This appeal concerns a criminal defendant’s Sixth Amendment right to the effective assistance of counsel. Appellant Richard Cosío asserts that he was convicted of multiple felony counts of child sexual abuse because his trial counsel failed to discover readily available evidence that would have impeached his accuser. A divided three-judge panel of this Court rejected appellant’s claim. We went en banc not only to reconsider the panel’s decision, but also to clarify the legal principles governing ineffectiveness claims based on counsel’s alleged investigative shortcomings. In this case, those principles make it necessary for us to vacate appellant’s convictions and grant him a new trial.
I.
A. Summary of the Proceedings
An eight-count indictment filed on February 3, 1998, charged appellant with three counts of first degree child sexual abuse, three counts of second degree child sexual abuse, and one count each of carnal knowledge and taking indecent liberties with a minor child. The alleged victim of these offenses, which occurred between January 1994 and November 1997, was appellant’s younger half-sister. The case was tried in *1111July 1998, and the jury found appellant guilty on all counts. The trial court sentenced him to imprisonment for 33 years to life. Appellant noted an appeal.
In March 2002, while his direct appeal was still pending, appellant moved pursuant to D.C.Code § 23-110 (2001) to set aside his conviction on grounds of ineffective assistance of counsel. After holding an evidentiary hearing, the trial court denied the motion. .Appellant’s appeal from that denial was consolidated with his direct appeal.
On July 8, 2004, with one judge dissenting, a division of this Court rejected appellant’s claims and affirmed his convictions in Cosio v. United States, 853 A.2d 166 (D.C.2004) (Cosio I). Several months later, on February 3, 2005, the full Court granted appellant’s petition for rehearing en banc and vacated the judgment of the division. Cosio v. United States, 867 A.2d 967 (D.C.2005). The Court limited its grant of rehearing to the issue of ineffective assistance of trial counsel, requested full briefing, and invited the participation of amicus curiae counsel. Following the April 25, 2005, en banc argument, the Court requested supplemental briefing addressing the applicability of the June 20, 2005 decision of the United States Supreme Court in Rompilla v. Beard, 545 U.S. 374, 125 S.Ct. 2456, 162 L.Ed.2d 360 (2005). The supplemental briefing was completed in August 2005.
B. The Trial
1. The Prosecution Case
The government’s case against appellant rested chiefly on the testimony of the alleged victim, appellant’s fifteen-year-old half-sister, A.A. (his mother also was her mother). According to A.A., appellant came to the United States from Peru to live with her family a few years after her father died. A.A., who was a little girl at the time, imagined that her new big brother would be “an angel,” but she was disappointed. A.A. testified that appellant was mean and physically abusive to her and the other members of her family. “He would hit us and stuff’ with his hand or a belt, A.A. stated, and he even would “push[ ]” his mother around. A.A.’s testimony about appellant’s uncharged acts of violence was admitted to show why A.A. later failed to report appellant’s sexual abuse of her. Long after appellant stopped beating A.A., the prosecutor told the jury in his opening statement,
[A.A.] was still too scared of [appellant], too scared of what might happen. She had no idea what he might do. He had a short temper, he had beat her before, and she decided to just keep quiet about it.
A.A. testified that appellant began sexually molesting her in 1990, not long after he joined the household. She was only seven and he was eighteen years old. The first occasion, which took place at night in their small apartment while other family members were sleeping nearby, was very painful for A.A. As appellant held his hand over her mouth to prevent her from crying out, he fondled her and then put his penis in her vagina. It hurt “a lot,” A.A. testified, and she was crying during the ordeal. A.A. did not report what happened because, she said, she was “afraid” of appellant. “If I went to my mom,” A.A. explained, “I don’t think she will believe me, and then after she will say something to [appellant] and then after he will like hit me or something like that.”
Appellant continued to molest A.A. over the next several months, she testified, until she had the good fortune to leave home and go to a charitable boarding school in Virginia for disadvantaged children. The school was an eight hours’ drive from the *1112District of Columbia. For the next few years, A.A. remained at the boarding school, where she was happy and “tried not to think of’ what appellant had done to her. She told no one at the school about it because she still was “too afraid.” Eventually, A.A., said, she “forgot all about it.”
In 1993, however, a few months before her eleventh birthday, A.A. left the boarding school and returned home. The family then moved to a larger apartment, in which appellant had his own bedroom. At this time, A.A. testified, appellant resumed his sexual abuse of her, though he no longer beat her. Over the course of the next four years, according to A.A., appellant repeatedly compelled her to endure painful vaginal penetration and other acts of sexual molestation. He would call her into his room, and she would submit. During these incidents, A.A. testified, she cried and asked herself “why ... the bad things [are] happening to me,” but she told no one what she was going through because she was “too afraid” of appellant:
Q. Did you tell anybody that the sexual abuse had started again?
A. No, sir.
Q. Why not?
A. I was too afraid.
Q. Afraid of what?
A. Afraid of him [appellant].
Q. What were you afraid he might do?
A. I don’t know, go insane.
The turning point came in November 1997, when A.A. was in the ninth grade. As she explained, she had begun meeting with two volunteer tutors, young attorneys named Michael Ledesma and Roseanne Medina, who were husband and wife. Appellant was opposed to the arrangement. Objecting that A.A. was staying out too late, he would not let her leave the house in the evenings to attend the tutoring sessions. One afternoon, A.A. testified, she went to visit Ledesma at his home to discuss how she could continue to be tutored despite appellant’s interference. When they could not find a suitable time to reschedule their meetings, A.A. broke down and tearfully confided that “Richard [appellant] was bothering with me, and just bothering me, touching me and stuff.” This was the first time in over seven years that A.A. asserted to anyone that she was being sexually abused. When asked at trial why she “finally” told Ledesma, A.A. answered,
Because I didn’t like what was going on in my life. I just — I didn’t like anything that was going on in my life. I always wished my life would be like one of my friends or anything. I didn’t want it like that, I didn’t want it to come true and stuff.
She decided to tell Ledesma, A.A. added, because he and his. wife were lawyers, and she thought “they could try to find [her] help or something.”
The police were called, and A.A. was taken to Children’s Hospital. There she was examined by Dr. Beverly Lindsay, a pediatrician in the Division of Child Protection. Dr. Lindsay, who was qualified at trial as an expert in child sexual abuse, testified that A.A.’s hymen was “interrupted,” i.e., a portion of the hymen was missing. In Dr. Lindsay’s opinion, “something was introduced into her vagina that caused this complete transection of the hymen.” Dr. Lindsay considered the hymenal injury to be strongly indicative of sexual abuse; however, when asked whether she could “rule out” other causes, she responded that “[t]he only thing I can say is [that] something entered into her vagina that was large enough that would cause hyme-nal cleaves and tears.”
Dr. Lindsay further testified that child victims commonly do not report sexual abuse promptly because they have been *1113threatened. As a “very typical” example, Dr. Lindsay described the case of an HIV positive child who steadfastly refused to admit that anyone had molested her until her relatives caught the girl’s own father in the act. The girl had kept her father’s abuse secret because he had threatened to kill her if she told on him. “Many times children are threatened, their parents are threatened if they make a disclosure, and this is quite typical,” Dr. Lindsay concluded. She did not identify any other reasons why child sex abuse victims may keep silent about the abuse they endure.
The government’s third and only other witness at trial was James Swiney, the president of the boarding school that A.A. had attended. Mr. Swiney, who had been appointed A.A.’s legal guardian, confirmed her attendance at the school and recalled appellant’s visit there in June 1993 with his mother to pick A.A. up and take her home. Mr. Swiney also recollected that Michael Ledesma visited A.A. on three occasions, once with his wife, after she returned to the school in 1998.
In closing argument, the prosecutor emphasized A.A.’s fear of appellant and Dr. Lindsay’s “threats” testimony to explain why it took so long before A.A. “finally found the courage to tell somebody.” Addressing the period from 1990 to 1993 when A.A. was at boarding school, the prosecutor asked the jury,
Do you think she was wondering if she told it would get back to Mr. Cosio? Could she have any idea whether if he found out he could get his hands on her again?
