¶1 This case concerns the standard for assessing prejudice in a personal restraint petition alleging ineffective assistance of counsel. Hoyt Crace was convicted of attempted second degree assault with a deadly weapon. This was his third strike offense, resulting in a life sentence without the possibility of early release. Crace brought a timely personal restraint petition asserting he received *837ineffective assistance of counsel when his trial counsel did not request an instruction on the lesser-included offense of unlawful display of a deadly weapon, a nonstrike offense. A divided Court of Appeals applied the analysis for an ineffective assistance of counsel claim set forth in Strickland v. Washington, 466 U.S. 668, 104 S. Ct. 2052, 80 L. Ed. 2d 674 (1984), concluding that a showing of prejudice under this analysis satisfies the “actual and substantial prejudice” showing required on collateral attack. Without the benefit of our decisions in State v. Grier, 171 Wn.2d 17, 246 P.3d 1260 (2011) and State v. Breitung, 173 Wn.2d 393, 267 P.3d 1012 (2011), the court then granted Crace’s petition, holding that trial counsel’s performance was deficient and that counsel’s error prejudiced Crace.
¶2 We reverse the Court of Appeals. While the court correctly analyzed prejudice under the Strickland standard, Crace cannot show prejudice arising from the omitted instruction.
FACTS AND PROCEDURAL HISTORY
¶3 According to Crace, on August 16, 2003, he was at home watching television and consuming alcohol, cocaine, heroin, and the prescription pain medication Dilaudid. At around 2:00 p.m., he fell asleep, or partially overdosed, while watching a movie. When he awoke, it was dark outside and Crace began hearing and seeing things that led him to believe he was being stalked by murderous demons. Panicked and screaming, he ran from his trailer home and entered nearby trailers, frightening neighbors as he shouted, “ ‘They’re after me, they’re after me, they’re after me.’ ” Report of Proceedings (RP) at 119-21. Believing he was about to be killed, he returned to his trailer and took a “sword thing” off the wall. Id. at 123. He then ran into the street with the sword, screaming, “ ‘Help, help, they’re after me.’ ” Id.
¶4 While this scene was unfolding, Pierce County Deputy Sheriff Hardesty arrived. As he was talking to a *838neighbor, Hardesty’s attention was drawn to Crace, who was about two blocks down the street screaming and carrying on. Deputy Hardesty testified that as soon as he made eye contact with Crace, Crace began sprinting toward Hardesty, sword in hand, screaming. As Crace got closer, Hardesty pulled his duty weapon and instructed Crace to drop his sword. After Hardesty repeated this directive several times, Crace did drop the sword about 50 feet from Hardesty but kept running toward him. Still pointing his weapon, Hardesty told Crace to get to the ground. Crace finally complied when he was about 5 to 7 feet away from the deputy. According to Crace, he recognized Hardesty was a police officer but continued running toward Hardesty with the sword because he feared that if he stopped or threw down his weapon, the demons would attack. Hardesty testified that when Crace rushed him with the sword, he feared for his life, and that if Crace had come much closer — -just steps — with the sword, Hardesty would have shot him.
¶[5 Crace was charged with second degree assault with a deadly weapon. At trial, each side presented expert testimony from psychologists on the subject of Crace’s state of mind that night and his ability to form a criminal intent. The State’s witness opined that Crace could have formed intent that night, while Crace’s witness testified that the defendant was in a delusional state and could not have formed the requisite intent. At the trial’s conclusion, the jury was instructed on the lesser-included offense of attempted second degree assault. Am. Opening Br. in Supp. of Pers. Restraint Pet., App. F (Instruction 16). The jury deadlocked on the second degree assault with a deadly weapon charge but returned a verdict of guilty on the attempted second degree assault with a deadly weapon charge.
¶6 The assault conviction was Crace’s third strike, and he was sentenced to life without the possibility of early release. Am. Opening Br. in Supp. of Pers. Restraint Pet., App. A (J. and Sentence). His conviction was affirmed on direct appeal, State v. Crace, noted at 128 Wn. App. 1021, *8392005 WL 1540894, at *7, 2005 Wash. App. LEXIS 1540, at *18, and this court denied his motion for discretionary review, State v. Crace, 160 Wn.2d 1010, 161 P.3d 1026 (2007).
