concurring and dissenting.
The court has determined to vacate the sentence imposed on Heathcock by the three-judge panel and remand the case to Judge Serdahely to impose the applicable presumptive sentence. I agree. The conclusion that imposition of the presumptive sentence would result in “manifest injustice” is clearly mistaken. See McClain v. State, 519 P.2d 811 (Alaska 1974) (establishing the clearly mistaken standard to govern appellate review of trial court sentencing). The presumptive sentence adjusted to reflect all permissible and applicable aggravating and mitigating factors is not manifestly unjust because it is too severe,1 nor is *1159it too lenient.2 As Judge Coats makes clear in his opinion for the court, the legislative goal of ensuring uniformity and predictability in sentencing multiple offenders would *1160be defeated if recourse to the three-judge panel became routine.3 As I understand the concept of “manifest injustice,” before a ease can be referred to the three-judge panel, the record must support a finding that the presumptive sentence adjusted to reflect all permissible and applicable aggravating and mitigating factors either would unduly depreciate the seriousness of the defendant’s offense in the light of his past criminal experience and is therefore too lenient, or would approximate cruel and unusual punishment in light of the defendant’s conduct and his criminal history and therefore would be too severe. Since neither finding is sustainable on this record, I concur in the court’s decision to remand for imposition of the presumptive sentence.
The court, unnecessarily in my opinion, goes beyond this simple disposition and purports to address the proper relationship between the sentencing court and the three-judge panel. I would await an appropriate case before entering this murky thicket. As a majority of this court disagrees and proceeds, I, therefore, dissent.
In order to clarify our dispute, it is necessary to identify points upon which we apparently agree. I assume general agreement on the following basic propositions. First, that prior to considering an application by either the state or the defendant to refer a sentencing to a three-judge panel, or referring it sua sponte, the trial judge must thoroughly consider all of the facts that normally would enter into the determination of an appropriate sentence. Thus, the presentence report would have been filed and both parties given an opportunity to contradict and supplement the facts contained within it. If the facts were in dispute, the trial court would hold necessary hearings and provisionally resolve the dispute, drawing all inferences necessary to a full understanding of the defendant’s background, experience, character and the offense which he committed.
Second, the trial court would have considered all aggravating and mitigating factors, again, making necessary fact findings and would provisionally determine the effect of any aggravating and mitigating factors found. See Juneby v. State, 641 P.2d 823 (Alaska App.1982), modified on rehearing, 665 P.2d 30 (Alaska App.1983). For it is only after applying all applicable aggravating and mitigating factors and determining their significance that the trial court can determine what presumptive sentence the individual under consideration would receive. See AS 12.55.155. I call this sentence the “adjusted presumptive sentence.”
Third, the trial court would then look at the adjusted presumptive sentence in light of the facts previously found and determine whether it was manifestly unjust, i.e., either unduly depreciates the seriousness of the offense or approximates cruel and unusual punishment. In so doing, the trial judge would compare the adjusted presumptive sentence to other sentences customarily imposed for similar conduct and *1161for crimes involving similar degrees of seriousness within and without the state. Cf. In re Reed, 33 Cal.3d 914, 191 Cal.Rptr. 658, 663 P.2d 216, 220-23 (1983) (utilizing this analysis to determine whether application of a statute in general, considering typical conduct, constitutes cruel and unusual punishment). Accord Solem v. Helm, - U.S. -, 103 S.Ct. 3001, 3010-11, 77 L.Ed.2d 637 (June 28,1983) (utilizing this analysis to find sentence under statute unconstitutional as cruel and unusual punishment). I consider the same analysis appropriate in determining whether imposition of an adjusted presumptive term on a defendant, in light of his specific conduct and background, would be so severe or so lenient that it constitute manifest injustice. Obviously, a court could refer cases to the three-judge panel on a record that would not support a finding that the statute was unconstitutional. Ultimately the court would determine whether defendant’s conduct, character and offense are typical or substantially atypical, and if atypical, whether the statutory aggravating and mitigating factors adequately address this atypicality. It is only where the court determines that the presumptive sentence adjusted for all aggravating and mitigating factors still unduly depreciates the seriousness of the offense or alternatively approximates cruel and unusual punishment that he may properly determine manifest injustice and properly refer the case to a three-judge panel. The referral must be accompanied by the trial court’s complete findings of fact and conclusions of law and the appropriate record. See AS 12.55.165.
