1 2 3 4 5 6 7 8 UNITED STATES DISTRICT COURT 9 CENTRAL DISTRICT OF CALIFORNIA 10 ALAN WILLIAMS, Case No. 2:12-cv-08287-MCS-RAO 11 Petitioner, ORDER GRANTING PETITION FOR 12 WRIT OF HABEAS CORPUS 13 v.
14 FIDENCIO N. GUZMÁN,1 15 Respondent. 16
18 19 20 21 22 23 24 25
26 1 Fidencio N. Guzmán, the current warden of the facility in which Petitioner is 27 incarcerated, replaces Raymond Madden as the respondent to this action pursuant to 28 Federal Rule of Civil Procedure 25(d). 1 Petitioner Alan Williams petitions the Court for a writ of habeas corpus under 28 2 U.S.C. § 2254. After over a decade of litigation, one subclaim within Petitioner’s 3 second ground for relief remains: Petitioner asserts his trial counsel rendered 4 constitutionally defective assistance by failing to perform any investigation into victim- 5 witness Dayon Garrison and witness Christopher Ward. Upon review of a report and 6 recommendation by United States Magistrate Judge Rozella A. Oliver, pursuant to 28 7 U.S.C. § 636(b)(1)(C) and Federal Rule of Civil Procedure 72(b)(3), the Court received 8 further testimony from Mr. Garrison, Mr. Ward, Deputy District Attorney Frank 9 Santoro, and Detective John Duncan toward the prejudice element of the remaining 10 subclaim. For the reasons set forth below, the Court rejects Judge Oliver’s 11 recommendation as to the prejudice element of the claim, otherwise accepts and adopts 12 Judge Oliver’s report and recommendation, and grants the petition. 13 14 I. BACKGROUND 15 A Los Angeles County Superior Court jury convicted Petitioner of two counts of 16 premeditated attempted murder. (1st R. & R. 1, ECF No. 75.) In brief, the prosecution 17 asserted at trial that Petitioner shot Mr. Garrison and Khafra Akbar outside a barber 18 shop in Compton in 2006. (See generally id. at 4–10.) In 2008, the trial court sentenced 19 Petitioner to 80 years to life in state prison. (Id. at 1.) The state courts affirmed 20 Petitioner’s conviction on appeal and denied Petitioner’s subsequent habeas petitions. 21 (Id. at 2.) Petitioner initiated this § 2254 proceeding in 2012, asserting four grounds for 22 relief. (Pet., ECF No. 1.) On Judge Oliver’s recommendation, the Court dismissed most 23 of the grounds in 2020. (Order Accepting 1st R. & R. 1, ECF No. 87; 1st R. & R. 48.) 24 The Court also accepted Judge Oliver’s recommendation that she hold an 25 evidentiary hearing regarding a subclaim in Ground Two, in which Petitioner contends 26 his trial counsel provided ineffective assistance by failing to conduct any investigation 27 into Messrs. Garrison and Ward, both of whom maintain that Petitioner was not the 28 shooter. (See Order Accepting 1st R. & R. 1–2; 1st R. & R. 23–34, 48; Pet. 14–17.) 1 Judge Oliver heard evidence over three days in 2021 and 2022. (Mins., ECF Nos. 141– 2 42, 154.) 3 In 2025, Judge Oliver issued a further report and recommendation that the 4 remaining subclaim in Ground Two be denied and that the entire action be dismissed 5 with prejudice. (2d R. & R. 45, ECF No. 186.) Judge Oliver reasoned that Petitioner 6 had met his burden to show that trial counsel’s performance was not objectively 7 reasonable, but that Petitioner had not demonstrated prejudice from counsel’s failure to 8 perform any investigation into Messrs. Garrison and Ward. (Id. at 31–45.) Her analysis 9 of the prejudice element rested in part on adverse credibility findings as to Messrs. 10 Garrison and Ward with respect to their availability and willingness to testify. (Id. at 11 37–40.) On the same issue, Judge Oliver credited the testimony of DDA Santoro, the 12 prosecutor in Petitioner’s criminal trial, and Det. Duncan, the lead investigator of the 13 2006 shooting, who both provided testimony tending to show that they were unable to 14 locate and subpoena Messrs. Garrison and Ward for the criminal trial. (Id. at 39–40.) 15 Judge Oliver also opined that there is no reasonable likelihood the outcome of trial 16 would have been different had Messrs. Garrison and Ward testified, again resting her 17 position in part on adverse credibility findings as to Messrs. Garrison and Ward. (Id. at 18 40–45.) 19 Petitioner objected to Judge Oliver’s second report and recommendation, arguing 20 that Judge Oliver erred in her analysis of the prejudice element. (Pet.’s Objs. 3–26, ECF 21 No. 190.) Respondent Fidencio N. Guzmán did not respond to Petitioner’s objections 22 or file his own objections.2 In response to the objections and the oral argument of 23
