Stringer v. United States

Notice: This opinion is subject to formal revision before publication in the Atlantic and Maryland Reporters. Users are requested to notify the Clerk of the Court of any formal errors so that corrections may be made before the bound volumes go to press.

DISTRICT OF COLUMBIA COURT OF APPEALS No. 24-CO-0163 BARRY D. STRINGER, APPELLANT, V. UNITED STATES, APPELLEE.

Appeal from the Superior Court of the District of Columbia (2005-FEL-004970) (Hon. Julie H. Becker, Motions Judge) (Argued December 11, 2025 Decided April 30, 2026) Gregory M. Lipper for appellant.

David P. Saybolt, Assistant United States Attorney, with whom Edward R. Martin, Jr., United States Attorney at the time the brief was filed, and Chrisellen R. Kolb, Nihar R. Mohanty, and Caroline R. Burrell, Assistant United States Attorneys, were on the brief, for appellee.

Before DEAHL, HOWARD, and SHANKER, Associate Judges.

SHANKER, Associate Judge: This post-conviction matter relates back to the murder of Tilford Johnson in 2003. Appellant Barry D. Stringer and his nephew Roderick Charles were charged with the murder and related offenses and tried separately. Both men were convicted in 2006 of armed robbery, first-degree felony murder while armed, and several related offenses. We affirmed Mr. Stringer’s and Mr. Charles’s convictions on direct appeal but remanded in both cases for resentencing on merged offenses. Mr. Charles was resentenced in October 2014 to months of imprisonment; Mr. Stringer was resentenced in January 2021 to 432 months of imprisonment.

In 2014, based on new evidence, Mr. Stringer moved under the Innocence Protection Act (IPA), D.C. Code § 22-4131 et seq., for vacatur of his convictions and dismissal of the charges or a new trial. The new evidence consisted of an affidavit by Mr. Charles in which he claimed that he had acted alone in killing Mr. Johnson and that Mr. Stringer was not involved. Following an evidentiary hearing in 2018 that included testimony from Mr. Charles, the trial court denied Mr. Stringer’s motion on the ground that Mr. Charles was not credible. Mr. Stringer appealed, arguing that the trial court clearly erred in discrediting Mr. Charles.

Because we were “uncertain about the bases for some of the court’s findings and conclusions,” we remanded the case for reconsideration. Stringer v. United States, 301 A.3d 1218, 1220 (D.C. 2023) (Stringer I).

On remand, the trial court again declined to credit Mr. Charles’s testimony, relying on “two critical facts”: first, that Mr. Charles’s testimony could not be reconciled with the physical evidence at trial, and second, that a letter Mr. Stringer wrote to Mr. Charles when they were in jail supported the conclusion that the two men committed the crime together, undermining Mr. Charles’s claim that he acted alone and that Mr. Stringer was innocent.

Mr. Stringer now appeals a second time, asking us to reverse because the trial court’s credibility determination was premised on clear error and a failure to adequately scrutinize the evidence at trial. We agree that the trial court erred, but we conclude that another remand, rather than reversal, is appropriate. We cast no doubt on the principle that the ultimate responsibility to determine a witness’s credibility and whether an IPA movant is more likely than not actually innocent lies with the trial court. But we are compelled to again say that a trial court’s factual findings, including its credibility determinations, cannot rest on unsupported speculation or a manifest misunderstanding of the record evidence.

I. Background

The factual background prior to remand following Mr. Stringer’s first appeal is set forth in our opinion in Stringer I. We set forth here only the facts relevant to Mr. Stringer’s challenges to the court’s order following remand.

A. The Murder of Tilford Johnson

Tilford Johnson was murdered during the early morning hours of June 3, 2003.

Stringer I, 301 A.3d at 1220. His body was discovered later that morning in a car parked in an alley behind 28th Street in Southeast Washington, D.C. Id. When officers with the Metropolitan Police Department (MPD) arrived at the scene, they found “Mr. Johnson’s body slumped in the driver’s seat, with a single gunshot wound to the head.” Id. The doors to the car were locked, with the keys in the ignition and the parking brake engaged. Id. The rear driver’s side window was shattered. Id. Mr. Johnson’s empty wallet was in the car. Id. Officers found a “vast” amount of blood pooled on the rear seat and floor behind the driver’s seat. Id. at 1220-21. Blood was also on the outside of the car, with a trail of blood leading away from the car toward a property bordering the alley on Buena Vista Terrace, Southeast. Id.

Mr. Stringer and his nephew, Mr. Charles, were charged with Mr. Johnson’s murder and related offenses. Id. at 1221. They were tried separately in 2006. Id. The evidence against Mr. Stringer consisted of evidence implicating Mr. Charles (who had already been convicted but not sentenced); phone records showing calls between Mr. Charles and a number associated with Mr. Stringer close in time to the murder; testimony by Mr. Stringer’s half-brother (and Mr. Charles’s uncle) inculpating both men; and a letter Mr. Stringer sent Mr. Charles while both were incarcerated. Id. (citing Barry Stringer v. United States, No. 06-CF-1515, Mem. Op. & J. (D.C. July 20, 2009)).

The jury found Mr. Stringer guilty of first-degree murder while armed, second-degree murder while armed (as a lesser-included offense of first-degree premeditated murder while armed), armed robbery, and three counts of unlawful possession of a firearm during a crime of violence or dangerous offense. Id. at 1223.

The trial court sentenced him to 432 months of imprisonment. Id. This court affirmed Mr. Stringer’s convictions in 2009, rejecting his challenges to the admission of certain testimony, the sufficiency of the evidence, and the trial court’s aiding-and-abetting jury instruction. Id. We remanded for resentencing on merged offenses, and in January 2021, the trial court resentenced Mr. Stringer to the same aggregate term of 432 months of imprisonment. Id.

