with whom

The panel’s decision has effectively written out of the law the requirement that prejudice be pleaded and proved to meet the test for ineffective assistance of counsel. See Strickland v. Washington, 466 U.S. 668, 687 , 104 S.Ct. 2052 , 80 L.Ed.2d 674 (1984); Turner v. Calderon, 281 F.3d 851, 879 (9th Cir.2002). In fact, Hoffman alleged only that had his ineffective counsel made Hoffman competent, Hoffman would then have been able to assess his position and then decide whether to take the plea offer. 1 He did not allege that if *687 competent and if counsel had advised him to take the plea offer, he would have done so.

Further, in what may be a new high in self-effacing candor, the panel holds that it is ineffective assistance of counsel to rely on Ninth Circuit precedent with respect of federal constitutional law applicable in states located in this Circuit. Therefore, I respectfully dissent from the order denying rehearing en banc.

Hoffman and an associate, Ron Wages, killed Denise Williams, a police informant, after Williams made a controlled drug buy which resulted in the arrest of Richard Holmes. State v. Hoffman, 123 Idaho 638 , 851 P.2d 934, 935-36 (1993). On Holmes’s orders, Hoffman and Wages kidnaped Williams and drove her to a cave outside Silver City, Idaho. Id. at 936.

Id.

Hoffman was charged with first degree murder and offered a plea bargain: if Hoffman pleaded guilty to murder, the state would not seek the death penalty. Hoffman v. Arave, 455 F.3d 926, 929 (9th Cir.2006). However, Hoffman’s attorney, William Wellman, advised Hoffman to reject the plea. Id. This court had held Arizona’s death penalty scheme unconstitutional because in Arizona, as in Idaho, the death sentence was imposed by a judge, not a jury. Adamson v. Ricketts, 865 F.2d 1011, 1023-28 (9th Cir.1988) (en banc), abrogated by Walton v. Arizona, 497 U.S. 639 , 110 S.Ct. 3047 , 111 L.Ed.2d 511 (1990). Wellman thought it was only a matter of time before Idaho’s death penalty scheme would also be held unconstitutional. Hoffman, 455 F.3d at 929 . Hoffman let the plea offer lapse, and was tried and sentenced to death. Id. at 930 . A year after the sentence was imposed, the Supreme Court upheld Arizona’s judge-sentencing death penalty scheme and implicitly overruled Adamson. Walton, 497 U.S. at 647-49 , 110 S.Ct. 3047 .

The panel held Wellman provided ineffective assistance of counsel when he advised Hoffman to reject the plea agreement. Hoffman, 455 F.3d at 939-41 . Ineffective assistance of counsel requires both a showing that counsel’s performance was deficient and that such deficient performance caused defendant prejudice. Strickland, 466 U.S. at 687 , 104 S.Ct. 2052 . The panel found that Wellman’s performance was deficient because, while he was correct that Idaho’s death penalty scheme was “materially indistinguishable” from Arizona’s, Wellman did not discover that the Supreme Court of Arizona consistently upheld Arizona’s death penalty scheme. Hoffman, 455 F.3d at 940 . But Adamson was good law while the plea bargain was available. The Idaho Supreme Court had not rejected our Adam-son position, and, in a pre-AEDPA world, our decision would have made federal ha-beas relief for Hoffman likely. 2 There is *688 no evidence that had Wellman researched Arizona and Idaho state court rulings he would still not have relied on our Adam-son decision. The panel opinion rather astonishingly holds that it is deficient performance for an attorney to rely on the decisions of this court with regard to federal constitutional rights in states located in the Ninth Circuit.

By holding that failing correctly to predict future court decisions or trial court actions constitutes deficient performance, the panel is opening this court up to a cavalcade of challenges. Every defendant whose attorney reasonably predicted a likely sentence which turned out to be wrong, or who erroneously predicted the direction of the court’s constitutional holdings, has a claim for deficient performance. And yet, how often does an attorney give advice that does not in some way predict future court action?

After erroneously finding deficient performance, and in the absence of any allegation of prejudice, the panel finds prejudice because “had Wellman fully presented Hoffman’s options and told Hoffman that he was giving up very little in exchange for the security of the death penalty being off the table, Hoffman probably would have gone along with Wellman’s suggestion and would have accepted the plea agreement.” Hoffman, 455 F.3d at 942 (emphasis added). The panel also noted that “Hoffman’s desire to have the State prove its case was not a principled stand against accepting a plea agreement.” Id.

The panel opinion supplants the requirement of pleading and proof of prejudice, ie., that but for Wellman’s incorrect advice he would have pleaded to a life sentence, by creating a new vehicle to establish prejudice: that defendants’ rejections of plea agreements are invalid unless the state proves defendant based his rejection on a “principled stand.” The panel did not favor us with any authority for this invention, nor with a description of what principles erect an adequate foundation to effect a “principled stand.” Does the “principle” upon which defendant “stands” in his rejection have to be rationally probable, vaguely possible, or can it be simple defiant disdain for the authorities?

Up to now, under our jurisprudence, Hoffman would have to demonstrate that “but for counsel’s errors, he would have pleaded guilty and would not have insisted on going to trial.” Turner, 281 F.3d at 879 (emphasis added). In fact, “[wjhile a guilty plea may be tactically advantageous for the defendant, the plea is not simply a strategic choice; it is itself a conviction, and the high stakes for the defendant require the utmost solicitude.” Florida v. Nixon, 543 U.S. 175, 187 , 125 S.Ct. 551 , 160 L.Ed.2d 565 (2004) (internal quotation marks and citations omitted). Requiring a “principled stand” to validate a refusal to enter a plea agreement hardly demonstrates “the utmost solicitude” for a defendant’s choice. To the contrary, it creates a barrier to effective pleading out.