Fourth, it seems clear that the three-judge panel is not bound by the trial judge’s findings of fact since it can grant a hearing de novo and take oral testimony.4 The three-judge panel’s power to redetermine the facts presents the primary problem of appellate review where trial court and panel disagree.5
Fifth, once the three-judge panel has reevaluated the facts and made its own fact findings, it then proceeds to a determination of whether imposition of the presumptive sentence, adjusted to reflect all applicable aggravating and mitigating factors, would nevertheless result in manifest injustice. It appears that we are all in agreement that the three-judge panel is free to exercise its independent judgment to the extent permitted by Juneby in determining which aggravating and mitigating factors are applicable and what weight should be given the factors found in redetermining the adjusted presumptive sentence in order to determine whether its imposition would work a manifest injustice. Three results are possible. If the three-judge panel agreed with the trial court’s determination that imposition of the adjusted presumptive sentence would result in a manifestly unjust too lenient sentence or, alternatively, a manifestly unjust too severe sentence, I believe we all agree that the three-judge panel would then go on to sentence the offender based on its evaluation of the facts governed only by the criteria established in State v. Chaney, 477 P.2d 441 (Alaska 1970). See AS 12.55.005. A dissatisfied party could then appeal that decision to this court challenging (1) the finding of “manifest injustice” which, if sustained, would result in remand to the superior court to impose the presumptive sentence or, alternatively, (2) the resulting sentence under the Chaney criteria which, if sustained, would result in a remand to the three-judge panel to impose an appropriate sentence. Otherwise, the sentence would be affirmed. See Alaska R.Crim.P. 32(d)(7) (if three-judge panel imposes sentence, it becomes “sentencing court” for all further proceedings).
*1162If the three-judge panel disagrees with the trial court that imposition of the adjusted presumptive sentence would work a manifest injustice, the case is then remanded to the trial court to perform the ministerial act of imposing the presumptive sentence. Since the trial court has already made its findings of fact and conclusions of law, evaluated all aggravating and mitigating circumstances and adjusted them to arrive at an adjusted presumptive sentence, there is nothing left for the trial court to do but impose that sentence. The case is at an end. A dissatisfied party may then appeal that decision to this court. On appeal, it would appear that this court could consider whether “manifest injustice” in fact existed for the reasons found by the trial court but rejected by the three-judge panel. If it concluded that manifest injustice did exist for the reasons outlined by the trial court, it would remand the matter to the three-judge panel to impose an appropriate sentence. Alternatively, if this court agrees with the three-judge panel that manifest injustice was not established, it would then undertake the limited review contemplated by Juneby, to determine whether the trial court adequately handled aggravating and mitigating factors. I do not understand this alternative to present any unique problem except determining whose fact findings control in the event of a disagreement.
It is the third situation that poses the problem addressed by the majority in this case. The three-judge panel agrees with the trial court that imposition of an adjusted presumptive sentence would be manifestly unjust but concludes that a just sentence would operate in a direction opposite from that contemplated by the trial judge. The statutes are ambiguous regarding the respective authority of the trial judge and the three-judge panel under these circumstances. Resolution of the ambiguity requires interpretation of the word “agrees”6 in the legislative commentary to AS 12.55.-175. The word does not occur in the statute itself, though it is not unreasonable to imply it.
