with whom
Judges RYMER and KLEINFELD join, dissenting:The California Supreme Court twice considered Pinholster’s claim that his death sentence be set aside because his lawyers performed ineffectively, and twice rejected that claim on the merits. Under *685AEDPA, those determinations come to us encased in a double layer of deference: first, the substantial deference to which lawyers are entitled under Strickland in making judgments during the course of their representation; and, second, the deference to which the state court is entitled in determining whether the lawyers’ performance was ineffective and prejudicial. The first layer of deference may be overcome only if counsel’s performance was objectively unreasonable under prevailing norms at the time and place of trial. The second layer may be overcome only if the state supreme court’s determination is contrary to or an unreasonable application of clearly established Supreme Court authority. Pinholster comes nowhere close to flipping this “doubly deferential” presumption. Knowles v. Mirzayance, — U.S. -, 129 S.Ct. 1411, 1420, 173 L.Ed.2d 251 (2009).
The majority reaches the contrary conclusion through a series of mistakes that have, unfortunately, become far too common in our circuit. First, the majority relies on evidence never presented to the state courts and that we may therefore not consider in federal habeas proceedings governed by AEDPA. Contra Williams v. Taylor (Michael Williams), 529 U.S. 420, 437-40, 120 S.Ct. 1479, 146 L.Ed.2d 435 (2000). Second, the majority applies retrospectively a standard for counsel’s performance that bears no relationship to that prevailing in California at the time of Pin-holster’s trial in 1984. Contra Bobby v. Van Hook, — U.S. -, 130 S.Ct. 13,18-19 (2009) (per curiam). Third, and perhaps worst of all, the majority accords no deference to the California Supreme Court’s superior expertise in determining what constitutes competent representation among the members of its bar and the likely consequences (or lack thereof) of any deficient performance. Contra Schriro v. Landrigan, 550 U.S. 465, 473-74, 127 S.Ct. 1933,167 L.Ed.2d 836 (2007).
Few state court judgments can withstand even one such error, see, e.g., Jones v. Ryan, 583 F.3d 626 (9th Cir.2009); Libberton v. Ryan, 583 F.3d 1147 (9th Cir.2009); Gilley v. Morrow, 246 Fed.Appx. 519 (9th Cir.2007) (unpublished); Stankewitz v. Woodford, 365 F.3d 706 (9th Cir.2004); see also cases cited pp. 692, 711 infra, but in combination they are deadly. I had hoped that our en banc court would sweep away these mistakes and bring our caselaw into conformity with AEDPA. Instead, the majority repeats and magnifies the errors in these prior cases so that they will be very difficult, probably impossible, for us to correct. This perpetuates a habeas regime where few death sentences are safe from federal judges who know ever so much better than those ignorant state judges and lawyers how capital trials ought to be conducted. Because I don’t believe we are the ultimate font of wisdom on such matters, I must dissent.
Background
Following his conviction and sentence, Pinholster (represented by two new lawyers) took a mandatory appeal to the California Supreme Court. In a painfully thorough opinion that takes up 62 pages of the Pacific Reporter, Justice Mosk, writing for a unanimous court, affirmed. People v. Pinholster, 1 Cal.4th 865, 4 Cal.Rptr.2d 765, 824 P.2d 571 (1992). One of the numerous issues addressed in that appeal was whether the state had given adequate notice of its intent to present aggravating evidence during the penalty phase. Id. at 618-20. The court was thus well aware of Pinholster’s claims that his lawyers weren’t given notice and failed to ask for a continuance when the trial court offered it. Id. at 620.
*686Leonard Tauman was appointed habeas counsel for Pinholster in February 1990, when the direct appeal was still being briefed. We know that Tauman began his investigation long before the California Supreme Court affirmed the conviction in February 1992, because we have declarations in the record dating back to 1991. Yet Tauman did not file Pinholster’s first state habeas petition until August 1993, more than three years after he was appointed. This petition relied on many of the witnesses that were eventually presented in the federal habeas petition and raised many of the facts the majority now says Pinholster’s counsel should have discovered in preparing a mitigation case. The centerpiece of the petition was a report by a psychiatric expert, Dr. Woods, who diagnosed Pinholster as suffering from a mental disease, namely bipolar disorder, and a seizure disorder.
The California Supreme Court did not summarily brush aside this petition. Rather, it ordered that the state show cause why Pinholster ought not be granted relief “on the ground that petitioner received ineffective assistance of counsel [ (IAC) ] at the penalty phase of trial,” in part because of counsel’s failure “to ascertain that notice pursuant to Penal Code section 190.3 [of intent to provide aggravating evidence] had been provided and to move for a continuance.” [Ex. C-4]* The state responded at length, and Pinholster then filed a traverse. The California Supreme Court thus had extensive evidence and briefing on the central issue in our case and, as demonstrated by its show-cause order, must certainly have focused its attention on the IAC claim. Yet, after seeing it all, in July 1995 the court unanimously denied the petition “on the substantive ground that it is without merit.” [Ex. C-7] A majority of the justices also denied various claims (not relevant here) on procedural grounds.
At that point, Tauman took a bow and exited stage left. The scene moved to federal court which, in April 1996, appointed Pinholster his second team of post-conviction lawyers, Michael Snedeker and Michael Abzug (the Michaels). A year later, the Michaels filed Pinholster’s federal habeas petition, relying on a new mental health expert named Dr. Stalberg — a psychiatrist who just happened to be the same expert Pinholster’s counsel had consulted at trial and who had found Pinholster to be sane and sober on the night of the crime. Dr. Woods, with his discredited theory that Pinholster suffers from bipolar disorder, was jettisoned, never to be seen again.1
Because Dr. Stalberg disavowed Dr. Woods’s opinion (and vice versa), the parties recognized that the state court had to be given first crack at Dr. Stalberg’s evidence. So, pursuant to stipulation, the federal petition was held in abeyance while, in August 1997, the Michaels filed a second habeas petition in the California Supreme Court. Its major difference from the first state petition was the omis*687sion of Dr. Woods and the inclusion of Dr. Stalberg, who said that he now saw lots of mitigating evidence that hadn’t been brought to his attention when he was consulted at trial. Notably, Dr. Stalberg stopped well short of recanting his diagnosis that petitioner is a sane psychopath.
The state supreme court denied the second petition “on the substantive ground that it is without merit.” [Ex. B] In addition, a substantial majority of the justices denied most of the claims on various procedural grounds (that they were untimely, successive or barred by res judicata). This is not surprising, as the second state petition presented nothing new — not even a new psychiatric opinion purporting to absolve Pinholster of moral responsibility for the heinous acts he had committed.
The matter then went back to federal court where it was litigated for four more years, including cross-motions for summary judgment and preparations for an evidentiary hearing (generating over 110 docket entries), all on the assumption that Dr. Stalberg would opine that Pinholster was mentally impaired — which Dr. Stalberg would have figured out at trial if Pinholster’s lawyers had only provided him with the information habeas counsel dug up. This plan backfired in July 2001 when the state deposed Dr. Stalberg, who testified that none of the new evidence changed his diagnosis that Pinholster does not suffer from a mitigating mental illness. Oops.
Two months later, the Michaels followed Tauman off-stage and the Federal Public Defender’s office was substituted as Pin-holster’s third team of post-conviction lawyers. New counsel fired the radioactive Dr. Stalberg and found two new psychiatric experts (Drs. Olson and Vinogradov) who, in September 2002 — 18 years after the trial — came up with a diagnosis that Pinholster suffers from a mitigating mental illness — “organic personality syndrome” — as a result of head trauma he sustained as a child.
All the while, the district court and the parties labored under the mistaken impression that Pinholster’s case was not covered by AEDPA because he had filed his request for counsel and a stay of execution before AEDPA’s effective date. See Calderon v. U.S. Dist. Ct. for the Cent. Dist. of Cal., 163 F.3d 530, 533, 540 (9th Cir.1998) (en banc). The parties proceeded to an evidentiary hearing and the district court made findings while Calderon was still law of the circuit. This means that the district court did not limit petitioner to evidence that he first presented in state court, as required by AEDPA 28 U.S.C. § 2254(e); nor did the district court apply AEDPA’s standard of substantial deference to state court determinations of law and fact. See, e.g., Mirzayance, 129 S.Ct. at 1420; Landrigan, 550 U.S. at 473-74, 127 S.Ct. 1933; Yarborough v. Gentry, 540 U.S. 1, 11, 124 S.Ct. 1, 157 L.Ed.2d 1 (2003) (per curiam); Woodford v. Visciotti, 537 U.S. 19, 27, 123 S.Ct. 357, 154 L.Ed.2d 279 (2002) (per curiam).
As fate would have it, on the very day the district court issued its ruling, the Supreme Court cut the ground from under it by holding that AEDPA does apply to cases such as Pinholster’s. Woodford v. Garceau, 538 U.S. 202, 206-07, 123 S.Ct. 1398, 155 L.Ed.2d 363 (2003). Belatedly, the district court filed an order “finding” that petitioner had been diligent in developing the record in the state court and that he was therefore entitled to present evidence in federal court that he had not presented in state court. The district court also erroneously held that it owed no deference to the state supreme court under AEDPA.
*688Discussion
A claim that defense counsel in a criminal case was ineffective under Strickland calls for an inquiry into whether the lawyer’s performance was worse than would have been rendered by minimally competent counsel practicing in the same community at the same time. The Supreme Court has repeatedly cautioned that we must step into that lawyer’s shoes before attempting to judge his performance. We can’t use hindsight or import standards from a different time and place. Even if we find that a lawyer was incompetent, we may not vacate the conviction unless it’s clear that the defendant was actually prejudiced by the lawyer’s performance; that is, unless we are convinced there is a reasonable probability that the outcome would have been different.
When AEDPA governs, we are constrained by yet another measure of deference, one that we owe to the state courts which first examined and ruled on the issue. Deference in the IAC context is particularly appropriate because the state courts, and state supreme courts in particular, are most familiar with the type of inquiry we must undertake under Strickland. State court judges have an intimate familiarity with the local standards of practice and know far better than federal judges what could reasonably have been expected of competent counsel at the time and place of trial. State courts are also far more likely to understand the behavior of local juries, and thus can best figure out whether a hypothetical strategy, invented by habeas counsel years or decades after the trial, would have changed the outcome.
It is for such reasons AEDPA requires that we defer to the determinations of the state court unless they are contrary to or an unreasonable application of Supreme Court authority. 28 U.S.C. § 2254(d)(1). This extraordinarily high standard has teeth. See, e.g., Brown v. Payton, 544 U.S. 133, 148-49, 125 S.Ct. 1432, 161 L.Ed.2d 334 (2005) (Breyer, J., concurring) (“[T]his is a case in which Congress’ instruction to defer to the reasonable conclusions of state-court judges makes a critical difference. Were I a California state judge, I would likely hold that [the] penalty-phase proceedings violated the Eighth Amendment.... Nonetheless, in circumstances like the present, a federal judge must leave in place a state-court decision .... ” (citations omitted)). It is not merely a magic spell, the sing-song invocation of which can make the state court’s decision disappear into thin air.
AEDPA limits federal courts in another important way: We may only consider evidence that was first presented to the state courts. 28 U.S.C. § 2254(e). This, too, makes perfect sense: Whether the state court has correctly — or reasonably — interpreted Supreme Court precedent depends on the facts. It makes no sense to say that a state court unreasonably applied clearly established Supreme Court law to facts it didn’t know existed. The state court might well have ruled differently had petitioner presented different facts.
Through a pernicious combination of these errors, the majority runs roughshod over the perfectly reasonable determination, twice unanimously made by the state’s supreme court, that Pinholster did not suffer prejudicial Strickland error. And it does so based on facts Pinholster never presented to the state court. This is nothing like deference.
A. Our review is limited to the record presented in the state habeas petitions
According to the majority, 28 U.S.C. § 2254(e)(2), which requires development of the record in state court, is inapplicable because “Pinholster exercised diligence in *689pursuing an evidentiary hearing in state court regarding his mitigation ineffective assistance claim. By withdrawing its order to show" cause and dismissing Pinholster’s habeas petition on the merits, the state court denied Pinholster any further opportunity to develop the factual record in state court.” Maj. op. at 668.
The majority double-faults. First, Pin-holster has not been diligent in presenting the diagnosis of Drs. Olson and Vinogradov — the two experts on whom he now relies — in his state habeas petitions. If Pinholster’s trial counsel could have presented such expert opinions (or any expert opinion to the same effect) at the penalty phase, then Pinholster’s habeas counsel could easily have presented such declarations in his first state habeas petition some 9 years after trial. And they certainly should have done so in his second state habeas petition, which was filed 13 years after trial.
Maybe the majority believes that was impossible because Drs. Vinogradov and Olson weren’t available. But if not these particular doctor, there must have been some doctor who could have come up with the same diagnosis in 1993 or 1997 when Pinholster brought his state habeas petitions. If that was not possible, it would destroy Pinholster’s claim that his trial counsel were ineffective by failing to come up with a Vinogradov and Olson-like diagnosis in 1984.
Pinholster is thus caught in a finger trap: He cannot claim trial counsel were incompetent in 1984 for failing to do that which diligent habeas counsel didn’t do in 1993 or 1997. If competent trial counsel should have come up with this diagnosis at trial, then diligent habeas counsel should easily have come up with it in the state habeas petitions. But if diligent habeas counsel couldn’t do it 9 years after trial, then trial counsel certainly couldn’t have been incompetent in failing to do so at trial. The 18 year delay in presenting the diagnosis of “organic personality syndrome” must mean either that habeas counsel was not diligent or trial counsel was not ineffective. There’s no escape.
The majority seems to think that Michael Williams, 529 U.S. at 430-32, 120 S.Ct. 1479, Bradshaw v. Richey, 546 U.S. 74, 79, 126 S.Ct. 602, 163 L.Ed.2d 407 (2005), and Holland v. Jackson, 542 U.S. 649, 652-53, 124 S.Ct. 2736, 159 L.Ed.2d 683 (2004) (per curiam), authorize Pinholster’s habeas-by-sandbagging, but they don’t. Michael Williams excused petitioner’s failure to present evidence to the state courts only when “the factual basis of the claims was not reasonably available to petitioner’s counsel during state habeas proceedings.” Id. at 442,120 S.Ct. 1479. The claim in Michael Williams that could not have been presented to the state courts was based on information that only the prosecutor possessed, and petitioner’s lawyers couldn’t have discovered it until the case got to federal court. Id. at 440-43, 120 S.Ct. 1479. By contrast, a different claim, one based on evidence that was available to petitioner while in state court — but which he failed to present there — was held precluded. Id. at 438-40, 120 S.Ct. 1479.
