ROSEBORO v. TICE

IN THE UNITED STATES DISTRICT COURT

FOR THE EASTERN DISTRICT OF PENNSYLVANIA

KARL ROSEBORO,

Petitioner,

CIVIL ACTION

v. NO. 22-3377

KEN HOLLIBAUGH, et al.,

Respondents.

Pappert, J. February 14, 2025

MEMORANDUM

Karl Roseboro, currently serving a life sentence for first-degree murder, seeks a

writ of habeas corpus under 28 U.S.C. § 2254. Magistrate Judge Lynne A. Sitarski

issued a Report and Recommendation recommending denial of Roseboro’s petition, to

which Roseboro objected. After thoroughly reviewing the record, the Court overrules

the objections, adopts the R&R and denies the petition.

I

A

At around 2 a.m. on August 4, 2012, Karl Roseboro walked past a corner store at

the intersection of Wayne Avenue and Brunner Street in the Nicetown neighborhood of

Philadelphia. (Jury Trial Tr. Sept. 9 at 40:9–41:19.) A store security camera recorded

him and Rhonda Williams, a resident of the neighborhood, walking south on Wayne,

and then out of the camera’s view. (Id.; id. at 156:12–157:10; Jury Trial Tr. Sept. 10 at

26:17–27:11.) Less than thirty seconds later, four popping sounds are heard on the

surveillance video. (Prelim. H’rg Tr. at 83:22–84:9; Jury Trial Tr. Sept. 10 at 46:19–

50:13.)1 At 2:09:44, a 911 caller reported gunshots and a woman screaming in an alley

near Wayne and Brunner. (Jury Trial Tr. Sept. 10 at 23:12–24:16.) Police arrived in

the area at 2:13:37 and at 2:29 reported finding Williams’s body in an alleyway on the

west side of Wayne, fifty-five feet beyond view of the security camera. (Id. at 23:21–

25:21; Jury Trial Tr. Sept. 9 at 108:23–109:6, 156:5–157:14.) Williams had been shot

four times: once in the right forearm, and three times in the head. (Jury Trial Tr. Sept.

11 at 10:14–24.)

1

Four lay witnesses testified for the Commonwealth: Tyheem Williams, Rhonda

Williams’s son, (Jury Trial Tr. Sept. 11 at 39:14–40:8);2 Roseboro’s girlfriend, Shaquilla

Harmon, who was also a cousin of Williams’s, (Jury Trial Tr. Sept. 10 at 99:13–101:24);

Dominique Jackson, Williams’s daughter’s best friend, (id. at 145:4–146:17); and Lydia

Negron, a friend of Williams who lived in the neighborhood, (id. at 174:23–174:16,

197:21–24). Harmon and Tyheem testified that Roseboro sold crack on the 1800 block

of Brunner and that Williams, who was a crack addict, sometimes obtained drugs from

him. (Id. at 103:12–106:17, 110:16–111:9; Jury Trial Tr. Sept. 11 at 41:8–21, 44:24–

45:13.)

Tyheem also told the jury about a conversation with Roseboro in the afternoon of

August 4 outside of Negron’s house. According to Tyheem, he, his sister, Jackson and

1 The corner store had two exterior cameras, one across the street on the east side of Wayne

and one on 1900 block of Brunner on the same side of the street as the store. (Jury Trial Tr. Sept. 10

at 26:13–25.) Both captured Roseboro and Williams and were played at trial, but due to the manner

in which the audio surveillance was set up, only the video from the camera on Wayne included audio.

(Id. at 35:6–36:3.)

2 To avoid confusion between family members, the Court refers to the victim as Williams and

her son as Tyheem, to whom witnesses at trial often referred by his nickname, Randy. See (Jury

Trial Tr. Sept. 10 at 194:16–18.)

Roseboro were outside the house when someone asked Roseboro where he was at the

time of Williams’s murder. (Jury Trial Tr. Sept. 11 at 55:14–60:5.) Roseboro first said

he “wasn’t around,” then said he had seen Tyheem on Germantown Avenue at the time

of the murder. (Id. at 60:6–18.) But Tyheem, who admitted that he’d been drinking

and smoking PCP the night of August 3, told the jury he’d only seen Roseboro sometime

between 10 p.m. and midnight, not around 2 a.m. on August 4. (Id. at 46:12–49:1,

49:19–22, 61:17–62:12.) Tyheem also explained that this conversation was prompted by

his hearing a rumor that Roseboro was involved in his mother’s murder and that in

addition to denying it, Roseboro told Tyheem that he’d lost his own mom. (Id. at 60:24–

61:15.) 3 Tyheem, Jackson and Negron described Roseboro’s demeanor as scared or

nervous and said they never again saw him in the area. (Id. at 63:5–64:8.)4

Harmon wasn’t present for the August 4 afternoon conversation, but she told the

jury that when she picked Roseboro up from a neighborhood bar to go home sometime

between midnight and 1 a.m. that morning, they argued because he didn’t want to leave

yet. (Jury Trial Tr. Sept. 10 at 113:18–116:7.) He was “aggressive” during the

conversation, so she left him in the neighborhood and made the five- to ten-minute

drive to their home without him. (Id. at 108:2–110:11, 116:11–17.)5

3 Jackson, Negron and Tyheem’s accounts all differed slightly. Jackson recalled Roseboro

saying first that he was at one neighborhood bar, then at a different bar, and finally that he was

with Tyheem. (Jury Trial Tr. Sept. 10 at 150:7–154:6.) And Negron recalled Jackson being across

the street at the time Roseboro told Tyheem he wasn’t in the neighborhood at the time of the murder.

