The question presented in these two cases is whether the right to a jury trial announced in Blakely v. Washington1 should be retroactively applied to two state defendants, Troy Smart and Henry Douglas, whose sentences were final before June 24, 2004, when Blakely was decided. Blakely requires that any fact—except a fact admitted by the defendant or the fact of a prior conviction—necessary to increase a sentence above the statutory presumptive maximum be proved to a jury beyond a reasonable doubt. We decline to give Blakely full retroactivity. We conclude that the purpose of Blakely does not raise serious questions about the accuracy of past sentences and must be weighed against the state's reliance on the old rule for over twenty years and the administrative burden of implementing the new rule retroactively. We therefore reverse the rulings of the court of appeals in these two cases and remand.
II. FACTS AND PROCEEDINGS
A jury convicted Troy Smart of second-degree assault. Because Smart was a second felony offender he faced a presumptive sentence of four years with a maximum sentence of ten years.2 The sentencing judge found an aggravating factor and in May 2002 sentenced Smart to serve ten years, with six of those years suspended.3 The court of appeals affirmed Smart's sentence in January 2003.4 Smart did not seek further review of his sentence at that time.
Henry Douglas pleaded no contest to first-degree robbery. As a second felony offender Douglas faced a presumptive sentence of ten years and a maximum sentence of twenty years.5 In August 1999 the sentencing judge found multiple aggravating factors and sentenced Douglas to serve eighteen years, with eight of those years suspended.6 Douglas appealed but his appeal was dismissed in April 2000.7
In June 2000 the United States Supreme Court held in Apprendi v. New Jersey that defendants have a constitutional right to have a jury decide any disputed fact—other than the fact of a prior conviction—that increases a sentence beyond the statutory maximum, and to have each such fact proved beyond a reasonable doubt.8 In June 2004 the Supreme Court in Blakely v. Washington for the first time applied Apprendi to a sentence imposed under a presumptive sentencing scheme.9 It held in Blakely that a defendant's sentence was invalid because the aggravating facts, which supported an increase above the sentence that was authorized by the jury's verdict alone, were neither admitted by the defendant nor found by a jury.10
After Blakely was issued, Douglas filed an Alaska Criminal Rule 35(a) motion claiming that his sentence was illegal under Blakely. Superior Court Judge Ben J. Esch denied Douglas's motion, holding that "Blakely is not retroactive and does not apply to this case because it is not on . . . direct appeal."
Likewise, Smart filed an Alaska Criminal Rule 35(a) motion in 2004, claiming that his sentence was illegal under Blakely. Superior Court Judge Larry R. Weeks concluded that Smart's request could only be presented as an Alaska Criminal Rule 35.1 petition for post-conviction relief. Treating the motion as a Rule 35.1 petition, Judge Weeks denied Smart's motion, holding that Blakely did not apply retroactively to cases that have been decided on appeal.
Smart and Douglas both appealed. The court of appeals decided Smart's appeal first. It first held that the state retroactivity test announced in Judd v. State11—not the federal retroactivity test announced in Teague v. Lane12—applied.13 It then held that Blakely should be fully retroactive in Alaska.14 It therefore vacated Smart's sentence and remanded for further proceedings on the disputed aggravating factor.15 In resolving Douglas's appeal, the court of appeals relied on its decision in Smart and remanded for further proceedings on Douglas's contested aggravators.16
The state filed petitions for hearing challenging both decisions. We granted the petitions, ordered full briefing, and consolidated the cases for decision.
A. Standard of Review
These cases require interpretation of the United States and Alaska Constitutions and federal and state case law concerning the retroactivity of appellate decisions. We use our independent judgment in reviewing rulings turning on federal and state constitutional law.17
B. Apprendi and Blakely
Because this case turns on whether Blakely applies retroactively to cases on collateral review,18 we begin with an overview of Blakely and its predecessor, Apprendi.
