394 U.S. 831, 832 (1969) (per curiam). What might seem like needless formalities and truisms to a learned jurist could be vital protections and revelations for an otherwise unknowing layman. District courts best determine whether a guilty plea is knowing and voluntary by dotting their i’s and crossing their t’s during the colloquy. Rule 11, though lengthy and frequently trodden, is “designed to assist the district judge in making th[is] constitutionally required determination.” McCarthy v. United States, 394 U.S. 459, 365 (1969).
III.
All that said, we have also “repeatedly refused to script the Rule 11 colloquy, relying rather on the experience and wisdom of the district judges below.” United States v. Wilson, 81 F.3d 1300, 1307 (4th Cir. 1996). Indeed, even if a district court departs from Rule 11’s text, we generally must deem the departure “harmless error if it does not affect substantial rights.” Fed. R. Crim. P. 11(h). Put otherwise, “[t]he overarching inquiry is whether there is a reasonable probability that the defendant would not have pleaded guilty had there been no Rule 11 error.” United States v. Hairston, 522 F.3d 336, 341 (4th Cir. 2008); see also United States v. Thorne, 153 F.3d 130, 133 (4th Cir. 1998).
Harmless-error review makes good sense in this context. For one thing, trial judges naturally make mistakes even when trying earnestly to follow Rule 11’s many terms—they are only human, after all. See United States v. Reckmeyer, 786 F.2d 1216, 1221 (4th Cir. 1986) (quoting United States v. Dayton, 604 F.2d 931, 939–40 (5th Cir. 1979) (en banc)). For another, the “finality of guilty pleas” is of “particular importance.” United States v. Dominguez Benitez, 542 U.S. 74, 82 (2004). Such importance derives in part from the
7
vulnerability of a guilty plea to buyer’s remorse; that is, if guilty pleas were malleable, courts would soon find themselves swamped by defendants who, despite once admitting guilt, have convinced themselves during the long hours of confinement that a jury would have acquitted them. Finality also best preserves the benefits of guilty pleas for defendants and prosecutors alike, including leniency for the former and prompt convictions for the latter. Accordingly, while district courts must strive for comprehensive and punctilious plea hearings, appellate courts must not vacate and remand upon spotting just any misstep.
We emphasize harmlessness here because Smith, in his motion to withdraw his guilty plea, challenged only the effectiveness of his plea counsel—not the adequacy of the colloquy. See J.A. 102–09. These are not the same claim. True, both pertain in some fashion to whether Smith’s plea was knowing and voluntary. But more than that amorphous resemblance is needed. Were the rule otherwise, a litigant could preserve every due process claim merely by tendering one of them, however attenuated. Smith has thus forfeited his challenge to the plea colloquy, so we review it only for plain error. See United States v. Edgerton, 408 F. App’x 733, 735–36 (4th Cir. 2011) (per curiam) (deeming Rule 11 claim forfeited after defendant challenged only plea counsel’s effectiveness below). This means the burden of proof shifts from the government to Smith. United States v. Olano, 507 U.S. 725, 734–35 (1993).
With these guideposts in mind, consider Smith’s challenge to the Rule 11 colloquy.
Despite his qualms with its various omissions, none materially altered his decision to plead guilty. Indeed, the district court covered the lion’s share of Rule 11 through a series of clearly phrased questions to Smith. And in the instances where it fell short, the surrounding
8
circumstances indicate that a more by-the-book colloquy would not have changed anything.
For instance, though the district court failed to amply explain how the Sentencing Guidelines might affect Smith’s sentence, his written plea agreement—which he confirmed he had reviewed—thoroughly detailed the sentencing process. J.A. 43, 51–53. Smith, as a represented party with some college education, thus understood the role of the Guidelines, consistent with his admission to that effect during the colloquy. J.A. 42–43. The same goes for the appeal waiver; that is, despite the district court’s failure to bring it up during the hearing, Smith gleaned how the waiver worked when reviewing the comprehensive description of it in his plea agreement. J.A. 42, 52–53; see United States v. General, 278 F.3d 389, 400–01 (4th Cir. 2002) (upholding appeal waiver for represented defendant with ninth-grade education even though district court did not expressly ask about it). At any rate, “the proper remedy for an invalid appeal waiver is to sever [it],” leaving the rest of the plea agreement intact. United States v. Smith, 134 F.4th 248, 260 (4th Cir. 2025).
The other criticisms of the plea colloquy, assuming they concern Rule 11 violations in the first place, fall equally short. Smith laments how the district court did not extensively inquire into his competency during the colloquy, yet he has never raised a question regarding his competent state of mind. (Nor does the record cast any doubt over his competence.) Smith likewise criticizes the district court for failing to ask whether every plea offer had been presented to him, yet he does not allege the prosecution made any prior offers. (Nor, as before, does the record suggest any offers were made or left undisclosed to him.) And although the district court never went over any excerpts of the plea agreement
9
with Smith verbatim, much of the colloquy was premised on his reading the agreement. To reiterate, when asked by the district court whether he had “reviewed the plea agreement,” Smith answered “[y]es.” J.A. 42.