24 2 The Court declines to review portions of the second report and recommendation to which no party lodged objections—including Judge Oliver’s findings and 25 recommendations on the deficient performance element of the claim. See United States 26 v. Reyna-Tapia, 328 F.3d 1114, 1121 (9th Cir. 2003) (“The statute makes it clear that the district judge must review the magistrate judge’s findings and recommendations de 27 novo if objection is made, but not otherwise.”). The Court renders no opinion on 28 whether Respondent preserved for appeal the positions he presented to Judge Oliver on 1 counsel on November 24, 2025, (Mins., ECF No. 192), the Court exercised its discretion 2 under 28 U.S.C. § 636(b)(1)(C) and Federal Rule of Civil Procedure 72(b)(3) to receive 3 further testimony, (Order Re: 2d R. & R., ECF No. 195). The Court held a further 4 evidentiary hearing on January 28, 2026. (Mins., ECF No. 211.) On February 9th, 2026, 5 the parties filed post-hearing briefs. (Pet.’s Br., ECF No. 215; Resp.’s Br., ECF No. 6 216.) 7 8 II. LEGAL STANDARDS 9 A. Habeas Corpus and Review of Report and Recommendation 10 District courts are authorized to “entertain an application for a writ of habeas 11 corpus in behalf of a person in custody pursuant to the judgment of a State court only 12 on the ground that he is in custody in violation of the Constitution or laws or treaties of 13 the United States.” 28 U.S.C. § 2254(a). In evaluating such an application, where a 14 federal court has determined that “the last reasoned state court decision was contrary to 15 or an unreasonable application of clearly established law,” as this Court did when it 16 accepted Judge Oliver’s 2020 report and recommendation, the court may “evaluate the 17 claim de novo, and . . . may consider evidence properly presented for the first time in 18 federal court.” Crittenden v. Chappell, 804 F.3d 998, 1010 (9th Cir. 2015) (internal 19 quotation marks omitted). 20 Pursuant to 28 U.S.C. § 636(b)(1)(C), on a magistrate judge’s report and 21 recommendation, the district judge must “make a de novo determination of those 22 portions of the report or specified proposed findings or recommendations to which 23 objection is made.” The judge “may accept, reject, or modify” the recommendation, 24 “receive further evidence[,] or recommit the matter to the magistrate judge with 25 instructions.” Id.3 Ninth Circuit authority suggests that a de novo evidentiary hearing 26
27 the deficient performance element. 28 3 Petitioner asserts that the Court cannot rely on evidence admitted at the hearings 1 before the district court should precede reversal of a magistrate judge’s credibility 2 findings after an evidentiary hearing. See, e.g., Johnson v. Finn, 665 F.3d 1063, 1075– 3 76 (9th Cir. 2011); United States v. Thoms, 684 F.3d 893, 895 (9th Cir. 2012).4 4 5 B. Ineffective Assistance of Counsel 6 “The Sixth Amendment guarantees criminal defendants the effective assistance 7 of counsel.” Yarborough v. Gentry, 540 U.S. 1, 5 (2003) (per curiam); accord Missouri 8 v. Frye, 566 U.S. 134, 138 (2012) (“The right to counsel is the right to effective 9 assistance of counsel.”). To succeed on an ineffective assistance of counsel claim, a 10 defendant must demonstrate: (1) counsel’s performance was deficient, and (2) the 11 deficient performance prejudiced the defense. See Strickland v. Washington, 466 U.S. 12 668, 687 (1984). The prejudice element “focuses on the question whether counsel’s 13 deficient performance renders the result of the trial unreliable or the proceeding 14 fundamentally unfair.” Williams v. Taylor, 529 U.S. 362, 393 n.17 (2000). A defendant 15 may establish prejudice by showing “there is a reasonable probability that, but for 16 counsel’s unprofessional errors, the result of the proceeding would have been different. 17