Apprendi. Apprendi fired bullets into the home of an African-American family.19 He entered a plea agreement and pleaded guilty to three of the twenty-three counts charged.20 Under Apprendi's plea agreement, the sentences for two counts would run consecutively and the sentence for the third count would run concurrently with the other two.21 Apprendi faced a maximum aggregate sentence of twenty years on the two counts (ten-year maximum for each count) if the judge found no basis for a hate-crime enhancement.22 But if the hate-crime enhancement applied to one count, a New Jersey statute authorized a twenty-year maximum sentence on that count alone.23 Applying a preponderance of the evidence standard, the sentencing judge found that the hate-crime enhancement applied to one count.24 He sentenced Apprendi to a twelve-year term on that count and to shorter concurrent sentences on the other two counts.25
The question before the Supreme Court was whether a jury had to find on the basis of proof beyond a reasonable doubt that there had been a hate crime.26 This question would determine whether the twelve-year sentence on the single count was "permissible" given the ten-year maximum for that offense.27 The Court held that: "Other than the fact of a prior conviction, any fact that increases the penalty for a crime beyond the prescribed statutory maximum must be submitted to a jury, and proved beyond a reasonable doubt."28 The Court reversed the judgment of the New Jersey Supreme Court, which had upheld Apprendi's sentence.29
Blakely. In Blakely the Supreme Court applied Apprendi to a presumptive sentencing system for the first time. Blakely pleaded guilty in Washington State to kidnaping his estranged wife.30 Per state statute, the "standard range" for Blakely's conviction was forty-nine to fifty-three months.31 The statute allowed a sentence above the standard range if the judge found "substantial and compelling reasons justifying an exceptional sentence."32 The judge imposed an exceptional sentence of ninety months (thirty-seven months above the maximum of the standard range) because he found that Blakely acted with "deliberate cruelty"—a statutory aggravating factor.33 Reviewing Blakely's sentence, the Supreme Court concluded that "the `statutory maximum' for Apprendi purposes is the maximum sentence a judge may impose solely on the basis of the facts reflected in the jury verdict or admitted by the defendant."34 "In other words," the Court held, "the relevant `statutory maximum' is not the maximum sentence a judge may impose after finding additional facts, but the maximum he may impose without any additional findings."35 The Court held that because Blakely's sentence exceeded the presumptive sentence, absent a jury finding of the enhancing factor under the reasonable doubt standard, the sentence violated the Sixth Amendment right to a jury trial.36 The Court made it clear that prior convictions and facts admitted by the defendant do not fall within the requirements of Apprendi and Blakely.37
The question facing state courts after Apprendi and Blakely was whether the Supreme Court's interpretation of the Federal Constitution in those cases should be given full retroactivity by a state reviewing its own criminal convictions. One part of that question was whether the issue of retroactivity should be decided by applying an individual state's retroactivity standard or the federal retroactivity standard set out in Teague v. Lane.38 That issue was squarely presented to the Court in Danforth v. Minnesota.39
The Court in Danforth v. Minnesota summarized the Teague rule as follows:
New constitutional rules announced by this Court that place certain kinds of primary individual conduct beyond the power of the States to proscribe, as well as "watershed" rules of criminal procedure, must be applied in all future trials, all cases pending on direct review, and all federal habeas corpus proceedings. All other new rules of criminal procedure must be applied in future trials and in cases pending on direct review, but may not provide the basis for a federal collateral attack on a state-court conviction.40
Thus, under Teague, to use the terms we identified in footnote eighteen, limited retroactivity must be given as a matter of course to new constitutional rules announced by the United States Supreme Court. In other words, when a new ruling is announced by the United States Supreme Court it must be applied to any case still pending on direct review in the state system. Because neither of the cases before us was still pending on direct review in the state system when Blakely was decided, the question in these cases is whether full retroactivity should be afforded to the Blakely ruling. Under the Teague standard for full retroactivity, full retroactivity would be required if Blakely was a ruling that placed a certain kind of primary individual conduct beyond the power of the states to proscribe or if it was a "watershed" rule of criminal procedure. Until Danforth was decided, it was an open question whether a state court could grant full retroactive effect to a United States Supreme Court ruling in cases in which the Teague standard for full retroactivity was not met. That issue was decided by the Court in Danforth.