In Heathcock’s ease, the three-judge panel decided that so long as they agreed with the trial judge that imposition of the adjusted presumptive sentence would work a manifest injustice, they could then go on to sentence Heatheock pursuant to AS 12.55.-175(c) to any sentence authorized by AS 12.55.125, subject to the Chaney criteria, AS 12.55.005. This is a plausible reading of the statutes and would permit a party disagreeing with the result reached by the three-judge panel to obtain relief in a single appeal. The appeals court would initially consider the total record and determine whether the facts warranted a finding of “manifest injustice” from imposition of the adjusted presumptive term. In so doing, we would preliminarily determine whether aggravating and mitigating factors were properly found and used in the adjustment of the presumptive term. If we concluded that any finding of manifest injustice was clearly mistaken, we would remand to the trial court to perform the ministerial act of imposing the adjusted presumptive term. That is essentially the result we have reached in Heathcock’s case. If we concluded that the trial judge’s finding, that the adjusted presumptive term was manifestly too severe, was not clearly mistaken but that the panel's decision that it was too lenient was clearly mistaken, we would so hold and remand to the three-judge panel to impose a sentence less than the presumptive term. Alternatively, if we accepted the three-judge panel’s finding that the adjusted presumptive term was manifestly too *1163lenient, we would then go on to review the three-judge panel’s sentence under the Chaney criteria. In any event, a single appeal would dispose of the case in all but the most unusual circumstances.7
In contrast, consider the effect of the majority’s interpretation of the statute. Assume that the trial court determines that the adjusted presumptive term is too lenient and sends the case to the three-judge panel expecting a more severe sentence. The three-judge panel agrees that the adjusted presumptive term would work a manifest injustice but concludes that it would be too severe. Under the majority’s analysis, the three-judge panel could not at that point go on to impose what it considered an appropriate sentence but must close the case and remand to the trial court. It must, however, make findings of fact and enter conclusions of law fully explaining its position. Alaska R.Crim.P. 32(d)(5). Under the circumstances, on remand the defendant will no doubt ask the trial court to reconsider and accept the three-judge panel’s determination that the presumptive sentence is too severe, and retransfer the case to the three-judge panel. Conceivably, the trial judge, impressed by the three-judge panel’s reasoning, will agree and transfer the case back to the panel to impose a sentence less than the adjusted presumptive term. More likely, the trial judge will grit his teeth, adhere to his original decision, and with a blast at the three-judge panel, impose the adjusted presumptive term. In either event it is likely that both the state and the defendant will appeal. If this court, is satisfied on the facts found, that a finding of manifest injustice in either direction is clearly mistaken, it may affirm imposition of the presumptive term and the case is at an end. But, if the court concludes that the trial judge in the hypothetical was correct, that the adjusted presumptive sentence is too lenient, it must transfer the case back to the three-judge panel for imposition of an appropriate sentence which then could be appealed again. If, alternatively, the court agrees with the three-judge panel that imposition of the adjusted presumptive sentence would be manifestly unjust because too severe, it must remand the case to the trial court with directions to transfer the matter back to the three-judge panel for imposition of an appropriate sentence, and presumably that sentence would also be subject to appeal.
In summary, given the majority’s interpretation of the statute, I can envision future cases bouncing from court to court like a ping-pong ball, resulting in unnecessary cost and substantial delay. In contrast, adoption of the three-judge panel’s interpretation of the statute would have the merit of a final decision within sixty days of referral, see Alaska R.Crim.P. 32(d)(5), resulting in a single appeal with a final determination within a reasonable time.8
The majority errs,9 in my view, in seeing the panel as essentially a reviewing court *1164similar to this one, rather than as a sentencing panel. (A sentencing panel substitutes for the trial judge as initial decision maker in unusual cases). I believe the legislature viewed the panel as a trial court. It is clear that they intended the three-judge panel to have primary sentencing authority in sentencing repeat offenders in atypical cases while a single judge would sentence in typical cases within the presumptive sentencing framework. The question of what cases are exceptional is a mixed question of fact and law in which the factual elements predominate. Who decides what eases are exceptional and why? While the trial court must preliminarily make the determination, it is clear that the legislature intended that the panel ultimately and independently decide the question de novo, and that it be able to consider evidence that was not before the trial court. See AS 12.55.175. Consequently, the panel’s role cannot be viewed as a reviewing court in deciding the issue of manifest injustice. Thus, it appears that the trial judge stands in the shoes of an administrative law judge recommending a decision to an agency. The agency must consider the recommendation as part of the record but is not bound by it and can, in deciding the matter de novo, reach the opposite conclusion. See Universal Camera Corp. v. NLRB, 340 U.S. 474, 492-97, 71 S.Ct. 456, 466-469, 95 L.Ed. 456, 469-72 (1951), 3 K. Davis Administrative Law § 17:16, 327-30 (2d ed. 1983).
It appears that the Alaska Supreme Court recognized that the panel is a trial court and not a reviewing court when it gave the parties the right to peremptorily challenge a member of the panel. Alaska R.Crim.P. 32(d)(4). The remand requirement when the panel disagrees with the trial court regarding manifest injustice does not change the analysis since the trial judge’s function on remand is purely ministerial — to impose the adjusted presumptive sentence which he had previously determined.10
*1165It is clear that the legislative committee viewed the trial judge’s function on remand as ministerial:
If the three-judge panel does not agree with the sentencing court’s finding, it is required to remand the case to the sentencing court with a written statement of its findings and conclusions for sentencing under § 12.55.125. The sentencing court must then sentence the defendant to the presumptive term, adjusted for any relevant aggravating or mitigating factors established under § 12.55.155.