(Id. at 179:17–181:11.)

4 Harmon, Jackson and Negron also testified that Roseboro was in the neighborhood nearly

every day before Williams was killed, and Harmon said she still dropped him off there a few times

after August 4. (Jury Trial Tr. Sept. 10 at 104:18–105:2, 147:22–149:21, 176:17–178:23; 120:4–

121:5.)

5 Harmon further testified that Roseboro usually woke her up at 2:30 a.m. so she could leave

for her 3:30 a.m. shift, but on August 4 she woke up late, between 3:00 and 3:30 a.m.; as she left the

house, she saw Roseboro asleep. (Id. at 116:22–117:22.)

Roseboro’s trial counsel Stephen Patrizio cross-examined each witness. All four

admitted they knew of no problems between Roseboro and Williams, who had a positive

relationship and even treated each other like family. (Id. at 132:20–133:8, 135:18–

136:5, 173:14–174:3, 207:15–23; Jury Trial Tr. Sept 11 at 103:18–104:11.) To

undermine the credibility and effect of testimony about Roseboro’s conduct and

statements on August 4, counsel highlighted Tyheem’s use of PCP, a drug that affects

memory and perception.6 He also cross-examined Tyheem, Jackson and Negron on

their prior statements to police.7

2

The prosecution called multiple police officers, including Philadelphia Police

Detective James Dunlap, who presented the Commonwealth’s timeline of events.

Dunlap, who is trained to extract and store video data from DVRs and was responsible

for retrieving the security footage, explained that DVR clocks are often inaccurate.

(Jury Trial Tr. Sept. 10 at 17:3–20:16, 25:22–26:9, 29:7–16, 29:14–30:3.) As such,

6 None of the witnesses said they thought Tyheem was high during the August 4 conversation.

He denied being “out of it,” (Jury Trial Tr. Sept. 11 at 104:16–105:13), and although Jackson once

described him as “zoned out,” she also stated, when directly asked, that he didn’t seem high, (Jury

Trial Tr. Sept. 10 at 170:23–171:10, 172:20–23).

But counsel got Tyheem and Jackson to acknowledge that it would not be unusual to bar-hop

between the two neighborhood bars that Jackson said Roseboro named. (Jury Trial Tr. Sept. 11 at

96:6–99:22, 106:19–108:5; Jury Trial Tr. Sept 10 at 170:11–172:11).

7 Jackson saw the security footage at a police station on August 4 but didn’t identify Roseboro.

(Jury Trial Tr. Sept. 10 at 162:23–167:3.) Nor did Tyheem or Negron, who were at the station on

August 6, tell the police about the August 4 conversation with Roseboro. (Jury Trial Tr. Sept. 10 at

194:19–199:16; Jury Trial Tr. Sept. 11 at 91:21–92:14.)

Tyheem admitted on direct examination that he hadn’t identified Roseboro on August 6, only

doing so a month later after one of his uncles told him that he should tell the police the truth. (Jury

Trial Tr. Sept. 11 at 65:4–68:3, 71:7–78:5.) He later explained that he hadn’t done so because a

different uncle told him the Williams family would retaliate and he shouldn’t tell the police

anything. (Id. at 62:15–93:19.) And Negron, on cross-examination, explained that she hadn’t been

shown the video or asked any questions on August 4 but identified Roseboro a month later when she

was asked to watch the video. (Jury Trial Tr. Sept. 10 at 196:17–198:23, 202:5–203:7.)

whenever he retrieves video footage, he uses his smartphone to check the time on the

surveillance against the Naval Observatory Atomic Clock. (Id. at 29:25–30:12.) When

he retrieved the video showing Roseboro and Williams together on August 4, he

determined that the timestamp on that surveillance footage was 9 minutes and 55

seconds slow. (Id. at 30:13–17.) In other words, when Roseboro and Williams walked

off-screen at approximately 1:57:43 a.m., according to the timestamp, see (Prelim. H’rg

Tr. at 83:22–84:9), the video actually displayed what occurred at 2:07:38 a.m. And the

four popping sounds, heard around timestamp 1:58:08, happened around 2:08:03 a.m.

— almost two minutes before the 911 call at 2:09:44 a.m. (Jury Trial Tr. Sept. 10 at

23:12–14, 46:19–50:13.) The time of the 911 call was recorded in the RADQ (Radio

Assisted Data Query), a computer-generated list of all calls involving the police radio,

dispatch or patrol car terminals, which Dunlap described as “extremely accurate” in its

timekeeping. (Id. at 22:8–25.) The RADQ also recorded officers immediately

responding to the dispatch at 2:11:28, arriving on scene at 2:13:37 and finding

Williams’s body at 2:29. (Id. at 23:21–25:21; Jury Trial Tr. Sept. 9 at 108:23–109:6.)

Patrizio attempted to undermine Dunlap’s proffered timeline in two ways. He

first focused on Detective Ron Dove’s January 30, 2013 preliminary hearing testimony.