As to the period from 1994 to 1997, when A.A. was living at home again, the prosecutor argued that
Mr. Cosio still ran the show.... She was still scared of him. Her mother still agreed with everything he said. She felt she wasn’t safe.... You can imagine the thoughts that were running through her mind, what will he do if he finds out? ... How could she have known what he would do to her if she had told?
Furthermore, the prosecutor reminded the jury,
Dr. Lindsay told you that it is common in her practice and in the practice of child sexual abuse specialists to encounter kids who do not reveal child sexual abuse until long after it has begun.... And she told you some of the reasons. Children often don’t feel safe. They are scared, particularly when the abuser is a family member. So you know [A.A.] is not alone in keeping this a family secret.
2. The Defense
In his opening statement, appellant’s trial counsel forecasted that he would discredit A.A. by showing that she was an “immature,” “attention-seeking,” “deeply troubled teenager” who had “fabricated” her “horrible allegations” in retaliation for appellant’s “constant attempts to correct and discipline” her and her siblings. Counsel explained that appellant had antagonized his younger brothers and sisters by “constantly prodding” them to do well in school, complete their homework, help with household chores, keep their curfews, behave themselves, and live up to his “high standards.” And “when his family would not hold up the high standards Mr. Cosio had,” counsel stated, “Mr. Cosio would sometimes discipline his younger siblings.” Counsel asserted that this unwanted exercise of quasi-parental authority gave rise to “a very deep friction” between appellant and his brothers and sisters. A.A., counsel said, was “easily influenced” by her other siblings’ “jealousy, anger and hatred” of appellant to bring false accusations of sex abuse against him. “Little did Mr. Cosio know,” counsel told the jury, that
*1114his constant attempts to correct and to discipline his younger siblings would result in these horrible allegations. You will learn about [A.A.], and about what a troubled teenager she was, and how she sought to seek [sic ] attention, ladies and gentlemen, how she was easily influenced by others in order to make this tangled web of deception and fabrication, ladies and gentlemen.
The theory outlined in the defense opening statement never became a reality in the course of trial. There simply was no evidence that A.A. was a troubled teenager; no evidence that she was motivated by spite and resentment over appellant’s role as an overbearing family disciplinarian; no evidence that A.A. was starved for attention; no evidence that A.A. had been influenced by jealous or angry siblings.1 There likewise was no evidence that appellant had “high standards,” that he was “constantly prodding” his brothers and sisters to study hard and improve their behavior, or that he meted out corrective discipline when he felt it necessary to do so.
Appellant himself did not take the stand, and his counsel called no family member, friend, or other witness to testify about A.A., her relationship with appellant, or the family dynamics. The only witness who testified at all to such matters was A.A. herself. As a witness, A.A. did not appear to be immature, troubled, attention-seeking, easily influenced, or resentful,2 and her substantive testimony did not support any of those characterizations. In her direct examination, A.A. stated that she did not “hate” appellant because she was guided by the Biblical injunction “to love your enemies and pray for those who persecute you.” On cross-examination, A.A. rejected defense counsel’s depiction of appellant as the well-meaning, if heavy-handed, older brother who “constantly” monitored his younger siblings’ behavior and disciplined them when he thought it necessary. When counsel inquired whether appellant had upbraided her for not doing well at school, for instance, A.A. responded that appellant did not know what her grades were. And A.A. flatly denied counsel’s suggestion that appellant hit her “to correct her”; rather, she answered, “he just hit me whenever he wanted to.”3 In short, appellant’s trial counsel had no substantial evidentiary support for his theory that A.A. had fabricated her charges of sexual abuse against appellant out of resentment. The prosecutor fairly asked the jury in closing if it had heard “one shred of evidence in anything she or any other witness said” that supported the characterization of A.A. in the defense opening statement.
Unable to demonstrate a strong motive for A.A. to lie about appellant, defense counsel sought to discredit her accusations by highlighting her prolonged failure to report that she had been molested. A.A. admitted that for over seven years, she had not said anything about being sexually *1115abused to her mother, her older sister (whom A.A. trusted, and who disliked appellant), her brothers, or any teacher or school counselor, even when she was away from home at boarding school. But while appellant’s counsel relied heavily on A.A.’s silence to impeach her, he never seriously challenged the “fear” explanation given for that silence. When he cross-examined A.A., counsel did not question her assertions that appellant had beaten her when she was little and that she was afraid of him. When he cross-examined Dr. Lindsay, counsel did not question her expert opinion that fear-induced silence was “typical” of child sex abuse cases. Nor did counsel present any evidence in the defense case to undermine A.A.’s testimony that she was fearful of appellant. As a result, though appellant’s counsel asked the jury in closing argument to disbelieve A.A.’s claims of beatings and intimidation, his exhortation was perfunctory and unpersuasive. In his entire closing, he devoted just two sentences to A.A.’s testimony that “she remained silent because she is afraid of Mr. Cosío,” as follows:
Does her story make sense about being hit for no reasons, ladies and gentlemen? It just simply doesn’t make sense, that she would be eight hours away and afraid of Mr. Cosío and that is her reason, ladies and gentlemen.
Appellant’s counsel had nothing else to say on the subject.
The defense did call three witnesses, but those witnesses said little in their brief appearances on the witness stand to undermine A.A.’s credibility. Jose Garcia and Nora Carnathan had been appellant’s co-workers and were called as character witnesses, solely to say that appellant was a good worker and a law-abiding citizen. Appellant’s counsel did not ask either witness about A.A. On cross-examination, both co-workers acknowledged that they did not know what appellant did in the privacy of his own home. The third defense witness, Dr. John Adams, was a forensic pathologist who opined that A.A.’s hymenal injuries were “nonspecific with regard to sexual abuse” because they could have been caused “by any kind of foreign object, be it a penis, be it a finger, be it any firm or rigid foreign object.” Dr. Adams’s testimony was consistent with the testimony of Dr. Lindsay.4
In his closing argument, appellant’s counsel shifted the theme of his defense. Instead of emphasizing the theory of A.A.’s motive to fabricate outlined in his opening, counsel focused on A.A.’s relationship with Ledesma (which he had not mentioned in his opening) and a few other, secondary points. Counsel contended that A.A. had lied about appellant to “assert her own will, to do what she wanted, .... to figure out a way to keep getting tutored.” Referring to Ledesma as a 29-year-old man whom A.A. had known for *1116only a few weeks, counsel argued that she had no good explanation for how their conversation, while they were “in his apartment alone,” went “from trying to get more tutoring to allegations of sex abuse.” “It shows,” counsel said, that A.A. “had something to hide.”5 Without saying so directly, counsel insinuated that A.A. had been infatuated with Ledesma and had falsely accused appellant of molesting her because appellant was preventing her from seeing the object of her desire. In addition to highlighting the supposed peculiarity of A.A.’s decision to confide in her tutor, counsel asked the jury to consider the apparent failure of A.A.’s other family members to notice the alleged abuse,6 the less than certain nature of the medical testimony, and the defense witnesses’ testimony of appellant’s good character.
In sum, the case against appellant turned on A.A.’s credibility. She proved to be an impressive witness. She emerged from her cross-examination with her credibility intact — the defense failed to impeach, contradict, or otherwise undermine her testimony in any significant way — and the medical evidence corroborated her claim that she had been sexually abused. At the end of the day, the jury believed A.A. and found appellant guilty on all counts.
C. The § 23-110 Motion
Represented by new counsel, appellant filed a motion for a new trial under D.C.Code § 23-110. The motion alleged that trial counsel had been constitutionally ineffective in failing to discover and present readily available testimony that would have discredited A.A. by rebutting her claim that she was intimidated by appellant. In affidavits and in person at the hearing on the motion, five of appellant’s former co-workers at a computer assembly company called International Data Products (“IDP”) — including the same Jose Garcia who was a character witness at appellant’s trial7 — testified that A.A. and appellant had had a “friendly,” “affectionate,” “comfortable,” “normal” brother-sister relationship. The witnesses said they had observed A.A. and appellant together on many occasions — at work, where A.A. would regularly visit with appellant after school and during the summer months when she was not in school, and outside of work, at picnics, a baby shower, a birthday party, and other events. On these occasions, the witnesses had seen A.A. hug appellant, kiss him on the cheek, put her arm around his shoulder, joke and laugh and play with him, seek out and enjoy his company, and never appear afraid or ill at ease with him.