¶7 Crace filed a personal restraint petition with Division Two of the Court of Appeals. In re Pers. Restraint of Crace, 157 Wn. App. 81, 236 P.3d 914 (2010). He argued that his trial lawyer’s failure to request the lesser-included offense of unlawful display of a deadly weapon constituted ineffective assistance of counsel. The Court of Appeals agreed, holding that counsel’s performance was deficient and that the error prejudiced Crace. The Court of Appeals rejected the State’s argument that Crace was required to prove a measure of prejudice greater than that required under Strickland in order to satisfy the “actual and substantial prejudice” requirement for a collateral attack. In so doing, the court retreated from its prior decision in In re Personal Restraint of Davis, 151 Wn. App. 331, 211 P.3d 1055 (2009), review denied, 168 Wn.2d 1043, 234 P.3d 1172 (2010), which applied a heightened standard. Judge Quinn-Brintnall, the author of Davis, dissented on both the measure of prejudice and whether Crace received ineffective assistance of counsel.
¶8 The State filed a motion for discretionary review. We requested additional briefing on the impact of Grier, 171 Wn.2d 17, involving a similar issue of ineffective assistance of counsel and a lesser-included offense instruction. We then granted review.
ANALYSIS
¶9 This case requires us to consider the intersection between the prejudice requirement on collateral attack of a judgment and the prejudice requirement on direct appeal. Specifically, we must decide whether a personal restraint petitioner alleging ineffective assistance of counsel must undermine our confidence in the trial more than an appellant must. We must then decide whether Crace has established prejudice sufficient to support his claim.
*840Does a showing of prejudice under Strickland meet the personal restraint petitioner’s requirement to show actual and substantial prejudice ?
f 10 A successful ineffective assistance of counsel claim requires the defendant to show that counsel’s performance was deficient and that the defendant was prejudiced by the deficient performance. Strickland, 466 U.S. 668; State v. Thomas, 109 Wn.2d 222, 743 P.2d 816 (1987). The Court in Strickland defined prejudice as the “reasonable probability that, but for counsel’s unprofessional errors, the result of the proceeding would have been different. A reasonable probability is a probability sufficient to undermine confidence in the outcome.” 466 U.S. at 694.
¶11 The State complains that this showing of prejudice is insufficient when a claim of ineffective assistance of counsel is raised in a collateral attack. The State insists that the “actual and substantial prejudice” showing generally required to prevail in a personal restraint petition must be superimposed on the Strickland showing, to require proof that the outcome of the trial “more likely than not” would have been different. Suppl. Br. of Pet’r at 10 (citing In re Pers. Restraint of Hagler, 97 Wn.2d 818, 826, 650 P.2d 1103 (1982)). The State’s argument is built on a simple logical construct:
• On direct appeal, when the defendant proves a constitutional violation, the burden falls on the State to show the violation is harmless beyond a reasonable doubt.
• On collateral attack, the burden of showing prejudice shifts to the petitioner, who must prove actual and substantial prejudice by a preponderance of the evidence (i.e., that he was more likely than not harmed by the error).
• Therefore, in the context of an ineffective assistance of counsel claim raised in a personal restraint peti*841tion, the Strickland test establishes the constitutional violation, and the petitioner must additionally show actual and substantial prejudice.
See Suppl. Br. of Pet’r at 9-10. Accepting that this argument has some logical appeal, at least insofar as it adds one and one to equal two, it nevertheless overlooks the foundation of Strickland.
¶12 Strickland itself involved a collateral attack, a federal habeas corpus claim. The Court thoroughly examined the appropriate standard for determining prejudice on both direct and collateral review. In discussing the prejudice prong of its test, Strickland articulated the different levels of proof a defendant might make to show prejudice resulting from counsel’s errors. On the one hand, Strickland rejected the notion that a defendant could show an error “had some conceivable effect on the outcome of the proceeding.” Strickland, 466 U.S. at 693. Such a burden would be too easy to carry and would not account for the fact that “not every error that conceivably could have influenced the outcome undermines the reliability of the result of the proceeding.” Id.
¶13 On the other hand, the Strickland Court rejected any requirement that the defendant show counsel’s deficient conduct “more likely than not altered the outcome [of] the case.” Id. Although such a test “reflects the profound importance of finality in criminal proceedings,” its high standard “presupposes that all the essential elements of a presumptively accurate and fair proceeding were present” in the challenged proceeding. Id. at 693-94.
An ineffective assistance claim asserts the absence of one of the crucial assurances that the result of the proceeding is reliable, so finality concerns are somewhat weaker and the appropriate standard of prejudice should be somewhat lower. The result of a proceeding can be rendered unreliable, and hence the proceeding itself unfair, even if the errors of counsel cannot be shown by a preponderance of the evidence to have determined the outcome.