18 before Judge Oliver in its de novo evaluation. (Pet.’s Br. 4.) The sole authority he cites for this proposition is inapposite. Plummer v. W. Int’l Hotels Co., 656 F.2d 502, 505 19 (9th Cir. 1981) (“[T]he plaintiff has a right to introduce an EEOC probable cause 20 determination in a Title VII lawsuit . . . .”). Subject to the qualification in the next note, the Court has considered the exhibits and testimony from the 2021 and 2022 hearings 21 in rendering its decision. 22 4 The Court acknowledges these cases arose in situations in which a district judge reversed credibility findings favorable to a criminal defendant or habeas petitioner. 23 Petitioner persuasively argues the Court was not required to hold a hearing here, where 24 the Court weighed whether to reverse adverse credibility findings. (Pet.’s Objs. 7–8; Joint Statement 5 n.2, ECF No. 193.) In an abundance of caution and in the interest of 25 justice, the Court exercised its discretion under § 636(b)(1)(C) to conduct an evidentiary 26 hearing and make its own credibility determinations, which are based solely on the Court’s own perception of the witnesses at the 2026 evidentiary hearing. See United 27 States v. Mejia, 69 F.3d 309, 315 (9th Cir. 1995) (“There can be no doubt that seeing a 28 witness testify live assists the finder of fact in evaluating the witness’s credibility.”). 1 A reasonable probability is a probability sufficient to undermine confidence in the 2 outcome.” Id. at 391 (quoting Strickland, 466 U.S. at 694). “The likelihood of a different 3 result must be substantial, not just conceivable.” Harrington v. Richter, 562 U.S. 86, 4 112 (2011); accord Strickland, 466 U.S. at 693 (“It is not enough for the defendant to 5 show that the errors had some conceivable effect on the outcome of the proceeding.”). 6 “[W]ith respect to defective investigations, the test for prejudice is whether the 7 noninvestigated evidence was powerful enough to establish a probability that a 8 reasonable attorney would decide to present it and a probability that such presentation 9 might undermine the jury verdict.” Mickey v. Ayers, 606 F.3d 1223, 1236–37 (9th Cir. 10 2010). When evaluating the prejudice resulting from failure to present the testimony of 11 uninvestigated or uncalled witnesses, courts also evaluate whether the witness would 12 have been available and willing to testify at trial. E.g., Day v. Quarterman, 566 F.3d 13 527, 538–39 (5th Cir. 2009); Lawrence v. Armontrout, 900 F.2d 127, 130–31 (8th Cir. 14 1990); see also, e.g., Alcala v. Woodford, 334 F.3d 862, 872–73 (9th Cir. 2003); Tinsley 15 v. Borg, 895 F.2d 520, 532 (9th Cir. 1990); United States v. Harden, 846 F.2d 1229, 16 1231–32 (9th Cir. 1988).5 17 18 III. DISCUSSION 19 Having conducted a de novo review of the portions of Judge Oliver’s 2025 report 20 and recommendation to which objections have been stated, and having heard testimony 21 directly from the witnesses, the Court respectfully rejects Judge Oliver’s findings and 22 conclusions as to the prejudice element of Petitioner’s ineffective assistance claim. 23 Broadly, the Court finds that all witnesses testified credibly at the 2026 hearing, but the 24 testimony of Messrs. Garrison and Ward is entitled to significant weight and is 25 sufficient to carry Petitioner’s burden on the element. 26
27 5 For the purpose of argument, the Court notes but overrules Petitioner’s objection to 28 the application of this principle to his claim. (Pet.’s Br. 4 n.1; Joint Statement 5–6.) 1 A. Messrs. Garrison and Ward Were Available and Willing to Testify 2 The record demonstrates that Messrs. Garrison and Ward could have been located 3 and made to testify at Petitioner’s 2007 trial with minimal diligence. 4 Mr. Garrison unequivocally stated that he would have testified at the trial, as he 5 “kn[e]w it wasn’t [Petitioner] who [he] saw running off that day.”(Tr. 244, ECF No. 6 213; see id. at 243–45.) Mr. Garrison was on probation at the time of the shooting in 7 2006 and remained on probation through the time of Petitioner’s trial in 2007. (Id. at 8 214, 228; see id. at 46–47; Ex. 203-4 to -5.)6 Mr. Garrison was required to attend 9 monthly check-ins in with his probation officer at the Compton Courthouse, the same 10 courthouse in which Petitioner was tried. (Tr. 214–15; see Ex. N-009 (identifying trial 11 location as “SOUTH CENTRAL DISTRICT”).) Mr. Garrison had testified as a witness 12 at a different trial in 2006 after he was served a subpoena in the driveway of his home 13 address in Compton. (Tr. 215–16.) The same address appears in medical records related 14 to