The opinion
In the
United States Court of Appeals
For the Seventh Circuit
____________________
No. 22-2393
TYLER A. GONZALES, formerly known as Tyler A. Montour,
Petitioner-Appellant,
v.
CHERYL EPLETT, Warden,
Respondent-Appellee.
____________________
Appeal from the United States District Court for the
Eastern District of Wisconsin.
No. 1:19-cv-01604-WCG — William C. Griesbach, Judge.
____________________
ARGUED MARCH 31, 2023 — DECIDED AUGUST 9, 2023
____________________
Before EASTERBROOK, RIPPLE, and WOOD, Circuit Judges.
WOOD, Circuit Judge. Tyler Gonzales 1 was convicted in
2015 of charges arising out of a shooting in a parking lot. He
is currently serving a 25-year prison sentence, which will be
followed by 15 years’ extended supervision. Believing that he
1 Throughout most of the proceedings, petitioner was using the name
Tyler A. Montour. He changed his name at some point, however, and is
now known as Tyler A. Gonzales. We use his current name.
2 No. 22-2393
received constitutionally ineffective assistance of counsel at
his trial, he has turned to federal court for a writ of habeas cor-
pus. The district court concluded, however, that Gonzales has
not satisfied the stringent requirements for such relief, and so
it denied his petition. This is one of those cases in which the
standard of review matters. We are deeply troubled by the
performance of defense counsel. But 28 U.S.C. § 2254 requires
us to defer to a state court’s decision unless it is not only
wrong, but unreasonable. We conclude that the state court did
not stray beyond that extreme limit, and so we affirm.
I
The events underlying this case unfolded during the early
morning hours of June 12, 2015. Petitioner Gonzales had got-
ten into an altercation with Adrian Valadez and Blake Kruiz-
enga at the Hawk’s Nest Bar. After a heated argument, Gon-
zales left the bar and got into a car with his brother-in-law,
Pedro Gonzalez. As Pedro Gonzalez drove away, Gonzales
shot from the passenger window of the car toward Kruizenga
and Valadez, who were standing in the parking lot. Gonzales
fired the gun about six or seven times and hit Kruizenga in
the leg.
Charged under state law with attempted first-degree in-
tentional homicide and being a felon in possession of a fire-
arm, Gonzales was offered an opportunity to plead guilty to
recklessly endangering safety and unlawful possession of a
firearm for a recommended ten-year sentence of confinement.
Under Wisconsin law, recklessly endangering safety is a
lesser-included offense of attempted first-degree intentional
homicide, meaning that a defendant who commits attempted
intentional homicide necessarily commits reckless
No. 22-2393 3
endangerment as well, but the lesser charge carries a milder
punishment.
Attempted first-degree intentional homicide requires the
intent to cause the death of another human being and steps
toward the commission of that crime. See Wis. Stat. § 940.01
(defining first-degree intentional homicide); Wis. Stat.
§ 939.32 (defining attempt). To show intent, the prosecution
must prove that the defendant “has a purpose to do the thing
or cause the result specified, or is aware that his or her con-
duct is practically certain to cause that result.” Wis. Stat.
§ 939.23. First-degree recklessly endangering safety is defined
as “recklessly endanger[ing] another’s safety under circum-
stances which show utter disregard for human life.” Wis. Stat.
§ 941.30. Attempted first-degree intentional homicide carries
a maximum prison sentence of 40 years, as compared with
first-degree recklessly endangering safety, for which the sen-
tence is capped at 7.5 years. The maximum sentence for un-
lawful possession of a firearm is five years’ confinement.
After conferring with his defense counsel, Melissa Frost,
Gonzales rejected the plea deal and requested a speedy trial.
Frost advised Gonzales that she believed they should seek a
full acquittal. Her assessment rested heavily on her prediction
that the state was going to have a hard time getting the central
witnesses, Valadez and Kruizenga, to testify, particularly if
Frost and Gonzales succeeded in securing an early trial date.
Kruizenga had absconded from probation and the state was
still looking for him. All the witnesses had lengthy felony rec-
ords, and their accounts of the evening varied. They were
drunk and there were inconsistencies in their stories about
where they were standing, the color of the car, how many
shots were fired, and whether there was a third passenger in
4 No. 22-2393
the car. Frost believed she could capitalize on witness unavail-
ability and the impeachment fodder to create reasonable
doubt about whether Gonzales was the shooter.
It turned out that Frost had been far too optimistic. At trial,
it quickly became clear that all the state’s witnesses had been
located, were cooperating, and were going to testify that Gon-
zales was the shooter. Worse yet, Pedro Gonzalez had been
offered immunity and was prepared to testify that he drove
the car while Gonzales shot at Valadez and Kruizenga. The
state’s case was thus impressive, featuring three eyewitnesses,
all of whom would identify Gonzales as the shooter.