*842Id. at 694. Thus, Strickland arrived at a measure of prejudice that requires the defendant to show a “reasonable probability” that but for counsel’s deficient representation, the outcome of the proceeding would have been different. Id. The Court noted this test finds its roots in the test for assessing the materiality of exculpatory information the prosecution fails to disclose or of testimony made unavailable to the defense by the government’s actions. See id. (citing United States v. Agurs, 427 U.S. 97, 112-13, 96 S. Ct. 2392, 49 L. Ed. 2d 342 (1976); United States v. Valenzuela-Bernal, 458 U.S. 858, 872-74, 102 S. Ct. 3440, 73 L. Ed. 2d 1193 (1982)).
¶14 While the State might be correct that Strickland’s “reasonable probability” standard and the “more likely than not” standard under a personal restraint petition are not precisely the same standard, we cannot agree that a Strickland showing does not satisfy the petitioner’s burden on collateral attack. As noted, Strickland took its prejudice standard from cases dealing with withheld or undisclosed evidence. In that context, the United States Supreme Court has said, “The question is not whether the defendant would more likely than not have received a different verdict with the evidence, but whether in its absence he received a fair trial, understood as a trial resulting in a verdict worthy of confidence” Kyles v. Whitley, 514 U.S. 419, 434, 115 S. Ct. 1555, 131 L. Ed. 2d 490 (1995) (emphasis added).1 Likewise, “[a]n ineffective assistance claim asserts the absence of one of the crucial assurances that the result of the proceeding is reliable.” Strickland, 466 U.S. at 694. Thus, Strickland suggests that a petitioner who shows there is a reasonable probability that his trial lacked one of the crucial assurances of fairness also necessarily shows actual and substantial prejudice.
*843|15 Moreover, we are mindful that constitutional claims arising from ineffective assistance of counsel on the one hand, and claims arising from withheld or undisclosed material evidence on the other hand, share another important characteristic, one that makes them natural companions in this analytical framework. In these types of claims, prejudice inheres in the violation. That is to say, a petitioner who proves a violation shows prejudice. Writing in the context of undisclosed exculpatory evidence, Justice Souter explained:
Assuming, arguendo, that a harmless-error enquiry were to apply, [an error arising from the nondisclosure of material evidence] could not be treated as harmless, since “a reasonable probability that, had the evidence been disclosed to the defense, the result of the proceeding would have been different,” [United States v. Bagley,] 473 U. S. [667,] 682[, 105 S. Ct. 3375, 87 L. Ed. 2d 481 (1985)] (opinion of Blackmun, J.); id., at 685 (White, J., concurring in part and concurring in judgment), necessarily entails the conclusion that the suppression must have had “ ‘substantial and injurious effect or influence in determining the jury’s verdict,’ ” Brecht v. Abrahamson, 507 U. S. 619, 623 [, 113 S. Ct. 1710, 123 L. Ed. 2d 353] (1993), quoting Kotteakos v. United States, 328 U. S. 750, 776[, 66 S. Ct. 1239, 90 L. Ed. 1557] (1946).
Kyles, 514 U.S. at 435. There is no principled reason to depart from this reasoning in the context of ineffective assistance of counsel claims when the authority available to us draws very clear links between these classes of claims. If there is a reasonable probability that had counsel not been deficient, the result of the proceeding would have been different, then that necessarily requires a finding that counsel’s error had a substantial and injurious effect on the proceedings. See Harrington v. Richter,_U.S._, 131 S. Ct. 770, 792, 178 L. Ed. 2d 624 (2011) (noting that “[t]he *844likelihood of a different result must be substantial, not just conceivable”).2
¶16 Strickland’s test is ultimately concerned with “the fundamental fairness of the proceeding whose result is being challenged.” 466 U.S. at 696. “In every case the court should be concerned with whether, despite the strong presumption of reliability, the result of the particular proceeding is unreliable because of a breakdown in the adversarial process that our system counts on to produce just results.” Id. The court emphasized that this concern is the same in a collateral attack as on direct appeal. Id. at 697-98.