B. The Initial IPA Motion

In June 2014, prior to resentencing and while he was incarcerated at the same facility as Mr. Stringer, Mr. Charles executed an affidavit in which he confessed to killing Mr. Johnson and asserted that he acted alone in doing so. Stringer I, 301 A.3d at 1223. This affidavit was provided to Mr. Stringer’s counsel around July 2014, and, in December 2014, Mr. Stringer filed a motion under the IPA, D.C. Code § 22-4131 et seq., for vacatur of his convictions and dismissal of the charges or a new trial based primarily on Mr. Charles’s affidavit. Id. Mr. Stringer argued that the affidavit established by clear and convincing evidence that he was innocent of the robbery and murder of Mr. Johnson. Id.

The trial court held an evidentiary hearing at which the sole witness was Mr. Charles. Id. Mr. Charles testified to the details surrounding the murder of Mr. Johnson and maintained that Mr. Stringer was not involved with the murder. Id. at 1223-24. Acknowledging that he was coming forward fifteen years later, Mr. Charles explained that he was young and selfish at the time of the murder and the trials, and he had since realized that “[y]ou’ve got an innocent man standing over there and he didn’t have anything to do with this. Absolutely nothing.” Id. at 1224.

On cross-examination, Mr. Charles admitted that his affidavit and testimony were inconsistent with statements he made to law enforcement after the murder, his defense at his trial, and his testimony at Mr. Stringer’s trial. Id.

As the trial court noted, the outcome of Mr. Stringer’s IPA motion “hinge[d] entirely on the credibility of Mr. Charles,” and after personally observing him testify, the court “did not find Mr. Charles to be a credible witness.” Id. at 1227. Mr. Charles’s testimony at the hearing was inconsistent, the court determined, with (1) the physical evidence, (2) the telephone records, and (3) his prior testimony at Mr. Stringer’s trial as well as his written affidavit. Id. at 1225. The trial court also found that in addition to these inconsistencies, there were “other reasons to doubt” Mr. Charles’s testimony—Mr. Charles’s relationship with Mr. Stringer, the timing of Mr. Charles’s affidavit, and the fact that Mr. Charles had “nothing to lose by lying.” Id. at 1226. Ultimately the trial court concluded that, “even viewed in connection with Mr. Stringer’s arguments about the weaknesses of the government’s case” against him at trial, Mr. Charles’s testimony “failed to establish Mr. Stringer’s innocence either by clear and convincing evidence or by a preponderance of the evidence.” Id. (citation modified).

C. Stringer I

Mr. Stringer appealed, arguing that the trial court abused its discretion in three ways. First, the trial court relied on clearly erroneous findings that (a) Mr. Charles had “nothing to lose” when he confessed to killing Mr. Johnson alone, because when he executed his affidavit he had not yet been resentenced, and (b) Mr. Charles and Mr. Stringer saw each other in jail before Mr. Charles executed the affidavit, because no evidence supported that finding. Second, Mr. Stringer contended that the trial court relied on faulty assumptions in finding inconsistencies between Mr. Charles’s testimony and the evidence at trial or his written affidavit. Third, Mr. Stringer claimed that the court abused its discretion by failing to meaningfully consider the weaknesses in the government’s case against him at trial. Id. at 1227.

We vacated the trial court’s order and remanded for further proceedings. We explained that because “[t]his is a case where, if Mr. Charles is believed, Mr. Stringer is innocent[,]” “why the trial court did not believe Mr. Charles” was critical. Id. at 1229. The trial court, rather than referring to aspects of Mr. Charles’s demeanor or behavior on the stand as “call[ing] into question his trustworthiness,” based its credibility assessment solely on the comparisons it drew between Mr. Charles’s hearing testimony and certain objective circumstances, including the physical evidence, his written affidavit and prior trial testimony, and Mr. Charles’s potential motivations for coming forward. Id. Expressing “significant reservations” about the trial court’s bases for finding Mr. Charles not credible, id. at 1227, we held that remand was required because we could not be confident that, had the court’s decision not been influenced by certain dubious considerations, the court would have reached the same conclusion about the likelihood that Mr. Stringer is “actually innocent of the crime.” Id. at 1229-32 (quoting Caston v. United States, 146 A.3d 1082, 1084 (D.C. 2016)).

The Physical Evidence and Phone Records. In its original order, the trial court noted that Mr. Charles’s testimony was not consistent with the testimony of law enforcement witnesses regarding the crime scene. Id. at 1229. These inconsistencies included whether the bullet that killed Mr. Johnson only left “a hole in the car window” or “shattered” it; whether the car doors were left locked or unlocked; and whether Mr. Charles had handled and taken money from Mr. Johnson’s wallet or left it untouched. Id. at 1225. We held, however, that the significance of these inconsistencies vis-à-vis Mr. Stringer’s involvement was unclear. Id. at 1229. We reasoned that if Mr. Charles’s account of the scene was wrong, that either suggested that he was not present for the murder at all (a theory that was never advanced) or indicated that his memory of a fifteen-year-old event was flawed. Id. (citing Caston, 146 A.3d at 1096-98 & 1097 n.38). We observed that, had Mr. Charles committed or been present for the murder, we could not see how his mistakes regarding minute details bore on whether Mr. Stringer was with him or not. Id.

We also disagreed with the trial court that Mr. Charles’s version of events was “significantly inconsistent” with the phone records. Id. at 1230. At a minimum, Mr. Charles’s account—that the phone he called after the murder belonged to an individual named Carlos Sly, not Mr. Stringer—was plausible, and we credited the fact that Mr. Charles’s hearing testimony as to whom he spoke with on the night of the murder was consistent with his testimony at Mr. Stringer’s trial that he had talked with Mr. Sly. Id. We recognized that it was reasonable for the trial court to assume that if Mr. Charles had a “partner in committing the murder,” he would have spoken to him in the days after the murder, but we held that it was circular for the trial court to assume that there was a partner and that it was Mr. Stringer simply because Mr. Charles spoke to someone numerous times after the murder. Id.

