¶31 (concurring) — This is a case about whether a personal restraint petitioner must show *604“actual and substantial prejudice” to prevail on a collateral challenge to a guilty plea where the petitioner was misadvised that the statutory maximum was 20 years when the actual maximum was life and where the sentence imposed was far lower than either 20 years or life. Under our prior case law, these are constitutional errors that would have been presumed prejudicial on appeal. The majority holds that for this particular constitutional error, the greater interest in finality that attaches postappeal militates in favor of an actual prejudice inquiry on a personal restraint petition (PRP). I agree.
¶32 But I disagree with the two pillars of the majority’s analysis. First, the majority completely redefines what “prejudice” means in this context. The majority holds that in the plea-bargain context, prejudice means only a sentence that is longer than the statutory maximum or longer than the maximum of which the petitioner was advised. But the due process clause actually guarantees a procedure that allows the defendant to make a knowing, intelligent, and voluntary plea decision. U.S. Const, amend. XIV, § 1; Wash. Const, art. I, § 3. “Prejudice,” therefore, occurs when an error undermines that procedure and the voluntariness of the plea — regardless of the sentence imposed. Second, the majority uses this case to undertake a radical reanalysis of the sort of prejudice that every personal restraint petitioner has been required to show in every other type of PRP. That radical reanalysis is not necessary to decide this case. In fact, the rule established in In re Personal Restraint of Richardson, 100 Wn.2d 669, 679, 675 P.2d 209 (1983), overruled on other grounds by State v. Dhaliwal, 150 Wn.2d 559, 568, 79 P.3d 432 (2003), State v. Kitchen, 110 Wn.2d 403, 413, 756 P.2d 105 (1988), and In re Personal Restraint of Gunter, 102 Wn.2d 769, 774, 689 P.2d 1074 (1984), and restated in In re Personal Restraint of St. Pierre, 118 Wn.2d 321,328,823 P.2d 492 (1992) — that errors which are presumptively prejudicial on direct appeal will generally be presumed prejudicial in a PRP — is still good law. So although I concur in the outcome, I disagree with the majority’s two main points.
*605ANALYSIS
¶33 The majority’s principal error is totally redefining what “prejudice” means in this context. The majority asserts that erroneous misadvice in the plea-bargain context does not cause prejudice unless the defendant gets a sentence higher than the maximum, majority at 603, or higher than what the plea agreement said, id. The majority even asserts that this prejudice inquiry “do[es] not require insight into the defendant’s state of mind and motivations, but simply look[s] at the practical effects that resulted from error.” Id. at 602 (emphasis added).
¶34 This flatly contradicts state and federal due process clause jurisprudence. When we are dealing with the voluntariness of a plea, an error causes harm if it undermines the voluntariness of the decision to plead guilty — the process that is supposed to ensure a knowing, intelligent, and voluntary decision, not a particular sentence. This is true under our seminal decisions on this topic. In re Pers. Restraint of Hews, 108 Wn.2d 579, 594, 597, 741 P.2d 983 (1987) (Hews II) (court must examine “ ‘totality of circumstances’ ” to determine whether petitioner understood nature of charge, elements and whether Hews “had discussed with his attorney alternative courses of action”); In re Pers. Restraint of Mendoza Montoya, 109 Wn.2d 270, 277, 744 P.2d 340 (1987).10 This is true under the United States Supreme Court’s seminal decisions on this topic. Boykin v. Alabama, 395 U.S. 238, 242, 89 S. Ct. 1709, 23 L. Ed. 2d 274 (1969). This is true under United States Supreme Court decisions on the related topic of what constitutes ineffective assistance of counsel in the plea-advice context. Lafler v. *606Cooper,_U.S._, 132 S. Ct. 1376,1390-91,182 L. Ed. 2d 398 (2012) (distinguishing inquiry applicable to ineffective assistance claim arising in the plea context from requirement that plea itself be “knowing and voluntary”; treating prejudice in the former context as having adverse effect on cost-benefit analysis involved in deciding whether to plead guilty); Padilla v. Kentucky, 559 U.S. 356, 369, 130 S. Ct. 1473,176 L. Ed. 2d 284 (2010) (defense counsel must advise defendant pleading guilty of the consequence of deportation to provide effective assistance).