In our case, nothing prevented Pinholster’s counsel from presenting expert declarations with the same diagnosis as Drs. Olson and Vinogradov to the state supreme court. There was no problem with paying for such experts: Petitioner did present reports from two other psychiatrists, Drs. Wood and Stalberg, so his lawyers obviously had sufficient funding. Compare Michael Williams, 529 U.S. at 442, 120 S.Ct. 1479 (state court denied funding for an investigator). Nor did the state hide anything from Pinholster; his own mental condition was hardly some*690thing the state could have concealed in any event. Compare id. at 441-42, 120 S.Ct. 1479. Nor did Pinholster need the court’s subpoena power to obtain the expert reports. Compare id. at 439-40, 120 S.Ct. 1479. Nor can Pinholster claim that such expert opinions weren’t available at the time of the state habeas petitions, because that would make them irrelevant for purposes of evaluating the trial lawyers’ performance back in 1984. The reports of Drs. Olson and Vinogradov are just like the evidence in Michael Williams that the Supreme Court said couldn’t be used because it was not first presented in state court.
Holland and Bradshaw help petitioner even less than Michael Williams. Holland states bluntly: “In this and related contexts we have made clear that whether a state court’s decision was unreasonable must be assessed in light of the record the court had before it.” 542 U.S. at 652, 124 S.Ct. 2736 (citing Yarborough, 540 U.S. at 6, 124 S.Ct. 1, Miller-El v. Cockrell, 537 U.S. 322, 348, 123 S.Ct. 1029, 154 L.Ed.2d 931 (2003), and Bell v. Cone, 535 U.S. 685, 697 n. 4, 122 S.Ct. 1843, 152 L.Ed.2d 914 (2002)). Bradshaw, like Holland, was another reversal of a court of appeals that had relied on extrinsic evidence without first determining whether the habeas petitioner had been diligent in developing it in state court. We could be next.
But there are two reasons the majority is wrong, not just one. The second is that petitioner hasn’t shown he couldn’t have returned to state court to present the Vinogradov and Olson evidence. He returned to state court once already after swapping out psychiatric experts. See pp. 686-87 supra. The California Supreme Court may not have been thrilled to receive the second petition, but it did decide it on the merits and did not preclude further filings. Pinholster could have gone back — could still go back, so far as we know — to present the evidence from Drs. Vinogradov and Olson to the California Supreme Court. Our case is thus different from Michael Williams, where “state post-conviction relief was no longer available at the time the [hidden] facts came to light, [and] it would have been futile for petitioner to return to the Virginia courts.” 529 U.S. at 444, 120 S.Ct. 1479.
Diligence under Michael Williams at least required Pinholster to try to go back to state court and present the expert opinions of Drs. Vinogradov and Olson. By failing either to present his newfangled theories to the state court or to show that such an effort would have been futile, petitioner has indulged in a double dose of non-diligence. Our consideration of his new evidence is clearly barred by section 2254(e)(2).
This is the most dangerous part of the majority opinion as it blots out a key component of AEDPA. The statute was designed to force habeas petitioners to develop their factual claims in state court. See Michael Williams, 529 U.S. at 436-37, 120 S.Ct. 1479. The majority now provides a handy-dandy road map for circumventing this requirement: A petitioner can present a weak case to the state court, confident that his showing won’t justify an evidentiary hearing. Later, in federal court, he can substitute much stronger evidence and get a district judge to consider it in the first instance, free of any adverse findings the state court might have made. I don’t believe that AEDPA sanctions this bait-and-switch tactic, nor will it long endure.
The majority also says that none of this matters because excluding the two expert reports “would not affect our result.” Maj. op. at 669. If the majority means that, it should avoid making such terrible law and reach its result without relying on Pinholster’s new psychiatric evidence. *691But I don’t believe the majority does mean it. The majority must rely heavily on the new experts, see, e.g., maj. op. at 661, 675-77, because everything else Pinholster’s lawyers managed to dig up — after sifting through the rubble of his life for close to two decades — is so piddling. It’s hardly the stuff that would justify finding the state courts unreasonable. See pp. 709-16 infra. The proof is in the pudding: If the expert declarations didn’t matter, the majority would leave them out and avoid making an obvious error under Michael Williams. That it won’t tells us something important.
B. The California Supreme Court did not twice unreasonably apply established Supreme Court precedent in concluding that Pinholster’s trial counsel were competent
Strickland v. Washington, which was decided shortly after Pinholster’s trial, had a number of important caveats that the majority today seems to have forgotten:
No particular set of detailed rules for counsel’s conduct can satisfactorily take account of the variety of circumstances faced by defense counsel or the range of legitimate decisions regarding how best to represent the criminal defendant. Any such set of rules would interfere with the constitutionally protected independence of counsel and restrict the wide latitude counsel must have in making tactical decisions.
Judicial scrutiny of counsel’s performance must be highly deferential. It is all too tempting for a defendant to second-guess counsel’s assistance after conviction or adverse sentence, and it is all too easy for a court, examining counsel’s defense after it has proved unsuccessful, to conclude that a particular act or omission of counsel was unreasonable. Cf. Engle v. Isaac, 456 U.S. 107, 133-134, 102 S.Ct. 1558, 71 L.Ed.2d 783 (1982). A fair assessment of attorney performance requires that every effort be made to eliminate the distorting effects of hindsight, to reconstruct the circumstances of counsel’s challenged conduct, and to evaluate the conduct from counsel’s perspective at the time. Because of the difficulties inherent in making the evaluation, a court must indulge a strong presumption that counsel’s conduct falls within the wide range of reasonable professional assistance; that is, the defendant must overcome the presumption that, under the circumstances, the challenged action “might be considered sound trial strategy.” ...
The availability of intrusive post-trial inquiry into attorney performance or of detailed guidelines for its evaluation would encourage the proliferation of ineffectiveness challenges. Criminal trials resolved unfavorably to the defendant tvould increasingly come to be followed by a second trial, this one of counsel’s unsuccessful defense.
466 U.S. 668, 688-90, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984) (emphases added). More prophetic words have seldom been spoken.
The opinion here illustrates just how far we’ve strayed from the Court’s wise cautions in Strickland. Rather than looking to the standards of practice applicable in the community at the time trial was held, we have now adopted a national standard embodied in the ABA Guidelines, which are read rigidly to require a certain kind of investigation and a certain kind of mitigation defense (what is known as “humanizing” the defendant) in every capital case. Contra Strickland, 466 U.S. at 689, 104 S.Ct. 2052. No attention is paid to whether these standards reflect the contemporary norms in the community. Rather, *692current notions of a proper mitigation defense are telescoped back across the decades and retroactively imposed on counsel who had no way of knowing that this is what was expected of them. Contra id. This is exactly what the Supreme Court summarily reversed the Sixth Circuit for doing in Van Hook: “Judging counsel’s conduct in the 1980s on the basis of [later ABA Guidelines] — without even pausing to consider whether they reflected the prevailing professional practice at the time of the trial — was,” the Court held, “error.” Van Hook, 130 S.Ct. at 17. It’s the same error the majority commits today.
Aso contrary to Strickland, the majority pays hardly any attention to the facts and circumstances of this particular case, or the reasons Pinholster’s counsel may have had for proceeding as they did. Strickland’s twice-iterated caution that “counsel is strongly presumed to have rendered adequate assistance and made all significant decisions in the exercise of reasonable professional judgment,” id. at 690, 104 S.Ct. 2052, is stitched into window dressing; the “strong presumption” is treated like the cantankerous relative who gets an occasional Christmas card but is never invited to dinner.
Perhaps most troubling is our total disregard of Strickland’s admonition that we not interfere with the independence of counsel and the latitude they must have in making tactical decisions. 466 U.S. at 688-89, 104 S.Ct. 2052. The current infatuation with “humanizing” the defendant as the be-all and end-all of mitigation disregards the possibility that this may be the wrong tactic in some cases because experienced lawyers conclude that the jury simply won’t buy it. Not all defendants are capable of rehabilitation, and not all juries are susceptible to such a plea. Counsel, who are in the courtroom and can observe the jurors and their reaction to various witnesses (including the defendant), may have good reason for pursuing other avenues of mitigation, such as sympathy for the defendant’s family.
According to our case law, as amplified by today’s opinion, any trial lawyer who fails to worship at the altar of “humanization” will be labeled an incompetent boob by later counsel and federal judges who all think they know how to represent the capital defendant better than his own trial lawyers. No self-respecting lawyer wants to be berated in a published opinion — as today’s opinion does to Pinholster’s trial counsel — so their self-interest will cause them to pursue the safe course for themselves rather than the course that best serves the client’s interests. Current and future capital defendants will pay with their lives for preserving Pinholster’s. Cf. Van Hook, 130 S.Ct. at 17.
The majority’s methodology, which reflects the received wisdom in our court, has become an unstoppable engine for setting aside death sentences. See, e.g., Hamilton v. Ayers, 583 F.3d 1100, 1118 (9th Cir.2009) (requiring counsel to retain a mental health expert); Belmontes v. Ayers, 529 F.3d 834, 857-58 (9th Cir.2008) (requiring counsel not merely to investigate mitigating mental health evidence, but to present it), rev’d sub nom., Wong v. Belmontes, 558 U.S. -, 130 S.Ct. 383, — L.Ed.2d - (2009) (per curiam); Lambright v. Schriro, 490 F.3d 1103, 1119 (9th Cir.2007) (requiring counsel to “humanize” the defendant at sentencing). These cases, like the majority here, systematically fail to take Strickland seriously. That Belmontes was unanimously'— and unceremoniously — reversed seems to have made no impression around here.
1. The Standard of Care. Petitioner has presented no evidence as to the standard of professional competence in capital cases that prevailed in Los Angeles in *6931984. Rather, he relies on — -and the majority accepts — the ABA Guidelines as the governing standard. The Supreme Court disagrees with this approach: “Restatements of professional standards ... can be useful as ‘guides’ to what reasonableness entails, but only to the extent they describe the professional norms prevailing when the representation took place.” Van Hook, at 16. For the ABA Guidelines to be relevant, they “must reflect prevailing norms of practice and standard practice.” Id. 130 S.Ct. at 17 n. 1 (internal quotation marks, alterations and citations omitted). Pinholster, who has the burden of proof, offers no evidence that the 1982 ABA Guidelines meet this standard for a capital trial in Los Angeles in 1984.
Moreover, at the time of Pinholster’s trial, the ABA had no guidelines specifically applicable to capital cases (those didn’t come about until 1989), and the guidelines pertaining to criminal cases were quite general. They certainly did not impose anything like the straightjacket the majority retroactively imposes on counsel today. The most relevant portion of the 1982 Guidelines consists of a single sentence buried in a paragraph of “Commentary.” This sentence does no more than point out that various aspects of the defendant’s background “will be relevant” to an effective defense. That’s a far cry from a commandment that lawyers leave no stone unturned when investigating the defendant’s background. As the Court recently explained:
The ABA standards in effect in 1985 described defense counsel’s duty to investigate both the merits and mitigating circumstances in general terms.... Quite different are the ABA’s 131-page “Guidelines” for capital defense counsel.... Those directives expanded what had been (in the 1980 Standards) a broad outline of defense counsel’s duties in all criminal cases into detailed prescriptions for legal representation of capital defendants.
Id. at 17. The Court here was discussing the version of the guidelines in effect during Pinholster’s trial, which occurred about a year before Van Hook’s.
The Supreme Court, our court and the California Supreme Court have all held that competent counsel need not, and often will not, exhaust every avenue of investigation. Strickland itself held that counsel need not conduct an exhaustive investigation of a defendant’s background. Instead, counsel only has “a duty to make reasonable investigations or to make a reasonable decision that makes particular investigations unnecessary.” 466 U.S. at 691, 104 S.Ct. 2052 (emphasis added). “[Tjhere comes a point at which evidence ... can reasonably be expected to be only cumulative, and the search for it distractive from more important duties.” Van Hook, at 19. At least in 1984, counsel was only required to “cover several broad categories of mitigating evidence,” id., something Pinholster’s trial counsel did by any measure.
Van Hook is only the latest in a long line of similar cases. Burger v. Kemp noted that “[t]he record at the habeas corpus hearing does suggest that [counsel] could well have made a more thorough investigation than he did,” but the Court held he wasn’t required to. 483 U.S. 776, 794, 107 S.Ct. 3114, 97 L.Ed.2d 638 (1987). We follow a slightly stricter rule: “[C]ounsel is not deficient for failing to find mitigating evidence if, after a reasonable investigation, nothing has put the counsel on notice of the existence of that evidence” — but even we acknowledge there is some limit on counsel’s duty to investigate. Babbitt v. Calderon, 151 F.3d 1170, 1174 (9th Cir.1998) (internal quotation marks omitted). The California Supreme Court is of the same mind. See, e.g., In re Ross, 10 *694Cal.4th 184, 40 Cal.Rptr.2d 544, 892 P.2d 1287,1304-05 (1995) (collecting cases).
Instead of rehashing the ABA Guidelines, the majority should be asking whether the investigation by Pinholster’s counsel comported with the standards for counsel in a capital case in California in the mid-1980s. Van Hook, 130 S.Ct. at 15. The clearest evidence that it did comes from Hendricks v. Calderon, a capital case where we discussed the standard of care prevailing in California around the time of Pinholster’s trial:
Hendricks argues “that by 1981 [the time of trial] it was the recognized duty of defense counsel in capital cases to obtain social history evidence relevant to a client’s mental condition where there was an indication of mental disturbance and to provide such evidence to an expert in order for the client’s condition to be properly evaluated.” Under Hendricks’ 1995 view, Berman was duty bound to interview Hendricks’ family members and friends, to obtain medical, school, and employment records, to otherwise verify Hendricks’ autobiography and pass this information on to his mental health experts. “Without such background material,” Hendricks continues, “the expert’s diagnosis of the client’s condition could not be meaningful.” Hendricks’ argument, then, is not that examination by two mental health experts was per se inadequate. Rather, Hendricks argues that trial counsel had a duty to provide the experts with his client’s family social history, even absent any request from the experts. He argues that counsel’s failure to do so undermines the expert’s conclusions to such an extent that they cannot provide an adequate basis on which the attorney could make strategic choices.
Neither reason nor the existing authority would lead one to conclude that an attorney in 1980 had such an affirmative constitutional duty.
70 F.3d 1032, 1038 (9th Cir.1995) (emphasis added). The state relied on Hendricks as proof of the standard of care in its brief before us and in the district court.
Pinholster asks us to adopt a different standard of care based on the ABA Guidelines, but the Hendricks panel, writing in 1995 — just about the time the California Supreme Court was ruling on Pinholster’s state habeas petitions — specifically rejected Pinholster’s interpretation of those guidelines:
Certainly, in 1981, Hendricks’ attorneys did not believe they had any duty to investigate Hendricks’ social history in the face of the unanimous opinions of their own experts that there was no basis for a mental defense.
The legal authority Hendricks cites offers meager support for his argument. Hendricks’ cited authority merely discusses the basic duty under Strickland to investigate potential defenses before making strategic decisions. See ABA Standards Relating to the Administration of Criminal Justice, Standard 4—4.1; U.S. v. DeCoster, 487 F.2d 1197 (D.C.Cir.1973); Rummel v. Estelle, 590 F.2d 103 (5th Cir.1979). Hendricks’ authority that touches upon the duty to investigate mental defenses for the most part deals with an attorney’s failure to investigate the possibility of a mental defense. See People v. Frierson, 25 Cal.3d 142, 162, 158 Cal.Rptr. 281, 599 P.2d 587 (1979); People v. Pope, 23 Cal.3d 412, 428, 152 Cal.Rptr. 732, 590 P.2d 859 (1979); Deutscher v. Whitley, 884 F.2d at 1160. At most, these cases establish a duty to seek out psychiatric evaluation of a client where the grounds of a mental defense are apparent, a duty Berman clearly discharged. Hendricks’ cases make no comment on an attorney’s *695alleged duty to investigate material relevant to a mental defense in the face of expert advice that there is no basis for such a defense. See also Evans v. Lewis, 855 F.2d 631, 637 (9th Cir.1988) (failure to investigate defendant’s mental condition when there is evidence of impairment constitutes deficient performance, and is prejudicial when it hampers later presentation of evidence of mental impairment).