Mr. Garrison’s treatment for the gunshot wounds he sustained in 2006, which records 15 the prosecution produced to the defense before trial. (Ex. 208-2; see Tr. 133.) This 16 evidence tends to show that the prosecution or defense team easily could have located 17 Mr. Garrison at his home address in Compton, an address which had been disclosed in 18 discovery materials and where he previously had been subpoenaed. Further, the 19 prosecution or defense team could have found Mr. Garrison through his probation 20 officer. The Court credits Mr. Garrison’s testimony that he would have accepted an 21 invitation or obeyed a command to testify at Petitioner’s trial. The Court also credits 22 Mr. Garrison’s assertion that he would be willing to testify as he was one of the victims 23 of the crime being prosecuted. The Court reasons that a victim of a crime would prefer 24 25
26 6 The parties lodged exhibits identified and admitted at the 2021, 2022, and 2026 27 evidentiary hearings as attachments to ECF Nos. 221 and 223. The Court cites the 28 exhibit and page numbering scheme the parties affixed to the bottom of these exhibits. 1 justice to visit the true perpetrator and not a person who the victim knew with certainty 2 had no involvement in the incident. 3 Mr. Ward similarly averred that he would have testified at the 2007 trial had the 4 prosecution or defense team asked him to do so. (Tr. 177–78.) He represented that “at 5 the time [he] wouldn’t want an innocent man . . . to lose his life for something he did 6 not do.” (Id. at 149.) Mr. Ward was enrolled in high school and was skipping class at 7 the time of the shooting. (Id. at 141.) Mr. Ward was on probation at the time of 8 Petitioner’s trial, and he had court dates at the Compton Courthouse at the same time as 9 Petitioner’s trial. (Tr. 149–50; compare, e.g., Ex. 201-3 (memorializing Mr. Ward’s 10 preliminary hearing on Sept. 12, 2007), with Ex. N-018 (memorializing Petitioner’s jury 11 trial in progress on Sept. 12, 2007).) Mr. Ward acknowledged that law enforcement 12 attempted to contact him at his grandmother’s house, where he resided from his youth 13 until well after the trial. (Tr. 147–48.) Like Mr. Garrison, Mr. Ward easily could have 14 been contacted and located using readily available public information. The fact that law 15 enforcement made contact with him at his residence supports this conclusion. The Court 16 also credits Mr. Ward’s representation that he would be motivated to testify to ensure 17 that an innocent person was not convicted for a crime he did not commit. 18 Relevant to the Court’s credibility analysis is that Messrs. Garrison and Ward 19 submit they would have given exculpatory, rather than inculpatory, testimony had they 20 been summoned to testify at Petitioner’s trial. The record supports an inference that 21 Messrs. Garrison and Ward would evade efforts to command their testimony in support 22 of a conviction. For example, Mr. Garrison averred that he withheld any description of 23 the shooter when he spoke with law enforcement officers at the hospital after the 24 shooting. (Tr. 202.) Mr. Garrison explained, “Back then, that was just like the mentality 25 I had. That’s how I just grew up. You know, it wasn’t a lie. I just withheld that part.” 26 (Id.) Mr. Garrison conceded that he did not ask any follow-up questions when an officer 27 told him that law enforcement already knew who shot him. (Id.) However, a crucial fact 28 of which Mr. Garrison was unaware was that law enforcement had identified Petitioner 1 as the shooter. Had law enforcement, the prosecution team, or (as is relevant to the 2 Strickland inquiry) the defense team informed Mr. Garrison that Petitioner was being 3 prosecuted, Mr. Garrison readily would have provided a negative identification: 4 THE COURT: . . . [L]et’s say the police in the hospital tell 5 you they know who did it and it was definitely Williams. 