C. Danforth and the Applicable Test for Full Retroactivity
The Supreme Court introduced the retroactivity issue in Danforth as follows: "The question in this case is whether Teague constrains the authority of state courts to give broader effect to new rules of criminal procedure than is required by that opinion. We have never suggested that it does, and now hold that it does not."41 The Court stated that "[s]tates are independent sovereigns with plenary authority to make and enforce their own laws as long as they do not infringe on federal constitutional guarantees."42 A state is thereby "free to give its citizens the benefit of [the Court's] rule in any fashion that does not offend federal law."43 Danforth clarified that it is acceptable for states to give broader protection to defendants by applying state retroactivity law in state post-conviction proceedings.44 The Court held in Danforth that Teague does not impose a binding obligation on state courts and that "a state court, when reviewing its own state criminal convictions, [may] provide a remedy for a violation that is deemed `nonretroactive' under Teague."45
Danforth therefore allows us to apply either the Teague test for full retroactivity or a state constitutional test so long as the state test is at least as comprehensive as the federal test. The Alaska Court of Appeals determined that the three-factor test used in Judd v. State46 should be employed. In Judd we identified three criteria which guide resolution of the question of the retroactivity of a new ruling: "(a) the purpose to be served by the new standards; (b) the extent of the reliance by law enforcement authorities on the old standards; and (c) the effect on the administration of justice of a retroactive application of the new standards."47
We borrowed these standards from those the Supreme Court discussed in Linkletter v. Walker.48 They were designed to answer the question whether the Court's decision in Mapp v. Ohio49—requiring that state courts exclude unconstitutionally seized evidence— would apply to state convictions that were on collateral review.50 The Linkletter Court in applying the three-factor test decided that Mapp should not be applied to cases on collateral review.51
Subsequently, in Williams v. United States the Supreme Court applied the Linkletter standards to a case on direct review.52 Justice Harlan dissented from the Williams decision, noting that he had come to regret the Linkletter decision in which he had joined.53 In his view new constitutional rulings should as a matter of course be applied to state cases which are still pending on direct review when a new ruling is announced.54 Justice Harlan decried the practice of prospective overruling as both inequitable to litigants who should benefit from a new ruling and inconsistent with the role of a reviewing court: "Simply fishing one case from the stream of appellate review, using it as a vehicle for pronouncing new constitutional standards, and then permitting a stream of similar cases subsequently to flow by unaffected by that new rule constitutes an indefensible departure from this model of judicial review."55 Justice Harlan also concluded that the Linkletter standards were inappropriate for determining whether a new ruling should be applied to cases on collateral review and suggested instead that retroactive application in such cases should be limited to new substantive due process rulings and "for claims of nonobservance of . . . procedures that . . . `are implicit in the concept of ordered liberty.'"56 Justice Marshall joined in Justice Harlan's dissent in Williams with respect to cases on direct review.57 As to cases on collateral review, Justice Marshall was of the opinion that the Linkletter standards were appropriate. He wrote that
[t]he method commends itself, once the point of finality after direct review is passed, as a careful and appropriate way of adjudicating the "procedural" rights of litigants in view of the purposes of a new decisional rule and the concerns of effective law enforcement. In particular, if the purposes of a new rule implicate decisively the basic truth-determining function of the criminal trial, then I believe the rule should be given full retroactive application, for the required constitutional procedure itself would then stand as a concrete embodiment of "the concept of ordered liberty."58
Use of the Linkletter standards for cases on direct review did not long survive in the United States Supreme Court. In United States v. Johnson, the Court adopted Justice Harlan's view, save for cases where the new ruling was said to be "a clear break with the past."59 Thus, a "clear break" rule was not to be applied retroactively to cases on direct review. In Griffith v. Kentucky the Court decided that the "clear break" exception was not tenable and held that "a new rule for the conduct of criminal prosecutions is to be applied retroactively to all cases, state or federal, pending on direct review or not yet final, with no exception for cases in which the new rule constitutes a `clear break' with the past."60 Griffith, however, left undisturbed the application of the Linkletter standards to cases pending on collateral review. Two years after Griffith, Teague decided that question and overruled the use of the Linkletter standards for cases on collateral review in favor of the approach advocated by Justice Harlan in his dissent in Williams.61
In light of the force of Justice Harlan's rationale, it may be difficult to defend the use of the Linkletter/Judd standards in cases involving direct review on the state level. No party argues that any standard other than those discussed in Teague and Judd should be applied here on collateral review. Absent meaningful argument on the issue, we decline to consider whether to apply some other standard in these cases.
