David Saylor v. Noah Nagy

NOT RECOMMENDED FOR PUBLICATION
File Name: 21a0524n.06

Case No. 20-1834

UNITED STATES COURT OF APPEALS
FOR THE SIXTH CIRCUIT
FILED
Nov 17, 2021
DAVID SAYLOR, ) DEBORAH S. HUNT, Clerk
)
Petitioner-Appellant, )
) ON APPEAL FROM THE UNITED
v. ) STATES DISTRICT COURT FOR
) THE EASTERN DISTRICT OF
NOAH NAGY, Warden, ) MICHIGAN
)
Respondent-Appellee. ) OPINION
)

Before: GILMAN, THAPAR, and NALBANDIAN, Circuit Judges.

NALBANDIAN, Circuit Judge. For almost two years, David Saylor sexually abused his

toddler daughter. He pled guilty to first-degree criminal sexual conduct. Now he seeks habeas

relief. Saylor asserts four grounds: (1) ineffective assistance of plea counsel; (2) Ex Post Facto

Clause violation; (3) judicial bias; and (4) ineffective assistance of appellate counsel. The district

court denied Saylor’s habeas petition. For the reasons below, we AFFIRM.

I.

Thirteen years ago, David Saylor walked into a police station and confessed to sexually

abusing his two toddler daughters. The State charged him with five counts of criminal sexual

conduct under Mich. Comp. Laws § 750.520b.

Confronted with these charges, Saylor decided to bargain. He agreed to plead guilty to a

single count of first-degree criminal sexual conduct with a person under thirteen. In exchange, the
No. 20-1834, Saylor v. Nagy

State dropped the remaining charges. And it agreed to recommend a sentence of twenty-five to

forty years’ imprisonment as well.

On May 22, 2008, Saylor entered his plea. At the hearing, Saylor confirmed that he made

one of his daughters “repeatedly” “perform fellatio on [him]” “from the time she was two until

four.” (R. 9-2, May 22, 2008 Hr’g Tr., PageID 86-87.) And he acknowledged that the abuse started

on June 1, 2006 and ended on April 26, 2008.

But Saylor changed his mind after the fact and moved to withdraw his plea about two weeks

later. Saylor argued that his plea was unknowing because his lawyer: (1) spent only five minutes

with him to discuss the plea; and (2) told him “you either take 25 to 40 or you do life.” (R. 9-3,

June 16, 2008 Hr’g Tr., PageID 94-95.) The trial court judge denied the motion. He then sentenced

Saylor to twenty-five to forty years’ imprisonment. He also ordered Saylor to “register on the

Michigan Sex Offender Registration” and “be on lifetime GPS.” (R. 9-6, July 9, 2008 Hr’g Tr.,

PageID 179-180.)

Saylor next applied for leave to appeal in the Michigan Court of Appeals. Through

appointed appellate counsel, he raised two issues: ineffective assistance of plea counsel and due

process. On the first, Saylor alleged that his plea counsel never mentioned lifetime electronic

monitoring (“LEM”) and sex-offender registration. On due process, Saylor asserted again that his

lawyer misled him into pleading guilty. The Michigan Court of Appeals denied the application

“for lack of merit in the grounds presented.” (R. 9-13, Mar. 30, 2011 Order, PageID 259.)

Saylor tried to appeal to the Michigan Supreme Court. But it rejected his application for

leave to appeal. People v. Saylor, 810 N.W.2d 35 (Mich. 2012) (“[W]e are not persuaded that the

questions presented should be reviewed by this Court.”).

2
No. 20-1834, Saylor v. Nagy

Following the rebuke from the Supreme Court, Saylor returned to the state trial court and

filed a motion for relief from judgment. This time, he had a longer list of arguments: (1) judicial

bias; (2) Ex Post Facto Clause violation; (3) ineffective assistance of plea counsel; and (4)

ineffective assistance of appellate counsel. The judicial-bias claim turned on the fact that the trial-

court judge and prosecutor were brothers. As for his second claim, Saylor contended that Michigan

amended the sentencing statute after he began abusing his daughter. Thus, Saylor argued, his

sentence violated the Ex Post Facto Clause. Saylor also brought back his ineffective-assistance-

of-plea-counsel claim. And he pointed to direct-appeal counsel’s failure to raise his judicial bias

and ex post facto claims.