More compelling still is the other side of the harmless-error equation: the most likely outcome had Smith gone to trial. In this counterfactual, Smith likely would have lost his three-level sentence reduction for acceptance of responsibility. He also would have exposed himself to twenty-three additional counts of making false statements to federal firearms licensees, raising the stakes exponentially. And the government’s prospects of winning were all but certain. Recall that one of Smith’s purported addresses did not even exist, and the long-term residents of the other planned to testify that Smith had never lived there. J.A. 17–18. Recall also that the prosecution had GPS evidence showing, for every morning and every evening of the at-issue gun purchases, Smith’s mobile phone pinged cell sites in Maryland. J.A. 20–21. With such incriminating evidence, any rational defendant would have opted for the favorable guilty plea that Smith received. It was a bargain.
IV.
We thus proceed from the plea colloquy to Smith’s distinct, albeit related, challenge to the denial of his motion to withdraw his guilty plea. “[A] defendant does not have an absolute right to withdraw a guilty plea.” United States v. Moore, 931 F.2d 245, 248 (4th Cir. 1991). Rather, it may be withdrawn if “the defendant can show a fair and just reason for requesting the withdrawal.” Fed. R. Crim. P. 11(d)(2)(B). We have long articulated six non-exhaustive “Moore factors” to guide district courts in determining whether a reason is
10
“fair and just,” and we review their judgment only for an abuse of discretion. E.g., United States v. Mayberry, 125 F.4th 132, 140–41 (4th Cir. 2025). Far from showing an abuse of discretion, the Moore factors support the denial of the motion to withdraw the plea.
First is “whether the defendant has offered credible evidence that his plea was not knowing or not voluntary.” Id. As already discussed at length, the omissions in Smith’s colloquy—while not ideal—fall short of the threshold needed. The fact that the district court overlooked some items in Rule 11 cannot invariably mean that a defendant who reviewed and professed to understanding his plea agreement acted without knowledge or volition.
Second, and crucially, is “whether the defendant has credibly asserted his legal innocence.” Id. As Smith admits, he has never once claimed to be innocent of fabricating his address when buying firearms. Oral Arg. at 10:44–11:08. Nor could he. Again, the government had damning evidence, from both independent witnesses and Smith’s cell phone, that he did not live in the addresses listed in his firearms transaction records.
Third is “whether there has been a delay between the entering of the plea and the filing of the motion.” Moore, 931 F.2d at 248. One hundred and ninety-nine days separated Smith’s plea hearing and motion to withdraw, J.A. 41, 99, a period far exceeding what we have already deemed untimely. E.g., United States v. Nicholson, 676 F.3d 376, 384 (4th Cir. 2012) (two months). True, this delay is partially mitigated by the fact that, during this over half-year wait, Smith was appointed new counsel and reviewed the full discovery in his case. Yet even the shortest interim alleged by Smith—fifty-two days, J.A. 327—
11
exceeds periods that we have already deemed too tardy. E.g., Moore, 931 F.2d at 248 (six weeks).
Fourth is “whether defendant has had close assistance of competent counsel.” Id. A defendant seeking to establish this factor must demonstrate “(1) that his counsel’s performance ‘fell below an objective standard of reasonableness’ and (2) that ‘there [was] a reasonable probability that, but for counsel’s error, he would not have pleaded guilty and would have insisted on going to trial.’” United States v. DeFreitas, 865 F.2d 80, 82 (4th Cir. 1989) (alteration in original) (quoting Hill v. Lockhart, 474 U.S. 52, 57, 59 (1985)). Vigorously trying his hand at this factor, Smith criticizes his plea counsel’s communications as ineffective and points out the attorney’s disciplinary history. But these observations pertain solely to the first half of our inquiry. By contrast, nowhere in his briefing does Smith argue he would have opted for trial but for his plea counsel’s purported incompetence. And that argument would be a mighty difficult one to make; again, Smith avoided an open-and-shut trial—and a longer prison sentence—by pleading guilty. Assuming plea counsel’s performance was deficient, then, it made no difference to the outcome of Smith’s plea.
Fifth is “whether withdrawal will cause prejudice to the government.” Moore, 931 F.2d at 248. Had the district court granted Smith’s motion to withdraw, the government would have needed to spend time reassembling its witnesses and repreparing its case. In short, the government had its entire case prepared for trial until Smith, at the last minute, decided to plead guilty. That was his prerogative, but his second change of heart in his
12
motion to withdraw the plea cannot obscure the fact that the government’s extensive foregone preparation suffices to show prejudice.
Sixth is “whether [withdrawal] will inconvenience the court and waste judicial resources.” Id. Like above, granting Smith’s motion would have required the district court to refamiliarize itself with, and reschedule for trial, a nearly year-old case. Such a step “would have been futile” considering that Smith offered no real rejoinder to the prosecution’s compelling evidence. Id. at 249. Indeed, this factor grows even more one- sided on appeal, as Smith has since been released from custody. See Oral Arg. at 2:17– 3:34. To reopen his plea and expose him to further incarceration “would be pointless, perhaps even unjust.” Moore, 931 F.2d at 249.
V.
Smith’s dual requests are therefore unavailing. He asks first that we forego the instruction in Criminal Rule 11(h) and allow omitted parts of the plea colloquy to supersede the more inclusive whole. He then requests that we disregard the mutuality of obligation that underscores a guilty plea’s resemblance to a negotiated contract. The district court did not pursue a perfect path here, but neither did it commit reversible error or an abuse of discretion. All things considered, we are led to affirm.
AFFIRMED