¶17 The Court of Appeals below appropriately recognized that the State’s argument would run counter to Strickland, as it would require adopting the outcome-determinative standard that Strickland expressly rejected. Crace, 157 Wn. App. at 113 n.18. It is difficult to conceive of how much more actually and substantially a petitioner could be prejudiced than by an unfavorable result that was caused by “a breakdown in the adversarial process that our system counts on to produce just results.” Strickland, 466 U.S. at 696. We cannot say it any more clearly than the Strickland Court did when it reasoned that because “fundamental fairness is the central concern of the writ of habeas corpus, no special standards ought to apply to *845ineffectiveness claims made in habeas proceedings.” Id. at 697-98 (citation omitted). Likewise, for a petitioner on collateral attack claiming ineffective assistance of counsel, no “double prejudice” showing above and beyond the prejudice showing required under Strickland should be imposed. In meeting his Strickland burden, a petitioner has necessarily met the burden of proving “actual and substantial prejudice.”
¶18 While we have not before addressed this issue directly, this court has at least tacitly rejected the State’s argument in a long line of personal restraint petition cases employing the Strickland prejudice test. In In re Personal Restraint of Davis, 152 Wn.2d 647, 700, 101 P.3d 1 (2004), for example, our analysis of prejudice equated a showing under Strickland with a showing of actual and substantial prejudice. We held, “Petitioner Davis cannot establish actual and substantial prejudice .... Because there was overwhelming evidence of his guilt, he cannot show there was a reasonable probability that, but for his counsel’s deficient performance by not objecting, the outcome of the trial would have been different.” Id.
¶19 In In re Personal Restraint of Woods, 154 Wn.2d 400, 114 P.3d 607 (2005), we considered both an ineffective assistance claim and a Brady3 claim, two areas, as noted, in which the Supreme Court has adopted the “reasonable probability” prejudice standard for both direct appeal and collateral review. See Strickland, 466 U.S. at 694, 697-98; Kyles, 514 U.S. 419. With respect to both claims raised in Woods’ personal restraint petition, we held that his burden in establishing prejudice was to show a reasonable probability that the outcome of the proceedings would have been different. Woods, 154 Wn.2d at 421 (ineffective assistance claim); id. at 428-29 (Brady claim, noting that “[t]he question to be answered is not whether the defendant would more likely than not have received a different verdict with *846the evidence but whether the absence of the evidence undermines confidence in the verdict”).
¶20 Woods is particularly instructive because it confirms our application of the same prejudice standard under Strickland and in the Brady context in which the Strickland standard “finds its roots.” Strickland, 466 U.S. at 694. So does In re Personal Restraint of Rice, 118 Wn.2d 876, 828 P.2d 1086 (1992), upon which the State heavily relies.4 Contrary to the State’s description of Rice, it is simply one of many cases in which this court employed the “reasonable probability” prejudice standard in resolving an ineffective assistance of counsel claim in a personal restraint petition. See Rice, 118 Wn.2d at 890; see also In re Pers. Restraint of Dalluge, 152 Wn.2d 772, 100 P.3d 279 (2004) (holding petitioner demonstrates actual and substantial prejudice by proving a claim under Strickland), In re Pers. Restraint of Brett, 142 Wn.2d 868, 883,16 P.3d 601 (2001) (granting personal restraint petition based upon Strickland showing). We have never imposed a “double prejudice” standard on a petitioner arguing ineffective assistance of counsel on collateral attack, and we decline the State’s invitation to do so now.5 We hold that if a personal restraint petitioner makes a successful ineffective *847assistance of counsel claim, he has necessarily met his burden to show actual and substantial prejudice.
Did Crace meet his burden to establish ineffective assistance of counsel under Strickland?
¶21 Although Crace need not show more prejudice on collateral attack than on direct appeal, he must of course satisfy the Strickland test in order to have his personal restraint petition granted. We need not consider both prongs of Strickland (deficient performance and prejudice) if a petitioner fails on one. Strickland, 466 U.S. at 697. We conclude that Crace cannot show prejudice under Strickland and therefore do not address the question of whether his counsel’s performance was deficient.6
¶22 In holding Crace met his burden to demonstrate prejudice, the Court of Appeals did not have the benefit of our decisions in Grier, 171 Wn.2d 17, and Breitung, 173 Wn.2d 393. There, we rejected claims of ineffective assistance of counsel based on counsel’s failure to seek jury instructions on lesser included offenses. With respect to prejudice, we noted in Grier that the court must assume “that the jury would not have convicted [the defendant] unless the State had met its burden of proof.” 171 Wn.2d at 43-44. And, we must assume that “the availability of a compromise verdict would not have changed the outcome of [the] trial.” Id. at 44 (quoting Strickland’s admonishment that “ ‘a court should presume . . . that the judge or jury acted according to the law’ ” 466 U.S. at 694 (alteration in original)).