The trial court denied the motion. Saylor tried to appeal to the Michigan Court of Appeals.

It denied his application. But Saylor stopped there and didn’t ask the Michigan Supreme Court for

leave to appeal.

Saylor then turned to federal court, filing a habeas petition in the Eastern District of

Michigan. He raised six claims: (1) trial-court abuse of discretion; (2) ineffective assistance of plea

counsel; (3) judicial bias; (4) Ex Post Facto Clause violation; (5) ineffective assistance of plea

counsel under state law; and (6) ineffective assistance of appellate counsel. The State moved for

summary judgment on statute-of-limitations grounds. The trial court granted the motion and

declined to issue a certificate of appealability (“COA”). Saylor v. Haas, No. 2:14-CV-14557, 2015

WL 6125267, *3 (E.D. Mich. Oct. 19, 2015). Saylor appealed. We granted a COA as to claims

one, two, three, four, and six. The order came with a conclusory, one-sentence explanation:

“Reasonable jurists could debate whether Saylor states a valid claim of the denial of a

constitutional right as to grounds one, two, three, four, and six.” (R. 15, Oct. 31, 2016 Order,

PageID 464.)

3
No. 20-1834, Saylor v. Nagy

On the merits, we reversed the trial court’s decision and remanded. We determined that

Saylor’s case was timely under Holbrook v. Curtin, which established that “AEDPA’s one-year

statute of limitations [is] tolled during the period in which [a petitioner] could have, but did not,

appeal the [state court’s] denial of his motion for post-conviction relief.” 833 F.3d 612, 619 (6th

Cir. 2016).

On remand, the district court rejected each of Saylor’s claims. So Saylor has appealed

again. This time, our COA was narrower. It covered: (1) trial court abuse of discretion; (2)

ineffective assistance of plea counsel; (3) Ex Post Facto Clause violation; and (4) ineffective

assistance of appellate counsel. We declined to extend the COA to Saylor’s judicial-bias claim

because “whether due process requires disqualification of a judge due to apparent bias based on

kinship with the prosecutor has not been clearly established by Supreme Court precedent.” (Doc.

11, Jan. 20, 2021 Order, at 4.) And relatedly, we excluded direct-appeal counsel’s failure to raise

judicial bias from Saylor’s fourth claim. This, because “[c]ounsel cannot be deemed ineffective

for failing to raise a meritless argument.” (Id.)

Saylor now advances four claims: (1) ineffective assistance of plea counsel; (2) Ex Post

Facto Clause violation; (3) judicial bias; and (4) ineffective assistance of appellate counsel. Saylor

raises the judicial bias claim despite the COA. And his ineffective-assistance-of-appellate-counsel

claim includes judicial bias, again despite the COA.

II.

We review the district court’s legal conclusions de novo and its findings of fact for clear

error. Ivory v. Jackson, 509 F.3d 284, 291 (6th Cir. 2007). And AEDPA governs “any claim that

was adjudicated on the merits in State court proceedings.” 28 U.S.C. § 2254(d). For claims in this

category, habeas relief is appropriate only if the state court’s adjudication of the claim:

4
No. 20-1834, Saylor v. Nagy

(1) resulted in a decision that was contrary to, or involved an unreasonable
application of, clearly established Federal law, as determined by the Supreme
Court of the United States; 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.

Id.

Under the “contrary to” clause, habeas relief is warranted “if the state court arrives at a

conclusion opposite to that reached by [the Supreme] Court on a question of law or . . . 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). And the “unreasonable application” clause triggers

relief “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. As for the

state court’s findings of fact, “[t]he prisoner bears the burden of rebutting [them] ‘by clear and

convincing evidence.’” Burt v. Titlow, 571 U.S. 12, 18 (2013) (quoting 28 U.S.C. § 2254(e)(1)).

That means “a state-court factual determination is not unreasonable 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). Suffice to say, “AEDPA sets forth a heavy burden for a petitioner to

overcome.” Tibbetts v. Bradshaw, 633 F.3d 436, 442 (6th Cir. 2011).

