3. Analysis: Legal Reasoning and Precedents Cited
A. Procedural Architecture: Postconviction Review Is Not a Second Appeal
The court began with the distinction between claims requiring factual development and claims that can be rejected without an evidentiary hearing. Under Morris v. State, 317 So. 3d 1054, 1071 (Fla. 2021), quoting Matthews v. State, 288 So. 3d 1050, 1060 (Fla. 2019), summary denial is appropriate when a claim is legally insufficient, procedurally barred, or refuted by the record. Henyard v. State, 992 So. 2d 120, 125 (Fla. 2008), supplied de novo review of such rulings.
Those authorities mattered because Craven sought hearings on several claims whose defects were already apparent from the existing record. The court did not regard an evidentiary hearing as an automatic consequence of alleging constitutional error.
Medina v. State, 573 So. 2d 293, 295 (Fla. 1990), provided the broader finality principle: postconviction proceedings cannot function as a second appeal, and an already litigated issue cannot ordinarily be revived through a different argumentative label. This principle controlled the jury-selection claim.
The direct appeal, Craven v. State, 310 So. 3d 891 (Fla. 2020), had already affirmed the conviction and sentence and resolved the relevant peremptory-strike dispute. The present opinion therefore examined not simply whether Craven had articulated a different legal theory, but whether he was attempting to relitigate the substance of a previously decided issue.
The background also referred to Spencer v. State, 615 So. 2d 688 (Fla. 1993), identifying the sentencing hearing at which Craven submitted additional mitigation before the trial court imposed sentence. That citation explained the original sentencing process; it was not an independent ground for denying postconviction relief.
B. Jury Selection: A Readback Would Not Cure Disparate Treatment
Craven claimed that trial counsel should have requested a readback of prospective juror Willie Ford’s answers. Ford, who is Black, had initially indicated that premeditation would warrant death, although he also qualified his answer by reference to the circumstances. The defense attempted a peremptory strike, and the State questioned whether the stated reason was genuine.
The court rejected the ineffective-assistance claim on both procedural and substantive grounds. On direct appeal, Craven had already argued that a readback would substantiate the proposed strike. Recasting that contention as counsel’s failure to request a readback did not create a materially different issue.
More importantly, Craven v. State had upheld the ruling because Craven treated Ford differently from similarly situated non-Black prospective jurors. The decisive issue was therefore not whether Ford had made the statement attributed to him, but whether that statement genuinely explained the defense’s treatment of him. Replaying his answer would not eliminate the comparative-juror problem.
The State had also already suggested a readback. Counsel was not deficient for failing to repeat a suggestion before the trial court, and there was no reasonable probability that counsel’s additional request would have changed the ruling.
This reasoning preserves an important distinction: a race-neutral explanation need not establish grounds for a for-cause strike, but it must be genuine. Accurate recollection of an answer does not, by itself, answer evidence of inconsistent treatment across jurors.
C. False Testimony: An Inaccurate Age Was Not Material to Sentencing
At trial, witnesses incorrectly stated that Craven was eleven months old when first removed from his biological mother’s home. He was actually approximately one year and ten months old. Craven alleged that the State knowingly presented false testimony.
Procedural bar
Under Allen v. State, 261 So. 3d 1255, 1286 (Fla. 2019), a claim alleging knowing use of false testimony is procedurally barred when its supporting facts were available on direct appeal. Craven argued that redaction of his birthdate had prevented an earlier challenge. The record, however, contained his date of birth multiple times and documented his November 1987 foster-care placement. Those materials supplied the necessary calculation.
The governing precedents
Giglio v. United States, 405 U.S. 150 (1972), and Napue v. Illinois, 360 U.S. 264 (1959), supplied the constitutional foundation for the claim. As applied in the opinion, relief required false testimony, prosecutorial knowledge of its falsity, and materiality.
Glossip v. Oklahoma, 604 U.S. 226, 246 (2025), explained that the prohibition encompasses both knowingly soliciting false testimony and knowingly allowing it to remain uncorrected. It also supplied the reasonable-likelihood formulation of materiality. Truehill v. State, 358 So. 3d 1167, 1183 (Fla. 2022), quoting Guzman v. State, 868 So. 2d 498, 505 (Fla. 2003), stated the corresponding three-part Florida test. Green v. State, 975 So. 2d 1090, 1106 (Fla. 2008), described materiality as any reasonable possibility that the false testimony could have affected the verdict.