Using Judd as the test for full retroactivity, we must first determine whether the disputed cases announced a new rule.62 If so, we then examine the purpose of the new rule and, depending on that purpose, balance multiple factors to determine full retroactivity.63 By contrast, Teague only treats two narrow categories of rules as retroactive: (1) those that render types of conduct "beyond the power of the criminal law-making authority to proscribe" and (2) "watershed" rules that "implicate the fundamental fairness of the trial."64 The Supreme Court has applied exceptionally few rules retroactively under Teague.65
To confirm that Judd is no less protective than the federal standard, and that it can therefore be applied to determine full retroactivity of Blakely, we consider whether Blakely would be fully retroactive under Teague. To do so, we look to how the federal courts have applied Teague to Blakely. So long as Judd at least meets the federal standard, it does not offend federal law.66 That is, if Blakely would not be fully retroactive under Teague, we can proceed to determining whether Blakely is fully retroactive under state law.
The federal courts of appeals that have carefully considered the issue have concluded that Blakely is not fully retroactive under Teague.67 These courts have held that Blakely announced a new procedural rule, that this rule does not place certain kinds of conduct beyond the power of the criminal law-making authority to prohibit, and that Blakely did not announce a watershed rule of criminal procedure.68 State courts applying Teague have reached the same conclusion.69 Because Smart and Douglas could not succeed on a Blakely challenge under the federal retroactivity standard, we conclude that our state standard is no less protective than the federal standard.70
D. Is Blakely a New Rule or Was It Compelled by the Holding in Apprendi?
Under Alaska law, as a preliminary matter we must decide whether Blakely announced a new rule or was instead a compelled application of then-existing precedent.71 We only apply the Judd retroactivity balancing analysis if there is a new rule. Smart argues that Blakely did not announce a new rule because it was compelled by the Supreme Court's earlier holding in Apprendi. This issue is critical to Smart because his case had not become final when Apprendi was decided on June 26, 2000.72
In Teague Justice O'Connor, who wrote the lead opinion, stated that "in general . . . a case announces a new rule when it breaks new ground" or, in other words, "if the result was not dictated by precedent existing at the time the defendant's conviction became final."73 Justice O'Connor noted that it is "often difficult to determine when a case announces a new rule," and she declined to "define the spectrum of what may or may not constitute a new rule for retroactivity purposes."74 The Supreme Court recently revisited the question of how to determine whether a case creates a new rule. In Beard v. Banks, the Court instructed: "We must. . . ask `whether the rule later announced. . . was dictated by then-existing precedent—whether, that is, the unlawfulness of [the] conviction was apparent to all reasonable jurists.'"75
The Alaska Court of Appeals did not address this issue before analyzing retroactivity in its opinion in Smart, but all three members of the court agreed that Blakely announced a new rule.76 Although Blakely relied on Apprendi, it fundamentally changed courts' understanding of "maximum sentence."77 Blakely defined the "maximum sentence" not as the maximum allowed by a presumptive sentencing scheme, but as the maximum allowed by the jury's verdict.78 Before Blakely was decided, courts consistently held that Apprendi did not apply to sentences within the statutory maximum.79 The rule in Blakely was clearly not apparent to all courts and was not dictated by precedent.80
We consequently conclude that Blakely announced a new rule of law and that it therefore does not automatically apply on collateral review. We must apply Alaska's full retroactivity test to determine whether this new rule applies retroactively to Smart and Douglas.