Here, the state court reached a merits determination on each of Saylor’s four claims. And

so AEDPA deference applies throughout. The Michigan Court of Appeals determined the merits

of Saylor’s ineffective-assistance-of-plea-counsel claim in its March 30, 2011 order. (R. 9-13, Mar.

30, 2011 Order, PageID 259 (“The Court orders that the delayed application for leave to appeal is

DENIED for lack of merit in the grounds presented.”).) The order amounts to a single sentence,

and it provides no explanation. But under Harrington v. Richter, this is enough. 562 U.S. 86, 98

(2011) (“Where a state court’s decision is unaccompanied by an explanation, the habeas

5
No. 20-1834, Saylor v. Nagy

petitioner’s burden still must be met by showing there was no reasonable basis for the state court

to deny relief.”); see also Johnson v. Williams, 568 U.S. 289, 298 (2013); Brown v. Bobby, 656

F.3d 325, 329 (6th Cir. 2011). Presented with an unexplained merits decision, our task is to

determine: (1) “what arguments or theories supported, or . . . could have supported, the state court’s

decision”; and (2) “whether . . . fairminded jurists could disagree that those arguments or theories

are inconsistent with the holding in a prior decision of [the Supreme Court].” Richter, 562 U.S. at

102. As for Saylor’s remaining claims, the trial court reached reasoned merits determinations in

its October 22, 2013 order rejecting Saylor’s motion for relief from judgment.

III.

With these review standards in mind, we assess Saylor’s claims. Each misses the mark.

A. Ineffective Assistance of Plea Counsel

First, Saylor claims that his plea counsel never told him about LEM and sex-offender

registration. He argues that this resulted in an “unconstitutionally unknowing plea.” (Appellant Br.

at 12.)

Strickland v. Washington sets the test for ineffective assistance. 466 U.S. 668, 687 (1984).

The convicted defendant must establish that: (1) “counsel’s performance was deficient”; and (2)

“the deficient performance prejudiced the defense.” Id. Once AEDPA applies, this standard

becomes “doubly deferential.” Abby v. Howe, 742 F.3d 221, 226 (6th Cir. 2014) (quoting Burt,

571 U.S. at 15). Or put another way, “[t]he pivotal question is whether the state court’s application

of the Strickland standard was unreasonable.” Richter, 562 U.S. at 101. Saylor says this doubly

deferential standard “is no barrier for habeas relief here.” (Appellant Br. at 24.) As he sees it, the

state court deserves no AEDPA deference because it unreasonably applied clearly established

federal law.

6
No. 20-1834, Saylor v. Nagy

But a state court can’t unreasonably apply clearly established federal law unless that law

exists. And here, we see no Supreme Court cases to indicate that Saylor’s plea counsel provided

constitutionally deficient representation. That means Saylor’s claim fails at AEDPA’s doorstep.

The best Saylor can offer is Mabry v. Johnson, 467 U.S. 504 (1984). But it’s not enough

here. Mabry says only that a defendant must be made “fully aware of the direct consequences” of

his guilty plea. Id. at 509 (emphasis added) (quoting Brady v. United States, 397 U.S. 742, 755

(1970)). But it says nothing about whether LEM and sex-offender registration are direct or

collateral consequences. That distinction matters because “[t]he defendant need only be aware of

the direct consequences of the plea,” not collateral ones. King v. Dutton, 17 F.3d 151, 153 (6th

Cir. 1994) (emphasis added); see also Parrino v. United States, 655 F. App’x 399, 403 (6th Cir.

2016) (“Typically, a defendant’s failure to consider . . . collateral consequences of a criminal

conviction . . . does not vitiate a guilty plea.”).1 A consequence is collateral rather than direct if it

“remains beyond the control and responsibility of the district court in which that conviction was

entered.” El-Nobani v. United States, 287 F.3d 417, 421 (6th Cir. 2002) (quoting United States v.

Gonzalez, 202 F.3d 20, 27 (1st Cir. 2000)).