Application
The court assumed falsity and prosecutorial knowledge for purposes of analysis; it did not find that prosecutors knowingly deceived the sentencer. It resolved the merits on materiality.
The jury and sentencing court already knew that Craven experienced substantial childhood instability, repeated removals, abuse, and neglect. The sentencing court expressly relied on that history and gave significant weight to his dysfunctional upbringing. Against that mitigation and the four aggravators, the corrected age did not create a reasonable possibility of a different recommendation or sentence.
The holding is record-specific. It does not establish that an eleven-month error in a child’s developmental history can never matter. It holds that this discrepancy did not materially alter the sentencing picture presented in this case.
D. Penalty-Phase Representation: Additional Detail Is Not Necessarily Constitutionally Significant Mitigation
1. The governing two-part test
Strickland v. Washington, 466 U.S. 668 (1984), governed the ineffective-assistance claims. Craven had to establish both objectively unreasonable performance and resulting prejudice. Courts must assess counsel’s conduct from the circumstances existing at the time, avoid hindsight, and presume reasonable professional assistance.
Hayward v. State, 183 So. 3d 286, 297 (Fla. 2015), supplied the capital-sentencing formulation: whether, absent counsel’s errors, there was a reasonable probability that the sentencer would have concluded that the balance of aggravation and mitigation did not warrant death. The reasonable-strategy presumption also drew on Michel v. Louisiana, 350 U.S. 91, 101 (1955), as quoted through the governing ineffective-assistance authorities.
The court could reject a claim upon an insufficient showing on either component. That point is particularly important for the incomplete-records claim: its rejection for lack of prejudice did not require a ruling that counsel’s investigation was adequate in every respect.
2. Failure to obtain the complete dependency file
Trial counsel possessed a ten-page dependency predisposition study, but not the complete records of Craven’s dependency proceedings and foster placements. Craven argued that the missing records would have disclosed his “true life history.”
The court compared the additional material with the mitigation actually presented. The jury had already heard extensive evidence of childhood abuse and neglect, placement instability, attachment disruptions, mental health difficulties, and trauma-related symptoms. The later records supplied greater detail and context, but largely supported themes already before the sentencers.
Covington v. State, 348 So. 3d 456, 471 (Fla. 2022), quoting Wheeler v. State, 124 So. 3d 865, 881 (Fla. 2013), supported the proposition that counsel is not ineffective for omitting cumulative evidence. Dufour v. State, 905 So. 2d 42, 61 (Fla. 2005), supplied the closely related prejudice principle: additional mitigation that does not substantially differ from the evidence presented may be insufficient to undermine confidence in the sentence.
The aggravating circumstances were central to the comparison. Hodges v. State, 55 So. 3d 515, 542 (Fla. 2010), identified the substantial weight associated with prior violent felony and heinous, atrocious, or cruel aggravation. Wade v. State, 41 So. 3d 857, 879 (Fla. 2010), similarly addressed heinous, atrocious, or cruel and cold, calculated, and premeditated aggravation. Those authorities reinforced the conclusion that the largely cumulative additions did not create a reasonable probability of life imprisonment.
The court also treated the separately framed failure to obtain the Volusia County dependency court file as part of this same claim. Different descriptions of the missing records did not require a separate prejudice analysis.
3. The mistaken removal age, later family contact, and repeated separations
Counsel’s incorrect calculation of Craven’s age produced the same factual discrepancy underlying the false-testimony claim. Postconviction experts explained that removal at twenty-two months could have different developmental consequences from removal at eleven months because of longer exposure and greater attachment formation.
The court nevertheless found neither constitutional deficiency nor prejudice. The experts described the general risks of longer exposure to a dysfunctional home, but did not explain specifically how the additional eleven months produced a worse outcome for Craven. Existing mitigation already conveyed serious childhood dysfunction.
Craven also argued that counsel should have emphasized his brief return to his biological mother’s home as a teenager and the open nature of his adoption. Those facts might have challenged the State expert’s account of limited post-adoption contact. But Craven did not adequately explain why their omission was constitutionally unreasonable or how the proposed impeachment probably would have changed the sentencing result.
The repeated-removal claim failed for a more direct reason: counsel had elicited testimony that multiple placements caused trauma and disrupted attachment during a critical developmental period. Banks v. State, 219 So. 3d 19, 29 (Fla. 2017), established that counsel cannot be deficient for failing to do what counsel actually did.