Commentary on the Alaska Revised Criminal Code, Senate J. Supp. No. 47, at 163, 1978 Senate J. 1399 (emphasis added).
Finally, regardless of the conclusions reached by the trial court and the panel, the issue of “manifest injustice” is reviewable by this court on appeal. We, not the panel, are the reviewing court. See Alaska Criminal Code Revision Part VI at 78 (Tent.Draft 1978) (denial of a request for extraordinary sentencing is an appealable matter).
It is only in the situation where the trial judge and the three-judge panel could both legitimately find “manifest injustice” resulting from the imposition of the presumptive sentence and yet disagree as to whether that manifest injustice results because the presumptive sentence is too lenient on the one hand or too severe on the other, that problems of statutory interpretation arise. This is so because the legislature failed to make the trial judge’s findings of fact binding on the three-judge panel and empowered it to conduct a hearing de novo and independently determine the facts.
Recent amendments to the statutes governing sentencing of those convicted of first-degree sexual assault, which impose an eight-year presumptive sentence on first offenders who do not cause serious physical injury or use a dangerous instrument, may frequently present situations where reasonable people could agree that an adjusted presumptive sentence would be manifestly unjust but disagree on whether it was too severe or too lenient. See AS 12.55.-125(i)(l). This is so because of the broad range of behavior covered by the offense of first-degree sexual assault, see Reynolds v. State, 664 P.2d 621, 623-27 (Alaska App.1983), and the wide disagreement among judges regarding appropriate sentences for offenders convicted of first-degree sexual assault, where the facts are virtually indistinguishable. See Langton v. State, 662 P.2d 954, 962-63 (Alaska App.1983). The eight-year presumptive term for a first offender convicted of first-degree sexual assault cannot be reduced below four years no matter how mitigated the offense. AS 12.-55.155(a)(2). There is, therefore, a substantial risk that (1) some trial judges will conclude that imposition of a presumptive eight-year sentence on a first-felony offender convicted of first-degree sexual assault, based upon the trial judge’s view of the facts, was so lenient as to result in manifest injustice, while (2) a three-judge panel looking at the same facts but viewing them differently will conclude that the adjusted presumptive sentence is manifestly unjust because too severe, and (3) an appellate court giving due deference to the proper fact-finders role, will not be able to say that either view is clearly mistaken.11 It is that case that presents the serious problems addressed in the majority’s dicta.
*1166Consider the facts in State v. Doe, 647 P.2d 1107 (Alaska App.1982). Under current law, Jane Doe’s offense would be first-degree sexual assault, see Langton v. State, 662 P.2d 954, 958-59 (Alaska App.1983). She would be vulnerable to an eight-year presumptive term which could not be reduced below four years’ imprisonment despite mitigating factors. A trial judge might make fact findings which would render Jane Doe’s case indistinguishable from Seymore v. State, 655 P.2d 786 (Alaska App.1982), justifying a sentence in excess of eight years, but still not be able to find by clear and convincing evidence that statutory aggravating factors so outweighed the statutory mitigating factors that the eight-year sentence could be increased. See Linn v. State, 658 P.2d 150, 153-54 (Alaska App.1983); cf. Woods v. State, 667 P.2d 184 (Alaska 1983) (consideration of other offenses which did not result in conviction as aggravating factors requires reference to three-judge panel). He might then send the case to the three-judge panel. In contrast, the panel might make the fact findings actually made by the court in Jane Doe and based on those findings conclude that an adjusted presumptive four-year sentence was manifestly unjust because it was too severe. Both sets of fact findings, while in conflict, could be sustained against a clearly mistaken challenge. Langton v. State, 662 P.2d at 959-60.
Reynolds is another case where different fact finders could review the same record and reach conflicting conclusions regarding whether a four-year sentence was manifestly unjust because too severe or an eight-year sentence was manifestly unjust because too lenient. In light of Doe and Reynolds, I do not think the concerns raised in this opinion are purely academic and, therefore, appropriately ignored by pragmatic jurists.