¶35 The majority’s approach departs from this constitutionally required focus on the voluntariness of the plea procedure. This flouts state and federal constitutional law. It is also illogical: since the defendant’s only claim is that his plea was involuntary, that should be the center of the court’s inquiry.
¶36 The majority supports its analysis with dicta from St. Pierre about a heightened prejudice standard ostensibly applicable in most — though St. Pierre did not say all11— PRPs. The majority suggests that St. Pierre established a one-size-fits-all “actual and substantial prejudice” prerequisite to relief for all PRPs.12
¶37 The majority does accurately report some of St. Pierre’s discussion. In the 20 years since St. Pierre was decided, however, its dicta has been substantially eroded. In fact, contrary to St. Pierre’s dicta, there is no single rule that personal restraint petitioners must show actual and substantial prejudice to obtain relief in all cases. Rather, under our controlling precedent, I identify four categories of PRPs triggering distinct analyses of prejudice.
*607¶38 First, we have PRPs alleging constitutional errors of the “trial”13 type (as opposed to the “structural” type). In those cases, the petitioner must generally prove actual and substantial prejudice by a preponderance of the evidence to prevail. E.g., In re Pers. Restraint of Haverty, 101 Wn.2d 498, 504, 681 P.2d 835 (1984); St. Pierre, 118 Wn.2d at 329.
¶39 Next, we have PRPs raising claims of nonconstitutional error. In those cases, the petitioner must prove a fundamental defect resulting in a complete miscarriage of justice, also by a preponderance of the evidence, to prevail. In re Pers. Restraint of Woods, 154 Wn.2d 400, 409,114 P.3d 607 (2005) (citing In re Pers. Restraint of Isadore, 151 Wn.2d 294, 298, 88 P.3d 390 (2004)), overruled in part on other grounds by Carey v. Musladin, 549 U.S. 70, 127 S. Ct. 649, 166 L. Ed. 2d 482 (2006). This category includes, for example, many errors in sentencing calculation. E.g., In re Pers. Restraint of Goodwin, 146 Wn.2d 861, 864-65, 50 P.3d 618 (2002). These first two categories make sense because the value of finality weighs more heavily in favor of repose at the collateral challenge — as opposed to the appeal — stage of criminal proceedings.
¶40 In addition, this court has clearly recognized a category of PRP where the petitioner need not prove harm in addition to that which is inherent in proof of the error itself. This category includes claims of ineffective assistance of counsel and prosecutorial withholding of material exculpatory evidence. In re Pers. Restraint of Crace, 174 Wn.2d 835, 843, 280 P.3d 1102 (2012) (explaining that claims of ineffective assistance of counsel and prosecutorial withholding of exculpatory evidence “share [an] important characteristic [in that] a petitioner who proves a violation [necessarily] shows prejudice,” without any further, secondary requirement of additional prejudice on collateral re*608view).14 This category makes sense because in these cases, proof of a harmful effect on the trial outcome inheres in the claim itself.