Jose Garcia, for example, recalled that A.A. “would come by IDP frequently to visit Richard [appellant].” As he elaborated in his affidavit (the paragraph numbers are omitted):
I had an opportunity to observe their interaction and relationship on those occasions. I also saw Richard and [A.A.] outside the office many times during the period when Richard was working at IDP. We would all have dinner together or go on outings. I recall specifically that we all went to Kings Dominion [an *1117amusement park] together in or about 1996. The videotape attached to this affidavit was taken by me on that trip to Kings Dominion. It shows a the [sic] typical interaction between Richard and [A.A.] during that time period. This videotape has been in my possession since about 1996.
On all occasions when I observed [A.A.] and Richard together, she showed a very positive, friendly and relaxed attitude toward Richard. They would laugh together, and she would often want to go along with him when he went places. I never saw [A.A.] act the least bit anxious, withdrawn or uncomfortable around Richard. Her attitude around him was like that of any normal young girl who enjoys being with her older brother. I observed that their relationship was this way the entire time that Richard worked at IDP.
Shortly before Richard’s arrest in this case, I was with Richard and [A.A.] and observed that she had the same attitude toward him that I have described above. [A.A.] did not seem at all intimidated by Richard on that occasion.
Roxanna Fuster, the Director of Management Information Systems at IDP and appellant’s supervisor, testified that A.A. had asked her for permission (which she granted) to visit with appellant at the laboratory and offices where he was working. Ms. Fuster understood that A.A. “was bored at home” and wanted to “hang around with Richard.” Ms. Fuster stated that A.A. came to IDP after school, sometimes every day of the week. During the summer months, A.A. occasionally spent entire days at IDP with appellant. After traveling by bus from Washington, D.C., to IDP’s facility in Gaithersburg, Maryland, A.A. would arrive “and kiss [appellant] on the cheek,” Ms. Fuster recalled. She saw “nothing that would indicate ... that [A.A.] was afraid or that she was uncomfortable .... [S]he told me she wanted to be there.” Ms. Fuster observed A.A. and appellant together outside of work as well — at company picnics and a Christmas party, and once at her own home. A.A. and appellant would laugh and joke with each other, and they appeared to Ms. Fus-ter to have a “normal relationship” as brother and sister. Ms. Fuster perceived A.A. to be “a very happy girl.”
A third co-worker, Rafael Villatoro, also saw A.A. with appellant on many occasions at work and elsewhere. “[A.A.] got along very well with Richard and seemed very comfortable with him and affectionate toward him,” Villatoro reported, adding that
[s]he clearly enjoyed being with him and appeared to have a happy, positive relationship with him. In fact, she often wanted to join Richard when he would go places. At no point do [sic] she appear to be the slightest bit afraid or intimidated by Richard.... She did not show any signs of fear or nervousness toward him.
Villatoro remembered that A.A. had displayed physical affection for appellant at a company picnic by “go[ing] next to him and put[ting] her arm on his shoulders.” The other two co-workers, Erick and Henry Estrada, provided similar reports.
The co-workers also testified that appellant’s trial counsel or his investigator had interviewed them before trial but had not asked them about appellant’s relationship with A.A. The co-workers recalled being questioned only about appellant’s employment, his personal character and reputation, and his relationship with his mother. As a result, the co-workers had not shared what they knew of appellant’s interactions with A.A., though they would have been ready and willing to do so had they been asked.
*1118Although the government did not contest the substance of the co-workers’ affidavits and testimony, it argued in opposition to the § 23-110 motion that appellant’s counsel had conducted an adequate pretrial investigation and that the new evidence proffered by appellant would not have helped his defense. In support of this argument, the government submitted an affidavit from counsel himself.
Appellant’s counsel was an experienced trial attorney with the Public Defender Service. He averred that he had obtained information from appellant prior to trial about “persons who were familiar with him and with [A.A.],” and that he or his investigator had interviewed “[A.A.’s] mother, brothers, sister, friends, and school teachers, as well as Mr. Cosio’s co-workers and friends, and asked them what they knew about the relationship between [Cosío] and [A.A.].” “To the best of my recollection,” appellant’s counsel stated, “I learned from these persons, particularly the family members, that Mr. Cosio looked out for his younger siblings, and was a responsible and authoritative figure in his family.” Counsel believed that he had investigated the relationship between A.A. and appellant “as fully as possible given the general constraints of time and resources.”8 He and his investigator had interviewed several of appellant’s co-workers, and he had called “the persons whom [he] considered to be the most credible, and who would have come across the best during cross-examination to testify” as character witnesses. Counsel “focused at trial on eliciting character testimony from Mr. Garcia and Ms. Carnathan because they were Mr. Cosio’s co-workers and as such were familiar with his work ethic and personality.” Counsel “sought to convey through this testimony that defendant was a hard working person with a good steady job.”
After receiving the government’s response, the trial court set the § 23-110 motion down for an evidentiary hearing. The court observed that the affidavit of appellant’s trial counsel neither stated what he and his investigator had learned from appellant’s co-workers about appellant’s relationship with A.A., nor “explain[ed] why he chose not to corroborate the fabrication defense with evidence that, prior to her relationship with Michael Le-desma, [A.A.] appeared not to have any fear of [appellant].”
At the hearing, appellant’s trial counsel testified in person. He stated that his theory at trial to explain why A.A. had lied about appellant “was basically that my client, who was a parental figure in the family and would provide limitations and discipline to his siblings, including [A.A.], that [A.A.] basically wanted to get my client out of the picture to be able to do whatever she wanted.” Counsel had hoped that his investigation would uncover a “romantic relationship” between A.A. and Ledesma so that he could “hook that up then and say that she was making these allegations because she wanted [appellant] out of the way to be able to go into a relationship that she knew [appellant] didn’t approve of’ — but, he acknowledged, *1119no evidence of such a relationship ever was found.
With respect to the first unanswered question identified by the trial court, the government introduced an April 1998 memorandum from trial counsel’s files summarizing the interviews his investigator had conducted with several of appellant’s co-workers at IDP.9 The interviews did not focus on A.A.’s interactions with appellant. However, the memorandum, which was prepared three months before trial, did report the following comments about A.A.:
[The co-workers] all met [A.A.] and almost all of them thought that she was acting like a little girl. They told me that [A.A.] was not acting like her age. They told me that [A.A.] would come up to them and want some affection from them.
They don’t understand why [A.A.] would accuse Richard of doing this.
The co-workers also shared their impressions of appellant and his other family members. According to the defense investigator’s report, they described appellant as “the father figure” and “head of his household” who paid the bills and “would look out for his family.” They “all agreed that Richard was a normal person,” but they viewed his family as “abnormal.” His mother was overprotective and had openly declared in front of her children that she loved appellant the best. The co-workers disliked A.A.’s other siblings, opining that they “had fabricated this he out of jealousy” to get appellant “out of the picture so that [he] wouldn’t be able to boss them around.”
“[I]n a case like this,” appellant’s trial counsel testified at the hearing, “it’s almost ... assumed ... that I would investigate the relationship between the complainant and Mr. Cosio; it’s just ... basic, ... it is just implicit in this type of case that we’re going to do that.” However, counsel added, while the relationship between A.A. and appellant was investigated, it “wasn’t [his] focus;” rather, his “focus was on trying to definitely give a reason for her to fabricate things.” Counsel admitted that he never considered seeking information from appellant’s co-workers about A.A.’s relationship with appellant. When asked why not, he answered that he was “focused” on the co-workers serving as character witnesses.10 The court speeif-*1120ically inquired of appellant’s trial counsel whether it had occurred to him to ask Jose Garcia, whom counsel had selected to testify as a character witness, about Garcia’s observations of appellant and A.A. “No, Your Honor,” counsel responded, “I was focused on him being a character witness for Mr. Cosío.”
