The governing law is clear.  The rule of Lockett v. Ohio gives the defendant the right to introduce any aspect of his character or record that he claims is mitigating.  (The complete absence of any trace of such a rule in the text or history of the Constitution does not matter to the high court.)  Further, the defendant has a right to effective assistance of counsel.  If counsel was ineffective, then the judgment is overturned if and only if the failure was prejudicial, which in most cases is measured by a reasonable probability of a different result.

The opinion of the Kentucky Supreme Court at issue in this case is here.  Hodge and his accomplice committed the murder in 1986.  The judgment was affirmed on appeal in 1990.  In 1992, Hodge filed his collateral challenge.  It was denied without a hearing, and the Kentucky Supreme Court remanded for a hearing in 2001.  [Why a case in which no hearing was held took nine years is not apparent from the decision.]  From the 2009 case number, it appears the case took another eight years in the trial court and another three years in the Kentucky Supreme Court, again for reasons not apparent.

Hodge's first claim was jury tampering.  That claim was rejected by the trial court based on its assessment of the credibility of his main witness, who can charitably be described as flaky.  No problem there.

The second claim is ineffective assistance in the trial court.  The state conceded the deficient performance prong of Strickland.  The defendant's childhood was horrific, and the Kentucky Supreme Court chastised the trial court for soft-peddling it.  So it comes down to the prejudice prong, a reasonable probability of a different result.  To decide that, the court reviews the aggravating side of the scale:

And, we must also consider the heinous nature of Hodge's crime. See Epperson and Hodge v. Commonwealth, 809 S.W.2d 835 (Ky. 1990). The assault on Dr. Acker and the murder of his daughter were not just brutal and vicious, but calculated and exceedingly cold-hearted. The sentencing jury was aware that Hodge and his two co-defendants carefully planned the robbery after learning of the large quantity of cash kept in the home safe, that they traveled from out of state to carry out the plan, and that they packed weapons and tools in advance. They posed as FBI agents to gain entry into the elderly doctor's home and followed him to the kitchen where they pretended to take his statement regarding a former business partner's supposed fraud. They had the doctor call his daughter to the room to witness the statement. At that point, Hodge brandished a handgun. They covered the heads of both the father and the daughter. They restrained Tammy, a young college student due to go back to school the next day, alone in a bedroom. She begged them not to hurt her father. After forcing Dr. Acker to open the safe, Hodge's accomplice strangled him with an electrical cord until he lost consciousness. Hodge went to Tammy's bedroom and stabbed her at least ten times, then stole a bracelet and watch from her wrist. Afterwards, he coolly told Epperson that he knew Tammy was dead because the knife had gone "all the way through her to the floor." Autopsy reports confirmed this boast.

Believing both victims were dead, they left the home. The three men then fled to Florida. Along with their girlfriends, they brazenly spent the stolen money on a lavish lifestyle and luxury goods, including a Corvette. A former cellmate testified that Hodge recounted spreading all the money out on a bed and having sex with his girlfriend on top of it.

Is there a reasonable probability that the bad-childhood evidence, if it had been introduced at trial, would have outweighed this aggravating evidence, resulting in a life sentence?  The Kentucky Supreme Court said (emphasis added):

Balancing all of the available evidence in mitigation and aggravation, we are compelled to reach the conclusion that there exists no reasonable probability that the jury would not have sentenced Hodge to death. There is no doubt that Hodge, as a child, suffered a most severe and unimaginable level of physical and mental abuse. Perhaps this information may have offered insight for the jury, providing some explanation for the career criminal he later became. If it had been admitted, the PTSD diagnosis offered in mitigation might have explained Hodge's substance abuse, or perhaps even a crime committed in a fit of rage as a compulsive reaction. But it offers virtually no rationale for the premeditated, cold-blooded murder and attempted murder of two innocent victims who were complete strangers to Hodge. Many, if not most, malefactors committing terribly violent and cruel murders are the subjects of terrible childhoods. Even if the sentencing jury had this mitigation evidence before it, we do not believe, in light of the particularly depraved and brutal nature of these crimes, that it would have spared Hodge the death penalty. We, therefore, affirm this portion of the trial court's judgment.

Reviewing this opinion for the purpose of deciding whether to grant a writ of certiorari, Justice Sotomayor writes (emphasis added):

But in discounting the countervailing effect of Hodge's proposed mitigation, the court misunderstood the purpose of mitigation evidence. The court reasoned that Hodge's mitigation evidence might have altered the jury's recommendation only if it "explained" or provided some "rationale" for his conduct.

Huh?  That would indeed be error if the state court had decided that, but it did not.  The state court opinion did not say or imply that only mitigating evidence with a connection to the crime can be considered.  The court noted the lack of a connection in explaining its balancing of the aggravating and mitigating evidence, a process that necessarily involves assigning weight to each.  Mitigating evidence with a connection to the crime quite properly is given greater weight than unconnected background evidence.  To see that, one need only look at the lists of mitigating evidence enacted by state legislatures before Lockett took away their authority to decide what would be considered mitigating.  Not a single one listed bad-childhood evidence as a statutory mitigator.  The Model Penal Code draft on which many post-Furman state statutes were based did not list it.  Taking the lack of connection into account in assessing the weight to be given to bad-childhood evidence, the state court decided it was unlikely to have altered the jury's decision under the facts of this case.  This is a reasonable (and, IMHO, correct) application of the principles of Lockett and Strickland.

Justice Sotomayor's interpretation of what the state court held is an entirely unjustified extrapolation from what the state court actually said.  The connection of mitigating evidence to the crime, or lack of connection, is quite properly considered as to weight, not whether the evidence is considered at all.

What happens when this case goes over to federal habeas?  Well, if the wrong panel of the Sixth Circuit gets it, they will make the same erroneous interpretation of the state court decision as Justice Sotomayor.  That will enable them to announce that the decision is "contrary to" controlling Supreme Court precedent, free them from the strictures of §2254(d), and allow them to decide the ineffective assistance question de novo.

But to what end?  Hasn't the Kentucky Supreme Court, in its consideration of Hodge's ineffective assistance claim, already given him the weighing of aggravating and mitigating evidence to which he was entitled?  That weighing does not have to be done by a jury.  An appellate court can reweigh if it finds the trial was insufficient.  See Clemons v. Mississippi, 494 U.S. 738 (1990).

Regardless of the theoretical basis, the Kentucky Supreme Court has decided that the balance of aggravating and mitigating circumstances clearly tilts toward aggravating.  The sentence in this case is just, and it should be carried out.  It is already very long overdue.