Inconsistencies Between Mr. Charles’s Affidavit and Prior Testimony. The trial court also had “trouble crediting” Mr. Charles’s testimony at Mr. Stringer’s IPA hearing because it was inconsistent with both his 2014 affidavit and his testimony in 2006 at Mr. Stringer’s trial. Disagreeing with the trial court, we found some of these discrepancies to be minor and others to be “overstate[d],” reasoning that the “tension” in Mr. Charles’s accounts was “more sensibly attributed to sloppiness or memory lapses than to a lack of credibility as to Mr. Stringer’s involvement.” Id. at 1231.

The Timing of Mr. Charles’s Affidavit. The trial court pointed to the timing of Mr. Charles’s affidavit, accepting that Mr. Charles signed it in June 2014 and finding it notable that this was when Mr. Charles and Mr. Stringer were “being held together in the D.C. Jail awaiting resentencing.” Id. at 1231-32. The court stated that “Mr. Charles testified that this was the first time he and Mr. Stringer had seen each other since their respective convictions in 2006” and observed that it was “likely” that the two “crafted the plan for the affidavit together when they met in jail in 2014.” Id. at 1232. We disagreed, explaining that the record “does not reflect Mr. Charles testifying that he saw or met Mr. Stringer in jail in 2014” and, indeed, that there was no evidence that Mr. Charles and Mr. Stringer were even in the same unit or area of the jail during that time. Id. Although we acknowledged that it was “certainly possible” that the two saw each other in the D.C. jail in 2014, we held that “to question Mr. Charles’s credibility on that basis require[d] . . . some affirmative basis to find or infer that they did.” Id.

Mr. Charles Having “Nothing to Lose.” The trial court concluded that Mr. Charles had “nothing to lose by lying in this case” because, at the time of the hearing, his conviction was final and he had been resentenced. Id. According to the court, that did not mean that Mr. Charles was lying, but it “also [did] not offer any reason to conclude he [was] telling the truth.” Id. We agreed with the trial court that, at the time of the hearing, Mr. Charles faced no further risk by confessing to having committed the crime alone. Id. We noted, however, that whether Mr. Charles faced a risk at the time of the hearing did not entirely resolve the “nothing to lose” question if, as the trial court appeared to accept, Mr. Charles signed the affidavit in June 2014. Id. Whether Mr. Charles had anything to lose depended on what exposure, if any, he had—or believed he had—when he executed the affidavit in June 2014 ahead of his resentencing that October. Id. Potential Weaknesses in the Government’s Case. In Stringer I, we concluded by impressing upon the trial court that it was “‘incumbent on the court’ to ‘consider the potential weaknesses in the government’s case’ to a greater extent than it did.” Id. at 1233 (quoting Caston, 146 A.3d at 1099) (citation modified). In its initial order, the trial court alluded to these weaknesses only briefly, concluding that Mr. Charles’s testimony “even viewed in connection with Mr. Stringer’s arguments about the weaknesses of the government’s case . . . fail[ed] to establish Mr. Stringer’s innocence either by clear and convincing evidence or by a preponderance of the evidence.” See id. at 1226. Recognizing that the IPA motion judge did not preside over Mr. Stringer’s (or Mr. Charles’s) trial and was thus at a disadvantage when it came to assessing the weight of the trial evidence, we nonetheless held that the court’s credibility determination could not “occur in a vacuum” and required a closer examination of the strength of the trial evidence against Mr. Stringer. Id. at 1233.

D. Remand

On remand, the trial court again denied Mr. Stringer’s IPA motion without holding another hearing or requesting supplemental briefing. In accordance with our instruction, the trial court “reconsidered Mr. Charles’s testimony” and set aside several of the bases on which it originally found him not credible, including “the placement of [Mr. Johnson’s] wallet; the telephone records; inconsistencies between his affidavit and his hearing testimony; the timing of his affidavit; and the idea that Mr. Charles had nothing to lose by testifying as he did.” The trial court nonetheless reached the same conclusion: that it could not “find Mr. Stringer innocent, either by a preponderance or by clear and convincing evidence.” In doing so, the court based its credibility finding on two “critical facts”: first, that Mr. Charles’s testimony that he was alone when he shot Mr. Johnson could not be reconciled with the blood evidence presented at trial, and second, that Mr. Charles’s testimony that Mr. Stringer was not involved was undermined by a letter Mr. Stringer sent to Mr. Charles when the two were incarcerated together, in which Mr. Stringer acknowledged that the two of them had “fucked up” and advised Mr. Charles to “live by the code.”

With respect to the physical evidence, the trial court concluded that “the only way to square Mr. Charles’s testimony with the physical evidence is for someone else to have been in the car at the time of the shooting.” Summarizing the testimony of law enforcement officers at trial, the court noted that Mr. Johnson’s car contained a “combination of wet and dried blood.” On the front passenger side near the floorboard, “there was only a little bit of blood that had dried”; on the rear seat and floorboard, however, “there was a vast amount of blood that was still wet.” Officers also found blood on the exterior of the car and a “blood trail” that led to Buena Vista Terrace Southeast. To reconcile (a) the lack of blood on the front passenger seat with (b) Mr. Charles’s hearing testimony that he was sitting in the front passenger seat when he shot Mr. Johnson and the blood trail leading to Buena Vista Terrace, the court determined that the “logical conclusion” was that “there were two people in the car with Mr. Johnson; one on his right who did the shooting, and one in back who trailed the blood away from the car.” As a result, the court found that Mr. Charles’s account that he was alone in the car with Mr. Johnson “[did] not withstand scrutiny.”