If anything, the available caselaw cuts in the opposite direction of Saylor’s argument. It

puts sex-offender registration firmly in the collateral column. See, e.g., Chaidez v. United States,

1
Sometimes, a consequence can be neither direct nor collateral. The Supreme Court emphasized
in Padilla v. Kentucky that “[d]eportation as a consequence of a criminal conviction is . . . uniquely
difficult to classify as either a direct or a collateral consequence.” 559 U.S. 356, 366 (2010). And
it held that withholding advice about the deportation consequences of a guilty plea constitutes
deficient representation. Id. But the applicability of Padilla beyond the deportation context is
doubtful. See, e.g., Chaidez v. United States, 568 U.S. 342, 355 (2013) (finding that the “special”
nature of deportation renders Padilla an exception to the “direct collateral-divide”). In any event,
Saylor remains unable to identify any clearly established federal law from the Supreme Court that
encompasses LEM and sex-offender registration.

7
No. 20-1834, Saylor v. Nagy

568 U.S. 342, 349 n.5 (2013) (“[E]ffects of a conviction commonly viewed as collateral include

civil commitment, civil forfeiture, sex offender registration, disqualification from public benefits,

and disenfranchisement.”); Leslie v. Randle, 296 F.3d 518, 523 (6th Cir. 2002) (“[T]he

classification, registration, and community notification provisions [of an Ohio sexual-predator

statute] are more analogous to collateral consequences . . . .”). This makes sense, since “whether

[the defendant] must actually register is controlled by state and federal law and is beyond the

control or responsibility of the court.” United States v. Cottle, 355 F. App’x 18, 20 (6th Cir. 2009).

That same logic applies just as well to LEM.2 With no guidance from the Supreme Court to say

otherwise, Saylor’s ineffective-assistance-of-plea-counsel claim fails on arrival. And so we need

not reach prejudice.

Saylor also advances a second argument. He says the trial court’s decision turns on

unreasonable determinations of fact. More specifically, he hangs his hat on the trial court’s failure

to hold an evidentiary hearing before denying his claim in its October 22, 2013 order.3 But this

argument is problematic for several reasons.

As an initial matter, Saylor fixates on the wrong ruling. In its October 22 order, the trial

court explicitly declined “to address [Saylor’s] claim of ineffective assistance of trial counsel.” (R.

2
We have yet to consider this specific question. But a pair of unpublished district court cases are
instructive. See Smith v. Bauman, No. 5:10-CV-11052, 2018 WL 1399312, at *5 (E.D. Mich. Mar.
19, 2018) (“[LEM] . . . will be beyond the control or responsibility of the trial court and, therefore,
it is a collateral consequence . . . .”); Noonan v. Hoffner, No. 1:14-CV-830, 2014 WL 5542745, at
*10 (W.D. Mich. Oct. 31, 2014) (“[B]ecause the electronic monitoring requirement is governed
by state law, it is beyond the control or responsibility of the court.”).
3
Saylor also bootstraps a separate argument without identifying it as such. He takes issue with the
fact that the federal district court below “held with no evidentiary hearing that Saylor’s ‘rejection
of the plea offer’s single count conviction and twenty-five to forty-year recommended sentence
would have not been rational.’” (Appellant Br. at 26 (quoting R. 29, July 28, 2020 Op. & Order,
PageID 642).) But this argument fails for the simple reason that federal habeas review “is limited
to the record that was before the state court that adjudicated the claim on the merits.” Cullen v.
Pinholster, 563 U.S. 170, 181 (2011).
8
No. 20-1834, Saylor v. Nagy

9-11, Oct. 22, 2013 Order, PageID 240.) It determined that the claim “was already raised on

[direct] appeal by [Saylor] and the court of appeals found no merit.”4 (Id.) This means we must

look instead at the Michigan Court of Appeals’ March 30, 2011 order. That decision provides no

explanation. And so Richter instructs us to read in a reasonable explanation, assessing “what

arguments or theories supported, or . . . could have supported, the state court’s decision.” 562 U.S.

at 102 (emphasis added). This should have been the focus of Saylor’s arguments, not the trial

court’s subsequent ruling on collateral appeal.