The panel relies on Hoffman’s “compliant personality” and tendency to defer to Wellman throughout the trial to establish that “Hoffman probably would have gone along with Wellman’s suggestion” to accept the plea agreement. Hoffman, 455 F.3d at 942 . However, Hill v. Lockhart, 474 U.S. 52, 59-60 , 106 S.Ct. 366 , 88 L.Ed.2d 203 (1985), requires more: (1) an allegation in the habeas petition that but for counsel’s ineffective assistance, petitioner would have pleaded guilty and (2) *689 proof — not just post-hoc appellate court speculation — that the petitioner indeed would have accepted the plea agreement had the attorney correctly advised the petitioner.

In Hill , the Supreme Court affirmed the district court’s denial without an evidentia-ry hearing of Hill’s habeas petition because Hill did not satisfy the “prejudice” requirement for ineffective assistance of counsel claims. Hill claimed ineffective assistance of counsel because his attorney informed Hill he would be required to serve only one-third of his sentence before becoming eligible for parole. Id. at 54-55 , 106 S.Ct. 366 . In fact, as a second offender, Hill was required to serve at least one-half of his sentence before becoming eligible for parole. Id. at 55 , 106 S.Ct. 366 . The Court held that Hill “failed to allege the kind of ‘prejudice’ necessary to satisfy the second half of the Strickland v. Washington test.” Id. at 60 , 106 S.Ct. 366 .

Id. The Court concluded, “[bjecause petitioner in this case failed to allege the kind of ‘prejudice’ necessary to satisfy the second half of the Strickland v. Washington test, the District Court did not err in declining to hold a hearing on petitioner’s ineffective assistance of counsel claim.” Id.

In this case, Hoffman did not meet the Hill requirement that he allege that he would have accepted the plea agreement if Wellman had “properly” advised him as to the law and the judge’s propensity on death sentences. In fact, were the allegations of Hoffman’s habeas petition taken as true to test their legal sufficiency, it remains unclear whether Wellman’s allegedly deficient advice actually caused Hoffman to turn down the plea offer. First, Hoffman alleged that he “proceeded to trial due to the advice of trial counsel that he would not be executed regardless of the outcome.” But then Hoffman alleges:

Note that Hoffman does not allege that had he become competent through adequate care and treatment, he would have taken the plea offer. All he alleges is that, if so cured, he would have been “sufficiently competent to determine ” whether to take the plea.

It appears that Hoffman has here pleaded himself into a corner. 3 Even if he had been given “good” advice by Wellman, Hoffman alleges he wasn’t competent to take it. And of one thing we can be sure: Wellman did not render ineffective assistance by failing to challenge Hoffman’s competency: “Hoffman has not proved an ineffective assistance claim based on his counsels’ failure to challenge his competency because there is not a reasonable proba *690 bility that Hoffman would have been found incompetent.” Hoffman, 455 F.3d at 938 . So, there is no causal connection between Wellman’s supposedly ineffective assistance in predicting constitutional law and Hoffman’s refusal to plead guilty. Hoffman claims he was incompetent to follow Wellman’s advice, let alone that he could have rejected it and pleaded guilty.

This pleading has a further effect. Before pleading he was incompetent to take the plea, through counsel’s ineffective assistance, Hoffman had alleged that he had rejected the plea agreement “due to the advice of trial counsel.” Hence, he was (1) competent to reject the plea but (2) not competent to take the plea, all during the same ten-day period in February 1988.

Moreover, Hoffman did not offer any proof that he was prejudiced by Wellman’s advice. Hoffman could have offered an affidavit, deposition testimony or eviden-tiary hearing testimony that he would have taken the guilty plea had he been given the correct advice. He did not do so. If the bar for defective performance is as low as the panel sets it, a showing — that is, pleading and proof — of prejudice is absolutely required lest we be forced to find ineffective assistance of counsel in every case where an attorney makes a prediction about law or judges and the prediction turns sour.

En banc rehearing was necessary because not only does the panel find deficient performance in an attorney’s failure correctly to predict future court decisions (the Crystal Ball Rule), the panel finds prejudice in a mere supposition by this court that the defendant would have taken the plea had his attorney guessed correctly. Accordingly, I respectfully dissent from the denial of rehearing en banc.

. Hoffman alleged in his habeas petition: Had Petitioner been properly advised regarding his exposure to the death penalty and had defense counsel fully developed and understood Petitioner's mental state including his brain damage, dyslexia, mental illness including psychosis, and mental retardation, counsel could have secured to Petitioner adequate care and treatment that would have rendered Petitioner sufficiently *687 competent to determine to take advantage of the plea bargain offered and plead guilty.

. Of course, Wellman’s advice ultimately proved to be correct. Only his timing was wrong. The Supreme Court reversed Walton in 2002 in Ring v. Arizona, 536 U.S. 584 , 122 S.Ct. 2428 , 153 L.Ed.2d 556 (2002), ruling that a jury must find aggravating circum *688 stances necessary for the imposition of the death penalty. However, Hoffman is not helped by Ring ; it is not retroactive. Schriro v. Summerlin, 542 U.S. 348, 358 , 124 S.Ct. 2519 , 159 L.Ed.2d 442 (2004).

. A petitioner can plead inconsistent theories. But here Hoffman has not done so. He never pleads that if he had become competent, he would have taken the plea, nor that, in the alternative, he was competent to take the plea.