The second “critical” fact that the trial court relied on was the 2005 letter from Mr. Stringer to Mr. Charles. Noting that its original analysis had gone “unquestioned” by this court in Stringer I, the court relied solely on that analysis, which centered on a view that the letter, written “[a]fter they were both arrested for the murder,” “not only reflect[ed] the close relationship” between the two but also “strongly suggest[ed] that they had ‘fucked up’ together and were [ ] facing the same consequences” as a result, cutting against the notion that Mr. Charles acted alone.

Following our instruction, the trial court also reviewed the strength of the evidence at Mr. Stringer’s trial and its observations of Mr. Charles’s demeanor. The court concluded that while there were weaknesses in the government’s case, and that those weaknesses, along with Mr. Charles’s new testimony, may have been enough “to preclude a finding of guilt beyond a reasonable doubt,” they were not enough to satisfy Mr. Stringer’s burden to show that he was actually innocent of the crime even under a preponderance-of-the-evidence standard.

Finally, acknowledging that its prior assessment of Mr. Charles’s credibility did not rest on any observations about his demeanor during the IPA hearing, the court explained that it had chosen to focus instead “on the objective evidence” due to the “inherently subjective nature” of demeanor considerations. The court nonetheless recalled that its “subjective impression” of Mr. Charles’s testimony at the IPA hearing was that he was “unconvincing.” Citing his “flat” affect and “expressionless demeanor,” the court reasoned that while “nothing in Mr. Charles’s demeanor clearly suggested he was lying, there also was nothing that persuaded the [c]ourt he was telling the truth rather than relating a story that would help clear his uncle’s name.” At the same time, the court reiterated that “judgments about demeanor are highly subjective,” and it observed only that “[t]o the extent that this factor does play into the [c]ourt’s determination,” “it contributes to the decision not to credit Mr. Charles.”

This appeal followed.

II. Legal Background and Standard of Review

“In relevant part, the IPA provides that at any time, a person convicted of a criminal offense in the Superior Court may move to vacate the conviction or to grant a new trial on the grounds of actual innocence based on new evidence.” Caston v. United States, 146 A.3d 1082, 1089 (D.C. 2016) (citation modified); see D.C. Code § 22-4135(a), (b). “The motion must ‘set forth specific, non-conclusory facts’ and must identify the specific new evidence, establish how it demonstrates the movant’s actual innocence, and establish why the evidence is ‘not cumulative or impeaching.’” Caston, 146 A.3d at 1089 (quoting D.C. Code § 22-4135(c)(1)-(3)).

As relevant here, “new evidence” is evidence that “[w]as not personally known and could not, in the exercise of reasonable diligence, have been personally known to the movant at the time of the trial or the plea proceeding.” D.C. Code § 22-4131(7)(A). “‘Actual innocence’ or ‘actually innocent’ means that the person did not commit the crime of which he or she was convicted.” Id. § 22-4131(1).

“In determining whether to grant relief, the court may consider any relevant evidence, but shall consider the following: (A) The new evidence; (B) How the new evidence demonstrates actual innocence; [and] (C) Why the new evidence is or is not cumulative or impeaching.” Id. § 22-4135(g)(1). The movant must include an affidavit with the motion stating, under penalty of perjury, that they are “actually innocent of the crime that is the subject of the motion, and that the new evidence was not deliberately withheld by the movant for purposes of strategic advantage.”

Id. at § 22-4135(d)(1). If, after considering these factors, “the court concludes that it is more likely than not that the movant is actually innocent of the crime, the court shall grant a new trial.” Id. § 22-4135(g)(2). And if the court “concludes by clear and convincing evidence that the movant is actually innocent of the crime, the court shall vacate the conviction and dismiss the relevant count with prejudice.” Id. § 22-4135(g)(3); see also Caston, 146 A.3d at 1089-90.

“We review the denial of a motion to vacate under the IPA for abuse of discretion, giving great deference to the trial court’s role as the trier of fact on the ultimate issue of ‘actual innocence’ under the IPA.” Williams v. United States, 187 A.3d 559, 562 (D.C. 2018) (citation modified). “Thus, we apply the clearly erroneous standard of review to the trial judge’s rejection of allegedly newly discovered evidence offered to prove ‘actual innocence.’” Id. at 562-63 (citation modified). “As such, the scope of our review is narrow on the question of whether that new evidence establishes appellant’s ‘actual innocence.’” Id. at 563. (citation modified).

The “ultimate responsibility to determine [a witness’s] credibility and whether [the] appellant is more likely than not actually innocent lies with the Superior Court judge.” Caston, 146 A.3d at 1099. As such, “the Superior Court judge’s factual findings anchored in credibility assessments derived from personal observations of the witnesses are beyond appellate reversal unless those factual findings are clearly erroneous.” Id. (citation modified) (quoting Hill v. United States, 664 A.2d 347, 353 n.10 (D.C. 1995)). While the trial court is “certainly entitled to take into account internal inconsistencies,” in assessing a witness’s credibility, the ultimate significance of those inconsistencies “is a determination of law, subject to appellate scrutiny.” Id. at 1095-96. “Minor inconsistencies and omissions will not support an adverse credibility determination.” Id. at 1096-97 (citation modified) (quoting Zhang v. Holder, 737 F.3d 501, 504 (8th Cir. 2013)).

Factual findings are “clearly erroneous when although there is evidence to support [them], the reviewing court on the entire evidence is left with the definite and firm conviction that a mistake has been committed,” Johnson v. United States, 232 A.3d 156, 167 (D.C. 2020) (quoting United States v. United States Gypsum Co., 333 U.S. 364, 394-95 (1948)), or they reflect “a manifest misunderstanding of the record evidence,” Ashrafi v. Fernandez, 193 A.3d 129, 133 (D.C. 2018). Given this, we “may well find clear error even in a finding purportedly based on a credibility determination.” Dorsey v. United States, 60 A.3d 1171, 1205 n.120 (D.C. 2013) (quoting Anderson v. City of Bessemer City, 470 U.S. 564, 575 (1985)).