E. Is Blakely Fully Retroactive Under Judd?
The full retroactivity standard adopted in Judd v. State requires us to consider three factors when deciding whether to apply a new constitutional rule retroactively: (1) the purpose to be served by the new standards; (2) the extent of the reliance by law enforcement authorities on the old standards; and (3) the effect on the administration of justice of a retroactive application of the new standards.81
1. The purpose of the rule announced in Blakely
The purpose of the Blakely rule is to guarantee that a jury finds beyond a reasonable doubt the facts that determine the maximum sentence. The Supreme Court explained that "[w]hen a judge inflicts punishment that the jury's verdict alone does not allow, the jury has not found all the facts `which the law makes essential to the punishment,' and the judge exceeds his proper authority."82 A violation of Blakely does not call into question the guilt of the defendant, but it "undermines the factual foundation of the sentencing court's authority to impose an increased punishment."83
Applying the Judd standard, the court of appeals concluded that Blakely should be applied fully retroactively.84 The court parsed the two elements of Apprendi: the right to a jury, and the right to demand that the government prove the disputed issue of fact beyond a reasonable doubt.85 Focusing on the requirement of proof beyond a reasonable doubt,86 the court concluded that the "`reasonable doubt' component of Apprendi and Blakely is the type of rule . . . whose primary purpose is `related to the integrity of the verdict', and whose function is to cure a flaw that `raises serious questions about the accuracy of guilty verdicts.'"87
The court relied on our opinion in Rutherford v. State, which quoted from Justice White's plurality opinion in Williams v. United States88: "Where the major purpose of new constitutional doctrine is to overcome an aspect of the criminal trial which substantially impairs its truth-finding function and so raises serious questions about the accuracy of guilty verdicts in past trials, the new rule [is to be] given complete retroactive effect."89 The court of appeals also relied on our statement in Rutherford that the first Judd factor (purpose) will supersede other considerations whenever "the purpose of the new rule is primarily related to the integrity of the verdict."90 Because "[t]he Blakely requirement of proof beyond a reasonable doubt is . . . designed to prevent unconstitutional punishment," the court of appeals concluded that full retroactivity should be granted "even if the other two parts of the Alaska retroactivity test . . . would militate against retroactive application."91
The Rutherford language relied on by the court of appeals in holding that the first factor was determinative sets a high standard. That standard ensures that only the most exceptional new rules are not subject to the three-part balancing test. The new rule must "raise serious questions about the accuracy" of previous verdicts and "overcome an aspect of the criminal trial which substantially impair[ed] its truth-finding function."92 We stated in Rutherford that where the purpose of the new rule is "primarily related to the integrity of the verdict," the new rule is generally applied to all cases.93 But we also recognized that "where the purpose of a new constitutional standard is not to minimize arbitrary or unreliable fact findings . . . retroactive application has generally been denied."94 We acknowledged that "[d]ifficult analytical problems arise [that] necessitate reference to the other two criteria, because with many new constitutional doctrines, (t)he extent to which a condemned practice infects the integrity of the truth-determining process at trial is a `question of probabilities."95 In Rutherford we gave limited retroactive effect to the speedy trial right, but we did so only after weighing all three Judd factors.96
Here we must decide whether to balance the three Judd factors, as we did in Rutherford, or whether Blakely's purpose is so closely related to the accuracy of guilty verdicts that it supersedes the other factors. The proper comparisons here are (1) between judge and jury as fact-finder, and (2) between the clear and convincing proof standard and the proof beyond a reasonable doubt standard. The Rutherford standard asks whether permitting a judge to find facts under a lower standard of proof "substantially impairs [the criminal trial's] truth-finding function" and "raises serious questions about the accuracy of guilty verdicts in past trials."97 If the answer is yes, Blakely must be given retroactive effect regardless of the outcome of the other two factors. If not, all three factors must be weighed. This question turns on that "question of probabilities"98: whether the likelihood of accurate verdicts is seriously diminished if the requirements of Blakely — both the right to a jury and the standard of proof — are not met.
Jury Trial. The Supreme Court, in Schriro v. Summerlin, has already resolved the first comparison — judge to jury — by holding that judicial fact-finding, instead of jury fact-finding, does not "so seriously diminish accuracy as to produce an impermissibly large risk of injustice."99 We think the Court's analysis is persuasive on this point. In Summerlin, decided the same day as Blakely and written by the author of the Court's opinion in Blakely, the Court determined that a new rule requiring that aggravating factors be proved to a jury rather than a judge was procedural, not substantive, and concluded that the rule does not apply retroactively to cases already final on direct review.100 There an Arizona statute authorized the death penalty if a judge found certain aggravating factors.101 The Court noted that this judicial fact-finding violated Apprendi.102 But the Court concluded that the evidence of whether judges or juries were better fact-finders was too equivocal to conclude that judges were less accurate fact-finders or to hold that "judicial factfinding so seriously diminishes accuracy that there is an impermissibly large risk of punishing conduct the law does not reach."103 The Court held that the rule that a jury find aggravating factors does not apply retroactively to cases already final on direct review.104 We think it is highly likely that the Court would reach the same conclusion if it were deciding whether the jury fact-finding rule of Blakely is fully retroactive. We therefore conclude that judicial fact-finding instead of jury fact-finding does not substantially impair the truth-finding function of the criminal trial and does not raise serious questions about the accuracy of fact-finding during sentencing such that this factor must supersede the other considerations of our Judd test.