¶41 Then there are PRPs raising claims of so-called “structural” error. Structural errors do not really trigger a presumption of harm at all. Instead, they so fundamentally undermine the adversarial process that they “defy analysis by ‘harmless-error’ standards.” Arizona v. Fulminante, 499 U.S. 279, 309, 111 S. Ct. 1246, 113 L. Ed. 2d 302 (1991) (Rehnquist, C.J., concurring); see also United States v. Gonzalez-Lopez, 548 U.S. 140, 149 n.4, 126 S. Ct. 2557, 165 L. Ed. 2d 409 (2006). “Structural” errors resulting in automatic reversal on direct appeal include courtroom closure, State v. Wise, 176 Wn.2d 1, 15, 288 P.3d 1113 (2012); complete lack of counsel, Gideon v. Wainwright, 372 U.S. 335, 342-44, 83 S. Ct. 792, 9 L. Ed. 2d 799 (1963); judicial bias, Tumey v. Ohio, 273 U.S. 510, 535, 47 S. Ct. 437, 71 L. Ed. 749 (1927); race discrimination in grand jury selection, Vasquez v. Hillery, 474 U.S. 254,106 S. Ct. 617, 88 L. Ed. 2d 598 (1986); and defective reasonable doubt instructions, Sullivan v. Louisiana, 508 U.S. 275, 113 S. Ct. 2078, 124 L. Ed. 2d 182 (1993). Our court has treated some errors of this sort as requiring automatic reversal when raised in a collateral attack, though without using the label “structural” error. E.g., Richardson, 100 Wn.2d at 673-74. This category makes sense because, as discussed above, structural defects “defy analysis by ‘harmless-error’ standards.” Fulminante, 499 U.S. at 309 (Rehnquist, C.J., concurring).
¶42 To be sure, this court has not yet decided whether all structural errors, or to use different language, all errors that result in automatic reversal on direct appeal without proof of prejudice, must also result in automatic reversal in a PRP. But it has certainly held that some do. See Kitchen, *609110 Wn.2d at 413 (££[t]hose types of constitutional errors which can never be considered harmless on direct appeal will also be presumed prejudicial for purposes of personal restraint petitions”). This court has, in effect, treated double jeopardy clause violations as falling into this category. In re Pers. Restraint of Orange, 152 Wn.2d 795, 820-22,100 P.3d 291 (2004); In re Pers. Restraint of Francis, 170 Wn.2d 517, 524, 242 P.3d 866 (2010); State v. Mutch, 171 Wn.2d 646, 663-64, 254 P.3d 803 (2011). It has ruled that the failure to require the State to prove its case beyond a reasonable doubt falls into this category. Gunter, 102 Wn.2d at 774. It has placed conflict of interest resulting in deprivation of counsel into this category. Richardson, 100 Wn.2d at 679 (counsel’s actual conflict of interest falls into this category; query whether proof that the conflict was actual rather than potential in fact places this in the category with Crace-type errors).
¶43 Into which category does the plea-advice error alleged in this case fall? On direct appeal, we have presumed prejudice from an error in counsel’s advice so apparently important that it could be presumed to have affected the knowing, intelligent, and voluntary nature of the plea. State v. Weyrich, 163 Wn.2d 554, 557,182 P.3d 965 (2008); State v. Mendoza, 157 Wn.2d 582, 590, 141 P.3d 49 (2006); Isadore, 151 Wn.2d at 302.15 On PRP, however, we traditionally put such errors into the first category described above, for constitutional errors of the “trial type,” and required some proof of an effect on the petitioner’s decision about whether to plead guilty. In re Pers. Restraint of Hews, 99 Wn.2d 80, *61088, 660 P.2d 263 (1983) (Hews I) (“An invalid plea of guilty constitutes actual prejudice.”); Hews II, 108 Wn.2d at 588-89 (“actual prejudice” established where petitioner shows that when he pleaded guilty, “he did not possess the requisite understanding of the law in relation to the facts” (citing Hews I, 99 Wn.2d at 87)). When the petitioner claimed such prejudice and the State disputed the existence of prejudice the petitioner would get a reference hearing to resolve that factual issue. Hews I, 99 Wn.2d at 88 (citing RAP 16.11(a), 16.12); cf. Isadore, 151 Wn.2d at 300 (where prosecutor neglected to inform petitioner of a direct consequence of his plea — a mandatory one-year community placement — petitioner was not required to show that the misinformation was material to his decision to plead guilty; given peculiar procedural posture, however, we applied the direct appeal prejudice standard).