III. Analysis

Mr. Stringer again appeals the denial of his motion under the Innocence Protection Act, arguing that the trial court clearly erred in discrediting Mr. Charles and in evaluating the evidence against Mr. Stringer at trial. With respect to the court’s credibility determination, he contends that it is undermined by a pair of clear errors: first, the trial court’s conclusions about the blood evidence rested on unsupported assumptions and a clear error of logic, and second, the court’s analysis of Mr. Stringer’s 2005 letter depended on a mistaken chronology, leading the court to infer a motive for the letter that is not supported by the facts. Mr. Stringer also asserts that the trial court again failed to thoroughly consider the weaknesses in the government’s case against him.

The government disagrees, arguing that the trial court did not clearly err in discrediting Mr. Charles or in evaluating the evidence against Mr. Stringer. The trial court’s credibility findings, the government emphasizes, were also firmly grounded in its observations of Mr. Charles’s demeanor at the hearing, to which we owe significant deference.

Because we again have significant reservations about the trial court’s bases for discrediting Mr. Charles, we again remand for reconsideration consistent with this opinion.

A. Inconsistencies Between Mr. Charles’s Testimony and the Evidence Presented at Trial

As a general matter, trial courts make witness credibility determinations on two bases: observations of the witness’s demeanor during testimony and any perceived inconsistencies in that testimony when compared with objective facts. See Anderson, 470 U.S. at 575 (explaining that “factors other than demeanor and inflection go into the decision whether or not to believe a witness,” including “[d]ocuments or objective evidence [that] may contradict the witness’[s] story” or an account that is “so internally inconsistent or implausible on its face that a reasonable factfinder would not credit it”). While we review all credibility determinations for clear error, our assessment of findings based on demeanor differs somewhat from how we view findings based on factual inconsistencies. See Stringer I, 301 A.3d at 1227-28.

As we explained in Stringer I, “we are particularly unlikely to find clear error with respect to credibility determinations based on the witness’s demeanor—if for no other reason than that we have no appraisal of the witness’s comportment to compare against the trial court’s.” Id. at 1228; see Turner v. United States, 116 A.3d 894, 927 (D.C. 2015) (“A credibility determination, made after the judge had the opportunity to hear the recanting witnesses’ live testimony and observe their demeanor, may be overturned only if it is wholly unsupported by the evidence.”) (citation modified); In re Temple, 629 A.2d 1203, 1208-09 (D.C. 1993) (“The factfinder who hears the evidence and sees the witnesses is in a better position to make such determinations, having the benefit of those critical first-hand observations of the witness’[s] demeanor or manner of testifying which are so important to assessing credibility.”); see also Anderson, 470 U.S. at 575 (“[W]hen a trial judge’s finding is based on his decision to credit the testimony of one of two or more witnesses, each of whom has told a coherent and facially plausible story that is not contradicted by extrinsic evidence, that finding, if not internally inconsistent, can virtually never be clear error.”).

Conversely, when “documents or objective evidence [ ] contradict the witness’s story; or the story itself [is] so internally inconsistent or implausible on its face that a reasonable factfinder would not credit it,” we “may well find clear error even in a finding purportedly based on a credibility determination.” Stringer I, 301 A.3d at 1228 (citation modified) (quoting Anderson, 470 U.S. at 575). Thus, where the trial court found a witness not credible on the ground that their story was inconsistent with objective facts, we would still subject that finding to clear-error review, but “we might be more likely to find clear error based on our own comparison between the witness’s version of events and the objective facts and our assessment of the significance of any inconsistencies.” Id. (citing Caston, 146 A.3d 22 at 1096 (“[T]he significance of inconsistencies between a witness’s pre-hearing and hearing statements is a determination of law, subject to appellate scrutiny.”)).

When assessing a credibility determination based on a witness’s inconsistent statements, courts must consider both the circumstances in which the inconsistent statements were made and the witness’s explanations for the apparent discrepancies.

Caston, 146 A.3d at 1096. Importantly, minor or trivial inconsistencies will not provide an “adequate basis for concluding that [a witness’s] testimony was not credible.” Id. at 1096-98.

On remand, the trial court based its credibility determination primarily on what it understood to be two “critical” factual inconsistencies between Mr. Charles’s hearing testimony and his prior affidavit and his testimony at Mr. Stringer’s trial.

The court touched briefly on its impressions of Mr. Charles’s demeanor, but it noted that, due to the “highly subjective” nature of these judgments, it “did not place great weight” on demeanor “in concluding that his testimony was not credible,” and stated only that “[t]o the extent” it was considering demeanor, it “contributed” to its determination. More on that below.

Because the court again relied on the inconsistencies it found between Mr. Charles’s testimony and the trial evidence, we review the court’s findings for clear error, but, as in Stringer I, “we do not find ourselves constrained on review by an inability to assess ‘factors that could only be ascertained after observing the witness testify.’” 301 A.3d at 1229 (quoting Caston, 146 A.3d at 1099).

1. The Blood Evidence

The first inconsistency the trial court relied on in its credibility determination was the discrepancy between Mr. Charles’s testimony and the testimony of law enforcement witnesses regarding the crime scene.

In its first order denying Mr. Stringer’s IPA motion, the trial court identified the following inconsistencies as relevant. Mr. Charles testified at the hearing that he shot Mr. Johnson while sitting in the passenger seat of the car, with one foot in the car and the other hanging out. According to Mr. Charles, he then got out of the car without locking the door, left the engine running, and walked directly back to his own car on 28th Street. At trial, MPD officers who had reported to the scene contradicted Mr. Charles, testifying that the car doors were locked and that there was a trail of blood leading away from the car toward Buena Vista Terrace, in the opposite direction from where Mr. Charles testified his car was waiting.