The legislature adopted presumptive sentencing to limit disparity and encourage uniformity in sentencing. Wisely, it included a safety valve to deal with situations where the statute rigidly applied would work an injustice. To prevent trial judges from judicially repealing presumptive sentencing, the legislature limited the safety valve by allowing a trial judge to refer a case to a sentencing panel but ensured that no sentence would deviate from the adjusted presumptive sentence unless it was debated by at least three judges and joined in by two. Today’s decision makes use of the three-judge panel so cumbersome that I fear it will cease to function. This will deprive the statute of a needed safety valve. I therefore dissent.
. Judge Serdahely’s findings and conclusions are clearly before us in this appeal and should be expressly addressed to avoid multiple appeals. The forgery prohibited by AS 11.46.-505(a)(1) is a class C felony, punishable by up to five years’ imprisonment. Presumptive sentences are respectively two years for second felony offenders and three years for third felony offenders. AS 12.55.125(e)(1). Since a presumptive term of four years or less can be reduced to no time to serve by factors in mitigation, see AS 12.55.155(a)(1), and a finding that the defendant’s conduct was among the least serious conduct included within the definition of the offense would warrant mitigation, AS 12.55.155(d)(9), it would be rare for a person convicted of a class C felony to show that imposition of the adjusted presumptive sentence would be manifestly unjust. Given the trial court’s power to impose a five-year sen*1159tence if aggravating factors are found, the contrary conclusion that the adjusted presumptive term was manifestly unjust because it was too lenient would also be rarely sustained.
The trial judge thought Heathcock’s $550 forgery was insignificant. It is unclear why he did not find it subject to AS 12.55.155(d)(9). AS 11.46.505(a)(1) is analogous to AS 11.46.280 (issuing a bad check), and AS 11.46.100 (theft). Both punish an offense involving $500 to $25,-000 as a class C felony. AS 11.46.280(d)(2), AS 11.46.130(a)(1), (b). A $550 transaction would thus seem on the low end of the conduct covered by the offense and warrant mitigated treatment. Forgery was viewed by the Criminal Law Revision Commission as a form of preparatory theft. See Alaska Criminal Code Revision Part III, at 82 (Tent.Draft 1977). It is true that AS 11.46.500- 580, punishing forgery, do not make the seriousness of the offense turn on the amount to be gained by the forgery since the penalty in part reflects “the strong community interest in being able to rely on written instruments to be what they appear to be.” Id. The legislature has, however, concluded that there is a distinguishing factor between misdemeanor forgeries and felony forgeries:
[Fjorgery will be a felony when the written instrument falls into one of the four categories described in paragraphs (l)-(2) of AS 11.46.500 or .505, the forging of these instruments merit felony classification since the conduct will usually be preliminary to a large scale fraud.
Commentary on the Alaska Revised Criminal Code, Senate J. Supp. No. 47, at 54, 1978 Senate J. 1399 (emphasis added); AS 11.46.980(c) (in determining the degree or classification of a crime under this chapter, amounts involved in criminal acts committed during one course of conduct, whether from the same person or several persons, shall be aggregated). The trial judge did not seem to consider Heathcock’s acts preliminary to a large scale fraud.
The majority thinks that the trial judge rejected the mitigating factor because Heathcock failed to raise it prior to the hearing. See AS 12.55.155(f). If so, the trial court was in error. It has discretion to notice aggravating and mitigating factors established in the record sua sponte so long as the parties are given a reasonable opportunity to be heard. See Hartley v. State, 653 P.2d 1052, 1055-56 (Alaska App.1982).
A finding of a mitigating factor does not mandate a mitigated sentence, however, the factors established must warrant mitigation in light of the totality of the circumstances. See Linn v. State, 658 P.2d 150, 154 (Alaska App. 1983); AS 12.55.155(d). The trial court could have found the mitigating factor but concluded that in light of Heathcock’s committing a crime while on probation and committing a crime which was part of a pattern with his earlier offenses, mitigation was not warranted. Linn v. State, 658 P.2d at 154. But see AS 12.55.-155(d)(13) (defendant’s offenses consistently minor). The trial judge should clarify this ambiguity in his findings on remand.