Hendricks, 70 F.3d at 1039.
Since our own court, considering the very issue only 11 years after Pinholster’s trial, said that no such duty existed in California at the time, how can we now — a quarter century removed from the trial and with no evidence on point — say that such a duty did exist after all? More, how can we say in 2009 that the California Supreme Court in the mid-1990s unreasonably applied established Supreme Court precedent by rejecting a standard of professional competence that we unanimously held at the same time did not apply to counsel in the 1980s?
The majority reaches its unjustified conclusion by misreading three oft-invoked and oft-misinterpreted Supreme Court cases: Williams v. Taylor (Terry Williams), 529 U.S. 362, 120 S.Ct. 1495, 146 L.Ed.2d 389 (2000), Wiggins v. Smith, 539 U.S. 510, 123 S.Ct. 2527, 156 L.Ed.2d 471 (2003), and Rompilla v. Beard, 545 U.S. 374, 125 S.Ct. 2456, 162 L.Ed.2d 360 (2005). My colleagues seem to think that these opinions create a nationwide, retroactive code of professional conduct for capital cases. (Never mind that Strickland says, and Van Hook confirms, there’s no such thing.) The majority’s exegesis on these cases is but one more in a series of inapt methodologies that this en banc court should have put to rest.
Terry Williams has little to say about the source or scope of the standard for professional competence in capital cases. While it cites the ABA Guidelines in passing, 529 U.S. at 396, 120 S.Ct. 1495, there was no real dispute that counsel there had been ineffective. The same state trial judge who sentenced Williams concluded that counsel’s performance was deficient, id. at 370, 120 S.Ct. 1495, a conclusion the Virginia Supreme Court accepted and that the State “barely disputed” before the U.S. Supreme Court. Id. at 395, 120 S.Ct. 1495. There was evidence, as there is not in this case, that counsel’s performance did not comport with the standard of care in the community at the time and place of trial.
The only real dispute in Terry Williams was whether the state court had properly applied the prejudice prong of Strickland. The Virginia Supreme Court -wrongly believed that Lockhart v. Fretwell, 506 U.S. 364, 113 S.Ct. 838, 122 L.Ed.2d 180 (1993), modified Strickland’s prejudice inquiry, and the U.S. Supreme Court set it straight. Insofar as the Supreme Court passed on the competency of counsel, it was to confirm the state court’s judgment, not overrule it as unreasonable. Terry Williams has nothing to say about the standard under AEDPA when the state court finds the lawyer has been competent and the federal court disagrees.
Wiggins is hardly better for Pinholster. First, the trial judge there “observed from the bench that he could not remember a capital case in which counsel had not compiled a social history of the defendant, explaining, ‘[n]ot to do a social history, at least to see what you have got, to me is absolute error.’ ” Wiggins, 539 U.S. at 517, 123 S.Ct. 2527. Thus, the Wiggins Court had a finding by a state judge who was experienced in local practices that Wiggins’ lawyer had performed incompetently. The Supreme Court expressly relied on this time- and place-specific determination: *696“As [the trial judge] acknowledged, standard practice in Maryland in capital cases at the time of Wiggins’ trial included the preparation of a social history report.” Id. at 524, 123 S.Ct. 2527.
Second, insofar as the ABA Guidelines are relevant, Wiggins’ trial took place after the ABA issued its 1989 Guidelines providing specific standards for capital cases, and it is that version of the ABA Guidelines on which the Wiggins Court relied. But the 1989 Guidelines are far more mandatory and specific than the 1982 Guidelines that were in force at the time of Pinholster’s trial. Whereas the 1982 Guidelines contained only a single sentence in the “Commentary,” pointing out certain types of inquiries that “will be relevant” in all criminal cases, the 1989 Guidelines outline capital counsel’s affirmative duties at some length, including an investigation that “ ‘should comprise efforts to discover all reasonably available mitigating evidence and evidence to rebut any aggravating evidence that may be introduced by the prosecutor.’ ” 539 U.S. at 524, 123 S.Ct. 2527 (quoting ABA Guidelines for the Appointment and Performance of Counsel in Death Penalty Cases 11.4.1(C) (1989)) (emphasis added by the Court). Although I doubt that Wiggins turned the ABA Guidelines into a constitutional code of conduct for lawyers, see Van Hook, 130 S.Ct. at 17 & n.l, even if Pm wrong, it’s irrelevant; in Wiggins a newer, more exacting edition of the guidelines applied and those standards cannot be ported back to Pinholster’s case.
Finally, there’s Rompilla, which says nothing about the general scope of investigation required for defense counsel in capital cases. The Rompilla majority found a single flaw in defense counsel’s performance, namely the failure to examine the case file of a prior conviction that counsel knew the prosecution would rely on in aggravation, 545 U.S. at 383, 387-90, 125 S.Ct. 2456, a point emphasized in Justice O’Connor’s linchpin concurrence. Id. at 393-94, 125 S.Ct. 2456. And Justice O’Connor saw no need to cite the ABA Guidelines at all. Rompilla’s holding does not support the view, adopted by the majority here and by a number of other recent cases, that some version of the ABA Guidelines is the minimum standard of practice the Constitution imposes on all 50 States. I would be surprised to learn that the Supreme Court in Rompilla had meant to give the ABA the constitutional authority to set the standard of care for all criminal cases across the nation. And I wouldn’t be alone. See Van Hook, 130 S.Ct. at 20 (Alito, J., concurring).
Terry Williams, Wiggins and Rompilla rely on the ABA Guidelines as background support where they overlap with local standards (Terry Williams and Wiggins) or where the proposition is so obvious that it’s a matter of common sense (Rompilla). See Van Hook, 130 S.Ct. at 17. These cases don’t establish the ABA as the final authority on how lawyers must conduct criminal trials, with the power to override contrary determinations by the state courts about the lawyers they admit to practice. And yet the majority relies entirely on the ABA Guidelines without any evidence that the Guidelines reflect local standards in the community at the time of Pinholster’s trial. This is what the Sixth Circuit got summarily reversed for in Van Hook.
Here we have more than petitioner’s failure of proof. We have evidence going the other way: a determination by our court in Hendricks and other cases, see p. 709 infra, that the duty to investigate at the time was not what petitioner now claims it to be. This is Van Hook on stilts.
2. The Facts of This Case. Justice O’Connor’s concurrence in Rompilla emphasized “our longstanding case-by-case approach to determining whether an attor*697ney’s performance was unconstitutionally deficient under Strickland.” 545 U.S. at 394, 125 S.Ct. 2456. The majority here does very little to determine what competent counsel could realistically have done to help Pinholster at the penalty phase. Not even lip service is paid to the “strong presumption that counsel’s conduct falls within the wide range of reasonable professional assistance.” Strickland, 466 U.S. at 689,104 S.Ct. 2052. Rather, the majority is satisfied with pointing out that counsel were surprised by the need to put on mitigation evidence, that they turned down an offered continuance and that they “billed a total of only 6.5 hours in preparation for the penalty phase of the trial.” Maj. op. at 658 (footnote omitted). The majority is wrong on all counts.
I might as well start with the most obvious error. The “only 6.5 hours” has been brandished by petitioner’s able habeas counsel like a smoking gun: concrete proof that Pinholster’s trial lawyers must have been incompetent because they spent less than a work-day preparing for such an important hearing. Problem is, the “only 6.5 hours” is a myth.2
The 6.5 hour figure comes from the time records of Brainard, who was just one of Pinholster’s trial lawyers, and includes only time that bore the notation “prep, for penalty phase.” See Maj. op. at 658 n.2. But the defense team clearly spent a lot more time on mitigation. Under California law, the prosecution didn’t have to put on penalty-phase evidence at all in order to ask for the death penalty; it could have relied entirely on the evidence presented during the guilt phase. Thus, it doesn’t matter whether the prosecution has given, or the defense has received, notice of an aggravation hearing. In preparing for trial, counsel in California must necessarily consider both guilt and penalty, and try to work as much mitigation evidence as possible into the guilt phase.
Pinholster’s trial counsel did precisely this. They put on Pinholster’s brother, Terry, to support his alibi defense, but also elicited from him the following mitigation evidence: (1) that Pinholster “was more or less in institutions all his life” [TR 6015]; (2) that Pinholster suffered from epilepsy and Terry had seen him have two seizures in a single evening [TR 6015-18, 6031-32]; and (3) that Pinholster was drunk on the night of the murders. [TR 6036]
Terry was examined at trial by Brainard, and Brainard must have prepped him or he couldn’t have learned about those mitigating facts to begin with. Yet an examination of Brainard’s time records makes no mention of speaking to Terry, and certainly doesn’t say that this was time spent on mitigation.
It is possible that the prep work on Terry was done by Dettmar, Pinholster’s other trial counsel, or that it was included in one of Brainard’s more cryptic notations, such as “Interviewing defense witnesses” (3/28/84), “Brainard & Dettmar interview of witness” (3/6/84) or “Brainard & Dettmar interview with” (3/7/84). [Ex. 71-2] Or, perhaps Brainard was not diligent about time records. One way or the other, some of the time spent preparing Terry must be counted as time counsel spent on mitigation and the 6.5 hours, all of which were spent after Terry testified, obviously don’t include that.
There are other hours logged on Brainard’s time sheets that were clearly devoted to mitigation. For example, on February 23, 1984, he billed for “Conf. with Burnice Brashear, Pinholster’s mother,” and two days later for “Research re; epi*698lepsy and conf. with nurse.” [Ex. 71-2] There was nothing Burnice could have possibly said about the crime, so the time spent with her must have been entirely on mitigation. The time spent on epilepsy was also obviously mitigation-related. Brainard’s time records show that he spent about 700 hours on Pinholster’s case and much of that time is described in fairly general terms. Just because there were only 6.5 hours that he specifically described as preparation for the penalty phase doesn’t mean that’s all the time he spent on mitigation.
The majority claims that Pinholster’s counsel admitted “that they spent almost no time preparing for the penalty phase hearing,” maj. op. at 671,3 but I’ve found no such admission in this very hefty record. Nor is the majority right in claiming that “[b]illing records confirm” such an admission. The only time sheets and declarations they reference are Brainard’s. Maj. op. at 658 n.2, 671. Yet, according to Brainard, it was Dettmar who was primarily in charge of the penalty phase. [ER 182, 337, 350, SER 122] We don’t have a complete set of Dettmar’s time records; we have only those through March 15, 1984, which does not include the 8-week period leading up to and including the penalty phase, when Dettmar most likely would have been preparing for that portion of the trial. What we do have of Dettmar’s time sheets indicates that he spent considerable time preparing for mitigation.
Thus, on January 13, 1984, we have the notation “phone call to defendant’s mother re medical history”; on February 21 there is “Penal Code research on capital punishment”; on February 23 there is “conference with defendant’s mother re childhood problems”; on February 25 there is “Research on Pen. C. 190.3”;4 on February 29, there are notations for various mitigation-related items, including “Further research on Pen. C. 190.3” and “Phone call to appointed psychiatrist.” On February 26, Dettmar spent six hours on “preparation argument, death penalty phase ” (emphasis added).
There are other, more cryptic notations that might or might not go to mitigation, such as “Visit to client, L.A. County Jail” (February 10), but such as we have of Dettmar’s time sheets shows that the defense team was keenly aware of the need to show mitigation and was vigorously investigating mitigation evidence. Dettmar’s time records stop abruptly on March 15, 1984, which was in the midst of the guilt phase, so we don’t know what else he did to prepare for the “death penalty phase,” but we do know he was active because he made further court appearances and his name is referenced in Brainard’s time sheets.
For all we know, Dettmar did just about everything Pinholster claims should have been done on his behalf, like examining his medical and school records, talking to school authorities, prison guards, friends and neighbors, all in a vain effort to find someone who would say a good word about him. We know that Dettmar was the contact point for Dr. Stalberg, the defense team’s psychiatrist, and that Dettmar pro*699vided him with the information the doctor used in forming his diagnosis. (Dr. Stalberg’s opinion letter is addressed to Dettmar, and Dr. Stalberg mentions in his declaration that his dealings were with Dettmar.) This activity clearly went to mitigation, yet (except for a phone call on February 29) the time spent is not accounted for on any time-sheet Pinholster has supplied. There is no lawyer’s case file documenting the activity, and there’s no testimony about it; we know virtually nothing about it. Dettmar’s activities in the two months leading up to and including the penalty phase are a complete blank. There may be mitigation-related activities about which we know nothing at all.
Which brings us back to the burden of proof, which rests on petitioner, and the “strong presumption that counsel’s conduct falls within the wide range of reasonable professional assistance.” Strickland, 466 U.S. at 689, 104 S.Ct. 2052. Pinholster fails to shoulder his heavy burden of overcoming that presumption. Inasmuch as Pinholster has been unwilling or unable to document the actions of half his trial team during a crucial two months in the proceedings, he has failed to carry the ordinary burden of any habeas petitioner. And he certainly comes nowhere near carrying the especially heavy burden in ineffectiveness cases, where he must overcome the “strong presumption” that counsel were diligent and competent in advancing his interests.
We know, based on the documentation we do have, that the lawyers engaged in activities that were highly relevant to mitigation, such as retaining a psychiatrist who examined Pinholster, talking to the defendant’s mother and brother and educating themselves about epilepsy. We also know from Dettmar’s time sheets that he was preparing for the “death penalty phase” as early as February. See 698 supra. It therefore isn’t remotely consistent with the record and the “strong presumption” to say, as the majority does, that Pinholster’s trial team spent “less than an average workday” preparing for the penalty phase. Maj. op. at 672.
Petitioner bears the burden of proof, so if he wants to claim that his lawyers didn’t spend enough time on mitigation, he must account for all the time they did spend before we can say they didn’t spend enough. It’s not fair — or consistent with AEDPA — for petitioner to present partial time records for the first time in federal court, p. 698 supra; p. 700 n. 6 infra, and then claim — Tada!—my lawyers only spent 6.5 hours on mitigation. This gambit didn’t work in Van Hook, 130 S.Ct. at 15, and it shouldn’t work here.
In addition to the supposedly short time Pinholster’s defense team spent on mitigation, the majority’s other “proof’ that his lawyers must have been incompetent is the fact that they were taken by surprise at the end of trial when they learned there would be a penalty phase and yet refused a continuance offered to them by the trial court.5 The decision to decline a eontinu*700anee only seems incompetent if the defense lawyers hadn’t already conducted an adequate investigation into mitigation. But if they had been looking for mitigation evidence all along, they would already have decided what mitigation strategy to pursue. And if their decision was to present only Pinholster’s mother in mitigation, then turning down a continuance was perfectly reasonable; “strategic choices made after thorough investigation of law and facts relevant to plausible options are virtually unchallengeable.” Strickland, 466 U.S. at 690, 104 S.Ct. 2052.