Dove, the lead detective in this case, testified that the timestamp was “a little slow in

comparison to realtime . . . . within minutes.” (Prelim. H’rg. Tr. at 82:5–18.) Counsel

asked Dunlap whether he knew about Dove’s testimony or whether Dove ever told him

that he separately calculated the time differential. (Jury Trial Tr. Sept. 10 at 66:13–

67:4.) Dunlap said no. (Id.) Second, counsel compared Dunlap’s calculation to the

RADQ times, pointing out that while the officers reported arriving at 2:13:37, the

timestamp on the video when they first appeared was 2:15:55. (Id. at 71:14–72:12.)

Dunlap agreed that this would create a nearly three-minute differential based on the

RADQ. (Id. at 72:13–25.) But he also explained that he didn’t know what “on scene”

meant, and the officers could have reported when they arrived in the area but outside

the camera’s view. (Id. at 72:25–73:5.)8

Patrizio also cross-examined Dunlap about the recorded audio accompanying the

video. Dunlap had earlier testified that the microphone was located inside the store, on

the side parallel to Brunner. (Id. at 36:3–21.) 9 Under cross-examination, Dunlap

agreed the microphone was likely intended to record indoor conversations, such as

during a robbery, and admitted that he didn’t know anything about the quality or

sensitivity of the microphone. (Id. at 67:20–70:3.) Counsel then played the video

footage for a minute beyond the time the four popping sounds were heard, revealing two

additional noises that Dunlap described as a “click” and a “light pop.” (Id. at 76:8–77:3,

96:10–98:18.) He then played various surveillance video clips, revealing that the

microphone didn’t pick up a bouncing basketball outside, passing cars or conversation

on the store’s front steps. (Id. at 77:4–91:15.) On redirect, the video of Roseboro and

Williams leaving the camera’s view and the subsequent six sounds were played again

for the jury. (Id. at 95:8–98:18.)

8 Officer Daniel Levitt, who along with his partner was dispatched to respond to the 911 call,

testified to that effect the previous day, explaining that the RADQ arrival time could have been

based on their use of the patrol car terminal or police radio and was not necessarily connected to

their appearance on the video. (Jury Trial Tr. Sept. 9 at 102:5–104:2, 107:7–25.)

9 The jury heard the day before that the south side of a clothing donation bin on the Wayne

Avenue side of the store was 119 feet from the alley where Williams’s body was found, (Jury Trial Tr.

Sept. 9 at 202:4–18), giving it a rough idea of how far the microphone was from the alley.

3

The jury deliberated for more than a day, twice requesting to watch the video

from the time Roseboro appeared on camera through the six sounds recorded after he

and Williams walked out of view. (Jury Trial Tr. Sept. 15 at 2:1–6:5; Jury Trial Tr.

Sept. 16 at 2:1–4:23.)10

On September 16, the jury found Roseboro guilty of first-degree murder and

related firearms charges. (Jury Trial Sept. 16 at 5:22–8:14.) He was sentenced the

same day. (Id. at 12:5–19.) The Pennsylvania Superior Court affirmed the judgment

on appeal, Commonwealth v. Roseboro, No. 2833 EDA 2014, 2016 WL 903949, at *1 (Pa.

Super. Ct. March 9, 2016), and the Supreme Court of Pennsylvania denied allocatur,

Commonwealth v. Roseboro, 145 A.3d 164 (Pa. 2016).

B

On March 9, 2017, Roseboro filed a pro se petition for relief under Pennsylvania’s

Post Conviction Relief Act. (PCRA Op. at 2, ECF No. 31-2.) The court subsequently

appointed counsel, who on October 10, 2018 filed an amended petition raising a variety

of ineffective assistance of counsel and due process claims. (Id. at 2–3.) After holding

an evidentiary hearing on November 25, 2019, to address two of the ineffective

assistance claims, the PCRA court denied Roseboro’s petition. (Id. at 2.) The Superior

Court affirmed, adopting in full the PCRA court’s decision, Commonwealth v. Roseboro,

No. 123 EDA 2020, 2021 WL 2012602, at *2 (Pa. Super. Ct. May 20, 2021), and the

10 There was some discussion at trial concerning the volume at which to play the video, given

that in-store conversations heard in the first clip played during Dunlap’s testimony were overly loud.

(Jury Trial Tr. Sept. 10 at 77:4–84:20.) Counsel and the court eventually settled on a volume they

considered “normal,” (id. at 84:21–85:6), and when the jury reviewed the footage during

deliberations, it was replayed at both the lowest and highest volumes used during trial, (Jury Trial

Tr. Sept. 15 at 4:6–11; Jury Trial Tr. Sept. 16 at 2:1–7).

Pennsylvania Supreme Court denied allocatur, Commonwealth v. Roseboro, 283 A.3d

176 (Pa. 2022).

C

On August 24, 2022, Roseboro filed this petition pro se, asserting four claims of

ineffective assistance of counsel. He alleged Patrizio (1) failed to object to Tyheem’s

testimony about Roseboro’s rumored killing of Williams as hearsay and to request a

curative instruction; (2) failed to keep a promise he made in his opening statement to

introduce evidence that Dove had previously testified to a different timeline; (3) failed

to call Dove to testify or have his preliminary hearing testimony admitted as the prior

statement of an unavailable witness; and (4) failed to investigate Dunlap’s testimony or

obtain an expert report from him prior to trial. (Habeas Pet., ECF No. 2.)