Seeing the writing on the wall at the end of the second day
of trial, Gonzales confidentially admitted to Frost that he was
the shooter. He asked her if he should testify and explain that
he was not trying to hit anyone and was just trying to scare
Valadez and Kruizenga. Frost advised Gonzales not to do
that. By that point in the trial, she thought that Gonzales’s tes-
timony would guarantee conviction; he would be caught
dead to rights on the unlawful possession count and, even if
he managed to undermine the state’s showing of intent to
commit attempted intentional homicide, he very well could
face conviction on that count as well. Frost had reserved her
opening statement until after the state’s case-in-chief, but she
did not make any adjustments to her presentation of the case,
despite Gonzales’s private confession to her. She proceeded
with their “all-or-nothing” strategy, pursuing acquittal rather
than trying to focus the jury on the reckless-endangerment
count. The gamble did not work: the jury convicted Gonzales
of the more serious crime.
Frost expressed discomfort with her strategy as early as
sentencing. She described the trial as bizarre and felt
No. 22-2393 5
responsible for not pursuing the lesser-included offense. And
our review of the record indicates that there is a great deal to
criticize in her performance. Her cross-examination of the
state’s witnesses failed to bring out material inconsistencies in
the testimony; worse, it invited the state’s witnesses to reiter-
ate their testimony that Gonzales was armed and shooting to-
ward them. In addition, rather than coming up with a revised
trial plan in the evenings, she wasted time reviewing jail calls
to see if there was evidence of a side deal or an undisclosed
police report. Her cross-examination of Pedro Gonzalez also
failed to shake his story.
After sentencing, the court appointed a new lawyer to rep-
resent Gonzales, and new counsel filed for post-conviction re-
lief as permitted by Wisconsin law, Wis. Stat. § 974.02, raising
a claim of ineffective assistance of counsel. See Strickland v.
Washington, 466 U.S. 668 (1984). The Wisconsin trial court held
an evidentiary hearing at which it examined Frost’s perfor-
mance. Both Gonzales and Frost testified at the hearing. Frost
fell on her sword. She testified that it “never even crossed
[her] mind” to argue for the lesser-included offense, that she
had tunnel vision about pursuing the acquittal, and that she
had felt no need to adjust her trial strategy even when it be-
came clear that the state’s witnesses were all available. Gon-
zales testified that he and Frost never seriously discussed the
lesser-included offense.
It also turned out that three jurors told Frost after the trial
that they did not understand the difference between at-
tempted first-degree intentional homicide and first-degree
recklessly endangering safety. They disclosed that the jury
just picked attempted intentional homicide for the conviction
because they knew Gonzales had been the one who pulled the
6 No. 22-2393
trigger. Gonzales wound up with a sentence of 25 years in
prison, to be followed by 15 years’ extended supervision. Of
that, 20 years was for the attempted first-degree intentional
homicide, twice what the state had offered before trial, and
nearly three times the statutory maximum Gonzales would
have faced if the jury had convicted on the lesser-included of-
fense.
The Wisconsin trial court concluded that Frost’s perfor-
mance, taken as a whole, did not fall below the constitution-
ally permissible minimum. Pursuing acquittal was reasona-
ble, it concluded, based on the character of the eyewitnesses,
and it thought that Frost’s decision not to shift her strategy
mid-trial fell within the boundaries of acceptable legal strat-
egy. It agreed with Frost that Gonzales’s suggested testimony
would have guaranteed a conviction. The court also sug-
gested that it would have been difficult for Frost to argue both
for acquittal and, in the alternative, for a conviction only on
the lesser-included offense. Even though inconsistent de-
fenses are not strictly forbidden, the court observed that they
are often incredible to a jury. The court also briefly addressed
prejudice and concluded that there was sufficient evidence to
convict Gonzales of attempted intentional homicide, and so
the outcome would not have changed even if Frost had ad-
justed her approach.
The Wisconsin appellate court affirmed the trial court’s
bottom line, but it rested its opinion solely on Frost’s perfor-
mance, declining to reach the issue of prejudice. Gonzales’s
lawyer then filed a no-merit petition with the Supreme Court
of Wisconsin pursuant to Wisconsin’s Rules of Appellate Pro-
cedure. See Wis. Stat. § 809.32(4). Gonzales personally did not
avail himself of the option of filing a supplemental petition.
No. 22-2393 7
The state supreme court denied the no-merit petition in a
standard order.