. The three-judge panel reasoned that Heath-cock’s committing the crime while on probation warranted aggravating his offense as a factor not specifically set out in the statute. AS 12.55.165; AS 12.55.175; see Woods v. State, 667 P.2d 184 (Alaska 1983) (if a trial judge feels probationary status of a defendant warrants enhanced sentence, he must refer case to a three-judge panel). Alternatively, the panel concluded that Heathcock’s recidivism required that he not receive flat time but be subject to probation after completing his sentence. Such a sentence can only be given by a three-judge panel in the absence of statutory aggravating factors. McManners v. State, 650 P.2d 414, 416 (Alaska App.1982). A defendant presumptively sentenced is not subject to parole or parole supervision. This is a close case in light of Woods and McManners but both factors would appear sufficiently routine to preclude their use as factors establishing manifest injustice. This is in accord with the views expressed in A. Dershowitz, Fair and Certain Punishment, 20-21 (1976) (the source of the aggravating and mitigating factors found in AS 12.55.155 as originally enacted):
The Task Force recommends that the legislature, or the body it designates, also define specific aggravating or mitigating factors, again based on frequently recurring characteristics of the crime and the criminal. If aggravating factors outweighed the mitigating factors, the judge could impose a sentence that exceeded the presumptive by a specified percentage. If mitigating factors outweighed aggravating factors, the judge could impose a sentence that fell below the presumptive, again by a specified percentage. The Task Force believes that sentencing hearings should be mandatory to establish any aggravating or mitigating circumstances and to have the sentence pronounced.
In imposing sentences, judges normally consider a wide variety of factors, some of these, such as the defendant’s race, appearance, or sex, are clearly improper; others, such as whether the defendant pleaded guilty or “cooperated” with the authorities, are debatable. It is the view of the Task Force, based on its own experience and on what it has learned about the system as a whole, that different judges — acting without legislative or appellate court guidance — have different views as to whether a given factor is appropriately considered in sentencing. It is our *1160conclusion that these issues should be openly debated, that, in situations where the factors are fairly typical and frequently recurring, the legislature (or delegated body) should decide whether these factors should be considered in sentencing.
Only in truly extraordinary and unanticipated circumstances would the judge be permitted to deviate from the presumptive sentence beyond the narrow range permitted by an ordinary finding of aggravating or mitigating factors. Any deviation would have to be justified in a reasoned opinion subject to a searching review on appeal. Absolute maximum or minimum sentences available to a judge in such extraordinary cases should also be established by the legislature.
(Emphasis added.) It is the “extraordinary and unanticipated circumstances” which our legislature apparently assigned to the three-judge panel.
See AS 12.55.155(c)(20) (probation status is now an aggravating factor). It appears that earlier omission of this factor was an oversight. To permit its retroactive use against a defendant whose crime preceded the amendment by means of the three-judge panel would create at least the appearance of an ex post facto problem.
. While referral should not be routine, atypical cases are sufficiently frequent that we should not discourage its use by making the procedure too cumbersome.
. The power to hold a hearing de novo does not require the panel to take new evidence if the matter can be resolved fairly on the existing record, see Keiner v. City of Anchorage, 378 P.2d 406, 410 (Alaska 1963), nor does the power to substitute judgment require a panel to disregard the trial court’s fact findings, see Pan Am Petroleum Corp. v. Shell Oil Co., 455 P.2d 12, 19-22 (Alaska 1969).
. While the panel emphasized factors differently than did the trial judge in Heathcock’s case, it does not appear that it differed with the trial court on the historical facts.
. Agreement could refer to: (1) the ultimate conclusion that imposition of the adjusted presumptive term would be manifestly unjust, the view taken by the panel; (2) the preliminary conclusion that it was manifestly too severe or too lenient, the view apparently taken by a majority of this court; or (3) the mixed questions of fact and law involved in determining applicable aggravating and mitigating factors and weighing their application to the facts of the particular case. As I read the majority opinion, the panel and the trial court would “agree” so long as they reached the same preliminary conclusion — too severe or too lenient — even though their reasons, including their fact findings and handling of aggravating and mitigating factors, differed.
.The statute clearly gives the three-judge panel power to hold a hearing de novo and by implication substitute judgment on the facts. If the trial court’s finding and the three-judge panel’s finding were both supported by substantial evidence and therefore not clearly mistaken, a reasonable interpretation of the statute would give preference to the three-judge panel. This view is also supported by the observations that when reasonable people could differ on the facts, a collegial decision, i.e., one by a jury or panel of judges, is better than an individual decision and that having a statewide panel will make the ultimate decision serve the interest of uniformity in the law. See S. Weiner, The Civil Non-Jury Trial and The Law Fact Distinction, 55 Calif.L.Rev. 1020, 1032-36, 1050-52 (1967).