The record is clear that the defense team was quite busy preparing mitigation evidence, and specifically preparing for the capital penalty phase, long before the jury came back with a guilty verdict. We know that they consulted a psychiatrist, but only after the trial started, suggesting that his testimony wasn’t intended for guilt-phase issues (which involved an alibi defense), but for mitigation. They talked to the mother about his emotional problems and medical history, and they educated themselves on epilepsy, all matters not relevant to the alibi defense. Pinholster’s lawyers clearly investigated the “several broad categories of mitigating evidence” they should have considered. Van Hook, 130 S.Ct. at 19. They were required to do no more than that.
For all we know, this is just the tip of the iceberg of what the defense team did in pursuing mitigation evidence since Dettmar’s time records for the crucial period from March 15 until the end of trial are missing. Dettmar’s case file is also missing, which could have given us pretty good insight into the activities of Pinholster’s “death penalty phase” lawyer as the defense team was in full swing preparing for the penalty phase. Dettmar was dead by the time of the first habeas petition, so we have no statements from him, but his case file would almost certainly still have been in existence at the time Pinholster’s first habeas lawyer commenced his investigations. Yet there is no explanation for why the case file wasn’t retrieved and presented in the state habeas petition. Most likely it’s because Pinholster’s first two habeas teams didn’t pursue this line of inquiry; even Brainard’s time records were never presented to the state courts.6
In the absence of any evidence as to what the defense team, and particularly Dettmar, actually did in investigating mitigation evidence, the majority relies on a series of non-denial denials7 from Pinhol*701ster’s surviving counsel, Brainard. Maj. op. at 658. These statements, stamped out by the dozen at the behest of habeas counsel 7 years after the trial, seem to say a great deal but, in fact, say nothing at all. Thus, Brainard had “no recollection” that Dettmar had reviewed any medical or school records; he did “not recall” interviewing or attempting to interview family members or any other persons regarding penalty phase testimony; he had “no recollection of seeing or attempting to secure Scott’s school records.” Within a single paragraph the majority quotes Brainard as not recollecting or having no recollection or stating “so far as[he] recolleet[ed]” no fewer than 5 times. Maj. op. at 658 (second alteration in original).8 There is not a single categorical statement attesting that Brainard and Dettmar failed to do any of these things.
Failure to recall can be very useful because it avoids the risk of contradiction or perjury. Who, after all, can dispute that someone else doesn’t recall something? But when evidence is needed to carry the burden of proof, a lack of recollection is worth absolutely nothing. Indeed, a failure to recall does not even satisfy the burden of production, because it does not prove or disprove the fact not recalled.9 Brainard’s accommodating statements, no doubt made to try to help a former client, prove zilch. See also pp. 700-01 n.7 supra.10
If we assume, as the majority plainly does, that Pinholster’s trial lawyers couldn’t walk and chew gum at the same time, then no degree of idiocy on their part would surprise us. But we must indulge in the contrary “strong presumption that [they made choices] for tactical reasons rather than through sheer neglect.” Yarborough, 540 U.S. at 8, 124 S.Ct. 1 (citing Strickland, 466 U.S. at 690, 104 S.Ct. 2052). What this means for us is that we must presume, consistent with the time records we do have, that Pinholster’s trial lawyers were fully aware that they would have to deal with mitigation sometime during the course of the trial, did spend con*702siderable time and effort investigating avenues for mitigation and made a reasoned professional judgment that the best way to serve their client would be to rely on the fact that they never got notice and hope the judge would bar the state from putting on them aggravation witnesses. Nevertheless, the time sheets show that both lawyers were busy preparing for the penalty phase long before they supposedly knew for sure there would be a penalty phase hearing. [CT 798, 844, 864,1160]
Had this strategy succeeded, it would have been quite a coup. It almost did succeed; the trial judge had to think long and hard before ruling against the defense’s motion to preclude aggravation evidence, and the California Supreme Court had to work hard to affirm. Pinholster, 4 Cal.Rptr.2d 765, 824 P.2d at 618-20. We’ll never know, of course, but the burden of proof, Strickland’s presumption of competence and AEDPA deference each require us to presume that the lawyers did the smart thing, not the dumb one.
The fact that Pinholster’s lawyers tried to bar the state from putting on aggravation evidence does not mean they were not also preparing for the alternate possibility; their time sheets (such as we have) clearly show they were. This would explain why, when offered a continuance, the defense chose not to accept the offer. They had already done the necessary investigation and knew how they were going to proceed. If we “strongly presume,” as Strickland commands, that the defense lawyers were competent, then they would have had no need for a continuance because they had already considered all the available possibilities and concluded that the best strategy was to put only Pinholster’s mother on the stand.
This obviously isn’t a view shared by the majority here, but that’s because my colleagues view the case through the “distorting lens of hindsight.” Sims v. Brown, 425 F.3d 560, 585 (9th Cir.2005) (internal quotation marks omitted); Turner v. Calderon, 281 F.3d 851, 873 (9th Cir.2002) (internal quotation marks omitted); Hendricks, 70 F.3d at 1036. If we step into the defense lawyers’ shoes and look at the case from ground zero, the decision to rely on Pinholster’s mother is entirely plausible, perhaps inevitable. It was certainly plausible enough that the California Supreme Court didn’t unreasonably apply Supreme Court precedent by determining the lawyers were competent.
The majority seems to overlook the very difficult position defense counsel found themselves in after the jury came back with guilty. This wasn’t a case where defendant sat doe-eyed at counsels’ table looking sad and contrite while others spoke for him. Instead, he took the stand and protested his innocence aggressively and at length. While we don’t have video of Pinholster’s testimony, the transcript gives strong indications that he was arrogant to the point of smirking at the prosecution’s questions. Equally important, the transcript reveals that he was lucid and entirely coherent. He presented himself as a violent and highly aggressive criminal. He claimed that he used only guns but then had to admit he had used knives in other crimes. He tried to snow the jury about his whereabouts on the night of the crime, but the jurors were not taken in. During the hour or two he testified, Pin-holster doubtless made a deep impression on the jury and this would have made it difficult or impossible to “humanize” him in their eyes. Brainard and Dettmar were there to see all of this and their mitigation strategy was constrained by this reality. My colleagues see only a paper record.
Trial counsel were limited in their choices in other significant ways. They had hired a competent psychiatric expert, *703Dr. Stalberg, and his opinion was devastating:
1. When examined, Mr. Pinholster did not manifest any significant signs or symptoms of mental disorder or defect other than his antisocial personality disorder by history. Although he allegedly has epilepsy, he has not been taking medication for the past year at County Jail and has not had a seizure. Additionally, it does not appear that he suffers brain damage as he was cognitively intact on mental status examination.
He has a history of hyperactivity as a youngster and hospitalization at Camarillo State Hospital at the age of 14 for incorrigibility. He was not medicated there and just received 90 day observation. He spent time in the Youth Authority and in various state prisons without requiring psychiatric treatment or medication.
He has a lengthy history of drug dependency of the narcotic variety, but his drug dependency has not resulted in any brain damage.
2. As for the alleged offenses, the defendant denies any involvement other than directing an individual to burglarize the residence with Mr. Pinholster collecting 1/3 of the money from sale of the stolen loot. He is able to describe his actions on the night in question in great detail, omitting of course any direct involvement with the victims. It would not appear therefore he was significantly intoxicated or impaired on the night in question.
Additionally, considering the statements attributed to the defendant by witnesses, it would also appear he was fully aware of what he was doing at the time of the offenses. He allegedly stated, “I’ll take care of finding out where the dope is. I’ll handle Kumar.” This was some days before the offenses occurred.
On the night of the offense Mr. Pin-holster apparently planned his robbery and was aware of what he was doing. This is by inference as he did not provide any subjective information other than his directing an individual to burglarize Kumar’s home.
3. As for mitigation, it does not appear Mr. Pinholster’s epilepsy, hyperactivity as a child, or incorrigibility were related to the offenses except the incorrigibility reflects upon his psychopathic personality traits. He was not under the influence of extreme mental or emotional disturbance, nor did he have impaired ability to appreciate the criminality or conform his conduct to the requirements of the law.
Also, because of Mr. Pinholster’s personality disorder it is likely he would be recalcitrant and a security problem while in custody.
[ER 797-98] (emphases added).
Years later, in 2002, when presented with the additional evidence gathered by habeas counsel, the same Dr. Stalberg found it of marginal significance and did not change his diagnosis:
As set forth in my declarations of April 19, 1997; January 24, 2000; and June 5, 2001, the additional documents I have reviewed since 1997 did contain some information that might conceivably be mitigating within the very broad definition of California Penal Code section 190.3(k). On the other hand, as I testified at my deposition in July 2001, the additional materials I reviewed did not alter my conclusion that Mr. Pinholster suffers from Antisocial Personality Dis*704order, as that term is defined in the DSM-III.
[ER 793] (emphases added).
Given Dr. Stalberg’s expert opinion, trying to develop a psychiatric mitigation case at the time of trial would have been extraordinarily difficult. He told the lawyers that Pinholster was entirely sane and sober on the night of the murder and that, in general, he’s a psychopath. Unlike unrebutted evidence of brain damage or post-traumatic stress disorder, Porter v. McCollum, 558 U.S. -, 130 S.Ct. 447, 451-52 & nn.4-5, — L.Ed.2d - (2009) (per curiam), what Dr. Stalberg had to say would hardly have been the stuff of mitigation; most likely it would have been aggravating. Daniels v. Woodford, 428 F.3d 1181, 1204-05 (9th Cir.2005); Gerlaugh v. Stewart, 129 F.3d 1027, 1034-35 (9th Cir. 1997); see also Graham v. Collins, 506 U.S. 461, 113 S.Ct. 892, 122 L.Ed.2d 260 (1993) (Thomas, J., concurring). Had Dr. Stalberg testified, see pp. 718-19 infra, Pinholster would be here arguing that his lawyers were incompetent for putting him before the jury. See, e.g., People v. Hines, 15 Cal.4th 997, 1064-65, 64 Cal.Rptr.2d 594, 938 P.2d 388 (1997). Moreover, counsel practicing criminal law in California back then would have been aware of a number of capital cases where psychiatric testimony backfired with devastating effect. See, e.g., Harris, 949 F.2d at 1505; People v. Williams, 44 Cal.3d 883, 934-35, 245 Cal.Rptr. 336, 751 P.2d 395 (1988); People v. Robertson, 33 Cal.3d 21, 44 n. 11, 188 Cal.Rptr. 77, 655 P.2d 279 (1982).11
With no realistic possibility of a psychiatric mitigation defense, what could Pin-holster’s lawyers do? Unlike habeas counsel, who have years of time, unlimited resources and the power to conjure imaginary mitigation cases with which to mesmerize federal judges, trial counsel are stuck with the hard realities that are the lot of the trial lawyer. One such reality is the client’s wishes and preferences. Pin-holster did not testify at his own evidentiary hearing in district court, so we don’t know first-hand what instructions he might have given his lawyers, but we do have a pretty good idea: The record contains the report of Sheryl Duvall, who was an investigator for Pinholster’s first habeas counsel, a document the majority wrongly refuses to consider.12
*705Duvall interviewed Pinholster in July-1991, about 7 years after the trial. She reports that Pinholster, at that time, “[wa]sn’t enthused about attempting to get a reversal on penalty alone” and “had instructed his trial attorney not to put on a penalty defense.” [Ex. 40-1] Of course, it is entirely rational not to want to spend decades as an inmate in a maximum security prison, and quite a few capital defendants have chosen not to try to have death sentences “reduced” to life behind bars. E.g., Demosthenes v. Baal, 495 U.S. 731, 732-34, 110 S.Ct. 2223, 109 L.Ed.2d 762 (1990); Gilmore v. Utah, 429 U.S. 1012, 1013 n. 1, 97 S.Ct. 436, 50 L.Ed.2d 632 (1976) (Burger, C.J., concurring). Pinholster’s lawyers would have been entitled— maybe required — to respect that choice.
Duvall also reported that Pinholster’s “most significant relationships appear to have been with his mother, step-father and maternal grandfather.” The report continues:
Scott told me he has always felt very close to his mother for whom he had nothing but praise. “She’s always been supportive of all the kids. She’s always the first there and the last to leave. She always had a hot dinner on the table.” Further, Scott said to his knowledge his mother has never abused drugs or alcohol. She always treated her children kindly, was never abusive.
Scott continues to stay in close contact with his mother. She visits him nearly every month in San Quentin. Scott told me he was upset by his mother testifying at his trial. He thought it was unnecessarily hard on her.
[Ex. 40-3] (emphasis added). Pinholster also reported that he had no recollection at all of his natural father, Garland.
And this is what Duvall revealed about Pinholster’s step-father:
Bernice [sic] [Pinholster’s mother] has told me Mr. Brashear [Pinholster’s stepfather] abused her boys, particularly Scott, as they were growing up. Scott doesn’t concur. He said while his stepfather was hard on them, in his opinion he and his brothers benefited from Mr. Brashear’s discipline — “it toughened us up.” He had a home made paddle which he used freely on the boys. According to Scott, this worked to their benefit as they lived in tough neighborhoods so they needed to get used to rough treatment.
[Ex. 40-3]
Pinholster also had positive things to say about his grandparents:
When Scott was small Mr. and Mrs. Baumbaek [sic] had a chicken farm in Sunland. Scott used to spend his summers there. He and his grandfather were very close. Mr. Baumbaek [sic] was affectionate with the boys. He made them work hard on the farm but he was quick to praise their efforts. It seems that he made them feel good about themselves.
[Ex. 40-4]
Seen through the looking-glass of a quarter-century of hindsight pigs might *706fly, but in the here and now, when a lawyer has to make hard decisions about his client’s case, not every theoretical possibility can be turned into a reality. Counsel can try to cajole or persuade, but experienced lawyers know that pushing a client too far can backfire. Pinholster was not enthusiastic about putting on a mitigation case, and he probably did instruct his lawyers not to do so, as he told Duvall. The lawyers may have persuaded him to let them put his mother on the stand, but he was reluctant and resentful. Still, the lawyers must have reminded him that everyone has a mother and the jurors may take pity on her even if they feel no pity for the son. Pinholster went along but, years later, he still didn’t like it. That tells us something about the stiff resistance counsel could have expected from Pinholster had they proposed trashing his mother in open court, as the majority does today.
Is it possible, is it even conceivable, that Pinholster, who was not enthusiastic about life in prison, would have allowed his lawyers to put his estranged — and strange — ■ relatives on the stand to tarnish his grandparents? To say that his step-father was a monster, to besmirch his beloved mother as a selfish, uncaring, neglectful crone who wore mink while the children had no food? Even if Pinholster had allowed this — and I can’t imagine he would have — what good would it have done? It certainly wouldn’t have worked for the mother to testify, and then to be followed to the stand by the sister, brother, son and daughter calling her a selfish liar. Such a family food fight would have done Pinholster only harm. Had the lawyers decided to go with the other relatives, and had Pinholster let them do so, they would have had to give up the mother, forfeiting the inherently sympathetic bond they hoped to establish between her and the jurors. It’s not clear to me that this would have been a wise trade-off. See pp. 716-17 infra.