On February 7, 2024, Judge Sitarski recommended denial and dismissal of all

four claims. (R&R, ECF No. 37.) On the first, she found neither deficient performance

nor prejudice, and on the remaining claims found that despite trial counsel’s deficient

performance, no prejudice resulted from any of his errors. (Id.)

On June 17, 2024, Roseboro objected to the R&R. (Objs., ECF No. 42.) He first

argues that the magistrate judge overlooked his argument that counsel should have

requested a curative instruction with respect to the rumor that Roseboro committed the

murder. (Id. at 2–3.) His remaining objections concern the magistrate judge’s alleged

errors in conducting the prejudice analysis given, as he sees it, that the weight of the

evidence against him was not overwhelming. (Id. at 3–11.)

II

A

28 U.S.C. § 2254 bars the Court from granting habeas relief on any claim that a

state court has already adjudicated on the merits unless the state court’s decision (1)

“resulted in a decision that was contrary to, or involved an unreasonable application of,

clearly established Federal law, as determined by the” United States Supreme Court; or

(2) “resulted in a decision that was based on an unreasonable determination of the facts

in light of the evidence presented in the State court proceeding.” 28 U.S.C. § 2254(d).

“Clearly established” federal law consists only of “the holdings, as opposed to the dicta,

of [the Supreme] Court’s decisions.” Andrew v. White, No. 23-6573, 604 U.S. ---, slip op.

at 5 (Jan. 21, 2025) (per curiam) (quoting White v. Woodall, 572 U.S. 415, 419 (2014)).

The “contrary to” clause permits a court to grant the writ if “the state court

arrives at a conclusion opposite to that reached by [the Supreme] Court on a question of

law or if the state court decides a case differently than [the Supreme] Court has on a set

of materially indistinguishable facts.” Williams v. Taylor, 529 U.S. 362, 413 (2000).

The “unreasonable application” clause permits a court to grant the writ if “the state

court identifies the correct governing legal principle from [the Supreme] Court’s

decisions but unreasonably applies that principle to the facts of the prisoner’s case.” Id.

Finally, a decision is not based on an unreasonable determination of facts “‘merely

because the federal habeas court would have reached a different conclusion in the first

instance.’” Wood v. Allen, 558 U.S. 290, 301 (2010).

If a federal habeas court determines that a petitioner meets one of § 2254(d)’s

exceptions, the court “must then resolve the claim without the deference [28 U.S.C. §

2254(d)] otherwise requires.” Panetti v. Quarterman, 551 U.S. 930, 953 (2007).

Similarly, “when a state court's rationale is less than clarion, it is permissible to

sidestep AEDPA deference if the claim would fail under de novo review.” Hannibal v.

Sec’y Pa. Dep’t of Corrs., 2024 WL 1422015, at *7 (3d Cir. Apr. 2, 2024) (citing Berghuis

v. Thompkins, 560 U.S. 370, 390 (2010)).

B

To succeed on an ineffective assistance of counsel claim, a petitioner must show

that (1) his “counsel’s performance was deficient,” and (2) he suffered prejudice as a

result. Strickland v. Washington, 466 U.S. 668, 687 (1984).

Counsel’s performance is deficient when it falls “below an objective standard of

reasonableness.” Id. at 687–88. Courts apply a “strong presumption” of reasonableness

and should endeavor “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. at 689. The presumption of reasonableness can

be rebutted “by showing either that the conduct was not, in fact, part of a strategy, or

by showing that the strategy employed was unsound.” Thomas v. Varner, 428 F.3d 491,

499–500 (3d Cir. 2005). Even where the presumption is rebutted, “a court must still

determine whether, in light of all the circumstances, the identified acts or omissions of

counsel were outside the wide range of professionally competent assistance.” Lewis v.

Horn, 581 F.3d 92, 113–14 (3d Cir. 2009) (cleaned up). And when “the record does not

explicitly disclose trial counsel’s actual strategy or lack thereof . . . the presumption [of

reasonableness] may only be rebutted through a showing that no sound strategy . . .

could have supported the conduct.” Thomas, 428 F.3d at 500.

To establish prejudice, the petitioner must demonstrate “a reasonable

probability that, but for counsel's unprofessional errors, the result of the proceeding

would have been different.” Strickland, 466 U.S. at 694. “A reasonable probability is a

probability sufficient to undermine confidence in the outcome.” Id. “The likelihood of a

different result must be substantial, not just conceivable.” Harrington v. Richter, 562

U.S. 86, 112 (2011). To determine the likelihood of a different outcome, the Court

“must consider the totality of the evidence before the judge or jury.” Strickland, 466

U.S. at 695.

C

The standard pursuant to which the Pennsylvania Superior Court assessed the

prejudice prong of Roseboro’s claims is unclear. The court first stated that PCRA

petitioners must establish, “by a preponderance of the evidence” that counsel’s errors

“so undermined the truth determining process that no reliable adjudication of guilt or

innocence could have taken place,” Roseboro, 2021 WL 2012602 at *2, then stated that

prejudice requires demonstrating “a reasonable probability that, but for counsel’s

unprofessional errors, the result of the proceeding would have been different,” id.