In Stringer I, we questioned the significance of these inconsistencies “vis-à- vis Mr. Stringer’s involvement,” observing that “[i]f Mr. Charles’s account of the scene was wrong,” it would suggest either that he was not present for the murder or that “his memory of a [fifteen]-year-old event was faulty,” and questioning “how his mistakes regarding details bear on whether Mr. Stringer was with him or not.” 301 A.3d at 1229. On remand, the trial court homed in on the blood evidence, explaining that the reason the above inconsistencies mattered was because “the only way to square Mr. Charles’s testimony with the physical evidence is for someone else to have been in the car at the time of the shooting.” To explain this conclusion, the trial court pointed to the blood evidence. At trial, the responding officers testified that when they reported to the scene, the car contained a combination of wet and dried blood. On the front passenger side near the floorboard, “there was only a little bit of blood that had dried”; “on the rear seat and on the rear floorboard, there was a vast amount of blood that was still wet.”

It was “theoretically possible,” the trial court reasoned, that Mr. Charles could have misremembered what he did after the shooting, and that instead of leaving the passenger seat immediately and heading to his car on 28th Street, he instead went into the building on Buena Vista Terrace. But the court then rejected this theory altogether, concluding that it was completely incompatible with the blood evidence.

According to the trial court, Mr. Charles could not have made the blood trail leading to Buena Vista Terrace if he had been sitting in the passenger seat as he claimed.

The absence of blood on his seat, taken together with the “vast amount” of blood on the rear seats and floorboard, led to only one “logical conclusion”: that “there were two people in the car with Mr. Johnson; one on his right who did the shooting, and one in back who trailed the blood away from the car.” The court thus concluded that Mr. Charles’s story—that he was alone with Mr. Johnson and that Mr. Stringer had nothing to do with the shooting—did not withstand scrutiny. According to the court, this was the “most important[ ]” reason for its finding that Mr. Charles was not credible.

The court’s analysis raises several concerns. First, the conclusion that the court drew from the blood evidence appears to be based on speculation. This theory was never argued at trial and was not raised in the court’s first order denying Mr. Stringer’s IPA motion. 1 We see nothing in the record that provides a foundation for the court’s conclusions, and, despite relying on this blood evidence as a new basis to question Mr. Charles’s credibility, the trial court did not seek any additional evidence, including expert testimony, on the significance of the blood patterns in the car.

Second, the trial court’s logic suffers from significant flaws. In its attempt to square the evidence of the blood trail with the blood evidence found in the car, the

The trial court raised this theory for the first time on remand and Mr. Stringer could not have been expected in his prior IPA appeal to challenge the trial court’s reasoning before it existed. We therefore reject the government’s argument that Mr. Stringer waived his challenge to this prong of the trial court’s reasoning. court treated blood on the seat as a proxy for blood on the seat’s occupant: no blood on the seat meant that there could be no blood on the occupant and thus no blood trail, and vice versa. It strikes us that the opposite conclusion is at least as, if not more, likely true: the reason for the absence of blood on the front passenger seat was that the blood from Mr. Johnson landed on the shooter, who was sitting there. See, e.g., State v. Fischer, 360 P.3d 105, 115 (Az. Ct. App. 2015), as amended (Oct. 8, 2015) (noting an uncontested expert determination that a “void, or absence of blood, on the chair located directly next to” the victim meant that the defendant was sitting in that chair “when the fatal shot was fired”), opinion vacated in part, 392 P.3d 488 (Az. 2017). And, if he were covered in blood, the person sitting in the front passenger seat would almost certainly have left a blood trail behind. This also provides a compelling explanation for the “vast amount” of blood in the rear seat that would be consistent with Mr. Charles’s claims: the blood covered the seat because no one was sitting there. It is of course possible that there was a second person in the car sitting in the back seat and that there was so much blood that both he and the seat got covered, accounting for the pool of blood left behind and the blood trail leading away from the car. This possibility, however, is far from the only logical conclusion.

Third, the trial court’s logic is in tension with the government’s theory at trial.

During closing argument, the prosecutor floated the following theory about the blood evidence to the jury: “Why do you think there’s not any blood on the passenger seat? . . . The blood that must have popped right out of his head should have been on [the front passenger] seat. But somebody was there and it covered them instead.

You will see the blood is all in the back seat.” The government insinuated that this was obvious: “You don’t need a blood spatter expert to tell you that.” Although the government did not explicitly tell the jury that the front seat passenger was the shooter, this exchange shows that the government understood that the absence of blood in the front passenger seat evidenced the presence of blood on the front passenger—and the government’s reference to the blood “all in the back seat” suggests that it thought that, conversely, no one was sitting there for the blood to cover.2 The trial court appears to have overlooked this trial theory in assessing the blood evidence, despite our instruction to “consider the potential weaknesses in the government’s case to a greater extent than it [previously] did.” Stringer I, 301 A.3d at 1233 (quoting Caston, 146 A.3d at 1099) (citation modified).

When we compare Mr. Charles’s version of events and the physical evidence, we agree that there are inconsistencies that could bear on Mr. Charles’s credibility. Id. at 1228. We are unpersuaded, however, that they are so significant as to entirely

As we noted in Stringer I, the bullet exited the car through the rear driver’s side window, which makes it unlikely that someone shot Mr. Johnson from a back seat toward the front driver’s seat. 301 A.3d at 1234. discredit Mr. Charles’s testimony because we disagree with the court’s conclusion that, based on the blood evidence, the only possible explanation for these inconsistencies is that Mr. Charles could not have acted alone and thus that he lied about Mr. Stringer’s lack of involvement in the murder. Id. Further, we are not confident that, “in light of the record viewed in its entirety,” the trial court’s conclusion is plausible. Anderson, 470 U.S. at 574. Not only is the court’s reasoning at odds with the government’s own theory at trial, but it is also unclear what, if anything, the analysis is based on beyond the court’s own speculation.