The cross-examination of appellant’s trial counsel centered on the second unanswered question raised by the trial court— why he had failed to present evidence countering A.A.’s asserted fear of appellant. Counsel acknowledged having received the government’s Notice of Intent to Introduce Evidence of Uncharged Misconduct four months before trial. The Notice had informed him that the government would introduce evidence that appellant inflicted physical beatings on A.A. when she was seven or eight years old, which “made her fear him and discouraged her from disclosing his acts of sexual abuse sooner.”11 The Notice also had stated that the government’s “need for this evidence is high in light of the importance juries tend to place on the absence of a contemporaneous report of rape.” Thus, appellant’s counsel knew well in advance of trial of A.A.’s claim that she did not report appellant’s sexual abuse for seven years because she was afraid of him. Nonetheless, although counsel conceded that “the fear issue [was] an important issue,” he had not tried to find witnesses who could testify about whether A.A. actually feared appellant, nor had he instructed his investigator to pursue that question in his witness interviews.
Appellant’s trial counsel was obliged to concede that he had not presented any testimony at trial to rebut A.A.’s claim of fear. Pressed further, he agreed that the genuineness of that claim, in conjunction with A.A.’s delay in charging appellant with sexually abusing her, went to A.A.’s credibility; “if it could have been proven, ... through witnesses, that [A.A. and appellant] had a cordial, friendly, open public relationship, ... that would have gone some distance in rebutting the picture of physical abuse and fear that [A.A.] had portrayed.” Such evidence, counsel acknowledged, “would not only throw into question the whole issue of her delay in disclosure, but also throw into question her credibility in that she had very forthrightly said she was afraid of him because of the beatings.”
The government did not present the testimony of the defense investigator or any other witness besides appellant’s trial counsel. In argument on the motion, the government conceded that the co-workers were not “lying when they said that [A.A.] didn’t appear afraid when she was with him.” The government also accepted the testimony of appellant’s trial counsel that he had not investigated A.A.’s claim of fear; that he had not “focused” on AA’s relationship with appellant; and that he had not asked appellant’s co-workers what they knew about that relationship.
The trial court denied appellant’s § 23-110 motion in a written opinion. The court found that appellant’s trial counsel had not considered using appellant’s co-workers to refute A.A.’s claim of fear because he had *1121not recognized the “fear factor” as “an important issue” at the time. In the court’s view, however, A.A.’s fear of appellant “was not critical to the government’s case or a central theme” at trial. In particular, the court stated, A.A.’s fear was “not a factor” in explaining her silence from 1994 through 1997.12 Therefore, the court concluded, trial counsel’s “judgment to pursue other lines of investigation and his choice of a different trial strategy [were] well within the range of effective representation.” And since A.A.’s fear of appellant “was not a significant issue in the case,” the court further concluded that the proffered testimony of appellant’s coworkers “would not have helped the defense” at trial in any event. “Being comfortable and affectionate with [appellant] in public was not inconsistent with a private sexual relationship of long duration,” the court stated. Moreover, the court added, “[h]ad the defense strategy been different, the prosecution strategy would have been different as well.” For example, “[t]he prosecutor could have used [the co-workers’] testimony about how much A.A. seemed to like her brother and seemed to enjoy spending time with him to argue that she had no motive to fabricate a story against him.” Or the prosecutor “might have offered expert testimony to explain how older children, who are otherwise truthful, may embellish the fear factor to avoid acknowledging their own sexual feelings and complicity despite the legal irrelevance of consent.”13 Accordingly, the court found “no reasonable probability” that presenting the co-workers’ testimony would have led to a different result at trial.
D. The Panel Decision on Appeal
In Cosio I, a division of this Court upheld the denial of appellant’s § 23-110 motion and affirmed his conviction by a 2-to-1 vote. Differing with the trial court, the panel majority recognized “that A.A.’s alleged fear of Cosio played a central role in the government’s case.” 853 A.2d at 170. Moreover, because appellant’s trial counsel was informed that the government intended to rely on A.A.’s fear of appellant, and that several of appellant’s co-workers had observed and interacted with A.A., the majority agreed that “it would have been logical for Cosio’s lawyer to have followed up that information by asking how much his fellow employees knew about his relationship with A.A.” Id. at 172. The observations of appellant’s co-workers “would have been the best evidence Cosio could have found for rebutting A.A.’s alleged fear of him.” Id. Hence, the majority *1122“assume[d]” that, in failing to discover and present the co-worker testimony, appellant’s trial counsel “did not, in fact, exercise the kind of conscious, professional judgment that a reviewing court, applying hindsight, should not second-guess.” Id. at 173.
Even so, the majority reasoned, trial counsel’s performance could not be held constitutionally deficient, because a hypothetical lawyer who actually had discovered the co-worker evidence could have made a reasonable tactical choice not to present it:
[SJuppose that trial counsel had, in fact, conducted a more thorough investigation of Cosio’s coworkers, obtained all the information presented at the § 23-110 hearing, and then consciously decided not to use it out of a concern that the “close relationship” testimony might undermine the fabrication defense based on sustained resentment. Strickland, we believe, would have required us to defer to that judgment. If that is so, then the fact that trial counsel presented a fabrication defense without actually coming to grips with the available, additional evidence on the fear issue offers no principled basis for revising that conclusion.
Id. at 174. Given, especially, what it perceived to be the difficulty of reconciling evidence of affection with the defense strategy of showing that A.A. deeply resented appellant, the majority also found no reasonable probability that the outcome of appellant’s trial would have been different had the jury heard the testimony of appellant’s co-workers. Id. at 174-75.14
II.
The basic principles governing our evaluation of Sixth Amendment ineffective assistance of counsel claims are well-settled. “The benchmark for judging any claim of ineffectiveness must be whether counsel’s conduct so undermined the proper functioning of the adversarial process that the trial cannot be relied on as having produced a just result.” Strickland v. Washington, 466 U.S. 668, 686, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984). An ineffective assistance claim therefore has two components:
First, the defendant must show that counsel’s performance was deficient. This requires showing that counsel made errors so serious that counsel was not functioning as the “counsel” guaranteed the defendant by the Sixth Amendment. Second, the defendant must show that the deficient performance prejudiced the defense. This requires showing that counsel’s errors were so serious as to deprive the defendant of a fair trial, a trial whose result is reliable.
Id. at 687, 104 S.Ct. 2052. In considering these two components, “the ultimate focus of inquiry must be on the fundamental fairness of the proceeding whose result is being challenged. In every case the court should be concerned with whether, despite the strong presumption of reliability, the *1123result of the particular proceeding is unreliable because of a breakdown in the adversarial process that our system counts on to produce just results.” Id. at 696,104 S.Ct. 2052.
“[B]oth the performance and prejudice components of the ineffectiveness inquiry are mixed questions of law and fact.” Id. at 698, 104 S.Ct. 2052. On appeal, we accept the trial court’s findings of fact unless they lack evidentiary support in the record. Byrd v. United States, 614 A.2d 25, 30 (D.C.1992). We review the trial court’s legal determinations de novo. Id.; see, e.g., United States v. Little, 851 A.2d 1280, 1287 n. 10 (D.C.2004) (“[W]e owe no deference to the trial court’s legal conclusion as to whether counsel was deficient.”); Chatmon v. United States, 801 A.2d 92, 110 (D.C.2002) (“[O]ur review of the prejudice prong is de novo”).