¶44 The one exception is In re Personal Restraint of Bradley, 165 Wn.2d 934,205 P.3d 123 (2009). In Bradley, we held that the petitioner was entitled to withdraw his plea where the prosecution failed to advise him that his juvenile convictions “should have ‘washed out’ of his offender score,” id. at 938, because the court “will not speculate” about that misadvice’s actual effect but will instead presume that the misadvice caused the plea. Id. at 940 (citing Isadore, 151 Wn.2d at 302). The Bradley decision simply cited Isadore for this rule without noting that Isadore was not subject to regular PRP requirements, given the peculiar procedural posture of that case.16 Bradley is therefore out of step with Hews I, Hews II, Montoya, and their progeny. Instead, even though we presume prejudice on appeal from misadvice like the misadvice in this case, we require a petitioner to show prejudice to prevail in a PRP.
*611¶45 But the majority’s erroneous redefinition of “prejudice” is completely out of step with all of our prior case law and all of the United States Supreme Court’s prior case law. It essentially treats claims of constitutional error in the plea-bargain process as nonconstitutional errors subject to a far more demanding prejudice inquiry. The majority does this by demanding proof of not just actual and substantial prejudice to the right at issue — the right to a knowing, intelligent, and voluntary plea process — but something more. That something more seems to be proof of a far longer sentence.
¶46 In sum, the majority has watered down the “prejudice” standard for this category of constitutional error. Its analysis is not compelled by precedent. In fact, it conflicts with our seminal PRP plea-advice cases and with state and federal authority holding that the due process clause protects the plea process, not just the plea outcome. The majority instead cites St. Pierre. But it cites St. Pierre's dicta — much of it now discredited — rather than St. Pierre's holding.17
CONCLUSION
¶47 Under our prior cases, a personal restraint petitioner can prevail only if he or she shows (1) a constitutional *612error that caused actual and substantial prejudice for constitutional errors of the “trial” type; (2) a nonconstitutional error that inherently caused a complete miscarriage of justice; (3) a limited number of constitutional errors where prejudice inheres in proof of the error itself — Crace explicitly placed Strickland,18 and Brady19 errors into this category; or (4) structural errors (e.g., biased judge, courtroom closure, complete deprivation of counsel, improper beyond a reasonable doubt instruction) resulting in automatic reversible error. Here, Stockwell raises a constitutional error of the first type. He pleaded guilty after being erroneously advised that he faced a maximum sentence of 20 years, when the maximum sentence was actually imprisonment for life. In re Pers. Restraint of Stockwell, 161 Wn. App. 329, 331-32, 254 P.3d 899 (2011). Stockwell did not receive a sentence as high as either 20 years or life; he received a 24 month special sex offender sentencing alternative sentence. He must therefore prove actual and substantial prejudice to his right to a knowing, intelligent, and voluntary plea-bargain process to prevail on his PRP. He need not prove a longer sentence than the maximum or a longer sentence than he expected; in short, he need not prove the sort of complete miscarriage of justice that the majority’s redefinition of “prejudice” would require.
¶48 He has not met his burden. He has made no allegation of prejudice at all. I would therefore deny Mr. Stock-well’s PRP because he fails to meet the actual and substantial prejudice standard. I would not silently overrule the definition of “prejudice” that we adopted as far back as Hews I and Montoya, as the majority seems to do. I would not silently overrule our prior precedent retaining automatic reversible error on PRPs for certain especially intractable errors, e.g., Richardson, Orange, Francis, Gunter, and Mutch, as the majority does. And I would not place this *613constitutional error into the nonconstitutional error category, as the majority does.
Stephens, J., concurs with Gordon McCloud, J.Reconsideration denied April 1, 2014.