6 Would you have told them that it wasn’t? 7 [MR. GARRISON7]: Sure. Yes. I would have said for sure 8 then. 9 (Tr. 245–46.) Mr. Ward testified that he told officers he was asleep at the time of the 10 incident “[b]ecause [he] was on probation, [he] was scared, [he] didn’t want to get 11 involved in it, [and he] didn’t want to go back to jail if [he] lied,” (id. at 170), but that 12 he would have told the defense team what he had seen if they had approached him 13 “because they’re not the police,” (id. at 176; see also id. at 149), and that he would have 14 testified on behalf of the prosecution if asked, (id. at 178). The Court acknowledges that 15 Messrs. Garrison and Ward likely would have been reticent to testify at the trial of an 16 individual who they sincerely believed shot or might have shot Messrs. Garrison and 17 Akbar. Providing a description of the shooter, asking questions of law enforcement 18 about potential suspects, and admitting to being an eyewitness carried a risk that Messrs. 19 Garrison and Ward would be swept into the prosecution’s case against the actual 20 shooter, whoever it was. But the dispositive question here is not whether Messrs. 21 Garrison and Ward would have testified at any trial concerning the shooting, but 22 whether they would have testified at Petitioner’s trial. Both testified that they did not 23 know before or during the trial that the prosecution had identified Petitioner as the 24 shooter. (Tr. 177, 206.) As detailed below, both witnesses unequivocally maintain that 25 Petitioner was not the shooter. Had defense counsel—or anyone else—informed them 26 that Petitioner was on trial for the 2006 shooting, they likely would have participated in 27
28 7 The transcript mistakenly identifies the Court as the responding to its own question. 1 the trial proceedings given their firm conviction that Petitioner does not fit the profile 2 of the shooter they saw. 3 Judge Oliver opined that Messrs. Garrison and Ward’s assertions that they would 4 have testified were “dubious at best” given that she observed “no evidence that either 5 witness made any attempts to find out what happened in this case or who was ultimately 6 prosecuted for it.” (2d R. & R. 39.) On the de novo evidentiary record, the Court 7 respectfully views Messrs. Garrison and Ward’s lack of interest in the prosecution of 8 Petitioner more charitably than Judge Oliver did. Tracking the status of criminal cases 9 before this Court is a challenge that requires the attention of numerous trained and 10 experienced civil servants; the Court is uncertain how or why laypeople such as Messrs. 11 Garrison and Ward might keep apprised of criminal proceedings if they were unaware 12 of who, if anyone, was being prosecuted for the shooting. Further, both Messrs. 13 Garrison and Ward were subject to the scrutiny of law enforcement at the time for 14 unrelated incidents and, accordingly, might have been discouraged from making 15 inquiries into the investigation or prosecution of the shooting. (See Tr. 219–20 (Mr. 16 Garrison stating law enforcement was “just always unfair”).) Indulging their curiosity 17 would have created an appreciable personal risk. 18 The Court credits Messrs. Garrison and Ward’s accounts about their availability 19 and willingness to testify over the countervailing evidence. Det. Duncan testified about 20 his general practices in attempting to locate witnesses, but he admitted that he did not 21 remember what he did to look for Messrs. Garrison and Ward specifically. (Tr. 26–27, 22 73, 83; see id. at 27–28, 77.) Similarly, DDA Santoro recounted that he had asked Det. 23 Duncan orally and in writing to find Messrs. Garrison and Ward, but he could not recall 24 what Det. Duncan had done to try to find them. (Id. at 88–91; see Ex. 300-01.) While 25 the Court concurs with Judge Oliver’s assessment that Det. Duncan and DDA Santoro 26 testified credibly on these issues, (see 2d R. & R. 39), the Court finds that their 27 testimony is minimally probative as to whether Messrs. Garrison and Ward were 28 available and willing to testify at Petitioner’s trial in 2007. First, Det. Duncan and DDA 1 Santoro do not remember any specific details about their efforts to locate Messrs. 