A. Trial Counsel’s Performance
“The proper measure of attorney performance remains simply reasonableness under prevailing professional norms.” Strickland, 466 U.S. at 688, 104 S.Ct. 2052. Accordingly, the defendant’s burden is to show that “counsel’s representation fell below an objective standard of reasonableness.” Id. “[C]ounsel is strongly presumed to have rendered adequate assistance and made all significant decisions in the exercise of reasonable professional judgment.” Id. at 690, 104 S.Ct. 2052. Recognizing that there may be “countless ways to provide effective assistance in any given case,” and that “it is all too easy for a court, examining counsel’s defense after it has proved unsuccessful, to conclude that a particular act or omission of counsel was unreasonable,” the Supreme Court has cautioned that “[j]udicial scrutiny of counsel’s performance must be highly deferential.” Id. at 689, 104 S.Ct. 2052. “A fair assessment of attorney performance requires that every effort be made to eliminate the distorting effects of hindsight, to reconstruct the circumstances of counsel’s challenged conduct, and to evaluate the conduct from counsel’s perspective at the time.” Id. “The court must then determine whether, in light of all the circumstances, the identified acts or omissions were outside the wide range of professionally competent assistance.” Id. at 690, 104 S.Ct. 2052.
In making that determination, “the court should keep in mind that counsel’s function ... is to make the adversarial testing process work in the particular case.” Id. The proper functioning of the adversary process demands appropriate investigation and preparation by counsel. See Monroe v. United States, 389 A.2d 811, 817 (D.C.1978). The presumptive deference that courts owe to fully informed decisions of counsel therefore is withheld from decisions that are inexcusably uninformed or under-informed. Under the Sixth Amendment guarantee, a criminal defendant is entitled to the benefits of counsel’s informed judgment and choice among reasonable alternatives; it is objectively unreasonable for defense counsel to make an uninformed decision about an important matter without justification for doing so. As the Supreme Court said in Strickland, “strategic choices made after thorough investigation of law and facts relevant to plausible options are virtually unchallengeable,” but “strategic choices made after less than complete investigation are reasonable precisely to the extent that reasonable professional judgments support the limitations on investigation.” Id. at 690-91, 104 S.Ct. 2052. Thus, “counsel has a duty to make reasonable investigations or to make a reasonable decision that makes particular investigations unnecessary.” Id. at 691, 104 S.Ct. 2052. “In any ineffectiveness case,” therefore, “a *1124particular decision not to investigate must be directly assessed for reasonableness in all the circumstances,” id., taking into account “not only the quantum of evidence already known to counsel, but also whether the known evidence would lead a reasonable attorney to investigate further.” Wiggins v. Smith, 539 U.S. 510, 527, 123 S.Ct. 2527, 156 L.Ed.2d 471 (2003).
The majority opinion in Cosío I took the position that a failure to investigate cannot be held to be deficient performance (or prejudicial, for that matter) unless any reasonably competent counsel necessarily would have presented at trial the evidence that the investigation would have discovered. In Wiggins, the Supreme Court made clear that no such stringent showing is required for either prong of Strickland. All that need be shown is a “reasonable probability” that a competent counsel would have utilized the undiscovered evidence; and strictly speaking, the likelihood that the evidence would have been used is part of the prejudice inquiry, not part of the performance evaluation.
The issue in Wiggins, as in this case, was whether defense counsel conducted insufficient investigation. The claim in Wiggins stemmed from “counsel’s decision to limit the scope of their investigation into potential mitigating evidence” for use at sentencing. 539 U.S. at 521, 123 S.Ct. 2527. Although counsel had conducted some investigation into mitigation — they had arranged for psychological testing and had reviewed the presentence investigation report and other social services records, see id. at 523-24, 123 S.Ct. 2527— they had failed to look further into the circumstances of the defendant’s life by commissioning a forensic social worker to prepare a social history report. Had they done so, counsel would have discovered “evidence of the severe physical and sexual abuse petitioner suffered at the hands of his mother and while in the care of a series of foster parents.” Id. at 516, 123 S.Ct. 2527. Reversing the lower courts, the Supreme Court held that counsel’s investigative performance was deficient and that the deficient performance was prejudicial — in other words, that counsel were constitutionally ineffective.
On the question of deficient investigative performance, the Court reiterated that “strategic choices made after less than complete investigation are reasonable precisely to the extent that reasonable professional judgments support the limitations on investigation,” and that “counsel has a duty to make reasonable investigations or to make a reasonable decision that makes particular investigations unnecessary.” Id. at 521, 123 S.Ct. 2527 (quoting Strickland, 466 U.S. at 690-91, 104 S.Ct. 2052). Applying those principles, the Wiggins Court explained that its “principal concern in deciding whether [counsel] exercised reasonable professional judgment is not whether counsel should have presented a mitigation case. Rather, we focus on whether the investigation supporting counsel’s decision not to introduce mitigating evidence of Wiggins’ background was itself reasonable.” Id. at 522-23, 123 S.Ct. 2527 (emphasis in the original; internal quotation marks and citations omitted).
Thus, in determining whether the investigation by Wiggins’ counsel was deficient, the Court did not consider whether counsel necessarily would or should have presented the undiscovered evidence at trial. Instead, the Court found the investigation deficient because (1) prevailing professional standards demanded a more thorough investigation into mitigation than counsel performed, id. at 524, 123 S.Ct. 2527; (2) the existing social services records that counsel did review contained leads that should have triggered further follow-up, *1125id. at 525, 123 S.Ct. 2527;15 (3) counsel had no countervailing grounds to conclude that further investigation would have been “counterproductive” or “fruitless,” id. at 525, 123 S.Ct. 2527; and (4) the “failure to investigate thoroughly resulted from inattention, not reasoned strategic judgment.” Id. at 526,123 S.Ct. 2527.
After concluding that counsel’s failure to discover the mitigating evidence of Wiggins’ history of deprivation and mistreatment was unreasonable, the Court turned to consider whether the deficient investigative performance prejudiced Wiggins’ defense. To establish prejudice, Wiggins’ burden was to show “a reasonable probability that, but for counsel’s unprofessional errors, the result of the proceeding would have been different.” 539 U.S. at 534, 123 S.Ct. 2527 (quoting Strickland, 466 U.S. at 694, 104 S.Ct. 2052). The core of the Court’s analysis makes unmistakably clear that Strickland does not demand a showing that counsel necessarily would have made use of the undiscovered evidence, but only a showing of a reasonable probability that counsel would have done so:
Given both the nature and the extent of the abuse petitioner suffered, we find there to be a reasonable probability that a competent attorney, aware of this history, would have introduced it at sentencing in an admissible form. While it may well have been strategically defensible upon a reasonably thorough investigation to focus on Wiggins’ direct responsibility for the murder, the two sentencing strategies are not necessarily mutually exclusive. Moreover, given the strength of the available evidence, a reasonable attorney may well have chosen to prioritize the mitigation case over the direct responsibility challenge, particularly given that Wiggins’ history contained little of the double edge we have found to justify limited investigations in other cases. [Citations omitted.]
The dissent nevertheless maintains that Wiggins’ counsel would not have altered their chosen strategy of focusing exclusively on Wiggins’ direct responsibility for the murder. [Citation] But as we have made clear, counsel were not in a position to make a reasonable strategic choice as to whether to focus on Wiggins’ direct responsibility, the sordid details of his life history, or both, because the investigation supporting their choice was unreasonable.