¶23 Assuming without deciding that counsel was deficient, consistent with Grier, we cannot say in all reasonable probability that counsel’s error — failure to seek the lesser included offense — contributed to Crace’s conviction on attempted second degree assault. There was sufficient evi*848dence from which a juror could conclude Crace committed this offense. Evidence established he intended to cause Deputy Hardesty fear and apprehension. RP at 143-45 (cross-examination of Crace suggesting that he rushed the deputy thinking Hardesty might be an assailant), id. at 208-17 (testimony from State’s psychologist suggesting Crace was not only capable of forming intent but was also malingering, which might have undermined his credibility with the jury). Indeed, if failing to request the lesser-included instruction was deficient performance, it occurred during an otherwise strategic and tactically driven presentation by counsel. In light of the presumptions we recognized in Grier, it would be difficult to show prejudice in such a context, and Crace has failed to do so here.
CONCLUSION
¶24 We hold that a petitioner who shows prejudice under Strickland necessarily meets his burden to show actual and substantial prejudice on collateral attack. However, we reverse the Court of Appeals because Crace has not shown that but for counsel’s alleged deficient performance, a reasonable probability exists that the outcome of his trial would have been different.
Madsen, C.J., and Owens, Fairhurst, J.M. Johnson, and González, JJ., concur.As with Strickland, Kyles was a habeas petition. It dealt with the withholding of exculpatory evidence, a constitutional claim that, like ineffective assistance of counsel, requires a showing of prejudice in order to establish a constitutional error. Kyles, 514 U.S. at 435-36 & n.9.
An additional word should he said about harmless error. At least one federal court has rejected a suggestion that harmless error analysis should be applied to a successful ineffective assistance of counsel claim on a habeas petition. Hill v. Lockhart, 28 F.3d 832, 839 (8th Cir. 1994). This is exactly right. A successful showing of ineffective assistance of counsel establishes actual prejudice — i.e., that the error was not harmless. This reflects the burden-shifting that applies generally to ineffective assistance of counsel claims. Unlike the standard under Chapman v. California, 386 U.S. 18, 87 S. Ct. 824, 17 L. Ed. 2d 705 (1967), where the State bears the burden to show a constitutional error is harmless beyond a reasonable doubt, here the defendant always bears the burden to show prejudice. See Strickland, 466 U.S. at 693-94. If he succeeds, there is nothing left for the State to rebut. Hill was cited with approval in Kyles, which rejected a harmlessness analysis in the context of withholding of exculpatory evidence. Kyles, 514 U.S. at 435-36 & n.9. The harmless error discussion becomes more involved with respect to constitutional errors that are per se prejudicial on direct appeal, but Strickland appropriately distinguished such errors from a claim of ineffective assistance of counsel. 466 U.S. at 692.
Brady v. Maryland, 373 U.S. 83, 83 S. Ct. 1194, 10 L. Ed. 2d 215 (1963).
Rice makes reference to a “prima facie case of ineffective assistance [of counsel].” 118 Wn.2d at 889. The State seems to read this language as confirmation of its argument that the Strickland test merely establishes error and that a further showing of prejudice is required, presumably reading “prima facie” to mean that a Strickland showing leaves open an opportunity for the State to rebut the petitioner’s evidence of a constitutional violation. Suppl. Br. of Pet’r at 11. As explained, the Strickland test does not operate like this. Once a petitioner meets his or her burden under Strickland, there is nothing left for the State to rebut.
At oral argument, the State offered our recent decision in State v. Sandoval, 171 Wn.2d 163, 249 P.3d 1015 (2011), as confirmation of the “double prejudice” standard. In Sandoval, we explained that a personal restraint petitioner who had not yet had an opportunity to appeal an ineffective assistance of counsel claim to a disinterested judge did not need to show actual and substantial prejudice but still had the burden of establishing prejudice under Strickland. Id. at 169 (quoting In re Pers. Restraint of Grantham, 168 Wn.2d 204, 214, 227 P.3d 285 (2010)). Any discussion in Sandoval that appears to draw a quantitative distinction between the burden on collateral attack and the burden under Strickland does not control here. The question we review today was not at issue in Sandoval.
Because we do not reach the issue of deficient performance, we do not consider counsel’s declaration filed in support of Crace’s personal restraint petition.