Standard of Proof. The more difficult question is whether finding an aggravator under the clear and convincing standard instead of Blakely's reasonable doubt standard seriously diminishes the likelihood of accurate fact-finding for sentences. We conclude that failure to fully retroactively impose Blakely's requirement that aggravators be found under the reasonable doubt standard does not impair truth-finding so substantially that the factors in Judd need not be weighed.
To begin, we review some of the constitutional principles that do impact truth-finding so substantially that the importance of their purposes precludes balancing other factors. Rutherford listed cases in which this standard was applied and the purpose of the new rule was held to be outcome-determinative. For example, the Supreme Court held that the right to counsel applied retroactively because "the `denial of the right must almost invariably deny a fair trial.'"105 The Court reasoned that the assistance of counsel was necessary for "marshaling the facts, introducing evidence of mitigating circumstances and in general aiding and assisting the defendant to present his case"106 and that "only the aid of counsel could ... enable the accused to know all the defenses available... and to plead intelligently."107 Likewise the Court held that the right to confront witnesses significantly affected the "integrity of the fact-finding process" and therefore applied retroactively because "one of the important objects of the right of confrontation [is] to guarantee that the fact finder ha[s] an adequate opportunity to assess the credibility of witnesses."108
Both the right to counsel and the right to confront witnesses are central to a fair and accurate trial.109 Without counsel, an indigent defendant would be severely disadvantaged at all stages of the process. Without the right to confront witnesses, the fact-finder may not be able to fairly assess the accuracy of the facts on which the case is decided. The rule announced in Blakely does not have as substantial an impact on the truth-finding function of a criminal trial.
The court of appeals relied heavily on the Supreme Court's discussion of the reasonable doubt standard in In re Winship.110 The Court there held that the Due Process Clause of the Fourteenth Amendment requires that every fact necessary to constitute a crime be proved beyond a reasonable doubt.111 The Court stated that the reasonable doubt standard "plays a vital role in the American scheme of criminal procedure" and that it is "indispensable."112 The Court later held that the rule announced in Winship must be given retroactive effect.113 But Winship focused only on the adjudicatory phase of the trial; the Court specifically noted that it did not "direct [its] attention to the post-adjudicative or dispositional process."114
Not every rule involving the reasonable doubt standard is as important as the rule announced in Winship.115 In a case holding that Apprendi did not apply retroactively, the Ninth Circuit Court of Appeals reasoned that "[t]he rule announced in Winship ... [was] given retroactive effect because [it was] to `overcome an aspect of the criminal trial that substantially impairs its truth-finding function and so raises serious questions about the accuracy of guilty verdicts[.]'"116 In contrast, and like the rule in Apprendi, the rule announced in Blakely affects the enhancement of a defendant's sentence after guilt is found beyond a reasonable doubt. Put another way, a defendant sentenced in violation of Blakely cannot complain that his entire conviction is tainted, only that the fact-finding underlying an aggravating sentencing factor might not have been accurate — a complaint that, given the heightened clear and convincing standard applied by judges before Blakely, does not raise serious questions about the accuracy of the sentence.