Id. at 535-36,123 S.Ct. 2527.16
Following Wiggins, we hold that in assessing the alleged shortcomings of the investigation performed by appellant’s trial counsel in the present case, the issue “is not whether counsel should have presented” at trial the evidence that ought to have been discovered. Id. at 523, 123 S.Ct. 2527. Rather, we must “focus on whether the investigation supporting counsel’s decision not to introduce [such] evi*1126dence ... was itself reasonable.” Id. (emphasis in the original). Similarly, the issue in evaluating counsel’s performance is not the reasonableness of the strategy counsel ultimately pursued at appellant’s trial, but “the reasonableness of the investigation said to support that strategy.” Id. at 527, 123 S.Ct. 2527.17 Deficient investigation cannot be excused on the ground that a competent attorney, aware of the evidence that an adequate investigation would have uncovered, could have made an informed judgment to pursue an alternative strategy and not utilize that evidence at trial.18 If we conclude that counsel’s investigation was unreasonable in its own right — after “eliminat[ing] the distorting effects of hindsight,” Strickland, 466 U.S. at 689, 104 S.Ct. 2052, “by pegging adequacy to ‘counsel’s perspective at the time’ investigative decisions are made,” Rompilla, 545 U.S. at 381, 125 S.Ct. 2456 (quoting Strickland, supra)—then the performance prong of an ineffectiveness claim under Strickland is satisfied.19
When we assess the reasonableness of counsel’s actions, we owe deference to counsel’s informed strategic choices. In the present case, however, such deference does not come into play. Appellant complains of his trial counsel’s failure to ask the readily available and apparently knowledgeable co-workers what they knew about A.A.’s relationship with appellant and whether she was afraid of him. Counsel “offered no strategic explanation for failing to pursue these avenues” of investigation. Poindexter v. Mitchell, 454 F.3d 564, 579 (6th Cir.2006). At the § 23-110 hearing, counsel agreed that A.A.’s relationship with appellant was a necessary subject of his investigation, and that A.A.’s alleged fear of appellant was an important issue. It simply did not occur to him to question the co-workers about those matters, counsel admitted, even after he had received his investigator’s report, because he “focused” on the co-workers only as possible character witnesses. The failure to make this inquiry of the co-workers thus “was not the sort of conscious, reasonably informed decision made by an attorney with an eye to bene-fitting his client that ... courts have denominated ‘strategic’ and been especially reluctant to disturb.” Pavel v. Hollins, 261 F.3d 210, 218 (2d Cir.2001). Rather, counsel’s investigative omission “resulted *1127from inattention, not reasoned strategic judgment.” Wiggins, 539 U.S. at 526, 123 S.Ct. 2527. Consequently, we owe no deference in the present case to counsel’s “judgment” as to the scope of his investigation; counsel made such no judgment.
That is only the beginning of our inquiry. Ultimately, “[t]he relevant question is not whether counsel’s choices were strategic, but whether they were reasonable.” Roe v. Flores-Ortega, 528 U.S. 470, 481, 120 S.Ct. 1029, 145 L.Ed.2d 985 (2000) (citation omitted). In the final analysis, our task under Strickland is an objective one. “Even where an attorney’s ignorance of relevant law and facts precludes a court from characterizing certain actions as strategic (and therefore presumptively reasonable), ... the pertinent question under the first prong of Strickland remains whether, after considering all the circumstances of the case, the attorney’s representation was objectively unreasonable.” Bullock v. Carver, 297 F.3d 1036, 1050-51 (10th Cir.2002) (citations omitted); accord Pavel, 261 F.3d at 217 n. 7.20 Thus, we must assess whether the challenged investigative omission was objectively unreasonable under the circumstances counsel confronted. If a competent defense attorney in trial counsel’s shoes reasonably could have decided to forego questioning appellant’s co-workers regarding their knowledge of his relationship with A.A., then trial counsel’s failure to conduct that inquiry as part of his investigation cannot be deemed constitutionally deficient performance. A reviewing court must take care, of course, not to slap the label of objective reasonableness on fanciful or unrealistic rationalizations for an attorney’s conduct.
We turn, therefore, to consider the objective reasonableness of the investigation carried out by appellant’s trial counsel under the circumstances he confronted in this case. The duty to conduct a reasonably thorough investigation “does not force defense lawyers to scour the globe on the off-chance something will turn up.” Rompilla, 545 U.S. at 383, 125 S.Ct. 2456 (citation omitted). However, defense counsel has a basic obligation to “conduct a prompt investigation of the circumstances of the case and explore all avenues leading to facts relevant to the merits of the case and the penalty in the event of conviction.” Pettiford v. United States, 700 A.2d 207, 216-17 (D.C.1997) (quoting American Bar Association Standards for Criminal Justice, The Defense Function 4-4.1(a) (3d ed.1993)). We take this as a succinct statement articulating the objective standard of reasonableness applicable to the duty to investigate — the standard of reasonableness that we must apply in this case. See Rompilla, 545 U.S. at 387, 125 S.Ct. 2456 (referring to the ABA Standards for guidance in determining what is reasonable investigation under prevailing norms of practice).
The following considerations persuade us that the failure of appellant’s trial counsel to ask appellant’s co-workers about his interactions with A.A. fell below the level of objectively reasonable performance. To begin with, in order to investigate the charges against his client, counsel unquestionably needed to investigate thoroughly the relationship between A.A. and appellant. As counsel himself testified at the § 23-110 hearing, “it’s just ... basic, it is just implicit in this type of case that [defense counsel is] going to do that.” At the most fundamental level, such investigation would be compulsory to *1128ascertain whether A.A.’s actual behavior with appellant was consistent or inconsistent with the charges of long-term abuse that she leveled against him, and whether anything in their relationship supplied AA. with a motive to fabricate those charges. See, e.g., Tucker v. Ozmint, 350 F.3d 433, 444 (4th Cir.2003) (“Trial counsel have an obligation to investigate possible methods for impeaching a prosecution witness, and failure to do so may constitute ineffective assistance of counsel.”) (citations omitted).21 The pretrial Notice from the government, alleging that appellant beat A.A. when she was young and that she delayed reporting his sexual abuse because she was afraid of him, only heightened the need for such investigation and sharpened its necessary focus. It would have been obvious to any competent defense attorney in trial counsel’s shoes that A.A.’s long failure to complain of abuse was a potential vulnerability that could be exploited if evidence were found to refute the government’s seemingly plausible explanation for that failure. Thus, we have no doubt that any competent defense attorney would have appreciated the need to investigate whether A.A. was afraid of appellant. Cf. Rompilla, 545 U.S. at 385-87, 125 S.Ct. 2456 (holding that when defense counsel is informed prior to trial that prosecution intends to introduce damaging evidence against defendant, counsel must make reasonable efforts to investigate that evidence); Bullock, 297 F.3d at 1050 (“clearly negligent treatment of a crucial deficiency in the prosecution’s case or an obvious strength of the defense will render an attorney’s overall performance defi-dent”) (internal quotation marks and citation omitted).
Appellant’s trial counsel conceded at the § 23-110 hearing that he did not undertake to investigate whether A.A. was afraid of appellant. Arguably, we could treat that concession as an admission that his investigation was deficient. Counsel did testify, though, that he investigated A.A.’s relationship with appellant generally, through inquiries of family members and friends. The scope of that general investigation is unclear, but perhaps it was broad enough to discover whether A.A.’s behavior was consistent with her claims. On the other hand, by all indications, the interviews of family and friends developed no materially helpful evidence regarding the relationship between A.A. and appellant — none, at least, that found its way into appellant’s trial or was revealed at the § 23-110 hearing. The real question before us, therefore, is whether counsel should have pursued the inquiry with appellant’s co-workers as well.
Ordinarily, defense counsel might not expect a defendant’s co-workers to be a source of information about the defendant’s intra-family relationships. At the outset of his representation of appellant, trial counsel might have had little reason to think that the co-workers could tell him anything useful about his client’s relationship with A.A. and whether she was afraid of him. Moreover, there is no evidence in the record that appellant told his trial counsel to look to his co-workers for such information.22
*1129“In assessing the reasonableness of an attorney’s investigation, however, a court must consider not only the quantum of evidence already known to counsel, but also whether the known evidence would lead a reasonable attorney to investigate further.” Wiggins, 539 U.S. at 527, 123 S.Ct. 2527. In our view, the report that appellant’s trial counsel received from his investigator months before trial should have led him to probe further what appellant’s co-workers knew about A.A.’s relationship with appellant. Counsel learned several things from his investigator of great potential interest. He learned that the co-workers knew A.A. and other members of appellant’s family. The co-workers perceived A.A. as an immature girl looking for affection outside her family. They did not “understand” why she was accusing appellant of abusing her. They viewed appellant as “the father figure” who “look[ed] out for his family.” They reported that appellant’s mother openly favored him over his siblings and told them so. They believed the siblings to be jealous of appellant. They suspected that the siblings had fabricated the accusations of abuse to get him out of the way so that he would not continue to “boss them around.”