But even if we ignore this point, Saylor’s argument still fails. That’s because it goes toward

prejudice. (Appellant Br. at 25-26 (“Though he took no evidence, Judge Borchard somehow found

that Saylor had failed to demonstrate a reasonable likelihood that he would not have pleaded had

his attorney advised him about LEM.”).) As we have already explained, Saylor’s claim fails at

Strickland’s first prong without reaching prejudice. So an evidentiary hearing on prejudice makes

no difference in the end.

Nor does Saylor cite any Supreme Court or Sixth Circuit cases for his argument. He

marshals a handful of nonbinding cases, but these say only that an evidentiary hearing can be

necessary if “all reasonable minds agree that the state court needed to hold a hearing in order to

make [its] factual determinations.” Smith v. Aldridge, 904 F.3d 874, 883 (10th Cir. 2018) (internal

quotation omitted). Indeed, “most of the time . . . it will be reasonable for a state court to make

4
Saylor cites the trial court’s prejudice analysis. But there, the trial court was assessing its own
failure to inform Saylor about LEM, not plea counsel’s. This all originates from an amended
version of Saylor’s motion for relief from judgment, where he briefly notes, while fleshing out his
ineffective-assistance-of-plea-counsel claim, the trial court’s failure to mention LEM during his
plea colloquy. Elsewhere in the October 22 order (and in the specific context of Saylor’s
ineffective-assistance-of-plea-counsel claim), the trial court did discuss prejudice some more. But
it did so only in the alternative.
9
No. 20-1834, Saylor v. Nagy

factual determinations based on the evidence before it without holding a hearing.” Id. (emphasis

added).

Here, the trial-court judge presided over Saylor’s proceedings from the start. And the

record was sufficient to resolve prejudice in the State’s favor. The State charged Saylor with five

counts of first-degree criminal sexual conduct. That means he faced a potential life sentence or

multiple consecutive sentences (with a mandatory twenty-five year minimum for each). See Mich.

Comp. Laws § 750.520b. What’s more, Saylor had already made a “full confession to having

numerous counts of [criminal sexual conduct] with both daughters.” (R. 9-2, May 22, 2008 Hr’g

Tr., PageID 79.) It stands to reason that Saylor realized he’d likely spend the rest of his life in

prison unless he took the deal. Squared against this context, LEM and sex-offender registration

would appear to be comparatively de minimis. Thus, even if Saylor’s cases were binding, they

wouldn’t apply to our facts.

And so no matter how we slice this analysis, the outcome remains the same. We defer to

the state court’s rejection of Saylor’s ineffective-assistance-of-plea-counsel claim.

B. Ex Post Facto

Saylor was charged and convicted under Mich. Comp. Laws § 750.520b. And Michigan

amended that statute effective August 28, 2006. The amendment introduced a twenty-five-year

mandatory minimum and an LEM requirement. The trial court sentenced Saylor to twenty-five to

forty years’ incarceration (plus LEM and sex-offender registration) under the amended version.

Saylor claims “intolerable ex post facto concerns” obtain because the amendment took place after

June 1, 2006, the start date of his offense. (Appellant Br. at 27.) Put another way, Saylor claims

“[t]he record is ambiguous regarding whether [he] pleaded guilty to an offense that occurred before

or after August 28, 2006.” (Id.)

10
No. 20-1834, Saylor v. Nagy

But Saylor procedurally defaulted on this claim.5 In seeking habeas relief, a petitioner must

exhaust state-court remedies by “fairly present[ing] his claim in each appropriate state court

(including a state supreme court with powers of discretionary review).” Woolbright v. Crews, 791

F.3d 628, 631 (6th Cir. 2015) (emphasis added) (quoting Baldwin v. Reese, 541 U.S. 27, 29

(2004)). And “[w]hen a petitioner has failed to fairly present his claims to the state courts” in this

way, “and no state remedy remains, his claims are considered to be procedurally defaulted.” Id;

see also O’Sullivan v. Boerckel, 526 U.S. 838, 848 (1999) (“[Petitioner’s] failure to present three

of his federal habeas claims to the Illinois Supreme Court in a timely fashion has resulted in a

procedural default of those claims.”). Once a petitioner procedurally defaults on a claim, he can

overcome it by showing “cause and prejudice.” Coleman v. Thompson, 501 U.S. 722, 747 (1991).