Gonzales then turned to federal court with a petition un-
der 28 U.S.C. § 2254 for a writ of habeas corpus. The state
moved to dismiss the petition for failure to exhaust his state
remedies. It contended that Gonzales’s failure to file a supple-
mental petition in the state supreme court was fatal to his re-
quest for habeas corpus relief. The district court denied the mo-
tion to dismiss, but it ultimately ruled in the state’s favor on
the ground that the state appellate court (the last state tribunal
to issue a fully reasoned opinion, see Wilson v. Sellers, 138 S.
Ct. 1188, 1192 (2018)), had not been unreasonable when it
found that Frost’s performance was not constitutionally defi-
cient. It also expressed skepticism that Gonzales could
demonstrate prejudice. Nonetheless, it found that reasonable
jurists could reach a contrary decision, and so it issued a cer-
tificate of appealability. This appeal followed.
II
In this court, the state begins by reiterating its exhaustion
argument, which if accepted would lead to a finding of pro-
cedural default for Gonzales. To reach the merits of Gonza-
les’s petition, we must ensure that he fairly presented the
claim “through one complete round of review in state court.”
Brown v. Eplett, 48 F.4th 543, 552 (7th Cir. 2022) (citing Baldwin
v. Reese, 541 U.S. 27, 29 (2004)). We assess de novo the district
court’s ruling on procedural default. Hicks v. Hepp, 871 F.3d
513, 530 (7th Cir. 2017).
The state argues that Gonzales defaulted by failing to com-
ply with the petition procedure established by Wisconsin law.
See Wis. Stat. § 809.32. If an attorney concludes that a direct
8 No. 22-2393
appeal to the Supreme Court of Wisconsin lacks “any argua-
ble merit within the meaning of Anders v. California, 386 U.S.
738 (1967),” the attorney must file a no-merit petition. That
petition must include a statement of the case, and counsel
must append the lower court opinions. If the defendant disa-
grees with that assessment and believes the appeal has merit,
he or she must then file a supplemental petition stating the
issues for review and an argument for why review is proper.
See Wis. Stat. §§ 809.32(1) & (4). Gonzales did not file a sup-
plemental petition; instead, he relied on his attorney’s no-
merit filing.
While this does not strictly comply with Wisconsin proce-
dural rules, the failure to file a supplemental petition does not
automatically doom a habeas corpus petition. The record as a
whole is what matters. The federal court should determine
“whether the petitioner has fairly presented his federal claim
to the state court,” looking at factors such as (1) the presence
of a federal constitutional analysis; (2) the citation to state
court cases that apply constitutional analysis; (3) the framing
of the claim in accordance with “a specific constitutional
right”; and (4) the use of a fact pattern “that is well within the
mainstream of constitutional litigation.” Brown, 48 F.4th at
552. “All four factors need not be present to avoid default … .”
Id. (quoting Whatley v. Zatecky, 833 F.3d 762, 771 (7th Cir.
2016)).
These considerations weigh in Gonzales’s favor. Even
without a supplemental petition, the state supreme court had
a comprehensive account of the case. The no-merit petition
filed by Gonzales’s attorney alerted the court to the potential
constitutional arguments in the case and thus did what an An-
ders-type brief is intended to do. The statement of facts
No. 22-2393 9
explained both the deficiencies in Frost’s performance and the
prejudice Gonzales faced as a result. The petition also cited
the relevant state-court cases, including State v. Machner, 92
Wis. 2d 797 (Wis. Ct. App. 1979), which establishes the Wis-
consin post-trial procedure for dealing with ineffective assis-
tance of counsel claims, and State v. Maloney, 2005 WI 74, a
case from the Supreme Court of Wisconsin that follows the
Strickland standard. Even though the no-merit petition did not
directly engage in a federal constitutional analysis, not every
factor needs to be present to preserve a petitioner’s claim. We
have considered the state’s assertions otherwise, including its
analogies to other cases involving Wisconsin no-merit peti-
tions, and find none persuasive. We thus reject the proce-
dural-default argument and move to the merits of Gonzales’s
petition.
III
The standard of review for a habeas corpus petition is estab-
lished by the Antiterrorism and Effective Death Penalty Act of
1996 (AEDPA). We may issue the writ only if the state-court
proceedings “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 “resulted in a decision that was based on
an unreasonable determination of the facts in light of the evi-
dence presented in the State court proceeding.” 28 U.S.C. §
2254(d). Though we must defer to any reasonable state court
decision, our review of the district court’s decision is de novo.