This information indicated that the coworkers actually had observed A.A.’s relationship with appellant. More than that— it indicated that the co-workers might well be witnesses who could support trial counsel’s contemplated defense theory that A.A. was an immature, attention-seeking, troubled teenager, easily influenced by her jealous siblings, who had fabricated her accusations against appellant out of longstanding resentment over his efforts to discipline them and control their behavior. And trial counsel had no other helpful witnesses through whom he could present that central defense theme. Under those circumstances, any competent defense attorney in trial counsel’s position surely would have taken advantage of the opportunity to explore with the co-workers what they could say about appellant’s interactions with A.A. (and the rest of his family). Did A. A. (and her siblings) complain of any mistreatment, chafe under appellant’s direction, and appear to resent him? How did A.A. act around appellant? Did she try to avoid or evade him? Did she fear appellant and act as if he had mistreated her? How did appellant behave toward A.A. (and other family members)? Did appellant seek to discipline her for misbehavior, and, if so, how did she react? Did appellant ever mistreat A.A.? How did appellant, A.A. and the other siblings react when their mother told them she loved appellant the best? The investigator’s report would have triggered these and similar questions in the mind of any reasonable defense counsel and impelled counsel to discover what the co-workers would answer.
There is a second, related reason why appellant’s trial counsel should have investigated the co-workers’ knowledge of A.A. and her relationship with appellant. Counsel viewed the co-workers as potential character witnesses, and he actually called two of them to testify at trial that appellant was a hard-working and law-abiding person. In selecting and preparing these witnesses, counsel needed to anticipate that the introduction of character testimony would open the door to wide-ranging cross-examination. On cross-examination, *1130the prosecutor might have probed the witness’s familiarity with appellant’s behavior toward A.A., including any specific bad acts inconsistent with the positive character trait asserted (whether or not such acts had culminated in a criminal conviction). See Michelson v. United States, 335 U.S. 469, 479, 482, 69 S.Ct. 213, 93 L.Ed. 168 (1948); Rogers v. United States, 566 A.2d 69, 73 (D.C.1989); see also Kenneth S. Broun, McCormick on Evidence § 191 (6th ed.2006) (noting that “almost any accusation” would seem relevant to a general trait like being law-abiding). The prosecutor also could have explored whether the witness was biased against A.A. See Clayborne v. United States, 751 A.2d 956, 962 (D.C.2000). In order to make an informed decision about whether to call a particular co-worker as a character witness, it therefore was incumbent on counsel to inquire what, if anything, that co-worker knew about A.A. and appellant that might come out on cross-examination. If that inquiry had been performed, counsel would have discovered that A.A. and appellant appeared to enjoy a friendly relationship.23
The government argues that trial counsel’s failure to ask the co-workers about A.A.’s relationship with appellant cannot be deemed deficient because counsel “conducted a thorough investigation which included relationships among the family members, devised a compelling defense theory [i.e., the “resentment” theory of A.A.’s motive to fabricate her accusations], and made a strategic decision to focus his limited remaining time and resources on developing that theory rather than pursue additional investigation of family dynamics.” Brief of Appellee at 27. As we already have indicated, however, the record belies the government’s description of trial counsel’s performance in almost every particular. While counsel supposedly investigated “relationships among the family members” through interviews of family and friends, he apparently did not learn anything about how A.A. actually behaved with appellant; he admittedly did not focus on whether she was fearful of him. And while counsel settled, at some point, on the “resentment” theory to explain why A.A. would falsely accuse appellant, that theory hardly deserves to be called “compelling” in view of the absence of evidence to support it at trial.
In addition, trial counsel’s failure to ask appellant’s co-workers about A.A.’s interactions with him and whether she was afraid of him cannot be explained as “a strategic decision” to focus his limited remaining time and resources on developing the “resentment” theory “instead of’ continuing to investigate “family dynamics.” Counsel testified to no such thing — he made no “strategic decision” to forego inquiry of the co-workers; he had the time and resources to interview them and he did so; and nothing prevented him from asking them in those interviews whether A.A.’s actual behavior was consistent or inconsistent with her claims of fear and abuse. Moreover, the “resentment” theory depended on the family’s dynamics, so it makes no sense to suggest that he chose to focus on that theory by looking elsewhere.24 Indeed, as we have discussed, *1131the co-workers appeared to be a possible source of much-needed support for the “resentment” defense, so it is doubly surprising that counsel did not pursue the question of A.A.’s putative resentment of appellant with them after he received his investigator’s report.25
Trial counsel’s investigative omissions cannot be justified with the argument that impeaching A.A. with evidence that she liked appellant would have been inconsistent with the defense theory that A.A. resented appellant. Such a “post-hoc rationalization,” Wiggins, 539 U.S. at 526, 123 S.Ct. 2527, puts the cart before the horse. Counsel committed to the unpersuasive “resentment” theory prematurely, without having thoroughly investigated the relationship between A.A. and Cosio. His adoption of the “resentment” theory therefore was not the kind of “reasonable professional judgment[]” that could support the curtailment of further defense investigation. See Strickland, 466 U.S. at 690-91, 104 S.Ct. 2052. Abandoning an investigation “at an unreasonable juncture” makes a reasonable professional judgment impossible. Wiggins, 539 U.S. at 527-28, 123 S.Ct. 2527. The mere fact that counsel “had some information with respect to” appellant’s family relationships does not mean that he was yet in a position to make a reasonable strategic choice not to present an alternative defense based on impeaching A.A.; “Strickland does not establish that a cursory investigation automatically justifies a tactical decision with respect to ... strategy.” Id. at 527, 123 S.Ct. 2527. What appellant’s counsel knew “would [have led] a reasonable attorney to investigate further” before deciding what strategy to pursue. Id.
In sum, this is not a case
where counsel made a reasonable decision to cease further investigation as a result of having “discovered ... evidence ... to suggest that” challenging the prosecution’s [fear] evidence “would have been counterproductive, or that further investigation would have been fruitless.” Wiggins, 539 U.S. at 525, 123 S.Ct. 2527. Nor is this a case of “diligent counsel ... drawing] a line when [he has] good reason to think further investigation would be a waste [of time or resources].” Rompilla, [545 U.S. at 383, 125 S.Ct. 2456].
Gersten v. Senkowski, 426 F.3d 588, 610 (2d Cir.2005). Instead, this is a case in which, objectively speaking, trial counsel had strong reasons to ask appellant’s coworkers about his relationship with A.A., and no good reason not to do so. Counsel likewise had strong reasons to investigate the prosecution’s claim that A.A. was afraid of appellant, and no good reason not to do so. We are compelled to conclude that counsel’s investigative omissions were objectively unreasonable, and hence that counsel’s performance was constitutionally deficient.
We next consider the second prong of a Strickland analysis, which is whether the deficient performance prejudiced appellant’s defense.
B. Prejudice
In order to obtain relief on grounds of ineffective assistance of counsel, “[t]he defendant must show that there *1132is a reasonable probability that, but for counsel’s professional errors, the result of the proceeding would have been different.” Strickland, 466 U.S. at 694, 104 S.Ct. 2052. Where “a defendant challenges a conviction, the question is whether there is a reasonable probability that, absent the errors, the factfinder would have had a reasonable doubt respecting guilt.” Id. at 695, 104 S.Ct. 2052. “A reasonable probability is a probability sufficient to undermine confidence in the outcome.” Id. at 694, 104 S.Ct. 2052. Under this standard, “a defendant need not show that counsel’s deficient conduct more likely than not altered the outcome in the case.... The result of a proceeding can be rendered unreliable, and hence the proceeding itself unfair, even if the errors of counsel cannot be shown by a preponderance of the evidence to have determined the outcome.” Id. at 693-94, 104 S.Ct. 2052; see also United States v. Dominguez Benitez, 542 U.S. 74, 83, 124 S.Ct. 2333, 159 L.Ed.2d 157 (2004) (“The reasonable-probability standard is not the same as, and should not be confused with, a requirement that a defendant prove by a preponderance of the evidence that but for error things would have been different.”) (citation omitted).