(citation omitted). However, the court did not expressly address the merits of

Roseboro’s claim, concluding only that it found “no legal errors” in the PCRA court’s

opinion and adopted the opinion as its own. Id. It is thus unclear whether the court

applied a preponderance of evidence or reasonable probability standard, so the Court

will review Roseboro’s claims de novo. See Hannibal, 2024 WL 1422015, at *7.

The Court reviews de novo the specific portions of the R&R to which Roseboro

objects. 28 U.S.C. § 636(b)(1); see also Cont’l Cas. Co. v. Dominick D’Andrea, Inc., 150

F.3d 245, 250 (3d Cir. 1998).

III

Roseboro’s first claim is that trial counsel failed to object or request a curative

instruction after Tyheem testified that he’d heard a rumor that Roseboro murdered

Williams. (Pet. at 8.)11

The possibility that Tyheem would so testify first arose during a motion in

limine hearing on September 10. The prosecutor explained that he intended to elicit

testimony about the August 4 conversation and that, although he was willing to avoid

the rumor testimony, he wanted the trial court to preclude Roseboro’s statements

denying the murder and explaining that he’d lost his own mother. (Jury Trial Tr. Sept.

10 at 2:9–4:11.) Defense counsel objected to the August 4 testimony in its entirety, but

the trial court ruled that any testimony concerning Roseboro’s own statements — either

his whereabouts or his self-serving denial and explanation — was admissible. (Id. at

4:12–6:6.) But the court deferred decision on what Tyheem could say about the rumor

that Roseboro killed Williams, stating that its admissibility could depend on the

context. (Id. at 6:7–8:12.)

In front of the jury, the court asked Tyheem to clarify when Roseboro made a

particular statement. (Jury Trial Tr. Sept. 11 at 60:24–25) Instead of answering the

question, Tyheem explained why the August 4 conversation happened: “We was asking

11 Roseboro’s petition included a claim that his PCRA counsel was ineffective for failing to raise

this issue as not just ineffective assistance but also a violation of Roseboro’s rights under the

Confrontation Clause, though he subsequently withdrew that argument. (Supp. Reply at 2–3, ECF

No. 34.)

him, we are hearing that you had something to do with my mom being killed. I asked

him.” (Id. at 61:1–4.) The court followed up by asking Tyheem for Roseboro’s response,

which he said was to deny the murder and explain that his own mom had died. (Id. at

61:5–15) Defense counsel did not object to or move to strike the testimony about the

rumor. Nor did he request an instruction, either to prohibit the jury from considering

the testimony at all or limit the jury’s consideration of the rumor to its effect on

Tyheem.

At the PCRA hearing, Patrizio admitted that his failure to object to the

testimony was not based on any strategy. (PCRA H’rg Tr. at 17:24–18:3) He and

counsel for the Commonwealth also indicated they viewed the trial court’s prior ruling

as a conditional one: if the testimony about the rumor was admitted, then Roseboro’s

self-serving statements would be as well. (Id. at 15:4–10, 27:6–13.)

A

Counsel’s admission of a lack of strategy rebuts the presumption of

reasonableness. See Thomas, 428 F.3d at 499–500. Nonetheless, the Court must

determine whether his conduct, “in light of all the circumstances,” falls within the

“wide range” of competent assistance. Lewis, 581 F.3d at 113–14 (citation omitted).

Under this deferential review of his performance, counsel’s failure to object or request a

jury instruction was not deficient.

As counsel testified, this portion of Tyheem’s testimony was a mixed bag for

Roseboro. If the trial court’s ruling had been conditional, it would not have been

unreasonable for an attorney to believe, at the time, that the favorable testimony

outweighed the harm. This is especially true in the circumstances of this case, where

counsel could reasonably have concluded that Roseboro’s denial and explanation

bolstered the strategic choice to emphasize the lack of any evidence of motive and the

testimony that Roseboro and Williams had a mother/son relationship. See (Jury Trial

Tr. Sept. 9 at 39:1–40:3; Jury Trial Tr. Sept. 12 at 47:8–49:3). Even if the trial court’s

ruling was not conditional, objecting to it may not have been successful, given the

court’s position that it might find the testimony was offered for its effect on Tyheem,

the listener. (Jury Tr. Transcript Sept. 10 at 7:15–18.) And Tyheem was testifying on

the fourth day of a very contentious trial. See (Jury Tr. Transcript Sept. 12 at 68:11–

15, 133:16–25.) Given this background, counsel could reasonably have decided not to

object.

And courts frequently acknowledge that not requesting a limiting instruction is a

reasonable strategy to avoid drawing attention to the issue. See, e.g., Albrecht v. Horn,

485 F.3d 103, 127 (3d Cir. 2007) (noting this strategy for evidence of prior bad acts);

Khan v. Gordon, No. 11-7465, 2013 WL 4957479, at *7–8 (E.D. Pa. Sept. 12, 2013)

(noting this strategy for evidence admitted under the state-of-mind hearsay exception).

Despite the strong presumption that juries follow instructions, see United States v.