2 Garrison and Ward. While the memories of Det. Duncan and DDA Santoro may have 3 faded in the two decades since the investigation and trial, the Court declines to conclude 4 that Messrs. Garrison and Ward were evading outreach efforts based on general 5 testimony describing methods for finding witnesses, which methods may or may not 6 have been used here. Second, at the time of the investigation and trial, Messrs. Garrison 7 and Ward had routine contact with public entities, including school, law enforcement, 8 court, and probation, that necessarily generated records that would have been available 9 to Det. Duncan and DDA Santoro. Even if “[t]he database and data systems were a lot 10 different in 2006,” the witnesses’ regular contacts with and supervision under public 11 authorities undercuts an inference that Det. Duncan and DDA Santoro, consistent with 12 their general practices, exhausted the resources at their disposal to locate the witnesses, 13 (Tr. 73), particularly given that the only documentary evidence memorializing Det. 14 Duncan and DDA Santoro’s efforts or requests to locate civilian eyewitnesses predates 15 the trial by a year, (Ex. 300-01 (dated Aug. 20, 2006)). Either Det. Duncan and DDA 16 Santoro should have been able to locate the witnesses using public records and resources 17 sometime in the year leading up to trial, or their efforts were inconsistent with their 18 general practices. Finally, there is no evidence that, in whatever search efforts they 19 undertook, Det. Duncan and DDA Santoro ever divulged that Petitioner was being 20 prosecuted for the shooting. As discussed, that detail likely would have materially 21 impacted Messrs. Garrison and Ward’s responsiveness to their inquiries. 22 Finally, the Court accords negligible weight to the two other strains of record 23 evidence Judge Oliver credited in her analysis of Messrs. Garrison and Ward’s 24 availability and willingness to testify: the lack of cooperation by two eyewitnesses who 25 testified at trial, Mr. Akbar and Richard Givens, and Petitioner’s codefendant’s motion 26 to continue trial based on counsel’s intent to investigate and subpoena unspecified 27 witnesses. (2d R. & R. 38–39.) Neither strain is particularly availing. The fact that two 28 testifying witnesses were uncooperative has little probative value as to whether two 1 different individuals would have testified if asked or compelled. If anything, the fact 2 that the uncooperative witnesses ultimately appeared and gave exculpatory testimony 3 after being subjected to process raises an inference that Messrs. Garrison and Ward 4 would have done similarly had anyone expended the resources to investigate and serve 5 them. (See 7 RT 1539, ECF No. 121-9 (Mr. Akbar stating that he was arrested and 6 required to come to trial); 8 RT 1948–49, ECF No. 121-10 (Richard Givens stating that 7 he was subpoenaed to testify at trial).) And the record does not demonstrate that the 8 codefendant’s counsel had difficulty locating Messrs. Garrison and Ward in particular, 9 let alone what efforts he made to find any witnesses. (See Ex. 301-02 (codefendant’s 10 counsel declaring that “there are witnesses who need to be interviewed and subpoenaed 11 for trial”); Tr. 91 (DDA Santoro: “I spoke with [the codefendant’s counsel] about me 12 trying to find the witnesses, all of them. I’m not sure I said Garrison versus Ward versus 13 Akbar, but just generally the witnesses.”).) 14 Ultimately, the Court credits Messrs. Garrison and Ward’s assurances that they 15 would have testified had they known Petitioner was being tried and had they been asked 16 or commanded to do so. 17 18 B. Messrs. Garrison and Ward’s Accounts Were Powerful Enough to 19 Establish a Probability that a Reasonable Attorney Would Decide to 20 Present Them and that Such Presentation Might Undermine the 21 Verdict 22 Messrs. Garrison and Ward provided assured, dispassionate testimony at the 23 evidentiary hearing unequivocally averring Petitioner was not the perpetrator of the 24 shooting. Notwithstanding the possibility that a jury could doubt or discount Messrs. 25 Garrison and Ward’s accounts and the overall strength of the prosecution’s case, the 26 record demonstrates there is a probability that a reasonable attorney would decide to 27 call them and a probability that their testimony might undermine the jury verdict. 