Although our recent decisions on this topic apply the prejudice inquiry applicable on direct appeal (as opposed to PRP), they clearly state that the due process clause protects the voluntariness of the decision — the cost-benefit analysis — involved in the guilty plea process, regardless of the ultimate sentence imposed. See State v. Weyrich, 163 Wn.2d 554, 556-57, 182 P.3d 965 (2008); State v. Mendoza, 157 Wn.2d 582, 587, 590-91,141 P.3d 49 (2006); In re Pers. Restraint of Isadore, 151 Wn.2d 294, 297-98, 88 P.3d 390 (2004).
In fact, St. Pierre explicitly stated that “some errors which result in per se prejudice on direct review will also be per se prejudicial on collateral attack ....” St. Pierre, 118 Wn.2d at 329.
I note the majority’s attempt to distance itself from this position. Majority at 598 n.3. The majority is correct to do that. As discussed below, however, the majority’s analysis could be used to conflate harmless error review of constitutional issues with harmless error review of nonconstitutional issues.
See generally Arizona v. Fulminante, 499 U.S. 279, 309-11, 111 S. Ct. 1246, 113 L. Ed. 2d 302 (1991) (Rehnquist, C.J., concurring) (holding admission of coerced confession to he “a classic ‘trial error’... similar in both degree and kind to the erroneous admission of other types of evidence” and thus reviewable for harmlessness).
Accord. Kyles v. Whitley, 514 U.S. 419, 435-36, 115 S. Ct. 1555, 131 L. Ed. 2d 490 (1995) (on federal habeas review of alleged Brady v. Maryland, 373 U.S. 83,83 S. Ct. 1194,10 L. Ed. 2d 215 (1963) violation, once petitioner shows prosecutorial withholding of exculpatory evidence and materiality “there is no need for further harmless-error review”).
The majority cites State v. Oseguera Acevedo, 137 Wn.2d 179, 203, 970 P.2d 299 (1999) (plurality opinion), as showing that “even this court’s views on materiality have changed over time.” Majority at 602 n.9. This is misleading. As this court unanimously recognized in Isadore, Acevedo was a plurality opinion stemming from “unique” facts and was never “intended to alter the longstanding rule... that a defendant must be informed of [all] direct consequences of his guilty plea.” Isadore, 151 Wn.2d at 302. It is inaccurate to imply that a majority of this court ever employed a materiality inquiry — as opposed to a “direct consequence” inquiry — is cases where the defendant alleges his plea was not knowing, intelligent, and voluntary.
Because the period for direct appeal had passed when the community placement at issue was added to Isadore’s original sentence, the court declined to apply “the heightened threshold requirements applicable to personal restraint petitions.” Isadore, 151 Wn.2d at 299. The court expressly noted, however, that Isadore’s claim would have succeeded even if he had been required to meet “the standard personal restraint petition requirements.” Id. at 300.
This is particularly unfortunate given the internal contradictions in St Pierre. In that case, this court considered whether an error in the charging document established per se prejudice on collateral review. St. Pierre, 118 Wn.2d at 328-29. The court concluded that it did not and that the petitioner was not entitled to relief. Id. at 329-30. While its holding was limited to that relatively narrow issue, the St Pierre opinion included several broad assertions about a petitioner’s burden on collateral review generally. Those assertions conflate several distinct categories of PRP, with the result that the St Pierre opinion is at times self-contradictory. For example, St Pierre affirms the “actual and substantial prejudice” standard for relief on collateral review but notes that that burden “may be waived where the error gives rise to a conclusive presumption of prejudice.” Id. at 328 (citing Richardson, 100 Wn.2d at 679). It also acknowledges that “some errors which result in per se prejudice on direct review will also be per se prejudicial on collateral attack” but nevertheless concludes that “the interests of finality . . . demand that a higher standard be satisfied in a collateral proceeding.” Id. at 329.
Strickland, v. Washington, 466 U.S. 668, 104 S. Ct. 2052, 80 L. Ed. 2d 674 (1984).
Brady v. Maryland, 373 U.S. 83, 83 S. Ct. 1194, 10 L. Ed. 2d 215 (1963).