Franz, 772 F3d 134, 151–53 (3d Cir. 2014), asking the jury to be instructed to consider

the rumor only for its effect on Tyheem — or requesting a specific instruction to

disregard the testimony entirely — risked emphasizing for the jurors an issue that had

its downside for Roseboro. Thus, although counsel could not articulate a strategic

reason for not objecting or requesting a jury instruction, his performance still fell

within the wide range of competent assistance.

B

Even if counsel’s failure to object or request an instruction constituted deficient

performance, Roseboro cannot demonstrate a reasonable probability that, but-for this

error, the trial would have turned out differently.

The jury heard undisputed evidence that roughly thirty minutes after she was

last seen with Roseboro, Williams’s body was found with four gunshot wounds in an

alley fifty-five feet away. The jury saw the video in which Roseboro and Williams

walked in the direction of that alley, and heard, thirty seconds after Roseboro and

Williams walked off-camera, four sounds consistent with gunshots. Jurors heard

Dunlap’s testimony about the time differential, which would place their walking off

camera together within two minutes of the 911 call reporting gunshots and a woman

screaming in an alley near Wayne and Brunner. And it heard a variety of

circumstantial evidence about Roseboro from the lay witnesses, including his mood

when Harmon last saw him; his lack of alibi and shifting responses when confronted;

and that he rarely, if ever, returned after Williams was killed to the block where he

previously dealt drugs.

IV

Roseboro next argues that his trial counsel failed to fulfill a promise he made to

the jury in his opening statement that it would hear evidence that Dove’s preliminary

hearing testimony offered a different timeline of the events surrounding the murder.

(Pet. at 10.) Roseboro claims that, had counsel kept this promise by introducing Dove’s

preliminary hearing testimony, the jury would have heard evidence that challenged the

Commonwealth’s theory that only two minutes passed between the four popping sounds

recorded in the video and the 911 call, which might have led a reasonable juror to doubt

that those sounds were gunshots. Cf. (Pet. at 70–71.)

In his opening statement, counsel made a number of representations about what

the evidence would and wouldn’t show. (Jury Trial Tr. Sept. 9 at 37:1–43:5.) One of

those concerned the Commonwealth’s timeline of the murder:

[W]hat has happened here, ladies and gentlemen, is that the evidence will

show [] that there is going to be a little shifting, and a retailoring, and a

little bit of overreaching by the Commonwealth in terms of the time that

is displayed on the video and the time that the police arrived in response

to the gunshots. So that they are going to try to suggest to you that the

period of time is within seconds of their being together and it is really

many, many minutes. I make that promise to you, as well. Why?

Because the assigned detective . . . testified at a prior hearing . . . and said

that he retrieved the videotape . . . and he checked the video-recorder for

his time and it was off only a minute or two. The problem is that doesn’t

fit within their scheme. So, they are going to have another witness come

in and tell you that the videotape time on the camera is 9 minutes and 55

seconds off to bring it in close proximity to the 911 call[.]

(Id. at 40:4–41:19.) In essence, counsel told jurors they would see that the

Commonwealth’s timeline of the murder was untrustworthy because Dunlap’s timeline

differed from Dove’s.12 But counsel never called Dove as a witness, nor sought to have

him declared unavailable so that his preliminary hearing testimony could be admitted

into evidence. He instead cross-examined Dunlap about Dove’s preliminary hearing

testimony and conversations Dove and Dunlap may have had:

Q: So you have no information that Detective Dove made his own

calculations of the time differential between realtime and the time on the

recorder?

A: No.

12 Precisely what he promised is unclear. For instance, based on the evidence at trial, the

“period of time” he referenced must be the one between Roseboro and Williams being together and

the gunshots, not their being together and the officers’ arrival: only the former period is the crucial

one, and the latter is at least six minutes, under even Dunlap’s timeline. But what is clear is that

counsel assured the jury it would see that a detective previously testified to a different timeline, one

that provides a longer gap between events than the Commonwealth was now arguing.

. . . .

Q: If he has testified previously, he said I don’t have the exact calculations

but it was a little slow in comparison to realtime, I would say a little, it

was within minutes off, he didn’t share any of that information with you;

correct?

A: Correct.

(Jury Trial Tr. Sept. 10 at 66:13–67:4.) Before closing statements, the prosecutor

moved to preclude defense counsel from talking about Dove’s preliminary hearing

testimony, arguing that Dove’s statements were never put into evidence; counsel

instead purportedly quoted Dove in asking his questions. (Jury Trial Tr. Sept. 12 at

11:15–12:6, 13:2–6, 16:19–17:2.) Patrizio argued the statements came into evidence

through his questions to Dunlap and that realistically, Dove would not respond to a

subpoena. (Id. at 12:7–16, 15:16–20, 19:24–20:2.) The trial court acknowledged that

typically, when a witness testifies at a prior proceeding, he must then testify at trial or

be declared unavailable before his prior testimony can be admitted as evidence. (Id. at

18:3–20:8.) Nonetheless, it allowed counsel to discuss Dove’s preliminary hearing

testimony in his closing argument. (Id. at 20:6–21:24, 22:18–23:12); see also (id. at

67:12–68:9).

A

Instead of calling Dove as a witness, subpoenaing him to testify or having him

declared unavailable such that he could get Dove’s preliminary hearing testimony into

evidence, Patrizio assumed instead he could rely on his own use of Dove’s statements

when cross-examining Dunlap. But, as the court instructed the jury before testimony

began, attorneys’ questions are not evidence. (Jury Trial Tr. Sept. 9 at 13:8–15:3.)