In determining whether there exists a reasonable probability that the outcome of the trial would have been different had counsel’s performance not been deficient, we must consider the totality of the evidence adduced at trial. Strickland, 466 U.S. at 695, 104 S.Ct. 2052. “Some errors will have had a pervasive effect on the inferences to be drawn from the evidence, altering the entire evidentiary picture, and some will have had an isolated, trivial effect. Moreover, a verdict or conclusion only weakly supported by the record is more likely to have been affected by errors than one with overwhelming record support.” Id. at 695-96, 104 S.Ct. 2052.
As previously discussed, the Supreme Court’s decision in Wiggins elucidates the dual aspect of the prejudice inquiry required when the error in question is an investigative omission resulting in counsel’s failure to discover evidence favorable to the defense. We must inquire, first, whether there is “a reasonable probability that a competent attorney, aware of [the favorable evidence], would have introduced it at [trial] in an admissible form”; if so, then we must ask whether, “had the jury been confronted with this ... evidence, there is a reasonable probability that it would have returned with a different [verdict].” Wiggins, 539 U.S. at 535-36, 123 S.Ct. 2527.
In the present case, the evidence that appellant’s trial counsel failed to discover was potentially powerful impeachment of the key witness for the prosecution. Actions often do speak louder than words. Several credible witnesses who had extensive personal contact with A.A. and appellant over the relevant time period, from 1994 through 1997, could have testified from personal knowledge that A.A. actively sought to be with appellant, enjoyed his company, was relaxed and comfortable with him, and was overtly friendly and affectionate toward him. The witnesses saw no sign that A.A. was intimidated by appellant or ill at ease with him; on the contrary, they perceived a happy girl in a normal brother-sister relationship. This evidence of A.A.’s actual behavior would have been in stark contrast with A.A.’s testimony that, throughout the same time period, appellant was tormenting her with sexual abuse; that she was physically afraid of him; that she was unhappy and “always” wishing she had a normal life. By undercutting not merely A.A.’s “fear” explanation for not telling anyone that appellant was molesting her, but also the believability of her very assertion that he *1133was doing so, the co-workers’ testimony-impeached A.A.’s credibility in toto. Perhaps A.A.’s claims could have been reconciled with her observed behavior, but they are not reconciled on the existing record of the trial and the § 23-110 hearing.
What appellant lost in this case was not merely the opportunity to call his co-workers as defense witnesses, but also — and perhaps even more critically — the opportunity to confront A.A. herself on cross-examination with the contradiction between her words and her deeds. In her direct examination, A.A. portrayed herself as appellant’s miserable, frightened victim. She referred to appellant as her persecutor. A.A. surely would have been hard pressed to explain how her testimony squared with the fondness she had displayed toward appellant, the delight she had exhibited in his company, and the lengths to which she had gone in order to be with him. Whether A.A. would have denied the accounts given by appellants’ co-workers, claimed that appellant had coerced her into visiting him and showing him affection, offered some other rationalization of her conduct, or had no explanation at all, her credibility would have been under intense scrutiny and strain, in a way and to an extent that it never was at appellant’s actual trial.
We need not go so far as to say that any competent defense attorney necessarily would have presented the co-workers’ testimony and cross-examined A.A. about it, or even that it is more likely than not that any competent attorney would have done so. The standard is reasonable probability, and given the probative value of the coworkers’ testimony, there is at least a reasonable probability that any competent defense attorney would have recognized the attractiveness of using the co-workers’ evidence to impeach A.A. and cast doubt on her veracity.
It is true that the co-workers’ “affection” evidence would have been at odds with the “resentment” theory of A.A.’s motive for lying about appellant.26 The tension between proving affection and proving resentment would have been the only apparent reason to forego impeaching A.A. with the co-workers’ testimony. But it is not a persuasive reason. Establishing that A.A. had a deep-seated and longstanding resentment of appellant was not much of a defense. One would expect A.A. to have resented appellant profoundly if he had been molesting her as she said he had. For the resentment defense to have worked at all, there needed to be substantial evidence that A.A. had very strong reasons for developing and nursing an abiding hatred for appellant other than the fact that he had sexually abused her. Merely showing that A.A. resented appellant without proving such strong reasons would not have helped the defense in the slightest — it only would have corroborated A.A.’s claim that appellant had mistreated her. Merely identifying trivial reasons for A.A. to have been annoyed with appellant would have been equally unhelpful. Trial counsel’s suggestion that comparatively minor familial grievances motivated A.A. to falsely accuse her own brother of the most heinous crimes likely would have been unconvincing even if the evidence had supported it.
*1134As we have discussed above, appellant’s trial counsel was not able to establish compelling reasons for A.A. to have fabricated her allegations. Lacking evidence that A.A. had a powerful motive to lie about appellant, counsel had all the more need to impeach her with the co-workers’ evidence that she had shown affection for her alleged abuser and persecutor. Because counsel’s theory of A.A.’s motive to fabricate was so weak, there was little if anything to be lost by de-emphasizing or jettisoning that theory in favor of impeaching A.A. with the affection evidence. By the same token, the helpfulness to appellant’s defense of the co-workers’ testimony would not have been reduced merely because that testimony would have undermined a resentment theory that was not credible and had scant value to the defense in any event. In view of the relative strengths of the alternative “affection” and “resentment” approaches, we conclude that “a reasonable attorney may well have chosen to prioritize” the former. Wiggins, 539 U.S. at 535, 123 S.Ct. 2527.
We also conclude that there exists a reasonable probability that the jury would have returned a different verdict had it heard the co-workers’ testimony. The prosecution case against appellant was not weak, but neither was it overwhelming; it was a “bare-bones” case that rested almost entirely on A.A.’s credibility and that had to overcome the potentially exculpatory implications of A.A.’s prolonged failure to report being mistreated.27 The introduction of substantial evidence impeaching not only A.A.’s accusations but also her explanation for keeping silent about them28 might well have had “a pervasive effect ... altering the entire evidentiary picture.” Strickland, 466 U.S. at 695-96, 104 S.Ct. 2052. To quote the amicus brief, “[a] reasonable jury could have found it incredible that the accuser went out of her way to spend time with, and maintained a close, friendly relationship with, the brother who had terrorized and abused her.” Brief of Amicus Curiae at 17. Had A.A.’s credibility been impeached, it would have been a different trial, at least in the sense that very different questions would have been put to the jury. Had the jury developed a reasonable doubt about her credibility, there would have been a different verdict.
It is no answer to argue that A.A.’s credibility might have survived the impeachment more or less intact. That cannot be said with any level of confidence, inasmuch as the record does not reveal how A.A. would have responded to the testimony that she was a happy teenager who was at ease with appellant and affectionate toward him. We cannot discount the force of her impeachment by speculating how A.A. and the government might have been able to meet it; we must confine *1135ourselves to the record before us, which means to the totality of the evidence actually adduced at trial and in the § 23-110 proceeding. Cf. Wiggins, 539 U.S. at 536, 123 S.Ct. 2527. The government had an opportunity at the hearing on appellant’s post-conviction motion to rebut the force of the co-workers’ testimony. It did not do so. Perhaps, as the trial court suggested, the evidence would have allowed the prosecution to argue that A.A. would not have made up lies about a brother she loved; perhaps, as the government suggests, it could have presented admissible, persuasive expert testimony to neutralize the evidence that A.A. outwardly had appeared to like appellant.29 Such possibilities merely underscore how different the trial might have been. They do not alter our judgment that the jury might well have disbelieved or doubted A.A., in which case the jury necessarily would have entertained at least a reasonable doubt of appellant’s guilt. The probability is high enough that it undermines our confidence in the outcome of a trial in which appellant’s sole accuser was not significantly impeached at all. In that crucial respect, the proceeding was “unreliable because of a breakdown in the adversarial process that our system counts on to produce just results.” Strickland, 466 U.S. at 696, 104 S.Ct. 2052.
III.
Having shown deficient performance on the part of his trial counsel and consequent prejudice to his defense, appellant has established that he did not receive the effective assistance of counsel guaranteed by the Sixth Amendment. We therefore reverse appellant’s convictions and remand the case for a new trial.
So ordered.