Additional dependency records would have permitted more individualized descriptions, including Craven’s distress when separated from a foster caregiver. The court regarded those additions as cumulative rather than a fundamentally different mitigation case. Hilton v. State, 326 So. 3d 640, 649 (Fla. 2021), reinforced that minor or cumulative mitigation does not undermine confidence in a penalty-phase outcome when aggravation is substantial.
4. Prenatal exposure to drugs and alcohol
The proposed prenatal-exposure theory faced a distinct problem: uncertain proof. A summary of an interview with Craven’s alleged biological father reported heavy drinking during pregnancy and possible crack-cocaine use. Whether trial counsel received that summary was disputed and remained unresolved.
A postconviction expert diagnosed Alcohol-Related Neurodevelopmental Disorder, relying on probable prenatal exposure and asserted developmental and cognitive difficulties. The court found the supporting record inadequate. The complete dependency file did not document substance use during pregnancy; the records available at trial placed significant substance abuse later; testing showed average or high-average intellectual functioning; and an evaluation for brain damage had yielded negative results.
Maharaj v. State, 778 So. 2d 944, 951 (Fla. 2000), supplied the controlling warning that postconviction relief cannot rest on speculation or possibility. On this record, the court concluded that the proposed pregnancy evidence depended on speculation or inadmissible double hearsay, rather than competent proof supporting the mitigation theory.
The court also found no prejudice. Even assuming admissible evidence of prenatal exposure could have been presented, the postconviction evidence did not establish an adverse effect on Craven’s cognitive or developmental functioning sufficient to create a reasonable probability of a different sentence.
This was not a categorical rejection of prenatal-exposure mitigation or the diagnosis itself. The decision turned on the evidentiary weaknesses of this particular claim.
5. Complex trauma and expert selection
At trial, the defense psychologist described cumulative trauma alongside multiple mental health conditions, childhood treatment, medication, and hospitalizations. During postconviction proceedings, she agreed that “complex trauma” was an appropriate description, but explained that her original testimony had essentially conveyed that condition.
The court therefore rejected the idea that using a different diagnostic label established ineffective assistance. It also found no reasonable probability that additional testimony about the lack of consistent, long-term trauma-specific treatment would have changed the sentence.
The expert-selection claim required a related but separate inquiry. Counsel selected the trial psychologist because she knew Craven, he did not respond well to male evaluators, and counsel had previously found her effective before a jury.
Hinton v. Alabama, 571 U.S. 263, 275 (2014), supported treating expert selection as a paradigmatic strategic choice when made after appropriate investigation of the facts and law. The qualification matters: the opinion does not make every expert-selection decision immune from review.
Covington v. State, 348 So. 3d at 468, quoting Diaz v. State, 132 So. 3d 93, 113 (Fla. 2013), supplied the complementary principle that finding a later expert with more favorable conclusions does not itself establish deficient trial representation. Craven also failed to identify material testimony that a child-development specialist could have provided but the trial psychologist could not. Much of the argument therefore returned to the already rejected incomplete-records claim.
6. Consultation with Dr. Garbarino and cross-examination of the State expert
Craven argued that counsel should have consulted Dr. James Garbarino to challenge the State expert’s assertion that research participants facing murder charges generally had childhood histories worse than Craven’s. Postconviction testimony indicated that the average research participant had an Adverse Childhood Experiences score of seven, while another expert assigned Craven a score of eight.
Consultation was not shown to be necessary because the relevant research information was available in publications. On appeal, Craven limited this claim to consultation for cross-examination; the court accordingly did not decide whether counsel should have called Dr. Garbarino as a witness.
The separate cross-examination claim failed because counsel articulated a strategic explanation. Her principal theory emphasized Craven’s moral code as the trigger for the murder. She also wanted the jury to hear the State expert minimize Craven’s childhood adversity, believing that contrast with the defense evidence would damage his credibility.
Whether that approach was ideal was not the constitutional question. Craven had to overcome the presumption of reasonable strategy and establish a reasonable probability of a different sentence. The court found that he did neither.
7. Cumulative error
Whitton v. State, 161 So. 3d 314, 333 (Fla. 2014), supported rejection of the cumulative-error claim because the individual claims did not establish the errors or ineffective assistance necessary for aggregate relief.