Counsel was allowed to cite to Dove’s “statements” in his closing, but he failed to deliver

what he promised when the trial began — that the jurors would hear evidence that

Dunlap offered a different timeline than the one the assigned detective, Dove,

previously testified to.

Breaking a promise made during an opening statement does not necessarily

constitute deficient performance. Elias v. Superintendent Fayette SCI, 774 Fed. App’x

745, 751 (3d Cir. 2019).13 Counsel’s broken promise, however, was not a strategic

decision based on an unforeseen development, see id.; it was a mistake based on an

error of law. And in light of the circumstances of this case, his conduct was not

reasonable. One of the defense’s primary strategies was to attack Dunlap’s timeline in

order to increase the time between when Roseboro and Williams were last seen on the

video and the 911 call. This strategy could be carried out in two ways: using the RADQ

record to suggest the timestamp is several minutes fast or calling Dove to testify that

he thought the timestamp was only a few minutes slow. Both offer different alternative

timelines, but they could, individually or together, be used to argue that the jury should

doubt Dunlap’s timeline. Once Patrizio chose to use both, there was no sound reason

for failing to introduce evidence he told the jurors they would hear and formed part of

that defense. Cf. Moore v. Sec’y Penn. Dep’t of Corrs., 457 Fed. App’x 170, 182 (3d Cir.

13 The statement in McAleese v. Mazurkiewicz, 1 F.3d 159, 166 (3d Cir. 1993) that a broken

promise is “sufficient of itself to support a claim of ineffectiveness” does not resolve Roseboro’s claim.

That blanket pronouncement is dictum, not supported by the cases it cites and contradicted

elsewhere in the opinion. See Elias v. Coleman, No. 14-1337, 2017 WL 5192476, at *21 n.12 (W.D.

Pa. Nov. 9, 2017) (discussing McAleese). The Third Circuit has since cited this part of McAleese for

the proposition that broken promises can constitute ineffective assistance, not that they necessarily

do. See Elias, 774 Fed. App’x at 751.

2012) (“Counsel’s failure to introduce evidence that contradicts a key witness’s trial

testimony is patently unreasonable.”).

B

Roseboro cannot, however, demonstrate prejudice. First of all, his lawyer never

promised the jury it would hear evidence from a specific witness, such as Dove himself.

See McAleese, 1 F.3d at 166–67 (noting the rationale for finding ineffective assistance is

that jurors might “think the witnesses to which counsel referred . . . were unwilling or

unable to deliver the testimony he promised”). The issue was the substance of Dove’s

preliminary hearing testimony and the effect it might have had on the jury. Accepting

Dunlap’s timeline creates a narrow window favorable for the Commonwealth: Roseboro

and Williams walk off-camera at 2:07:38 and two minutes later, at 2:09:44, the police

receive a 911 call reporting gunshots and a woman screaming in an alley near Wayne

and Brunner. Dove’s assessment increases the time between these two events to

approximately seven minutes, which, Roseboro argues, could affect whether a

reasonable juror would conclude the sounds on the video were gunshots.14

Roseboro’s argument is not implausible, but it does not demonstrate a

reasonable probability of a different outcome. See Harrington, 562 U.S. at 112 (“The

likelihood of a different result must be substantial, not just conceivable.”) No possible

variation on the timing of the 911 call changes the fact that, thirty seconds after

Roseboro and Williams walked off-camera, four sounds consistent with gunshots were

recorded, and less than thirty minutes after that, Williams’s body was found with four

14 Because Dove never provided a specific calculation, Roseboro’s argument is predicated on the

belief that Dove’s testimony would have shown the gunshots occurred four to seven minutes before

the 911 call, (Reply at 24, 38, ECF No. 18.)

gunshot wounds in an alley fifty-five feet beyond the camera’s view, in the direction she

and Roseboro had been walking. Although placing the 911 call within two minutes of

the recorded sounds does bolster the Commonwealth’s case, the other evidence against

Roseboro was substantial, consisting not just of the video but the witnesses’ testimony,

which indicated, inter alia, that Roseboro was aggressive when his girlfriend last saw

him between midnight and 1 a.m.; that he changed his story while talking to neighbors;

and that he largely, if not completely, stopped coming to the neighborhood, even though

he previously sold drugs there. Even assuming the jury accepted Dove’s testimony in

some manner — as evidence that the 911 call was placed five minutes later, relative to

the alleged gunshots, than Dunlap believed or instead, perhaps in combination with the

RADQ argument, as evidence that the timeline of events was too muddy to resolve —

there is no reasonable probability that, in light of the totality of the evidence, the

outcome at trial would have been different had the jurors learned what Dove testified to

at the preliminary hearing.15

V

Roseboro’s third claim is that his counsel’s failure to subpoena or call Dove as a

witness or have him declared unavailable so that his preliminary hearing testimony

could be admitted as evidence constituted ineffective assistance. (Pet. at 12.)

Roseboro’s argument here is effectively the same as it was with respect to the

prior claim: counsel failed to introduce Dove’s preliminary hearing testimony because

15 Not that Dunlap’s timeline was unimpeachable; no one at trial offered an explanation for the

seeming discrepancy between Levitt’s partner leaving the camera’s view at 2:19:58, which would be

2:29:53 under Dunlap’s calculations, and the RADQ report that Williams’s body was found at 2:29.