If we assume, as the majority does, that ■Pinholster’s lawyers were Laurel and Hardy, all such calculations might seem farfetched. But we are required to presume — strongly presume — that the lawyers were competent, and we must look at the situation through their eyes. We know that the lawyers talked to Pinholster and his mother; they also talked to Pinholster’s brother Terry and had a trial run at using him as a mitigation witness — -with no success. See pp. 697-98 supra. Not only did the jury disbelieve Terry’s testimony, but the prosecutor managed to insinuate that he was the conduit of the first death threat against Art Corona. [TR 6030] Maybe my colleagues in the majority would put Terry the Enforcer on the stand to help “humanize” Pinholster and soften the jury in his favor, but a competent lawyer might think better of it. The other erstwhile family witnesses had serious disabilities too. Burnice’s brother Keith was so estranged over an inheritance dispute that he didn’t even know Pinholster was on trial for murder. Pinholster’s sister had a serious criminal record.
One need only read the declarations and testimony of the various Pinholster relatives to see that they aren’t exactly the Osmonds. A competent lawyer talking to Pinholster, his mother and Terry would have quickly figured out that parading the family members through the courtroom to snipe at each other would do Pinholster no earthly good. See pp. 716-17 infra. Lawyers are hired to make those judgments and, best I can tell, Pinholster’s lawyers called this one right. At any rate, we certainly can’t say that the California Supreme Court was unreasonable for not second-guessing them. That one of Pinholster’s lawyers, years after the trial, wouldn’t offer tactical reasons for the defense team’s choices, maj. op. at 699-74, does not undermine the California Supreme Court’s judgment. See Strickland, *707466 U.S. at 688,104 S.Ct. 2052; Murray v. Carrier, 477 U.S. 478, 484, 487-88, 106 S.Ct. 2639, 91 L.Ed.2d 397 (1986); p. 701 n.9 supra.
The majority also forgets that there are no free-throws in criminal trials. See Belmontes, 130 S.Ct. at 386, 388-89. If the defense has a witness testify about Pinholster’s peaceable tendencies, then the state gets to cross-examine that witness about all of Pinholster’s misdeeds. If Pinholster puts on evidence about his medical or mental problems, this opens the door to evidence that he’s a remorseless psychopath. Pinholster’s counsel faced a serious risk that a mitigation case could turn out to be aggravating. See Id. at 384.
Burnice Brashear might not have been the ideal mitigation witness; then, again, Mother Theresa probably wasn’t available. But a mother- — one obviously devoted to her son, and vice versa — has a pretty good chance of arousing the jury’s sympathy while not giving the prosecution many openings to inflict damage through cross-examination or rebuttal. Burnice provided many details about Pinholster’s head injuries, abuse by his step-father, problems in school, problems with his siblings, his being beaten badly in jail and she confirmed Terry’s account of Pinholster’s epilepsy. Sure, Burnice could have been more effective in her testimony, but not everyone is lucky enough to have Joan Crawford for a mother.
My colleagues miss the point entirely when they refer to Burnice’s testimony as “misleading,” “self-serving,” “inaccurate” or “devastating.” Maj. op. at 672-73, 675, 677-78, 683. The main point of Burnice’s testimony was to create sympathy for herself and the other members of Pinholster’s family in the hope that the jury would take pity on them and spare them the agony of losing a son and brother to the executioner. That’s what’s known as the “family sympathy” mitigation defense and other lawyers in California used it at the time. People v. Cooper, 53 Cal.3d 771, 801, 844, 281 Cal.Rptr. 90, 809 P.2d 865 (1991); In re Visciotti, 14 Cal.4th 325, 336-37, 58 Cal.Rptr.2d 801, 926 P.2d 987 (1996). Is this a perfect strategy? Probably not, but there probably is no perfect strategy for mitigating a bloody double-murder by a smug, violent and remorseless psychopath.
Defense counsels’ strategy kept Pinholster’s jury from coming back with a swift death verdict: The jury deliberated for two days before making its decision. Do my colleagues really think that Pinholster would have done better if his lawyers had put on the child-molesting sister, the extortionist brother, the persnickety aunt and the greedy uncle to tell the jury that Burnice was a terrible mother who deserved to see her son executed? A competent lawyer might think otherwise.
3. Deference to the State Supreme Court. The district court did not defer in the least to the California Supreme Court for the simple reason that it, and the parties, mistakenly believed that AEDPA deference did not apply. See Calderon, 163 F.3d at 540. No sooner had the district court finished its work than the Supreme Court disabused it of this notion in Garceau. 538 U.S. at 206-07, 123 S.Ct. 1398. But the district court refused to be disabused: It did not reconsider its ruling in light of AEDPA’s clear mandate that we defer to the state courts — and their factual findings in particular — in all but the most unusual circumstances. We therefore owe the district court no deference at all. Rather, we should be suspicious of its findings as arrived at using the wrong standard and stubbornly persisted in despite clear contrary guidance from the Supreme Court. Unlike the Supreme Court in Porter, we don’t have a reliable finding by a federal or state trial court that counsels’ performance was deficient.
*708Perhaps led astray by the district court’s cavalier attitude, the majority doesn’t give much deference to the state court either. The opinion recites all the right verbal formulae, although it does twice denigrate the California Supreme Court’s action as a “postcard” denial, maj. op. at 660 n.3, 683, suggesting perhaps that the state court didn’t look at Pinholster’s petition very closely or carefully and therefore isn’t entitled to a full measure of deference.
In fact, it’s abundantly clear that the California Supreme Court looked at this case closely. By the time it got Pinholster’s first habeas petition, the justices were already familiar with his ease. Only three years earlier, they had written an exhaustive opinion dealing with a host of issues in Pinholster’s trial, including whether the state had given timely notice of a penalty-phase hearing. On receiving Pinholster’s voluminous first petition, the justices did not dismiss it out of hand. Rather, they asked for a response from the state on three specific claims of ineffectiveness, all dealing with counsel’s performance at the penalty phase. The state filed a response and petitioner filed a traverse. After this briefing was complete, the justices unanimously denied the petition on the merits, and some of the claims on procedural grounds. Justice Mosk would have denied the petition solely on the merits.
Although the process was more abbreviated for the second petition, the justices issued a reasoned order dismissing all counts on substantive grounds and some on various procedural grounds. This time Justice Mosk was joined by Justice Brown in believing that the petition should be dismissed on substantive grounds only.
In light of this record, we are bound to presume that the denial of Pinholster’s state petitions represented the reasoned view of the unanimous California Supreme Court. And we must then ask ourselves whether we can say that their collective judgment about the performance of two members of their bar was truly unreasonable. In doing so, we must keep in mind that the California Supreme Court not only is the ultimate arbiter of the conduct, performance and ethics of lawyers admitted to practice in the state, but that it has far more experience with IAC claims than we do. In the 15 years preceding Pinholster’s habeas petitions, the California Supreme Court resolved no fewer than 177 such cases.
Based on their vast experience with criminal trials in general and capital trials in particular, the California justices had any number of reasons for finding that Brainard and Dettmar weren’t incompetent. For example, they may have relied on cases where they’d held that “[a] defendant appearing in propria persona is held to the same standard of knowledge of law and procedure as is an attorney.” People v. Clark, 50 Cal.3d 583, 625, 268 Cal.Rptr. 399, 789 P.2d 127 (1990). If so, they would have held Pinholster responsible for failing to give his lawyers notice of the state’s intent to hold a penalty-phase hearing, along with any resultant failure to adequately prepare for it.
The justices may have also reasonably believed that whatever continuance the superior court would have allowed after Pin-holster’s counsel were notified of the penalty phase would amount to, at most, a few days or a week; the jury that had just rendered the guilty verdict could not be forced to return many weeks or months later. So the court could have concluded that there just wasn’t that much else a competent lawyer could have done in such a short time. Or the justices may have believed that a lawyer is entitled to rely on a competent expert to request information if the materials provided by the lawyer are *709insufficient. And, as Dr. Stalberg did not ask for further materials in making his diagnosis, the court may have concluded that competent counsel were not required to provide any such additional materials. E.g., Turner, 281 F.3d at 876; Wildman v. Johnson, 261 F.3d 832, 838 (9th Cir.2001); Harris, 949 F.2d at 1525. Or the court may have had in mind the bad experiences of other defense lawyers who put on psychiatric testimony in capital cases around that time, and concluded that a lawyer who got an aggravating psychiatric report would be foolish to raise a psychiatric mitigation defense. See, e.g., Thompson v. Calderon, 86 F.3d 1509, 1525 (9th Cir.1996) (since overruled); Hines, 15 Cal.4th at 1064-65, 64 Cal.Rptr.2d 594, 938 P.2d 388; Williams, 44 Cal.3d at 934-35, 245 Cal.Rptr. 336, 751 P.2d 395; Robertson, 33 Cal.3d at 43 n. 11, 188 Cal.Rptr. 77, 655 P.2d 279.
The California Supreme Court acted much closer in time to the trial and generally has a much better feel for what is reasonably expected of a competent criminal trial lawyer in California. Having twice looked at Pinholster’s claims of ineffectiveness, the justices — with not a single dissent among them — concluded that counsel were effective, or that Pinholster was not prejudiced, or both. It takes considerable hubris to say that all of the justices were not merely wrong, but so egregiously wrong that they unreasonably applied clearly established Supreme Court precedent. Justice Mosk, were he still with us, would be surprised and, I believe, hurt by such a suggestion.
C. The California Supreme Court did not twice unreasonably apply established Supreme Court precedent in concluding that Pinholster was not prejudiced by any deficiency on the part of his trial counsel
Under the heading “Available Mitigation Evidence,” the majority throws in every scrap of even potentially mitigating evidence gathered by Pinholster’s habeas counsel during the course of two decades of investigation and expert-shopping. Maj. op. at 675-78. It then concludes that Pinholster was prejudiced because his trial counsel didn’t present this evidence on his behalf. But much of the evidence the majority relies on isn’t properly considered, either because it was not presented to the state courts or because it could not possibly have been found by diligent trial counsel back in 1984. Almost all the rest of the evidence was disclosed to the jury during the penalty phase.
At best, the majority focuses on some minor differences in emphasis that don’t amount to a hill of beans. See Van Hook, 130 S.Ct. at 15. At worst, the majority dramatically overreads Terry Williams, Wiggins and Rompilla, cases that are irrelevant to the prejudice inquiry under AEDPA, in a way that will hamstring the states in our circuit for decades to come. The controlling case is Visciotti, and it clearly requires us to accept the California Supreme Court’s determination that Pin-holster wasn’t prejudiced.
Below I review the mitigation evidence in the order listed by the majority.
Organic Brain Damage. There are three separate and independent reasons we may not consider the expert reports that purport to diagnose Pinholster with brain-damage-related mental illness caused by childhood head injuries. First, evidence of this supposed relationship was neither presented nor diligently pursued in the state courts. It was not presented in the first state habeas petition, which relied on Dr. Woods as the psychiatric expert, and it was not presented in the second state habeas petition, where Pinholster relied on Dr. Stalberg — who later let him *710down and testified for the state. Nor, for the reasons I explained earlier, was Pin-holster diligent in developing the theories of Drs. Olson and Vinogradov. Thus, we can’t consider them. See 28 U.S.C. § 2254(e)(2); pp. 688-91 supra.
Second, petitioner has presented no evidence that a competent lawyer in 1984 could or would have found an expert to advance the theory that organic brain damage caused Pinholster to commit two cold-blooded murders. In fact, the record speaks loudly to the contrary: It wasn’t until Pinholster’s seventh legal team, 18 years after the trial, worked with a third set of psychiatric experts that they came up with this dubious theory.13
Moreover, Pinholster’s trial lawyers had an expert, and a very good one: Dr. Stalberg. Pinholster thought so highly of him that he re-hired him as the expert for his second state habeas petition. Even when he fired Dr. Stalberg for coming up with the wrong opinion, Pinholster didn’t dispute his competence. And what did Dr. Stalberg tell Pinholster’s trial lawyers in very clear terms? Their client is sane, wasn’t drug or alcohol-impaired at the time of the murder and is a garden-variety psychopath. See pp. 702-03 supra.
We and the California Supreme Court have clearly held that a competent lawyer may rely on the opinion of a competent expert and need not go expert-shopping until he finds one whose opinion he likes. Harris, 949 F.2d at 1525; In re Fields, 51 Cal.3d 1063, 1075, 275 Cal.Rptr. 384, 800 P.2d 862 (1990); People v. Grant, 45 Cal.3d 829, 863, 248 Cal.Rptr. 444, 755 P.2d 894 (1988); Williams, 44 Cal.3d at 945-46, 245 Cal.Rptr. 336, 751 P.2d 395. So, even if experts like Drs. Vinogradov and Olson could have been found in 1984, which petitioner has never proven, a competent lawyer would have had no duty to go looking for them. Until today, that was the settled law of our circuit. It should still be the law for AEDPA purposes because the United States Supreme Court has not held otherwise.
The majority might be under the mistaken impression that Dr. Stalberg would have come up with the organic-damage theory at trial if only he’d known about the head injuries, the epilepsy and all the other “mitigating” evidence that teams of habeas counsel have dredged up over the last two decades. But Dr. Stalberg did know about the head injuries (they are in the probation report the lawyers gave him) and about the epilepsy (he mentions it twice in his 1984 letter to Dettmar) at the time of trial, neither of which he thought was relevant. [ER 798; Ex. 10-2]
And what about the rest of the stuff that is supposedly so essential in assessing Pin-holster’s sanity? We need not speculate because Dr. Stalberg is here to tell us. He looked at it all and reached exactly the same diagnosis as in 1984: Pinholster is a sane psychopath. [ER 793] It’s a pipe dream to imagine that a competent trial *711lawyer in 1984 who went to his competent expert with all this evidence and got this diagnosis would have nonetheless searched for (much less found) a Dr. Vinogradov to reach the opposite conclusion — all in the week or two the trial judge might have been willing to continue the penalty phase.
There is a third reason the majority shouldn’t rely on Dr. Vinogradov’s and Dr. Olson’s evidence: It says it doesn’t need to. Maj. op. at 669. By representing that it would reach the same conclusion without their reports, the majority is holding that what remains is sufficient, standing on its own, to show that the California Supreme Court unreasonably applied Strickland. If this is what the majority means to hold, it should have the courage of its convictions and drop its discussion of the psychiatric evidence. It can only cause confusion in the law of the circuit for the majority to say in one part of its opinion that the psychiatric evidence is superfluous but then later rely so heavily on it.
In fact, the majority can’t possibly let go of the psychological evidence because, once we remove the diagnosis that Pinholster suffers from trauma-induced mental illness, there is very little left with which to impugn the judgment of the lawyers, far less the judgment of the California Supreme Court.