See Bell v. Hepp, 70 F.4th 385, 389 (7th Cir. 2023). And since
“AEDPA deference only applies to issues that the last rea-
soned state court decision reached on the merits,” we conduct
10 No. 22-2393
a de novo review of issues that were not reached on the merits.
Dunn v. Jess, 981 F.3d 582, 591 (7th Cir. 2020).
But before we turn to AEDPA, it is important to under-
stand Gonzales’s underlying claim. The Sixth Amendment
right to the assistance of counsel is a right to effective assis-
tance. Strickland, 466 U.S. at 686. In order to show ineffective-
ness, the defendant must prove that (1) “counsel’s represen-
tation fell below an objective standard of reasonableness …
under prevailing professional norms,” and (2) “the deficient
performance prejudiced the defense.” Id. at 687–88. Even
without AEDPA, this is a tough standard to meet, given the
Supreme Court’s admonition that “[a] court considering a
claim of ineffective assistance must apply a ‘strong presump-
tion’ that counsel’s representation was within the ‘wide
range’ of reasonable professional assistance. Harrington v.
Richter, 562 U.S. 86, 104 (2011) (quoting Strickland, 466 U.S. at
689). Layering AEDPA on top of that standard makes it even
harder to prevail on this type of claim.
The central question in this case is whether Frost provided
constitutionally ineffective assistance, taking her performance
as a whole. Gonzales argues that Frost exhibited plan-contin-
uation bias, or “tunnel vision”; she remained doggedly fo-
cused on acquittal even after it became impossible, never up-
dating her understanding of the evidentiary landscape or
adapting to the realities of the case’s developments, and her
cross-examinations were a disaster. Gonzales contends that
Frost’s decisions were unreasoned, rather than the product of
intentional strategy. This distinction is significant; the Su-
preme Court has told us to defer to an advocate’s “strategic
choices about which lines of defense to pursue,” but only if
those choices are “based on professional judgment” and
No. 22-2393 11
“assumptions [that] are reasonable.” Id. at 681 (internal quo-
tations omitted).
To evaluate Frost’s performance and her failure to pivot,
it is helpful to examine her decisions at three critical moments:
1) before trial, when she advised Gonzales against taking the
plea offer; 2) mid-trial, when she continued to pursue acquit-
tal even though she knew that all the state’s witnesses were
available, and she also had Gonzales’s confidential confes-
sion; and 3) at closing argument, when she did not argue for
the lesser-included offense.
For the first point, we now know in hindsight that it was a
mistake for Gonzales and Frost to pass on the plea deal that
was offered. But Frost’s choices at that time fell within the
wide range of professional judgment and reasonable assump-
tions. Frost considered the availability of the eyewitnesses,
their credibility, the inconsistencies in their accounts of the
shooting, and other available impeachment fodder such as the
eyewitnesses’ lengthy criminal records. As the district court
noted, Frost also accounted for the “prosecutor’s trial skills
and his potential for alienating the jury.” Her choice to pro-
ceed to trial and pursue full acquittal thus passed muster un-
der the applicable deferential standard.
Frost’s choices become less defensible as we move along
the timeline. As of mid-trial, she continued to pursue acquittal
even though she knew by then that the state’s case was much
stronger than she had anticipated. Her expectation that the
key eyewitnesses would be unavailable or impeachable was
foiled; all eyewitnesses appeared in court and named Gonza-
les as the shooter, including Gonzales’s own brother-in-law.
Gonzales himself sensed that things were not going well, and
so he offered his own testimony, which would have admitted
12 No. 22-2393
to reckless endangerment while undermining his criminal in-
tent for homicide. Since Frost had reserved her opening state-
ment, she was free to incorporate these changes into her
presentation to the jury. She had managed to elicit evidence
that would have helped her build a case for the lesser-in-
cluded offense. There were statements from a ballistics officer
that bullets were recovered from targets that were low to the
ground, and Kruizenga was hit low to the ground, just
slightly above his ankle. Another testifying officer explained
that someone firing a gun with the intent to kill would aim at
“center mass.” Frost could have emphasized this evidence to
illustrate that Gonzales was aiming low, with no intent to kill.