Here, Saylor raised his ex post facto claim in his motion for relief from judgment. After

the trial court and the Michigan Court of Appeals rejected the claim, Saylor could have filed for

leave to appeal in the Michigan Supreme Court. But he never did. And so Saylor ended up with a

procedural default. Nor can cause and prejudice save Saylor’s claim. Saylor says “the cause-and-

prejudice exception overcomes any ostensible procedural default” because: (1) “Saylor’s plea

counsel did not advise Saylor of the full consequences of his plea”; and (2) “appellate counsel

provided ineffective assistance in failing to raise on direct appeal his ex post facto claim, resulting

in prejudice.” (Appellant Br. at 38; Reply at 4.) But Saylor’s plea counsel had nothing to do with

appellate counsel’s failure to raise the ex post facto claim on direct appeal. And neither did that

5
Saylor emphasizes that the State failed to raise procedural default before the district court below.
But that doesn’t bar our review here. “We are . . . permitted to consider the procedural default issue
even when raised for the first time on appeal if we so choose.” White v. Mitchell, 431 F.3d 517,
524 (6th Cir. 2005).

11
No. 20-1834, Saylor v. Nagy

failure excuse Saylor’s choice to sidestep the Michigan Supreme Court’s review over two years

later, with different appellate counsel.

Saylor’s claim fails in any event. Saylor argues that his plea should have “pinpoint[ed] a

date on one side or the other of the amendment’s effective date.” (Appellant Br. at 30.) But Saylor’s

plea does just that. It “pinpoints” April 26, 2008 as the end date for his offense. This post-dates

the amendment by almost two years, which means that Saylor abused his daughter at least once

post-amendment. And it’s not as though the statute requires Saylor to abuse his child more than

once before the minimum sentence applies. See Mich. Comp. Laws § 750.520b. Thus, regardless

of any ambiguities about other instances of Saylor’s abuse, that end date hooks him onto the

amended statute. As the trial court rightly concluded, “the last offense was two years after the

amendment, and therefore, [Saylor] was properly sentenced under the amended statute.” (R. 9-11,

Oct. 22, 2013 Order, PageID 239.) Saylor can’t point to any Supreme Court precedents to the

contrary.

C. Judicial Bias

The trial court judge and prosecutor here were brothers. Based on this, Saylor argues that

the trial-court proceedings were unconstitutionally biased. Importantly, the COA does not cover

this claim. And so Saylor urges us to expand the COA.

It’s true that we have the “inherent authority to expand sua sponte the scope of the COA to

encompass additional issues.” Mitchell v. MacLaren, 933 F.3d 526, 539 n.4 (6th Cir. 2019)

(quoting Howard v. United States, 485 F. App’x 125, 127-28 (6th Cir. 2012)). But there is no need.

That’s because Saylor’s claim is not “adequate to deserve encouragement to proceed further.”

Slack v. McDaniel, 529 U.S. 473, 484 (2000) (quoting Barefoot v. Estelle, 463 U.S. 880, 893 n.4

(1983)). Saylor cannot point to a single case that says kinship between a judge and prosecutor rises

12
No. 20-1834, Saylor v. Nagy

to the level of unconstitutional bias. On the contrary, the Supreme Court noted in Tumey v. Ohio

that “matters of kinship, personal bias, state policy, [and] remoteness of interest” do not generate

constitutional concerns. 273 U.S. 510, 523 (1927). Although these grounds may “present prudent

grounds for disqualification as a matter of common sense, ethics, or ‘legislative discretion,’” they

do not generally “rise to a constitutional level.” Railey v. Webb, 540 F.3d 393, 400 (6th Cir. 2008)

(quoting Tumey, 273 U.S. at 523). We need something more, such as “evidence of, or even an

allegation of, any actual bias.” Id. at 401. Here, we see nothing to that effect. Saylor’s judicial-

bias claim can’t clear AEDPA’s first hurdle, and we will not expand the COA.6

D. Ineffective Assistance of Appellate Counsel

Finally, Saylor argues that his direct-appeal counsel provided ineffective representation by

“unreasonably fail[ing] to raise Saylor’s ex post facto and judicial-bias claims.” (Appellant Br. at

53.)