At the very least, the fact that the court reached the opposite conclusion about the blood evidence than the government did at trial suggests that any conclusion about it is not obvious and likely requires expert testimony. See Motorola Inc. v. Murray, 147 A.3d 751, 756-57 & n.8 (D.C. 2016) (explaining that expert testimony is appropriate where “the expert’s scientific, technical, or other specialized knowledge will help the trier of fact to understand the evidence or to determine a fact in issue” (quoting Fed. R. Evid. 702)); see also Diamond v. Davis, 680 A.2d 364, 379 (D.C. 1996) (explaining that if the area of knowledge “is one beyond the ken of the average juror (or judge), then the parties must produce expert evidence to assist them in understanding”). In our view, the trial court’s conclusion about the blood evidence lacks a foundation in the record that cannot be bridged without expert testimony and is thus insufficient to support its finding that Mr. Charles was not credible on this basis. See Stringer I, 301 A.3d at 1228-29; see also Anderson, 470 U.S. at 574.3

2. The Letter

The second critical fact that the trial court relied on was the letter from Mr. Stringer to Mr. Charles in June 2005. In that letter, Mr. Stringer told Mr. Charles that “we fucked up,” but “it won’t be long before we get home, all you have to do is ride this shit out.” Mr. Stringer urged Mr. Charles “to live by the code[:] see nothing, know nothing, hear nothing[.]” At trial, Mr. Stringer admitted to writing the letter, but testified that when he wrote “we fucked up” and that “all you have to do is ride this shit out,” he meant that he and his cellmate (who happened to be Mr. Charles’s cousin) “didn’t have no commissary money” and were “asking [Mr. Charles]” to “assist” them by sending items sold in the commissary like food, sodas, and chips.

In a head-spinning departure from its position at trial that expert testimony was not needed to analyze the blood spatter evidence, the government now asserts that this court should reject Mr. Stringer’s challenge to the trial court’s blood- evidence conclusion because his argument “requires expert testimony which Stringer has not provided.” Mr. Stringer could not have presented expert testimony on this issue on remand because the trial court did not hold another hearing on the IPA motion (or request supplemental briefing), let alone indicate before ruling that it was considering this theory.

He also testified that the portion of the letter directing Mr. Charles to “live by the code” was a warning “not to say anything about his legal work.”

The court relied on the analysis from its original order, which stated that the letter “not only reflects the close relationship between Mr. Charles and Mr. Stringer, but strongly suggests that they had ‘fucked up’ together and were now facing the same consequences” for the murder of Mr. Johnson. Opining that Mr. Stringer’s letter would not “make sense if, as Mr. Charles claims, Mr. Charles committed the murder alone and Mr. Stringer had nothing to do with it,” the court also rejected as incredible Mr. Stringer’s explanation at trial that the letter was “referring merely to commissary money and legal papers.”

Setting aside the content of the letter, the court appears to have believed that Mr. Stringer tried to send the letter to Mr. Charles “[a]fter they were both arrested for the murder.” That is incorrect. When Mr. Stringer wrote the letter in June 2005, he was in custody awaiting trial for a murder in a different case; he had not yet been charged with the murder of Mr. Johnson. The government did not charge Mr. Stringer until August 30, 2005. While this error was present in the trial court’s original analysis, we did not address it directly in Stringer I. We did, however, observe in the background section of the opinion that Mr. Stringer wrote the letter when he “was incarcerated on another charge, before he was charged with Mr. Johnson’s murder.” Stringer I, 301 A.3d at 1222. Our instructions on remand, moreover, were broadly constructed and to the extent that the trial court relied on the letter as an inconsistency supporting its finding that Mr. Charles was not credible, it was incumbent upon the court to reexamine the letter “in light of asserted weaknesses in the government’s case at trial and the evidence as a whole.” Id. at 1234.

Given this factual error, we turn to whether the trial court’s mistake as to the timing of the letter had an impact on its analysis of the letter’s content such that it was clear error. Mr. Stringer argues that the trial court’s entire analysis turns on this mistake because it permitted the trial court to perceive a motive that did not exist— namely, that Mr. Stringer was attempting to cover up his involvement in Mr. Johnson’s murder. The government does not contest the mistake in timing but suggests that it does not negate or contradict the court’s conclusions about the content of the letter.

As an initial matter, we agree with the trial court that it is difficult to square the language in the letter with Mr. Stringer’s explanation at trial “that he was referring merely to commissary money and legal papers.” We also agree, considering that Mr. Charles is Mr. Stringer’s nephew and that “he has known Mr. Stringer his entire life,” that the letter reflects the close relationship between the two men. We depart from the court’s reasoning, however, where the court concluded that the only possible “fuck[ ] up” that the letter could be referring to was a murder, let alone specifically Mr. Johnson’s murder.

According to Mr. Stringer, the fact that the letter was written roughly two years after the shooting, around nine months after Mr. Charles had been charged, and before Mr. Stringer was charged indicates that when he wrote it, he thought that he was unlikely to be charged. In support, Mr. Stringer points out that Mr. Charles’s trial counsel stated, in a hearing on a severance motion about a year after Mr. Stringer wrote his letter and nine months after he had been charged with Mr. Johnson’s murder, that he had “not anticipate[d]” that Mr. Stringer would be charged in the murder of Mr. Johnson. Although he would have received discovery about the government’s investigation of the shooting—including Brady disclosures about other suspects or participants—Mr. Charles’s trial counsel was nonetheless “very surprised that Mr. Stringer ended up [as a] co-defendant.” As Mr. Stringer points out, if Mr. Charles’s trial counsel did not think Mr. Stringer would be charged after reviewing the evidence, there is an argument that it would have been even less likely that Mr. Stringer would have feared otherwise.

Another fact at the time of the trial gives us pause as to the court’s conclusion.