28 1 Mr. Garrison testified that he knew Petitioner “from around the way,” in that they 2 were from close neighborhoods and had played basketball in the same area in the early 3 2000s. (Tr. 194; accord id. at 219.) Mr. Garrison recalled that Mr. Willaims was a 4 “[t]all, dark-skinned, slender dude.” (Id. at 194; see also id. (“6’5”, 6’6”, I’ll put 5 [Petitioner] at.”).) As to the shooting, Mr. Garrison stated that less than a minute after 6 he pulled up to a barber shop before noon, he heard shots ring out just as he was 7 shuffling “a couple of shoeboxes” in the back seat of his vehicle. (Id. at 195–96.) He 8 stayed in position until the shooting stopped, then “looked up and . . . saw the shooter 9 like running away basically.” (Id. at 196–97.) He had a three-to-four second window of 10 unobstructed visual observation. (Id. at 197.) Mr. Garrison described himself as having 11 medium brown skin and standing six feet, one inch tall, and the shooter as “a couple 12 shades lighter than me, shorter than I am.” (Id.; see also id. at 231 (stating the shooter 13 had “light brown” skin and stood “between 5’4” or 5’7”).) Mr. Garrison believed, based 14 on skin tone, that the individual could have been Black or Hispanic. (Id. at 231.) He 15 said the shooter wore darker-colored clothing, which made his skin tone more 16 noticeable given the contrast. (See id. at 198.) Mr. Garrison asserted that he knows that 17 Petitioner was not the shooter because the person he saw was not as tall as Petitioner— 18 “It’s like saying Shaq was accused of a crime. . . . It’s just like, it’s two different people 19 for sure.” (Id. at 219.) 20 Mr. Ward was familiar with Petitioner in that he had “seen [Petitioner] around,” 21 such as at a gym and barbershop, but he was not friends with Petitioner. (Tr. 139–40.) 22 Mr. Ward recalled that Petitioner looked “kind of like the rapper Snoop Dogg” in that 23 he was “tall and dark-skinned.” (Id. at 140.) Mr. Ward was in the back seat of Mr. 24 Garrison’s car when he heard gunshots less than a minute after arriving at the barber 25 shop. (Id. at 142–43.) Another passenger in the car (identified by counsel as Mr. Akbar, 26 (see id. at 160–61)) fell on top of him. (Id. at 143.) After the shots stopped, he popped 27 his head up and saw for “[a] few seconds” two men running away, one with a white T- 28 shirt and the other with a black hoodie. (Id. at 161; see id. at 143–45.) The men had skin 1 with “[l]ighter colored pigmentation” such that Mr. Ward was uncertain whether they 2 were “Hispanic or a light skinned African American.” (Id. at 162–63.) The men were 3 shorter than or of similar height to Mr. Ward, who is six feet tall. (Id. at 146, 164.) Mr. 4 Ward denied that either man looked like Petitioner. (Id. at 146–47.) 5 Benefitting from the “great advantage” of live testimony, Thoms, 684 F.3d at 905, 6 the Court finds the accounts of Messrs. Garrison and Ward wholly credible. Both 7 soberly recounted the events of the 2006 shooting with appropriate detail, 8 acknowledged the limits of their observations of the shooter, and made unequivocal 9 negative identifications based on those observations. Particularly salient to the 10 witnesses’ assertion that Petitioner was not the shooter was his dark complexion and 11 height. Both witnesses independently compared Petitioner to unusually tall celebrities, 12 and both said the suspected shooter was shorter than them—that is, shorter than six feet 13 tall. Both asserted, consistent with what the Court observed of Petitioner on 14 videoconference during the hearing, that Petitioner has a dark skin tone, whereas the 15 individual they saw fleeing the scene of the shooting had a relatively lighter skin tone. 16 The witnesses’ independent recollections of their perception of the shooter corroborate 17 each other and would have provided powerful evidence for the defense had the 18 witnesses testified at trial. A reasonable attorney probably would have called them. 19 Had they testified at Petitioner’s trial in 2007, Messrs. Garrison and Ward’s 20 testimony would have “altered significantly the evidentiary posture of the case.” Brown 21 v. Myers, 137 F.3d 1154, 1157 (9th Cir. 1998). A reasonable jury could credit the 22 testimony of Messrs. Garrison and Ward notwithstanding inconsistencies Respondent 23 perceives in their accounts and the possibility that the prosecution could have attempted 24 to impeached them. (Resp.’s Br. 8–16.) Notably, Messrs. Garrison and Ward 25 categorically denied that they provided any description of the suspected shooter to 26 police investigators in 2006, placing police reports stating they had done so into dispute. 