The analysis here bifurcates. That’s because the COA covers just half of the claim: the

direct-appeal counsel’s failure to raise the ex post facto issue. To consider counsel’s failure to raise

judicial bias, we would need to expand the COA. But that’s unwarranted. As we explained above,

Saylor’s judicial-bias claim is not “adequate to deserve encouragement to proceed further.” Slack,

529 U.S. at 484 (quoting Barefoot, 463 U.S. at 893 n.4). And here, the bar is even higher. Not only

does the judicial-bias claim need to have some merit, it has to be so meritorious that Saylor’s

6
Saylor makes much of “[t]he fact that this Court previously granted a COA on” the judicial-bias
claim. (Appellant Br. at 44.) He argues that this constitutes binding law of the case. But this is
incorrect. We have clarified that the law-of-the-case doctrine “does not apply to a certificate of
appealability.” Keahey v. Marquis, 978 F.3d 474, 481 (6th Cir. 2020). In ruling on a COA
application, “the issuing judge does not make binding legal determinations.” Id.

13
No. 20-1834, Saylor v. Nagy

counsel provided constitutionally deficient representation by failing to raise it. Saylor can’t meet

the lower bar, so the higher bar is beyond his reach as well.

As for the other half of the claim (ex post facto), procedural default applies. Saylor raised

his ineffective-assistance-of-appellate-counsel claim for the first time in his May 28, 2013 motion

for relief from judgment. But Saylor never took that claim to the Michigan Supreme Court. So just

as Saylor defaulted on his ex post facto claim, he defaulted here as well. See Woolbright, 791 F.3d

at 631. And Saylor offers no argument on cause and prejudice.

Even without procedural default, Saylor’s claim falls short. Strickland’s two-step applies.

Saylor must show both deficient performance and prejudice. See Roe v. Flores-Ortega, 528 U.S.

470, 476-77 (2000) (applying Strickland to the appellate-counsel context). Put another way, Saylor

must establish that his “counsel’s performance was [1] ‘so manifestly ineffective that [2] defeat

was snatched from the hands of a probable victory.’” Jacobs v. Mohr, 265 F.3d 407, 418 (6th Cir.

2001) (quoting United States v. Morrow, 977 F.2d 222, 229 (6th Cir. 1992) (en banc)).

No case, from the Supreme Court or otherwise, requires counsel to raise all non-frivolous

claims on appeal. In fact, the available caselaw says precisely the opposite. “Strategic decisions of

counsel, including whether to raise some non-frivolous claims over others, fall well within the

range of professional competence.” Hand v. Houk, 871 F.3d 390, 411 (6th Cir. 2017). Indeed,

“[the] process of ‘winnowing out weaker arguments on appeal and focusing on’ those more likely

to prevail, far from being evidence of incompetence, is the hallmark of effective appellate

advocacy.” Smith v. Murray, 477 U.S. 527, 536 (1986) (quoting Jones v. Barnes, 463 U.S. 745,

751-52 (1983)). And we have established that Saylor’s ex post facto claim lacks merit; it fits the

bill for a “weaker argument.” So declining to raise the claim was a “[t]actical choice[]” “properly

14
No. 20-1834, Saylor v. Nagy

left to the sound professional judgment of counsel.” United States v. Perry, 908 F.2d 56, 59 (6th

Cir. 1990). The claim fails at Strickland’s first prong, and we need not reach prejudice.

Saylor advances one final argument. He claims “the state court’s decision was based on an

unreasonable determination of the facts.” (Appellant Br. at 58.) This, because it made no factual

findings on prejudice. But this argument fails. That’s because the trial court’s determination turned

on Strickland’s first prong, not prejudice. Even if the trial court addressed prejudice, it did so only

in the alternative.

IV.

And thus all four claims presented on appeal fall short. Saylor is not entitled to the

extraordinary remedy of habeas corpus. We AFFIRM.

15