But that pivot would have been difficult, and we must re-
sist the lure of hindsight. Frost reasonably could have con-
cluded, in the exercise of her professional judgment, that such
a pivot would have been dangerous for Gonzales. It would
have guaranteed his conviction on at least two counts—reck-
less endangerment and unlawful possession of a firearm. And
through cross-examination she had brought out problems
with witness credibility and inconsistencies in eyewitness ac-
counts. As the state pointed out at oral argument, her cross-
examinations elicited several significant admissions from the
state’s eyewitnesses. Those admissions included Pedro Gon-
zalez’s concession that he lied to police when they questioned
him the day after the shooting, his forfeiture of an unlawfully
owned gun, and his deletion of text messages between him
and petitioner Gonzales from the night of the shooting. Frost
also elicited the facts that Pedro Gonzalez was offered im-
munity for his testimony, and that he had a motive to harm
Kruizenga and Valadez in retaliation for their involvement in
a home invasion at his house. Frost’s cross-examinations also
brought out Valadez’s admission that he told police that
No. 22-2393 13
Pedro Gonzalez was a passenger in the car, not the driver.
Though ultimately ineffective, these cross-examinations
aligned with Frost’s acquittal strategy by creating motive and
opportunity for Pedro Gonzalez, rather than petitioner Gon-
zales, to be the shooter. In sum, we can only speculate
whether Frost realistically could have shifted her strategy at
that point. She had only bad choices, and she may have cho-
sen the best of that bad lot.
The final stage, the closing argument, is the most vulnera-
ble part of Frost’s performance. Closing arguments can be sig-
nificant game changers. Indeed, “no aspect of [partisan] ad-
vocacy could be more important than the opportunity finally
to marshal the evidence for each side before submission of the
case to judgment.” Herring v. New York, 422 U.S. 853, 862
(1975). And we know that three jurors told Frost after the trial
that they did not understand the difference between at-
tempted intentional homicide and reckless endangerment
during their deliberations. Had Frost been able to clarify the
difference, it might have had an effect.
But the simple reality of the situation is that the state had
put on a strong case and boxed Frost into a difficult position.
The risks of conceding that Gonzales was the shooter were
huge, given the evidence supporting the attempted homicide
charge, including Kruizenga’s testimony that he saw straight
down the barrel of Gonzales’s gun. Even more damning,
Kruizenga was actually hit by a bullet. And as the state trial
court emphasized, juries are often skeptical about incon-
sistent defenses, and so any argument in the alternative about
the lesser-included offense might have weakened Gonzales’s
case. If we give Frost every benefit of the doubt, it is possible
that there is just enough to support her decisions at each turn.
14 No. 22-2393
Nonetheless, Frost’s overall performance is hard to justify,
and we are greatly troubled that the idea of strategic adapta-
tion to the state’s actual case “never even crossed her mind.”
Gonzales also makes a good point about plan-continuation
bias. An attorney’s choice rigidly to pursue a losing strategy
certainly can support an ineffective assistance of counsel
claim. If we were writing on a clean slate, this would be a close
case.
But we are not the primary decisionmakers. This is a habeas
corpus action, and our role is severely limited by AEDPA. For
ineffective assistance of counsel arguments, the Supreme
Court has said that the AEDPA layer makes our assessment of
counsel’s performance (and of prejudice, if that were at issue)
doubly deferential. See Richter, 562 U.S. at 105. First, as we al-
ready have noted, we presume that “counsel’s representation
was within the ‘wide range’ of reasonable professional assis-
tance.” Id. at 104. Second, we must defer to the state court’s
assessment of counsel’s performance unless “there was an er-
ror well understood and comprehended in existing law be-
yond any possibility for fairminded disagreement.” Id. at 103.
Gonzales cannot clear the second of those hurdles. Even if we
might have found that this is one of the unusual cases in
which counsel’s performance was constitutionally deficient,
we cannot say that there is no possibility for fairminded disa-
greement on that point.
It is worth noting, as we conclude, that the state trial court
(whose findings strongly influenced the state appellate court)
seems to have reached its decision in large part because of the
strength of the state’s case when all was said and done. It
thought that there was little Frost could have done, in the face
of that evidence. As we already have discussed, the record
No. 22-2393 15
showed, with little contradiction, that Gonzales shot in the di-
rection of the eyewitnesses. This undermines his insistence
that he was shooting at the ground and not trying to hit any-
one. And the state trial court reasonably concluded that the
act of shooting at a person supports a conviction for at-
tempted first-degree intentional homicide. The court put the
point bluntly, using language that mirrors the Wisconsin def-
inition of criminal intent: “Anyone with half a brain knows
that if you fire a gun in the direction of somebody, their death
could occur, that you are aware that their death could occur
and is probable to occur.”
Given the standards that bind us, we conclude that Gon-
zales has not advanced a successful claim for habeas corpus re-
lief based on ineffective assistance of counsel. Though Gonza-
les marshals strong arguments, we cannot say that the state
appellate court unreasonably applied Strickland or relied on
unreasonable determinations of fact.
IV
We AFFIRM the district court’s denial of Gonzales’s petition
for a writ of habeas corpus.