27 (Compare Tr. 147–48, 233, with Exs. 13-3, 14-2.) A jury could reason, just as the Court 28 did in its analysis of Messrs. Garrison and Ward’s willingness to testify, that the 1 witnesses would not have provided possibly inculpatory evidence to police 2 investigators. 3 In the context of all the trial evidence, the jury could have accorded Messrs. 4 Garrison and Ward’s accounts significant weight and acquitted Petitioner. Mr. Akbar, 5 the other victim, was the only testifying eyewitness to have purportedly seen the 6 gunman. Det. Duncan testified that Mr. Akbar said he saw a “tall six-three to six-five 7 thin male Black” person running from the scene, and that Mr. Akbar positively 8 identified Petitioner in a six-pack photographic lineup. (11 RT 2725, 2752, ECF No. 9 121-13.) In contrast, Mr. Akbar testified that he “didn’t see anybody shooting” at him, 10 and he denied having told police that he identified the shooter as “a tall, dark male, 11 Black, 6’3” to 6’5”.” (7 RT 1554.) The jury requested more than once to review trial 12 testimony pertaining to Mr. Akbar. (See 2 CT 166–67, 172, ECF No. 121-2.) Even 13 accepting arguendo that the prosecution’s case against Petitioner was strong for the 14 reasons persuasively articulated by Respondent and Judge Oliver, (Resp.’s Br. 16–22; 15 2d R. & R. 43–45), Messrs. Garrison and Ward’s negative identification of Petitioner 16 as the shooter could have given jurors a reasonable doubt as to his guilt considering the 17 apparent weight the jury gave to evidence pertaining to Mr. Akbar, the only trial witness 18 who might have seen the gunman. Further, testimony from Messrs. Garrison and Ward 19 disputing that they had provided positive identifications to law enforcement could have 20 impacted the jury’s perception of Det. Duncan’s credibility as to what Mr. Akbar (or 21 Messrs. Garrison and Ward) had told him about the shooter’s appearance. The jury 22 deliberated for three days, making it all the more likely that the outcome of trial could 23 have been different had two additional eyewitnesses testified. (2 CT 169–70, 173–74, 24 241–42.) See United States v. Caruto, 532 F.3d 822, 832 (9th Cir. 2008) (“Longer jury 25 deliberations weigh against a finding of harmless error because lengthy deliberations 26 suggest a difficult case.” (internal quotation marks omitted)); cf., e.g., Parker v. 27 Gladden, 385 U.S. 363, 365 (1966) (noting in prejudice evaluation that “the jurors 28 deliberated for 26 hours, indicating a difference among them as to the guilt of 1 | petitioner’); Jennings v. Woodford, 290 F.3d 1006, 1019 (9th Cir. 2002) (reasoning it 2 | was “reasonably probable that... [the jury] would have found a reasonable doubt” 3 | given that the jury had “deliberated for two full days’’). 4 Having assessed the credibility of the witnesses who testified in 2026 and 5 | considered the totality of evidence before the jury at trial in 2007, the Court determines 6 | that there is a probability that the jury could have acquitted Petitioner had Messrs. 7 | Garrison and Ward been called to testify, as any competent lawyer likely would have 8 | done. Strickland, 466 U.S. at 696. 9 10 | IV. CONCLUSION 11 For the reasons stated, the Court sustains Petitioner’s objections to Judge Oliver’s 12 | 2025 report and recommendation and respectfully rejects Judge Oliver’s report and 13 | recommendation on the prejudice element of Petitioner’s ineffective assistance 14 | subclaim in Ground Two. The Court determines that Petitioner has carried his burden 15 | to show his trial counsel provided deficient performance prejudicing his defense within 16 || the framework set forth in Strickland v. Washington, 466 U.S. 668 (1984), by failing to 17 || perform any investigation into witnesses Dayon Garrison and Christopher Ward. 18 The Court grants the petition for a writ of habeas corpus and orders Respondent 19 || to release Petitioner within seven days after entry of judgment, subject to appropriate 20 | release conditions, unless the State of California elects to retry Petitioner. Respondent 21 | shall file a report within 14 days of entry of judgment stating whether Petitioner was 22 | released or will be retried. Judgment shall enter forthwith. 23 24 | IT IS SO ORDERED. Hy, 26 | Dated: February 24, 2026 7 MARK C. SCARSI UNITED STATES DISTRICT JUDGE 28 16