The distinction is between acknowledged factual imperfections and established constitutional errors. Counsel’s age miscalculation was real, but the court held that it did not satisfy the constitutional standard. The cumulative-error doctrine did not convert the rejected claims into a basis for relief.
E. Newly Discovered Neurological Evidence: Discovery Concerns Facts, Not Merely a Later Evaluation
1. The one-year limitation and due diligence
Florida Rule of Criminal Procedure 3.851 generally requires a capital postconviction motion within one year after finality. The denial of certiorari in Craven v. Florida, 142 S. Ct. 199 (2021), marked finality for that purpose, making the ordinary deadline expire in 2022. The certiorari disposition served as a timing event, not as substantive approval of the conviction’s merits.
The relevant exception applies when the underlying facts were unknown and could not have been discovered through due diligence. A later expert report therefore does not automatically restart the limitations period. The defendant must identify genuinely new facts and establish timely presentation.
2. The seizure history was already known
Craven characterized a February 2025 neurological evaluation as the first confirmation of a “bona fide” seizure disorder. But the evaluator’s own report referred to a diagnosis more than two decades earlier. Craven reported recurrent seizures over approximately thirty years and another diagnosis in May 2024.
The court treated the report as corroborating an existing history rather than discovering a previously unknown condition. Even using the May 2024 diagnosis as the starting point, the claim had not been filed within one year. The seizure history had also been presented at trial, supporting the procedural bar.
The asserted new relationship between childhood trauma and seizures failed independently. The neurological report discussed head injuries, substance use, and family history, but did not supply the claimed opinion that psychological childhood trauma caused the seizure disorder. Trial testimony had already discussed an early EEG and possible biological and psychological explanations for Craven’s symptoms.
The court further concluded that the proposed connection would not probably produce a life sentence. Both the traumatic childhood and the seizure history were already known to the jury, and the aggravation remained substantial.
3. Risk of CTE was not a tentative diagnosis
The report stated that repeated head injuries placed Craven at high risk of developing CTE as he aged. It did not tentatively or definitively diagnose CTE. The court rejected the successive motion’s effort to equate these different propositions.
This distinction was decisive: vulnerability to a future condition does not establish that the condition presently exists. The court did not announce that CTE can never support postconviction relief.
There was also an independent timeliness defect. Even assuming a diagnosis, Craven did not credibly establish when the asserted condition became diagnostically discoverable in a living person and therefore when his claim could first have been raised.
Damren v. State, 397 So. 3d 607, 613 (Fla. 2023), quoting Mungin v. State, 320 So. 3d 624, 626 (Fla. 2020), placed the burden of establishing timeliness on the defendant. Those decisions supported denial where the movant failed to offer a credible discovery date.
The resulting rule is procedural and evidentiary: an asserted scientific development must be accurately described, supported by the expert material, and connected to a demonstrably timely claim. Calling an evaluation “new” does not establish that its factual subject was previously undiscoverable.
F. Execution-Method Challenges Require a Pleaded Alternative
Craven alleged that lethal injection could provoke and conceal a seizure and that electrocution could aggravate his condition. He challenged sections 922.10 and 922.105(3), Florida Statutes (2025), and requested a hearing to determine the safest execution method.
Baze v. Rees, 553 U.S. 35, 50 (2008), supplied the “objectively intolerable risk of harm” language on which he relied. But that phrase did not state the entire pleading burden.
Asay v. State, 224 So. 3d 695, 701 (Fla. 2017), applying Glossip v. Gross, 576 U.S. 863, 877 (2015), required both a substantial and imminent risk of serious suffering and identification of a known, available alternative method involving a significantly less severe risk of pain.
Craven identified no qualifying alternative. Electrocution did not fill that role because he also alleged that it created an intolerable risk. A request for a hearing to discover which method might be suitable did not satisfy the antecedent pleading obligation.
Valle v. State, 70 So. 3d 530, 549-50 (Fla. 2011), supported denial of speculative and insufficiently pleaded execution claims. The court therefore denied relief without deciding that either challenged method was medically risk-free for Craven.
G. Records and Testing: Independent Grounds and Particularized Need
1. Prison video recordings
Craven sought recordings of a January 2025 inmate altercation and a later body-camera conversation. The Department of Corrections invoked both facility-security protections under section 119.071(3), Florida Statutes (2025), and health-information protections under section 945.10(1). It also disputed the recordings’ connection to a colorable postconviction claim.