That a Blakely violation does not rise to the level of substantially impairing the truth-finding function of the criminal trial is also demonstrated by its remedy. A Blakely violation requires remand for resentencing. It does not require a new trial or vacating the conviction.117 We think the Supreme Court of Minnesota stated this point well in its consideration of Blakely: "We are hard-pressed to see how a rule that, when violated, merely requires re-sentencing can be said to be one without which the likelihood of an accurate conviction is seriously diminished."118
Because the standard-of-proof aspect of Blakely is dictated by the Federal Constitution, we also look to the federal courts for guidance. We find their logic persuasive even though the federal courts apply a different standard of full retroactivity. Every court of appeals that has dealt with the issue, whether calling it the Blakely or Apprendi rule, has concluded that the rule is procedural.119
A new procedural rule is only given fully retroactive effect under the federal standard if it is a "`watershed rule of criminal procedure' implicating the fundamental fairness and accuracy of the criminal proceeding."120 The procedural rule must be one "without which the likelihood of an accurate conviction is seriously diminished."121 If the beyond a reasonable doubt standard is a rule without which the likelihood of accurate sentences is seriously diminished, one would expect the federal courts to say so. To the contrary, every federal court of appeals which has addressed Apprendi has held that it did not create a watershed rule of criminal procedure.122 One federal court of appeals recognized that the reasonable doubt standard rule "will promote marginally more accurate results" but held that the rule did not "alter our understanding of the bedrock elements essential to the fairness of a proceeding" such that it seriously diminishes the accuracy of guilty verdicts.123 The courts addressing Blakely have likewise held that Blakely did not create a watershed rule of criminal procedure.124 The logical conclusion to be derived from their treatment of Apprendi and Blakely is that Blakely's reasonable doubt requirement is not a watershed rule or one without which the likelihood of an accurate sentence is seriously diminished.
We conclude that Blakely announced a procedural rule that does not raise serious questions about the fundamental fairness and accuracy of our criminal trials. The clear and convincing standard applied to Douglas and Smart did not seriously diminish the accuracy of fact-finding when they were sentenced. The purpose of Blakely is to ensure constitutional sentences — punishment matching the facts found by the jury — by requiring that aggravating factors be proved to a jury beyond a reasonable doubt. The right to jury findings and proof beyond a reasonable doubt is fundamental in our system of criminal procedure. But because the purpose of the rule announced in Blakely is not, under Rutherford, conclusive of full retroactivity, we must balance all three of the Judd factors.
2. The extent of reliance on the pre-Blakely rule
Although the court of appeals's decision did not balance the three factors of the Judd test, it nonetheless discussed the second and third factors.125 The court stated that the second factor — extent of reliance — favored non-retroactivity because the state "relied for more than two decades on the presumed constitutionality of the sentencing procedures ... for proving the aggravating factors that increased a defendant's maximum sentence under Alaska's presumptive sentencing law."126
The state points out that the court of appeals upheld judicial fact-finding over twenty years ago and asserts that, before Blakely was published, "hundreds, if not thousands" of appellate opinions were issued without questioning this rule.127 After Apprendi, the state argues, Alaska appellate courts continued to hold that presumptive sentencing was not affected by Apprendi.128 The state correctly points out that "Judd's reasonable-reliance factor does not require state officials to predict the outcomes of 5-4 Supreme Court splits in evolving lines of case law."129 Because the reliance on the pre-Blakely standards was reasonable and consistent for over twenty years, this factor militates against full retroactivity.
3. The effect on the administration of justice of applying Blakely retroactively
The court of appeals concluded that the third factor — the effect on the administration of justice of a retroactive application of Blakely — favored retroactivity.130 The court noted that many of the defendants who were sentenced in violation of Blakely have completed their sentences and stated that "giving full retroactivity to Blakely would not affect the administration of justice in those cases."131 The court then estimated that there were a "few hundred" cases that might be affected by Blakely, basing this estimate on the Appellate Court Clerks' Office statement that "slightly over 250 Blakely appeals [are] currently stayed and awaiting our decision of this retroactivity issue."132 But based on the Blakely cases it had already decided, the court of appeals concluded that only a few defendants would be entitled to relief.133 The court concluded that there would be a "relatively small impact on the criminal justice system" and that this factor favored retroactivity.134
Judd asks what effect retroactive application will have on the administration of justice, not how many appeals might be successful. Full retroactive application thus looks at the entire number of retroactivity claims that will be brought and asks how those cases will affect the administration of justice. Review of even 250 cases would take up a great amount of court and staff time and it seems certain that there may be hundreds more cases that have not yet been brought back to the trial courts. We agree with the observations of Chief Judge Coats, who stated in dissent that "[d]efendants who conceded aggravating factors ... can claim that they would not have conceded [the factor] had they known they were entitled to a jury trial."135 If a defendant's sentence is overturned, more time will be spent re-litigating aggravating factors. Prosecutors will have to reevaluate whether to bring the charges, old witnesses will have to be found, juries may have to be empaneled, court time will be spent relitigating, defense attorney and prosecutor time will be spent relitigating, and Department of Corrections time will be spent transporting convicted defendants to the courtroom. We think Chief Judge Coats correctly summarized the result as follows: "By making Blakely retroactive, we are not advancing an important constitutional principle. We are creating a major administrative problem. Whether a person gets his sentence reduced may turn more on good fortune than anything else."136 We conclude that the burden factor weighs against full retroactivity.