Even if it were proper to consider this evidence, the majority gives it far too much weight by relying on law that has no place in a federal court’s review of a state court decision under AEDPA. The majority’s infatuation with the supposedly magical effect of evidence of organic mental illness, as opposed to merely psychological mental illness, see maj. op. at 676-79, is another area where our court has gotten out in front of the Supreme Court. But see 28 U.S.C. § 2254(d)(1). I therefore need not engage the majority in a protracted debate about the wisdom of Caro v. Woodford, 280 F.3d 1247 (9th Cir.2002), and Douglas v. Woodford, 316 F.3d 1079 (9th Cir.2003). Nor need I argue that the en banc court should take this opportunity to overrule Caro and Douglas, although it should because there is no principled distinction between “organic” mental illness and “mental” mental illness. All I need point out is the undebatable: Under AED-PA, we may rely only on law clearly established by the Supreme Court. Circuit precedent, which is all Caro and Douglas are, isn’t enough for granting relief under AEDPA. That the majority is unable to cite a single Supreme Court case in support of this distinction is itself proof that it has no lawful basis for reversing the California Supreme Court.
Epilepsy. The majority spends some time discussing epilepsy, and whether it was caused in childhood by Pinholster’s head injuries or (as his mother speculated) as a result of a prison beating. It doesn’t matter: Epilepsy is a seizure disorder. If not controlled by drugs, it causes the afflicted person to drop to the ground with stiff arms and legs, clenched fists and often foaming at the mouth. To those who have seen an individual in the midst of an epileptic fit, it is an unforgettable, scary experience. But when not in the throes of a fit, the individual behaves normally; it is not a mental illness and his faculties are not impaired.
No one has suggested, and it would make no sense to suggest, that Pinholster robbed, beat and stabbed two people to death while having an epileptic fit. Indeed, Dr. Stalberg, who interviewed Pin-holster about the events on the night of the murders, remarked how good Pinholster’s memory was and concluded that he hadn’t been impaired or intoxicated. [ER 798] Dr. Stalberg knew of the epilepsy and mentioned it in his report, but observed it *712was irrelevant. And Dr. Stalberg tells us today that, even with all the new evidence, he stands by his original diagnosis. Trial counsel had educated themselves as to epilepsy as well, see pp. 697-98 supra, and must have been convinced it was a dead end. We’re in no position to second-guess that judgment.
Abusive and Deprived Childhood. The majority tries to make Pinholster’s childhood seem like hell on earth by using alarmist phrases like “abusive and deprived upbringing,” maj. op. 677, and “suffered extreme deprivation,” maj. op. at 679. The majority also tries hard to create the impression that the evidence presented on habeas was wildly different from what the jury heard, characterizing the mother’s testimony as “misleading,” maj. op. at 672, 675, 678, and the testimony of the other family members as standing “[i]n stark contrast,” maj. op. at 678, and being vastly different from hers. Maj. op. at 680. In fact, what’s remarkable is how little support the family members provide for Pinholster’s theory of extreme abuse and deprivation.
Pinholster’s mother testified at trial that the punishment inflicted by her husband, Scott’s stepfather, was “abusive or near abusive.” Maj. op. at 677. In what the majority claims is “stark contrast,” Pinholster’s siblings testified that the step-father “frequently beat Scott while Scott was a child---- Bud [the stepfather] would use his fists, a belt, or anything else available, including on at least one occasion a two by four board.” Maj. op. at 678 (Terry Pin-holster). And, “Scott as a child was frequently physically abused by Scott’s stepfather, Bud Brashear. Bud hit Scott with his fists as often as several times within one week. Bud’s punishments were unpredictable and severe.” Maj. op. at 678 (Tammy Pinholster). While Pinholster’s siblings give somewhat more detail than the mother, what they have to say is no different from' — much less “in stark contrast” to — the mother’s description. What they describe is abuse or near-abuse; their testimony confirms, rather than contradicts, the mother’s testimony.
Then there is the suggestion that Pin-holster’s grandmother abused him because he looked like his father, whom Pinholster’s grandparents despised. What’s remarkable here is just how weak this testimony actually is. Both Pinholster’s aunt and uncle make it clear that they are talking about nothing more than severe spankings. The aunt, Lois Fosberg, is quite explicit about this:
Q: “And at times, your mother would spank him?”
A: “Yeah.”
Q: “How many times did you see your mother spank Scott?”
A: “Quite a few.”
Q: “Can you give me an estimate?”
A: “No.... She was very rough with him and not so with the other ones.”
Q: “And how did she spank him? Did she punch him or hit him with her palm, backhand him, or what?”
A: “Hit him with her hand or jerk his arm or — ”
Q: “With her palm?”
A: “Yeah, things like that.”
[2 DT 53-54]
The uncle also leaves no doubt that what the grandmother supposedly did to Pinholster was no more than spanking:
A: “[I]t occurred — often she would pick him up out of the playpen by one or two arms and just spank him across the back and butt and the legs. And I mean, more than once. I don’t know how many times. It was enough to make an impression on me.”
*713[2 DT 125] Pinholster himself says he enjoyed his summers at the grandparents’ chicken farm; the grandfather “made [the children] feel good about themselves.” [Ex. 40-4]
That’s it — that’s all the evidence of “abuse” that Pinholster’s lawyers have been able to come up with after nearly two decades of digging and investigating.14 Is what the record does support really a “stark contrast” with what Pinholster’s mother had to say? To me, it seems like more of the same.
Pinholster himself describes the situation in terms much closer to those of his mother. See p. 705 supra (Duvall interview). Accusing Burnice of misleading the jury by providing an inaccurate picture of Pinholster’s home situation is wholly unjustified. Had Pinholster’s trial counsel decided to go that route, they might have been able to coax a bit more mileage out of some relatives, but would it really have given them something to suggest an “abuse excuse” for brutally murdering two strangers decades later? I don’t think so.
Let’s turn now to the majority’s claim that Pinholster suffered “extreme deprivation.” Maj. op. at 679. The entire support for this version of events, which the majority for some reason believes to be “the truth,” comes from the testimony of Pin-holster’s aunt and uncle, and mostly the aunt. Aunt and uncle were visitors to the household, but their declarations disclose that their observations were infrequent. The aunt makes a general statement that the children didn’t get enough to eat, and supports it with a single instance where she saw them mixing flour and water. In the same breath, however, she says that the children were fed “canned food like spaghetti” and the kids would “start throwing food at each other during meals”- — -awfully strange behavior for starving children. [ER 687] In general, the aunt’s testimony borders on the picayune, as she criticizes her sister for not disciplining the children sufficiently, for keeping a dirty house and for living in bad neighborhoods. One next expects her to accuse Burnice of forgetting to weed the lawn. The uncle testifies only that his sister (Pin-holster’s mother) took good care of herself while the children looked like ragamuffins; he says nothing about food. [ER 682] If this is “extreme deprivation” then there are few defendants in the criminal justice system who are not extremely deprived and deserving of a pass.
Neither of Pinholster’s siblings, who spent far more time in the house than his aunt and uncle, say that they didn’t get enough to eat or that they suffered economic deprivation. Indeed, Pinholster himself says exactly the opposite, praising his mother for always having a hot meal on the table. See p. 705 supra (Duvall interview).
The simple fact is, there’s nothing supporting the theory of abuse or deprivation, in stark contrast to the evidence in many other cases. There was no evidence of broken bones, concussions, bleeding, hospitalization or any kind of serious or lasting injury resulting from Pinholster’s so-called *714abuse. E.g., Wiggins, 539 U.S. at 517, 123 S.Ct. 2527; Terry Williams, 529 U.S. at 370, 120 S.Ct. 1495; Harris, 949 F.2d at 1505-06. Pinholster’s father didn’t try to shoot him. Porter, 130 S.Ct. at 449. None of the children were gang-raped. Wiggins, 539 U.S. at 517, 123 S.Ct. 2527. In fact, there was no evidence of incest or any kind of sexual abuse. E.g., id.; Hendricks, 70 F.3d at 1037. Nor was there evidence that the children had to be removed from the home and placed in foster care because they were abused or neglected. E.g., id.; Terry Williams, 529 U.S. at 395, 120 S.Ct. 1495. There was no evidence that either parent was convicted of child abuse or neglect, or was even charged with such behavior. Terry Williams, 529 U.S. at 395, 120 S.Ct. 1495; Harris, 949 F.2d at 1506. Home wasn’t a “combat zone.” Van Hook, 130 S.Ct. at 18. There was no evidence that the children were exposed to promiscuous sex by the parents, were encouraged or forced to commit crimes to sustain themselves or were denied the basic necessities of life. E.g., id.; Wiggins, 539 U.S. at 517, 123 S.Ct. 2527; Moore v. Johnson, 194 F.3d 586, 613 (5th Cir.1999). The children weren’t locked “in a small wire mesh dog pen that was filthy and excrement filled.” Rompilla, 545 U.S. at 392, 125 S.Ct. 2456. Pinholster didn’t grow up in a one-bedroom house described as a “chicken coop.” Belmontes, 130 S.Ct. at 387. The children didn’t live in a home where “someone had had a bowel movement” several places on the floor and “[u]rine was standing in several places in the bedrooms.” Terry Williams, 529 U.S. at 395 n. 19, 120 S.Ct. 1495. The parents didn’t violently assault each other in front of the children. In re Visciotti, 14 Cal.4th at 336, 58 Cal.Rptr.2d 801, 926 P.2d 987. Pinholster’s father didn’t beat his mother so severely that she was hospitalized and lost a baby. Porter, 130 S.Ct. at 450. There’s no evidence the children had to be hospitalized for drinking when they were very young. Terry Williams, 529 U.S. at 395 n. 19, 120 S.Ct. 1495.
It is true that the family was poor; the stepfather was abusive; the mother was more permissive than some would condone; and some of the familial bonds were not particularly warm or caring. But there is absolutely nothing in this record that suggests Pinholster’s experience growing up differed from that of millions of other young men from broken homes with parents who have a hard time making ends meet. Hyperbolic language cannot force from the record something it doesn’t contain. And what it doesn’t contain is anything that could conceivably have swayed a jury to go easy on Pinholster because of his childhood.
We must remember that this jury had just convicted Pinholster of two brutal murders and heard convincing evidence that this was just the last in a long pattern of brutal crimes. He had previously broken his wife’s jaw, seriously wounded someone with a straight-razor, kidnapped someone else at knife-point and on multiple occasions kicked, spit and thrown bodily fluids on police officers. He proudly admitted to having committed hundreds of robberies at gunpoint. The jurors also heard that Pinholster twice threatened the life of key prosecution witness Art Corona, most recently just a few days earlier. When Sgt. Barrett saw Pinholster after he returned from court, Pinholster boasted that he’d snowed the jury on the murder charge and “when he got out he’d have to kill [Art Corona].” [TR 7384-85] Van Hook reminds us that we must focus on the weight of the aggravating factors, not merely on their number. 130 S.Ct. at 20. Would these jurors, who had just convicted Pinholster of a double murder, take a chance that this uncontrollably violent defendant would make good his threat to *715arrange for the murder of Art Corona? Not bloody likely.
Family Criminal and Mental History. As the majority recognizes, Pinholster’s mother did disclose that her other children had difficulties with the law, difficulties with drugs and alcohol and that Tammy had engaged in self-destructive behavior. Beyond that, the record on habeas does go into more depth regarding the mental, substance-abuse and criminal problems of the Pinholster siblings and of his biological father.
Whether, and to what degree, this is mitigating is highly debatable. Perhaps a mental health expert might have argued from this evidence that Pinholster suffered from some genetic mental defect, but none here have. As noted, Dr. Stalberg always believed Pinholster was fundamentally a sane psychopath. The same is true of Drs. Rudnick and Geiger. Drs. Olson and Vinogradov, whose evidence we may not consider, see pp. 688-91 supra, believed that Pinholster’s “organic personality syndrome” was induced by childhood head trauma which, of course, cannot possibly be inherited and thus bears no relationship to whatever mental problems his siblings had. He never knew his biological father, so any bad behavior on the father’s part wouldn’t have affected the son. See p. 705 supra (Duvall interview). And, of course, providing evidence that Pinholster’s brother and sister are nuts would have undermined their usefulness as abuse and neglect witnesses.
Without a psychiatric expert to relate the mental and substance-abuse problems of Pinholster’s siblings to his behavior, this evidence could have just as easily been aggravating as mitigating- — more likely the former. As experienced trial lawyers know, jurors are not always in a forgiving or generous mood, especially after they have just convicted a defendant of multiple first-degree murders. Having learned that all of his siblings and his biological father suffered from mental problems, they might well have concluded that Pin-holster was a bad apple from a bad tree and there was no hope for rehabilitation or redemption. It’s not as if such ideas are unheard of, even in the pages of the United States Reports. See, e.g., Buck v. Bell, 274 U.S. 200, 207, 47 S.Ct. 584, 71 L.Ed. 1000 (1927).
Pinholster’s Substance Abuse. Much of Pinholster’s history of substance abuse was newly presented on habeas, although the jury did hear some of it through his brother Terry’s testimony during the guilt phase. We’ve held before that “juries are unlikely to favor defenses based on abuse of dangerous drugs in evaluating a defendant’s culpability for violent behavior.” Mayfield v. Woodford, 270 F.3d 915, 931 n. 17 (9th Cir.2001). Telling the jury a lot more about what a druggie Pinholster was would probably have hurt him rather than helped him.
There is no evidence, much less clear and convincing evidence, that Pinholster’s substance abuse had anything to do with his actions on the night of the murders. At the end of the trial, after the jury returned its death verdict, the trial court made certain findings, including the following: “[T]he defendant’s capacity to appreciate the criminality of his conduct and his capacity to conform his conduct to the requirements of law were in no way impaired as a result of mental disease, defect, or the effect of any intoxicants or drugs or a combination thereof.” [CT 1185] Under AEDPA, this finding is entitled to a presumption of correctness and can only be overturned if Pinholster rebuts it by clear and convincing evidence. 28 U.S.C. § 2254(e)(1).
*716So here is where we stand with the mitigation evidence:
Evidence Heard by the Jury
Pinholster’s stepfather was abusive, or nearly so.
Pinholster was in boys homes and juvenile halls after age 10 or 11.
Pinholster’s mother ran over him with a car at age 2 or 3, badly injuring his head and requiring treatment at a hospital.
At age 4 or 5, Pinholster was in a car accident where he flew through the window and hit his head badly.
Pinholster started failing at school in the first grade, and has learning problems dating back to kindergarten.
In third grade, Pinholster’s teacher suggested that he was “something more than just a disruptive child.”
In third or fourth grade, Pinholster was sent to a class for emotionally handicapped children where his performance improved.
Pinholster had a difficult childhood, often getting into fights with his brothers.
A psychiatrist recommended Pinholster be hospitalized at age 10.
At age 12, Pinholster was institutionalized for six months in a psychiatric hospital.
Throughout childhood Pinholster stole things and had a blustery personality.
At age 18, Pinholster was badly beaten in jail.
Pinholster’s personality changed after he was imprisoned in his late teens; he became withdrawn and had difficulty adjusting to life on the outside.
Pinholster has epilepsy, for which he’s taken medication, and has had several serious seizures.