At the time Mr. Stringer wrote the letter, he was in custody awaiting trial for murder in a different case, of which he would ultimately be acquitted in October 2005 and in which Mr. Charles’s father was a co-defendant. During the severance hearing, Mr. Stringer’s attorney clarified that the letter did not refer to the murder of Mr. Johnson but was instead written “in the context of him preparing for trial in the [other] murder case.” He argued that if the letter were introduced and the government told the jury that it was about Mr. Johnson’s murder, Mr. Stringer would be “put in a position where [he would] have to put forward evidence” of his other pending murder charge to “rebut” the government’s claim, which, Mr. Stringer’s attorney argued, would be “highly prejudicial.” Although the judge ultimately let the letter in, this argument strikes us as at least a possible explanation for both the letter itself and Mr. Stringer’s implausible explanation for it during his trial testimony. Based on the trial court’s conclusions, however, it is unclear whether the court was aware of, or considered, this other case in its analysis of the letter.

More generally, while we do not endeavor to interpret Mr. Stringer’s letter ourselves, we think it is difficult enough to interpret that it is far from the case that, as the trial court said, Mr. Stringer’s words do not “make sense” if “Mr. Charles committed the murder alone and Mr. Stringer had nothing to do with it.” Like the trial court, we are skeptical that Mr. Stringer’s explanation of the letter at trial as referring “merely to commissary money and legal papers” is plausible. But the letter covers a number of topics and contains language and references that could mean multiple things. While one possible reading of the letter is that Mr. Stringer was referencing the murder of Mr. Johnson and thus implicating himself in the shooting, to question Mr. Charles’s credibility on that basis requires, in our view, both a correct understanding of when the letter was written and a full accounting of the other plausible explanations, including the fact of Mr. Stringer’s other pending murder charge.

B. The Court’s Observations of Mr. Charles’s Demeanor

The government suggests that, even if we conclude that the trial court erred in basing its credibility determination on the inconsistencies between Mr. Charles’s testimony and the evidence at trial, we can still affirm the denial of Mr. Stringer’s IPA motion by relying on the trial court’s conclusions about Mr. Charles’s demeanor. We disagree. It is true that credibility findings based on first-hand observation warrant substantial deference, not only because the trial court had the opportunity to assess the witness’s demeanor but also because “[t]he trial judge’s major role is the determination of fact, and with experience in fulfilling that role comes expertise.” Anderson, 470 U.S. at 574; see Henderson v. United States, 276 A.3d 484, 489 (D.C. 2022) (“We also defer to trial judges because they are experienced fact finders and de novo review by an appellate court is unlikely to produce significantly more accurate factual determinations.”). Here, however, the extent to which the trial court’s assessment of Mr. Charles’s demeanor during the hearing influenced its overall credibility determination is unclear.

As an initial matter, it bears noting that the trial court omitted a demeanor assessment entirely in its initial order. See Stringer I, 301 A.3d at 1229 (noting the absence of any reference to Mr. Charles’s demeanor or behavior on the stand as impacting our analysis of why the trial court “did not believe Mr. Charles”). If demeanor had played an appreciable role in the trial court’s evaluation of Mr. Charles’s credibility, presumably the court would have mentioned it the first time around.

On remand, the court explained that it opted not to address Mr. Charles’s demeanor because, “mindful of the inherently subjective nature of observations based on demeanor, facial expressions, body language, and the like,” it instead decided to “focus on the objective evidence” related to his credibility. Then, evaluating Mr. Charles’s demeanor for the first time, the court found that while “nothing . . . clearly suggested he was lying, there also was nothing that persuaded the [c]ourt he was telling the truth rather than relating a story that would help clear his uncle’s name.”

The court described Mr. Charles’s affect as “flat” and noted that he “did not appear remorseful.” In concluding its analysis of his demeanor, the court again hedged, emphasizing that because “individuals react and show emotions in different ways,” his “matter-of-fact and expressionless demeanor” did not mean he was lying, and noting that these types of judgments are “highly subjective,” which is why the court “did not place great weight on Mr. Charles’s demeanor in concluding that his testimony was not credible.” As to whether these considerations impacted its analysis at all, the court said only that “[t]o the extent [demeanor] does play into the [c]ourt’s determination . . . it contributes to the decision not to credit Mr. Charles.”

In the final paragraph summarizing the bases for its credibility determination, the court reiterated the two “critical facts” discussed above but, notably, did not include any reference to its observations of Mr. Charles’s demeanor. The court further noted that the physical evidence issue was the “most important[ ]” factor and that the two critical facts were “sufficiently weighty . . . to defeat Mr. Stringer’s efforts to prove his innocence even by a preponderance of the evidence.” While we do not find clear error in the court’s observations about Mr. Charles’s comportment, we cannot conclude that the court found Mr. Charles’s demeanor sufficient by itself to deem him not credible, and so we decline to affirm on that basis.

In summary, we hold that the two factual bases for the trial court’s credibility determination—the blood evidence and the letter—require careful reconsideration.

With respect to the blood evidence, the court should first consider the alternative theory for the blood patterns, which is the very theory put forth by the government at trial. If the court still sees the blood evidence as weighing against Mr. Charles’s credibility, it must either explain why the government’s theory at trial supports that conclusion or request expert testimony to explore and support its own theory. With respect to the letter, the court must recognize that when Mr. Stringer wrote it he was not incarcerated for Mr. Johnson’s murder, and it must consider the fact and implications of Mr. Stringer’s other pending murder charge at the time the letter was written in determining whether the letter can be relied on as a basis to find Mr. Charles not credible. Finally, if the trial court wishes to say more about Mr. Charles’s demeanor, it may do so, recognizing that it has already observed that “nothing in Mr. Charles’s demeanor clearly suggested he was lying.”

IV. Conclusion

For the foregoing reasons, we conclude that a remand is necessary for the trial court to reconsider, in the full context of the record at trial, its analysis of the factual inconsistencies on which it relied in finding Mr. Charles not credible. Accordingly, the trial court’s order is vacated and the matter remanded for further proceedings consistent with this opinion.

So ordered.