This view also finds support in the respective treatment given decisions of administrative law judges and agency decisions on appeal when they conflict. See Universal Camera Corp. v. NLRB, 340 U.S. 474, 492-97, 71 S.Ct. 456, 466-469, 95 L.Ed. 456, 469-72 (1951); 3 K. Davis, Administrative Law § 17:16, 327-30 (2d ed. 1983).
. This assumes that both trial court and panel will have complied with the statute and completed their treatment of the case before it is passed along to the next court so that the trial court’s findings and conclusions and the panel’s findings and conclusions are before us at the time of the first appeal. That was the case here.
. While the parties have a right to be heard on the question whether the case should be referred to a three-judge panel, they have no *1164veto. The trial judge must refer sua sponte if he is satisfied that imposition of the adjusted presumptive term would work a manifest injustice. Alaska Criminal Code Revision Part VI at 78 (Tent.Draft 1978). Compare Hartley v. State 653 P.2d 1052, 1055-56 (Alaska App. 1982) (trial court may sua sponte notice aggravating and mitigating factors).
Heathcock argues for the interpretation adopted by the majority contending that the view of the statutes taken by the panel raises insurmountable problems of double jeopardy and procedural due process. This is an incorrect interpretation. There is no double jeopardy problem under federal law even if we view the panel as a reviewing court. See United States v. DiFrancesco, 449 U.S. 117, 101 S.Ct. 426, 66 L.Ed.2d 328 (1981) (reviewing court may increase sentence on appeal by prosecution without violating due process). Under Alaska law there would be no double jeopardy problem unless the trial court “meaningfully imposed” a sentence before transfer to the three-judge panel. See Dentler v. State, 661 P.2d 1098 (Alaska App.1983) (sentence not meaningfully imposed for double jeopardy purposes when suspended time added minutes later). Since the reason for referral is the trial court’s refusal to impose a sentence (meaningfully or otherwise) he considers manifestly unjust, there is no double jeopardy problem under Alaska law.
The due process problems are also not insurmountable. In the instant case, the three-judge panel increased the sentence without notifying Heathcock that this was a possibility. Heath-cock would be entitled to vacation of the sentence and remand to enable him to be heard on the issue. Hartley, 653 P.2d at 1055-56. Since we have determined that the panel’s conclusion was clearly mistaken, remand to the three-judge panel in this case is unnecessary. In the future, the problem could be avoided by notifying the parties that the panel is inclined to consider increasing the sentence to enable him to be heard, and by granting a brief continuance if necessary to enable preparation. This is similar to the procedure recommended where a person is charged with contempt which could be treated as either civil or criminal. Johansen v. State, 491 P.2d 759, 766 n. 27 (Alaska 1971).
Finally, Heathcock argues that the phrase “manifest injustice” renders the statute void for vagueness. I disagree. The phrase is no more vague than a number of other standards governing jury and judicial action: reasonable doubt, reasonable care, reasonable force under the circumstances. It is no more vague than the “shock the conscience” test associated with substantive due process, or the phrase “cruel and unusual punishment.” Use of the standard reinforces my view that the legislature intended a quasi-jury, i.e., a three-judge panel, to make the ultimate decision.
. If the trial judge and the three-judge panel disagreed on the adjusted presumptive sentence, i.e., calculated the aggravating and miti*1165gating factors differently, the trial judge could no doubt consider the three-judge panel’s treatment of those factors in a Criminal Rule 35(a) proceeding.
. Heathcock argues that ambiguous criminal statutes should be construed in favor of defendants. In his case, the trial judge would have been more lenient than the three-judge panel but, as hypothesized in the text, the converse could as easily have occurred. Which construction of the statute is therefore more favorable to defendants? Judge Frankel has extensively studied sentencing panels, which, unlike the panel here, include the trial judge. Judge Frankel has concluded that when a comparison is made between individual decisions, including those by members of the panel and panel decisions, the panel more often than not is more lenient. The panel tends toward “lenity” by giving shorter prison terms and making increased use of probation rather than confinement. M. Frankel, Criminal Sentences: Law Without Order, 70-71 (1973). I assume that this would be particularly true in the realm of sexual assault,