The Pinholster siblings had trouble with drugs and alcohol, and his sister was a self-destructive wild girl.
Pinholster’s brother Alvin died and was in considerable trouble with the law. Pinholster’s family “sticks close together like you would not believe.”
Pinholster was intoxicated on the night of the murders.
New Evidence on Habeas That Was Not Presented to the State Courts
The diagnosis of organic brain damage promoted by Drs. Vinogradov and Olson. Testimony from Pinholster’s uncle about his childhood and Pinholster’s mother and grandparents.
New Evidence on Habeas That Was Presented to the State Courts
Pinholster’s aunt thought that his mother was selfish and neglectful, a poor housekeeper, a poor disciplinarian and deprived the children of proper nutrition.
Pinholster’s grandmother spanked him, possibly for looking like his biological father.
Alvin’s death was by suicide.
Pinholster’s biological father may have been mentally ill.
A teacher found Pinholster’s mother unreceptive to guidance about how she should handle his learning disabilities and emotional problems.
Only the evidence in the third category was not presented to the jury but could have been.
There just ain’t much there. I suppose counsel could have presented this evidence, but one thing is clear: It would not have been possible to present the mother’s testimony, which painted a relatively rosy picture of the family, and particularly her role in it, and also the testimony of the brother, sister, aunt and uncle, which *717painted their family life as terrible and the mother as neglectful and selfish. Assuming that counsel had all this evidence available to present at the penalty phase, they would have had to make a choice: They could go with the mother and try to develop sympathy for her and the family, or they could paint Pinholster’s parents as villains and try to work up sympathy for him personally. They absolutely could not do both without having the sides collapse on the middle; the two approaches are not “consistent ... theories] of mitigation.” Porter, 130 S.Ct. at 455.
Now let’s consider the situation counsel found themselves in. The jury had just convicted their client — -a white supremacist gang member who boasted of having little regard for others and of committing hundreds of armed robberies — of two bloody murders. He had lied to the jury on the stand and generally behaved badly. His testimony would have given the jury a pretty good impression of who they were dealing with. They had also heard, and were about to hear more, about Pinholster’s long history of violent and abusive behavior towards just about everyone. The jurors had also heard that Pinholster had threatened prosecution witness Art Corona in an effort to keep him from testifying. Pinholster’s own competent psychiatrist was telling his lawyers that their client was sane and sober at the time of the crime and committed the crime because he’s a violent psychopath. For reasons I explained earlier, it’s not irrational or incompetent for lawyers in that situation to conclude that it would be best to have the mother paint a rosy picture of Pinholster’s family life and herself as a mother, in hopes that the jury would take pity on her and spare her son’s life to avoid causing her and the rest of the family additional pain.
My colleagues disagree. They believe that there is only one strategy here, namely trying to “humanize” the defendant by trashing the mother and stepfather. I think they’re wrong. But let’s apply the Belmontes methodology anyhow, 130 S.Ct. at 385, 387, turn the clock back to 1984 and try to figure out how far Pinholster’s lawyers could have gotten with trying to humanize him.
Imagine this: Counsel leave off the mother and lose whatever sympathy she may have gained for Pinholster. They put on the persnickety aunt who criticizes her sister for neglecting the kids, for failing to discipline them, for looking after herself while the kids were wearing rags and not giving them enough to eat. She also says that the grandmother spanked Pinholster very hard because he looked like his father. They put on Terry — the brother who may have delivered the threat to Art Corona — -to say that the stepfather was violent and Pinholster caught the worst of it, and they put on the sister to say the same. Then they put on evidence that their own mitigation witnesses (the brother and sister) suffer from mental illness and are drug addicts and child molesters. And they put on the teacher to say that Pinholster was bright and she recommended mental treatment, but the evil mother didn’t listen to her.
Beyond this it gets difficult. They have no commanding officer to testify to Pinholster’s decorations and valor during “two of the most critical — and horrific — battles of the Korean War.” Porter, 130 S.Ct. at 454. Pinholster’s personal troubles didn’t result from trauma he suffered while shedding blood for his county. Id. at 454. There is no evidence that Pinholster “struggle[d] to regain normality,” id. at 454, or that he ever did anything at all to turn away from a life of crime, violence and abuse of everyone he came in contact with. Pinholster’s *718is not “a crime of passion” or one motivated by “emotionally charged, desperate, frustrated” desire. Id. at 449. Pinholster did not commit these murders in a drunken stupor. Id. at 449. Rather, Pinholster’s murders grew out of a long-planned burglary, driven entirely by economic motives, and he robbed his victims before stabbing them to death. He doesn’t own up to the crimes by pleading guilty, id. at 449, but brazenly lies to the jury, claiming that this just isn’t his type of crime because he likes guns. Counsel can’t point to anything Pinholster has done in his life that was useful, constructive, generous or courageous — nothing at all that might redeem him in the eyes of a rational juror.
Nor do counsel have a psychiatric expert who can testify that Pinholster suffers from mental illness and there’s no evidence that such a mental health expert even existed in 1984. So they have to go forward without an expert, in which case they have very little. Or they can go forward with the expert they do have — Dr. Stalberg.
So let’s say they put on Dr. Stalberg to tell the jury that there is a bunch of mitigating evidence from Pinholster’s past. So far, so good. But the prosecution then gets to cross-examine Stalberg, and it goes something like this:
Prosecutor: Dr. Stalberg, based on your examination of the defendant, did you form an opinion as to whether he suffers from a mental illness?
Dr. Stalberg: Yes I did.
Prosecutor: Did you diagnose him as being bipolar?
Dr. Stalberg: No, I did not.
Prosecutor: Is he paranoid?
Dr. Stalberg: No.
Prosecutor: Is he a schizophrenic?
Dr. Stalberg: No, he’s not.
Prosecutor: Does he suffer from any other recognized form of psychosis?
Dr. Stalberg: No, he does not.
Prosecutor: Would you say that he is completely sane?
Dr. Stalberg: As sane as you and me.
Prosecutor: And do you believe he was sane on the night of the murders?
Dr. Stalberg: Yes I do.
Prosecutor: Did you form an opinion about whether he was impaired by drugs or alcohol on the night of the murder?
Dr. Stalberg: Yes I did.
Prosecutor: Was he?
Dr. Stalberg: It is my professional opinion that he was not.
Prosecutor: Dr. Stalberg, you testified that Mr. Pinholster suffers from epilepsy?
Dr. Stalberg: I believe he may have suffered from epilepsy sometime in the past. There have not been any recent manifestations.
Prosecutor: Do you believe he suffered an epileptic fit on the night of the murders?
Dr. Stalberg: I have no reason to believe this.
Prosecutor: Would suffering from epilepsy impair his mental faculties, the same as schizophrenia or paranoia might?
Dr. Stalberg: No it would not.
Prosecutor: Can epileptic fits be faked?
Dr. Stalberg: Very easily.
Prosecutor: Is it possible to tell a real epileptic fit from a fake one?
*719Dr. Stalberg: Perhaps a doctor could, but I doubt most lay people would be able to tell the difference.
Prosecutor: Did you form any other professional opinion as to Mr. Pin-holster’s mental condition?
Dr. Stalberg: Yes I did.
Prosecutor: And can you tell us what that is?
Dr. Stalberg: He has antisocial personality disorder.
Prosecutor: Is that what’s known, in common speech, as being a psychopath?
Dr. Stalberg: Yes it is.
Prosecutor: And what does this mean? Dr. Stalberg: It means he feels no empathy for the suffering of others. He has no conscience.
Prosecutor: Thank you very much. You may step down.
Counsel, of course, would have known that if they opened the door to psychiatric evidence, the state would be entitled to put in such evidence as well. So, if Pinholster put on Dr. Stalberg, the state would have presented its own expert, someone like Dr. Geiger, who examined petitioner on June 19, 1984, and gave his diagnosis as follows:
The subject was not impaired by mental disease or defect in such a way that he could not appreciate the criminality of his conduct or conform his conduct to the requirements of the law. This man’s conduct showed a high degree of cruelty, callousness and viciousness. ... DIAGNOSIS: Antisocial personality disorder.... CONCLUSION: This diagnosis is related to the commitment offenses in the sense that this man showed no responsible regard for the reasonable rights of other people and throughout the many years prior to this most recent conviction this character trait was apparent. [Ex. 41] (emphasis added).
Given this reality, competent counsel, as in Belmontes,15 would have been careful to avoid any possibility that the state would put in such damaging psychiatric evidence, and so would have been forced to make do without Dr. Stalberg. That would have left counsel with only the brother, the sister, the aunt, the uncle — all sniping at each other while trashing Pinholster’s mother, grandparents and step-father — and the school teacher.
This assumes, for the sake of argument, that Pinholster would have allowed a mitigation defense which publicly disgraced his mother and stepfather. Is this likely to have led even a single juror to change his mind about the penalty? I seriously doubt it. Far more likely, the jury would have seen this as just one more effort by Pinholster to manipulate them, like his lying and self-aggrandizing on the stand.
The Superior Court here made findings about the weight of the aggravating evidence: “The Court finds that the evidence concerning the truth of the Special Circumstances is overwhelming, and the jury’s assessment of the evidence that the aggravation outweighs the mitigation as to the selection of the proper penalty to be ‘death,’ is supported overwhelmingly by the weight of the evidence.” (emphases added). [CT 1184] “Considering all of the evidence, the Court finds that the factors in aggravation beyond all reasonable doubt outweigh those in mitigation.” (emphasis *720added) [CT 1186] These are the findings by the judge who saw the entire trial and the jury that convicted Pinholster; neither was set aside on appeal or otherwise. Under AEDPA these findings are entitled to a near irrebuttable presumption of correctness. In light of such overwhelming evidence, it would be impossible to conclude that even one juror would have been swayed by this paltry showing.
My colleagues disagree, but that’s not the test. What we should ask is whether the justices of the California Supreme Court unreasonably applied clearly established Supreme Court precedent in concluding otherwise. My colleagues hold that they did by comparing the mitigating evidence here with that in Terry Williams, Wiggins and Rompilla, but those cases are not on point because not a single one of them required the Court to hold that a state court’s determination on prejudice was unreasonable. Rather, in each case, the Court reviewed a state court decision that either applied a standard contrary to Supreme Court precedent or failed to address prejudice at all. Rompilla, 545 U.S. at 390, 125 S.Ct. 2456; Wiggins, 539 U.S. at 534, 123 S.Ct. 2527; Terry Williams, 529 U.S. at 395-96, 120 S.Ct. 1495. Comparison with the facts of those cases is irrelevant because the Court there wasn’t required to apply AEDPA deference. The same is true of Belmontes and Van Hook, both of which resulted in summary reversals even under pre-AEDPA law.
The case on point is Visciotti, which did involve a state court determination regarding prejudice, to which the Supreme Court unanimously deferred. 537 U.S. at 22-27, 123 S.Ct. 357. The majority mishandles Visciotti as well. It doesn’t matter whether we think the prejudice here is probably greater than in Visciotti; the Court did not say that Visciotti was a particularly close case, nor were there differing views about how Visciotti should come out, so nothing can be learned from that comparison. Indeed, Visciotti was a summary reversal, which tells us the Court didn’t even think the case was close enough to merit argument. Visciotti thus hardly marks the outer boundary of what deference the Supreme Court believes we owe state courts in deciding whether a criminal defendant was prejudiced by ineffective assistance of counsel.
Visciotti is actually remarkably similar to our case. The same California Supreme Court decided Pinholster’s second habeas petition as Visciotti’s state habeas petition; they were decided by the state supreme court about 10 months apart. In Visciotti, as in our case, the defense lawyer chose to rely on the family sympathy defense rather than “humanizing” the defendant because “[i]t was his opinion that any attempt to gain sympathy for petitioner would have failed.” In re Visciotti, 14 Cal.4th at 334, 58 Cal.Rptr.2d 801, 926 P.2d 987. Counsel therefore did not pursue — and was not aware of — a large “trove” of facts like the majority believes is the case here.
At the federal habeas hearing, Visciotti presented evidence of the
discordant atmosphere in the Visciotti family home created by an unending series of physical and verbal confrontations between petitioner’s parents; physical punishment of petitioner and his siblings; threats of violence; impermanence caused by the family’s numerous moves and its impact on school attendance and the ability to make lasting friendships; the children’s efforts to escape the household turmoil by hiding, leaving the house, early marriage, and resort to drugs as “self-medication.” Social workers, psychologists, and other witnesses testified regarding the impact *721of these events on petitioner’s development and ability to function in society.
Id. at 336, 58 Cal.Rptr.2d 801, 926 P.2d 987. Visciotti’s lawyer explained his strategy as follows:
He decided prior to jury selection in the Visciotti trial, when he saw petitioner’s videotaped reenactment of the murder, that he would attempt to elicit sympathy for petitioner’s family as his penalty phase strategy. He believed that, although sympathy for petitioner could not be expected, sympathy for petitioner’s parents might be. His defense would therefore suggest that the parents were nice people whose son should not be killed.
Id.
While the California Supreme Court justices were thrice unanimous as to Pinholster — unce on direct appeal, twice on habeas — they were twice divided in Visciotti’s. Notably, Justice Mosk, who wrote the opinion in Pinholster’s direct appeal, dissented on the IAC issue in Visciotti’s direct appeal and was in deep dissent on Visciotti’s habeas petition; he even invited the federal courts to set aside the death sentence. Id. at 359-62 & n. 1, 58 Cal. Rptr.2d 801, 926 P.2d 987. Justice Brown also dissented, and explained — in terms not so different from those employed by the majority today — the many ways in which Visciotti’s lawyer failed him, and how this must have prejudiced him:
The penalty phase proceedings against petitioner, the subject of this court’s order to show cause, are a textbook example of a process gone awry. Simply put, Agajanian [Visciotti’s lawyer] failed petitioner at every stage of the proceedings. I offer several of many, many examples that could be given.
During his pretrial preparation, Agajanian “did not send for the police report [of the Cusack incident] or go through the prosecutor’s file to read it in advance of trial and thus was surprised and unprepared to face that [aggravating] evidence.” Likewise, he “failed to investigate and discover mitigating evidence as a result of his ignorance of the types of evidence a jury might consider mitigating.”
During the penalty phase of the trial itself, Agajanian “failed to present readily available evidence that would have revealed to the jury the extent to which petitioner was subjected to psychological and physical abuse as a child, the impact the dysfunctional and peripatetic family life had on petitioner’s development, and the correlation between these events and petitioner’s resort to drugs.” Also during the penalty phase of the trial, Agajanian “delivered an unfocused closing argument, during which he undercut his client’s own case by telling the jury that the evidence of petitioner’s mental and emotional problems was not mitigating.”
In the context of the penalty phase of the trial, it is clear that Agajanian “entirely fail[ed] to subject the prosecution’s case to meaningful adversarial testing.” This court had it all wrong when, on direct appeal, it characterized Agajanian’s penalty phase closing argument as “a rambling discourse, not tied to particular evidence.” In fact, during the course of the so-called “rambling discourse,” Agajanian systematically conceded nine of the eleven aggravating and mitigating factors set forth in Penal Code section 190.3 to the prosecution.