(Jury Trial Tr. Sept. 9 at 70:1–71:5, 108:23–109:6.) But given the evidence as a whole, the jury did

not need to find that Dunlap’s timeline was correct, or choose between competing timelines, to find

Roseboro guilty beyond a reasonable doubt.

he misunderstood the law and there is a reasonable probability that introducing this

evidence would have affected the outcome. As discussed above, no reasonable strategy

supports failing to proffer evidence supporting a defense counsel believed critical. See

Section IV.A.2. Nonetheless, in light of all of the evidence against Roseboro, counsel’s

failure was not prejudicial. See Section IV.B.

VI

Roseboro’s final claim alleges that his lawyer’s failure to obtain an expert report

or otherwise investigate Dunlap’s testimony prior to trial constituted ineffective

assistance. (Pet. at 13–14.)

A

“[C]ounsel has a duty to make reasonable investigations or to make a reasonable

decision that makes particular investigations unnecessary.” Strickland, 466 U.S. at

691. The decision not to investigate is “assessed for reasonableness in all the

circumstances.” Id. Here, Patrizio learned that Dunlap would be testifying shortly

before a motion hearing on August 21, 2014, three weeks before trial. (Mot. H’rg Tr. at

35:1–38:9.) At the time, he told the trial court that he had received nothing from or

about Dunlap. (Id. at 38:13–17.) Yet his subsequent conduct suggests he took no steps

to investigate, waiting until Dunlap was testifying at trial to ask whether he’d prepared

a report and if he could have a copy of it. (Jury Trial Tr. Sept. 10 at 62:13–16.) Dunlap

explained that all he had was a single page of notes, typed in August of 2012, and the

court briefly recessed so counsel could review those notes before resuming cross-

examination. (Id. at 62:17–65:19.)

Counsel had made the timeline a key part of the defense and believed that

Dove’s testimony was “critical” to that strategy. He also intended to cast doubt on

Dove’s conduct during the investigation. See (Mot. H’rg Tr. at 39:11–18.) Given these

strategic choices, it was unreasonable for Patrizio to not take steps to inform himself, as

best he could, of Dunlap’s possible testimony and role in the investigation.

B

But no prejudice resulted from this shortcoming. Patrizio was not caught

unawares as to the substance of Dunlap’s testimony or how he might undermine it. His

opening expressly referenced the Commonwealth’s nine minute, fifty-five second time

difference, and Dunlap’s testimony was independent of counsel’s opportunity to rely on

other evidence he already knew about, including the RADQ or Dove’s testimony.

Moreover, counsel thoroughly cross-examined Dunlap on each aspect of his testimony.

Compare (Jury Trial Tr. Sept. 10 at 26:12–39:20; 22:8–25:21, 29:7–30:23; 35:6–36:21)

with (id. at 61:20–67:13; 66:13–67:4, 71:14–73:25; 68:5–70:3, 76:8–94:8.) And while

Dunlap was qualified at trial as an expert witness in the field of forensic video

technology, (Jury Trial Tr. Sept. 10 at 21:18–20), no expertise was necessary to rebut

his testimony concerning the time differential. Counsel’s failure to better prepare did

not impact his ability to address Dunlap’s testimony at trial, and no pre-trial

investigation into that testimony would have affected the jury’s evaluation of the key

component of the video evidence: Roseboro walked with Williams out of view of a

security camera thirty seconds before four sounds consistent with gunshots were heard,

and Williams’s body was shortly thereafter found with four gunshot wounds in an alley

fifty-five feet away from where she was last seen with him. In the face of this evidence

and all the other testimony, there is no reasonable probability that Patrizio’s failure to

investigate Dunlap’s opinion and role in the case affected the outcome of the trial. 16

VII

A certificate of appealability should only be issued if the petitioner “has made a

substantial showing of the denial of a constitutional right.” 28 U.S.C. § 2253(c)(2). The

petitioner must “demonstrate that reasonable jurists would find the District Court’s

assessment of the constitutional claims debatable or wrong.” Slack v. McDaniel, 529

U.S. 473, 484 (2000). Roseboro has made no such showing, so no certificate should

issue.

An appropriate Order follows.

BY THE COURT:

_/s_/_ G__e_r_a__ld__ J__. _P_a_p__p_e_r_t_ __

Gerald J. Pappert, J.

16 To the extent Roseboro’s fourth objection, which states that counsel’s errors “permeated”

trial, (Objs. at 8, 11); see also (Supp. Reply at 6) (referencing “accumulated errors”), can be read to

state a claim for cumulative error, he failed to exhaust this claim in state court and cannot raise it

now. See § 2254(b)(1); Collins v. Sec’y of Pa. Dep’t of Corrs., 742 F.3d 528, 541–42 (3d Cir. 2014)

(holding cumulative error is a separate claim that must be exhausted in state court). Even if

Roseboro had exhausted this claim, he could not obtain habeas relief. There is no indication that his

lawyer’s error in not investigating Dunlap prejudiced him, and there is no reasonable probability

that the combined effect of his errors, if any, with respect to Tyheem’s testimony and Dove affected

the outcome of trial given the evidence against Roseboro.