4. Conclusion: no full retroactivity
Weighing all three factors of the Judd test is a difficult exercise. It is clear that many defendants received sentences that are inconsistent with Blakely, because aggravating factors were not proved to a jury beyond a reasonable doubt. But defendants sentenced before Blakely were not deprived of procedural safeguards. The aggravating factors had to be proved by the elevated clear and convincing evidence standard. There is, as the Supreme Court stated in Summerlin, no reason to think judges are less accurate than juries in finding facts.137 Furthermore, all actors in the criminal justice system relied on the old standard for over twenty years with no warning that it was unconstitutional. The sentencing guidelines provided notice to the defendants of the consequences of their acts and helped to ensure equal protection by providing uniformity. Finally, the administrative burden of full retroactive application would be overwhelming and would severely impact current criminal procedure by affecting the time all participants in the criminal justice system have available to spend on current trials. After considering the purpose of the procedural rule announced in Blakely, the reliance on the pre-Blakely rule, and the burden of full retroactive application of Blakely, we conclude that Blakely does not merit full retroactive application under the Judd standard.
F. Other Issues
The parties and the amicus raise other related issues: whether we should adopt Blakely as a matter of state law; whether a Blakely claim is properly brought under Alaska Criminal Rule 35(a) or Criminal Rule 35.1; and whether post-conviction relief actions are governed by the retroactivity standard in AS 12.72.010. We briefly address each of these issues in turn.
1. Should we adopt Blakely under Alaska law?
Smart, Douglas, and the amicus argue that we should adopt the Blakely reasoning as a matter of state law under the Alaska Constitution's due process and right to a jury clauses. Even if we were to adopt Blakely as a matter of state law, Judd would govern full retroactivity. The new standard would consequently not apply to Smart and Douglas. It is therefore not necessary to decide the issue here.
2. Are Blakely challenges properly brought under Alaska Criminal Rule 35(a)?
The state argues that a Blakely challenge is a claim that the manner or procedure of imposing the sentence was illegal, not that the sentence itself was illegal. The state points out that the term "illegal sentence" has been narrowly construed and argues that Rule 35(a) does not permit challenges to sentences that were imposed through procedures that were legally defective.138 The state concludes that a Blakely challenge can only be brought under Rule 35.1.139
Smart and Douglas assert that Rule 35(a) is the proper means to bring a Blakely challenge. They argue that their sentences are "illegal" because each "exceeded the prescribed statutory maximum and were not authorized by either judgment of conviction."
We do not have to decide the issue. Because Blakely is not fully retroactive, neither the sentences nor the procedures used to impose the sentences were illegal.
3. Are post-conviction relief actions governed by the retroactivity standard in AS 12.72.010 and Rule 35.1?
The state argues that if a Blakely challenge must be brought under Rule 35.1, both Rule 35.1 and AS 12.72.010 adopt the Teague standard to determine whether Blakely is retroactive.140 Because we hold that Blakely is not fully retroactive, we do not reach this issue and express no opinion about it.141
Applying Alaska's standard first announced in Judd to the present question of full retroactivity, we hold that Blakely is not retroactive as to defendants whose sentences were final when Blakely was decided on June 24, 2004. We therefore REVERSE the holdings of the court of appeals in Smart v. State and Douglas v. State and REMAND for proceedings consistent with this opinion.