Uttecht is technically an AEDPA decision - the majority reversed the granting of the writ because the trial court's exclusion of Juror Z was not contrary to, or an unreasonable application of the Court's Witherspoon - Witt jurisprudence. However, this was really a case (for the majority anyway) about deference to the trial court:
Capital defendants have the right to be sentenced by an impartial jury. The State may not infringe this right by eliminating from the venire those whose scruples against the death penalty would not substantially impair the performance of their duties. Courts reviewing claims of Witherspoon-Witt error, however, especially federal courts considering habeas petitions, owe deference to the trial court, which is in a superior position to determine the demeanor and qualifications of a potential juror. The Court of Appeals neglected to accord this deference. And on this record it was error to find that Juror Z was not substantially impaired.The main dissent, written by Justice Stevens, had a different take. This quote is long, but worth reading (as is the entire dissenting opinion):
Today, the Court has fundamentally redefined—or maybe just misunderstood—the meaning of "substantially impaired," and, in doing so, has gotten it horribly backwards. It appears to be under the impression that trial courts should be encouraging the inclusion of jurors who will impose the death penalty rather than only ensuring the exclusion of those who say that, in all circumstances, they cannot. The Court emphasizes that "the State has a strong interest in having jurors who are able to apply capital punishment within the framework state law prescribes." Ante, at 6. But that does not and cannot mean that jurors must be willing to impose a death sentence in every situation in which a defendant is eligible for that sanction. That is exactly the outcome we aimed to protect against in developing the standard that, contrary to the Court’s apparent temporary lapse, still governs today. See Gray, 481 U. S., at 658 (explaining that to permit the exclusion of jurors other than those who will not follow their oaths "unnecessarily narrows the cross section of venire members" and "‘stack[s] the deck against the petitioner’" (quoting Witherspoon, 391 U. S., at 523)).
Judge Kozinski’s opinion for the Court of Appeals in this case is solidly grounded on the entire line of our cases recognizing the basic distinction dramatically illustrated by Justice Powell’s opinion in Darden and by Justice Rehnquist’s statement in Lockhart. He surely was entitled to assume that the law had not changed so dramatically in the years following his service as a law clerk to Chief Justice Burger that a majority of the present Court would not even mention that basic distinction, and would uphold the disqualification of a juror whose only failing was to harbor some slight reservation in imposing the most severe of sanctions.
So, what does this opinion mean going forward? It seems to indicate that habeas relief will rarely be granted on the merits when the allegation is a mixed question of fact and law best determined by the trial court. In the view of the majority, the exclusion for cause of the juror was best determined by the trial court because the trial court could view the juror's demeanor. Even when the juror's exclusion appeared to be contrary to clearly established Federal law based on the written record, the deference owed to a trial court on questions of fact, combined with AEDPA's deferential standard of review meant that habeas relief was inappropriate.
Of course, the distinguishing characteristic of Uttecht from a future case might be that the question at issue here was not guilt, but punishment. That may be why Justice Kennedy described the result in such detail at the beginning of the opinion (which was criticized by the dissent). Guilt was not at issue. As with many habeas decisions, both statutory and common law, actual innocence changes everything.