Agajanian conceded “[t]he facts and circumstances of the case in my opinion do not have to be reviewed. There is no way to make light of those types of things just like there’s no way to make light of any kind of murder, whether or *722not there’s a robbery involved.... ” He conceded “past violence” was a factor in aggravation. He conceded “[w]ith respect to the prior conviction for assault with a deadly weapon, there’s no way to make light of that either.” He conceded “[w]ith respect to emotional disturbance, there’s no evidence of that. That isn’t even a factor to be considered.” He conceded “[w]ith respect to the next one ... victim participated or consented. That’s not applicable. There’s no evidence of that.” He conceded “same situation” with respect to justification. He conceded “[ejxtreme duress, there was no evidence of that either. Although defense lawyers would like to have that present, it’s not fair.” He conceded “with respect to diminished capacity, when you ladies and gentlemen returned this verdict of first degree murder and found special circumstances, you indicated to all of us that you did not find diminished capacity. So if you did not find diminished capacity, how can I argue that as a factor of aggravation or mitigation? It just does not apply. It’s not there.” And he conceded “the indication here was that [petitioner] was not an accomplice or that his participation was minor-exactly the opposite. [Petitioner] is, as the People said, the trigger man.”
Certainly, as the majority states, “[t]he aggravating factors were overwhelming” and the mitigating factors were “minimal in comparison.” Even in such a case, though, counsel must hold the prosecution to its heavy burden. Agajanian did not rise to the occasion. Although his abortive attempts to construct a family sympathy defense exposed some of the mitigating evidence to the jury, Agajanian undermined its effectiveness by “conceding that the jury could find that all of the possibly aggravating factors were present, and none of the mitigating.” Indeed, the referee specifically found, and the majority agrees, that Agajanian “had no intention of introducing any evidence in an attempt to draw sympathy to his client.”
Id. at 364-66, 58 Cal.Rptr.2d 801, 926 P.2d 987 (internal citations and footnote omitted). According to Justice Brown, Agajanian did far more damage to Visciotti’s case than Brainard and Dettmar did to Pinholster’s.
No doubt swayed by the strong dissents of these well-respected state justices, the district court granted habeas and we affirmed. The United States Supreme Court was not impressed:
The [Ninth Circuit] Court of Appeals disagreed with [the majority of the California Supreme Court,] suggesting that the fact that the jury deliberated for a full day and requested additional guidance on the meaning of “moral justification” and “extreme duress” meant that the “aggravating factors were not overwhelming.” Perhaps so. However, under § 2254(d)(1), it is not enough to convince a federal habeas court that, in its independent judgment, the state-court decision applied Strickland incorrectly. The federal habeas scheme leaves primary responsibility with the state courts for these judgments, and authorizes federal-court intervention only when a state-court decision is objectively unreasonable. It is not that here. Whether or not we would reach the same conclusion as the California Supreme Court, we think at the very least that the state court’s contrary assessment was not unreasonable. Habeas relief is therefore not permissible under § 2254(d).
*723Visciotti 537 U.S. at 26-27, 123 S.Ct. 357 (internal citations and quotation marks omitted).
In our case, unlike Visciotti the same state supreme court justices were unanimous in denying the habeas petition. As Visciotti shows, this was not a court whose members played “follow the leader.” Justices Mosk and Brown clearly had minds of their own; Justice Kennard wrote separately to explain why she agreed with the majority. In re Visciotti 14 Cal.4th at 357-59, 58 Cal.Rptr.2d 801, 926 P.2d 987. But in Pinholster’s case they were unanimous, with the firebrand dissenter, Justice Mosk — who had spent decades on the California Supreme Court championing the rights of criminal defendants — fully on board. Indeed, Justices Mosk and Brown did show a difference of opinion in Pinholster’s case; they both would have denied the petition entirely on the merits and not on procedural grounds as well. [Exs. B, C-7] One can’t say this was a case where the justices brushed aside the habeas petitions without giving them their full attention.
Putting Pinholster’s case next to Visciotti’s, how can we possibly say that the California Supreme Court unreasonably applied established Supreme Court precedent in the former, when a unanimous Supreme Court summarily reversed us for saying that in the latter? The cases are remarkably similar, down to the fact that the trial lawyer in both cases was disbarred after the trial. Pinholster’s lawyers were no worse than Visciotti’s and probably better; at least they didn’t concede nine of the eleven aggravating and mitigating factors set forth in the death penalty statute. Can we truly say that we are giving Pinholster’s state habeas determination on prejudice the same degree of deference that the unanimous Supreme Court gave to Visciotti’s case? I think not.
Conclusion
The trial in this case took place over a quarter century ago. Pinholster’s lawyers are both dead. Justice Mosk, who wrote the California Supreme Court’s unanimous opinion in Pinholster’s direct appeal and participated in both of his habeas petitions, is also dead. Pinholster’s two victims are long dead and forgotten; whatever hopes and aspirations they may have had were cut short because they had the misfortune of getting in the way of Pinholster’s greed and anger.
Meanwhile, prison has been good to Pin-holster. He sits in his cell reading Machiavelli, Voltaire “and all the philosophers” [ER 704], drawing pictures to sell over the internet. He enjoys the gravitas, authority and mentoring opportunities that come with being an elder in his prison gang, and has surgery performed on his knees at taxpayer expense. He still stabs people whenever he can, without passion or regret; “it was just business,” he explains. [ER 704] His conscience doesn’t trouble him about the fact that he took the lives of two fellow human beings; he has never expressed the least remorse for his killings. The people of California are entitled to put an end to Pinholster’s paid vacation and insist that the punishment lawfully imposed on him be carried out.
I have no doubt that my colleagues sincerely believe they are following the Supreme Court’s directions. Admittedly, the Court has been less than clear in this area. See, e.g., Rompilla, 545 U.S. at 377, 125 S.Ct. 2456 (majority), 395, 125 S.Ct. 2456 (Kennedy, J., dissenting); Wiggins, 539 U.S. at 514, 123 S.Ct. 2527 (majority), 538, 123 S.Ct. 2527 (Scalia, J., dissenting); Terry Williams, 529 U.S. at 367, 120 S.Ct. 1495 (majority), 416-17, 120 S.Ct. 1495 (Rehnquist, J., dissenting). But I believe it’s been clear enough, and Pinholster’s death sentence must be reinstated. If we *724do not do it ourselves, it will surely be done for us.
Because the record is extensive, spanning multiple proceedings in different courts, I offer a legend for the cited sources. “ER” refers to (he excerpts of record filed in this court; lettered exhibits, e.g., “Ex. B,” are exhibits from the record of state habeas proceedings; numbered exhibits, e.g., “Ex. 40-1,” are exhibits from the federal habeas proceeding; "DT” refers to the transcript of the evidentiary hearing in federal court; "TR” refers to the transcript of Pinholster’s original trial in state court; and “CT” is the Clerk’s transcript from the original trial in state court.
. The reasons for dropping Dr. Woods, who maintains a cottage specialty in diagnosing criminal defendants as psychotic, are obvious. None of the other experts to examine Pinholster on behalf of either side thought much of his diagnosis.
. Building such mythology appears to be one of the tools of the trade for habeas counsel. See Van Hook, 130 S.Ct. at 8. There must be a CLE course by that name.
. In the same paragraph, the majority states: “One week before the penalty hearing, counsel told the judge that they ‘did not prepare a case in mitigation’ because they 'felt there would be no penalty phase hearing.’ ” Maj. op. at 671. These statements are not from the transcript of Pinholster’s trial. What the majority is quoting are Brainard’s self-inculpatory declarations, dozens of which he dutifully signed seven years after the trial. See pp. 700-01 & nn. 6-9 infra.
. Penal Code § 190.3 deals with mitigation in capital cases.
. I'll assume, like the majority, that the lawyers were in fact surprised that the state planned to put on aggravation evidence at the penalty phase, but the record can also be construed otherwise. As noted, Dettmar’s time records show him billing for the "death penalty phase” as far back as February. Brainard too has a notation in his time records, "Start prep, for penalty phase” on April 11, two weeks before the defense team was supposedly surprised that the prosecution planned to put on aggravation evidence. Maj. op. at 658 n.2. If the defense lawyers truly had no idea there would be a penalty phase, why were both of them billing time as preparing for it?
The defense team did claim in court they weren't aware of the state’s notice, but how else were they going to have a shot at precluding the state from putting on an aggrava*700tion case? Claiming that they had actual notice, but the notice was somehow defective because it had been sent directly to Pinholster, certainly wasn’t going to be persuasive. So the defense claimed they hadn't seen the notice — which may have been literally true. But what was certainly not true for two experienced criminal lawyers in California is that they expected the prosecution wouldn't try to put on aggravation witnesses in a capital murder case. Being experienced, the lawyers must certainly have known that if the jury came back with guilty, the prosecution would have witnesses lined up to support their demand for the death penalty. So they may have done their best to stop them by telling the judge they didn’t know it was coming.
. Which means we shouldn't be considering the time records at all because Pinholster was not diligent in presenting them to the state courts. See pp. 689-91 supra. Nor did he raise the “only 6.5 hours’’ argument in his state petitions. Had he done so, I doubt the state judges would have been gulled by the bogus "only 6.5 hours” figure.
. Non-denial denial is a phrase that became popular in the wake
of the Watergate scandal, referring to an equivocal denial, particularly one made by an official to the press. London’s The Sunday Times has defined it as “an on-the-record statement, usually made by a politician, repudiating a journalist's story, but in such a way as to leave open the possibility that it is actually true.”
A “non-denial denial” is a statement that seems direct, clearcut and unambiguous at *701first hearing, but when carefully parsed is revealed not to be a denial at all, and is thus not untruthful. It is a case in which words that are literally true are used to convey a false impression; analysis of whether or when such behavior constitutes lying is a longstanding issue in ethics.
Wikipedia, Non-denial denial, http://en. wikipedia.org/wiki/Non-deniaLdenial (last visited November 22, 2009).
. The majority doesn’t quote it, but Brainard also did not recall why he failed to ask for a continuance when the trial court offered it. [ER 367]
. The majority is right that Brainard's failure to recall also doesn’t prove he did perform a reasonable investigation, maj. op. at 671, but it doesn’t matter because the state need not prove that counsel performed a reasonable investigation. It is petitioner who must prove the opposite. Porter met this burden through the affirmative statements of his only trial counsel. Porter v. Crosby, 2007 WL 1747316, at *2, 25 (M.D.Fla.2007). Pinholster has no evidence at all from one of his lawyers and the other one doesn’t remember.
. Brainard’s forgetfulness is just another example of counsel “falling on his sword," something trial counsel are known to do to help their clients on habeas. See, e.g., LaGrand v. Stewart, 133 F.3d 1253, 1276 (9th Cir.1998); Hendricks, 70 F.3d at 1039. Because counsel usually suffer few if any consequences for being found ineffective, many see no harm in this tactic; some may even see it as their duty. See generally Lawrence J. Fox, Making the Last Chance Meaningful: Predecessor Counsel’s Ethical Duty to the Capital Defendant, 31 Hofstra L.Rev. 1181 (2003). Laudable or not, this tendency on the part of counsel seriously distorts our ineffective assistance jurisprudence; competent counsel are found ineffective because they effectively concede the point. Our jurisprudence then fills up with descriptions of perfectly adequate performance that is assumed to be deficient.
. Harris is especially instructive because of the uncanny parallels to Pinholster. Harris, like Pinholster, was charged with two murders and a robbery. Harris raised an alibi defense and testified during the guilt phase, admitting to robbery but denying the two murders. (Pinholster, too, admitted robbery.) At the penalty phase, Harris’s mother and sister testified that he’d been born prematurely and had suffered abuse from his father because the father believed Harris was someone else’s son. They testified that, as a result of this abuse, Harris suffered a head injury when his father knocked him off his high chair, an injury that resulted in “convulsions, blood coming out of his mouth, nose and ears.” 949 F.2d at 1506. The jury also heard that "Harris’s father[] tried to choke Harris with a table cloth ... [and that he] beat Harris and other children ‘into unconsciousness several times when they were kids,’ and that Harris was abused from the time he was a little baby.” Id. Harris’s father "was eventually sent to prison for child abuse and molestation.” Id. After hearing all this, the jury sentenced Harris to death and he was long ago executed.
. The majority shuts its eyes to this document because it is supposedly hearsay. Maj. op. at 673. But Pinholster’s own statements to Duvall aren't hearsay; they’re an admission by a party-opponent, which the federal rules specifically exclude from the definition of hearsay. Fed.R.Evid. 801(d)(2). As for Duvall herself, her report was admitted as Exhibit 238 by stipulation. [ER 1412] Pin-holster reserved no hearsay objection (or any other objection), so it is part of the record and there is no obstacle to considering it. See United States v. Foster, 711 F.2d 871, 877 (9th *705Cir.1983). Pinholster, in fact, never objected on hearsay or any other grounds to the state’s use of the Duvall report below; the hearsay argument was raised sua sponte by the district judge, who must have found the Duvall report inconvenient to deal with. [ER 1449] Petitioner does not defend the judge’s hearsay ruling on appeal. There is something very wrong with the court sua sponte — and selectively — raising an objection that both parties have bypassed. If the Duvall report is hearsay and cannot be considered, why didn’t the district court raise similar objections to other out-of-court statements, such as the declarations of Pinholster’s siblings and the reports of Drs. Vinogradov and Olson? The parties obviously relied on mutual waiver of hearsay objections in order to streamline the hearing; we have no business upsetting that arrangement just to help out one side.
. The diagnosis proffered by Drs. Vinogradov and Olson is vigorously disputed by the state's experts, who cast serious doubt on the notion that any reputable psychiatrist would have reached the same diagnosis at the time of Pinholster’s trial. Dr. Vinogradov's theory, in a nutshell, is that Pinholster must be suffering from brain damage caused by his childhood injuries because that is the simplest explanation for his behavior. [ER 730-41, 1273-1316] Neither she nor Dr. Olson performed a CT scan, MRI or any other test that actually showed the existence of brain damage, or suggested that such tests would have been available to reveal injuries at the time of trial. The state's experts held firmly to the view that Pinholster’s childhood injuries weren’t responsible for his antisocial behavior. [ER 754, 795] Indeed, Dr. Rudnick maintained that one of the two accidents caused no head trauma at all, injuring only Pinholster’s shoulder and ear. [ER 768]
. The majority also says that Pinholster’s stepfather used a paddle on him, occasionally knocking him out. Maj. op. at 678. This is not based on evidence in the record. It's a "fact” related by a psychiatrist’s report that is based entirely on what Pinholster told him. [ER 767] While the psychiatrist can use the information he gathers as a basis for forming his opinion (which can then be undermined, to the extent the facts on which he relies are unsupported), his report can’t be used to sneak into the record evidence that was not presented by the usual means, i.e., through sworn testimony subject to cross-examination. The majority is thus wrong to rely on this "evidence.”
. The opinion’s paean to counsel's performance in Belmontes, maj. op. at 683, must come as a surprise to certain members of the majority who are particularly familiar with that case. Then again, such recharacterizations of a matter in a habeas case are not so uncommon. Belmontes, 130 S.Ct. at 387.