On appeal, Craven challenged the health-information and relevance grounds, but not the independent security exemption. That omission was dispositive: even a successful challenge to the other grounds would leave an unchallenged basis supporting denial.
The Supreme Court therefore did not decide the disputed application of the health-information exemption or the recordings’ potential relevance. The opinion should not be read as a broad ruling that all prison video is categorically exempt, or that the asserted health-privacy rationale was necessarily correct.
2. Transport for imaging and neurological testing
Craven requested transport for a blood-flow PET scan, a non-contrast MRI with diffusion tensor imaging, a sleep-deprived EEG, and a prolonged EEG. He relied on Rogers v. State, 783 So. 2d 980, 999 (Fla. 2001), to argue that testing was necessary for experts to evaluate brain functioning and damage.
The court found his appellate explanation conclusory. Heath v. State, 3 So. 3d 1017, 1029 n.8 (Fla. 2009), established that vague and conclusory appellate allegations do not warrant relief. An expert’s recommendation to investigate possible abnormalities did not, without a more particularized explanation, establish entitlement to transport and testing.
Prejudice was independently absent. Even if the tests revealed abnormalities or epileptic activity, they would not change the longstanding nature of the seizure history or cure the untimeliness and procedural bar. The record also did not support Craven’s assertion that a PET scan could confirm his alleged CTE diagnosis.
The court did not establish a universal scientific rule about imaging or CTE. It held that the asserted diagnostic capability was unsupported in this record and that the requested testing was not shown to advance a viable claim.
H. Habeas Corpus: Appellate Counsel Need Not Raise Meritless or Nonfundamental Issues
1. Aggravating circumstances and the indictment
Craven argued that appellate counsel should have challenged the State’s failure to present aggravating circumstances to the grand jury and include them in the indictment. He invoked federal due process and article I, section 15, of the Florida Constitution.
Sanchez-Torres v. State, 322 So. 3d 15, 23-24 (Fla. 2020), had already rejected the same substantive argument. Craven attempted to distinguish that decision because of its manifest-injustice framing, lack of preservation, and the claim’s earlier novelty. The court found those distinctions unpersuasive: it had treated the earlier claim as ineffective assistance of appellate counsel, and preservation could not make a meritless argument successful.
The opinion identified a settled line of authority rejecting the requirement that the charging instrument allege death-qualifying aggravators: Cruz v. State, 320 So. 3d 695, 730 (Fla. 2021); Hall v. State, 246 So. 3d 210, 217 (Fla. 2018); Pham v. State, 70 So. 3d 485, 496 (Fla. 2011); Miller v. State, 42 So. 3d 204, 215 (Fla. 2010); and Rogers v. State, 957 So. 2d 538, 554 (Fla. 2007). Their collective significance was to demonstrate that the proposed appellate issue was foreclosed rather than overlooked.
Armstrong v. State, 862 So. 2d 705, 718 (Fla. 2003), supplied the dispositive representation principle: appellate counsel is not ineffective for failing to raise a meritless issue. Preservation does not alter that conclusion.
2. Contextual testimony was not victim-impact evidence
Craven also challenged guilt-phase testimony that Anderson was called “Rev,” preached the Bible, led Bible study, had been ordained in prison, and was pleasant. The court noted that the transcript did not support the additional characterization “easygoing” asserted in the petition.
Under section 921.141(8), Florida Statutes (2017), victim-impact evidence is designed to show the victim’s individual uniqueness and the community’s resulting loss. The challenged testimony instead identified Anderson, explained the circumstances of the killing, and established part of Craven’s motive. Favorable information about a victim is not automatically victim-impact evidence; its function matters.
The claim also lacked preservation. Chamberlain v. State, 881 So. 2d 1087, 1100 (Fla. 2004), required the specific legal ground to be raised at trial. No relevant objection was made.
Damas v. State, 423 So. 3d 811, 824 (Fla. 2025), explained that appellate counsel generally cannot be ineffective for omitting an unpreserved issue unless it amounts to fundamental error. Calhoun v. State, 312 So. 3d 826, 855 (Fla. 2019), defined that demanding category as error reaching the trial’s validity so deeply that the guilty verdict could not have been obtained without it.
Given the confessions and other evidence, the descriptive testimony could not satisfy that standard. The claim therefore failed because the evidence was not victim-impact evidence, the objection was unpreserved, and its admission was not fundamental error.