OPINION
DIETZEN, Justice.
Appellant Dylan Micheál Kelley was found guilty by a Benton County jury and convicted as an accomplice to first-degree aggravated robbery and third-degree assault.1 Kelley appealed his first-degree aggravated robbery conviction to the court of appeals, requesting a new trial based on an unobjected-to jury instruction on accomplice liability. He claimed the court of appeals could grant the requested relief under Minn. R.Crim. P. 31.02 because the jury instruction plainly violated the newly announced rule in State v. Milton, 821 N.W.2d 789 (Minn.2012).2 The court of appeals affirmed Kelley’s conviction, concluding the accomplice liability instruction given was error, but that he was not entitled to relief under Rule 31.02 because the law was unsettled at the time of the error and did not become settled in favor of Kelley until the time of appeal. We conclude Rule 31.02 is not limited to errors that were plain at the time of trial. Instead, it includes errors that are based on law, which although unsettled at the time of the error, was settled in favor of the defendant at the time of appellate review.3 Nevertheless, because the error alleged in this case did not affect Kelley’s substantial rights, we affirm as modified.
The victim, S.A., told police at a St. Cloud hospital that he was assaulted and robbed by two men in the early morning hours of January 4, 2011. The police investigated the case, and S.A. identified Kelley as one of his attackers. Kelley was arrested and charged with first-degree aggravated robbery in violation of Minn.Stat. § 609.245, subd. 1 (2012), and third-degree assault in violation of Minn.Stat. § 609.223, subd. 1 (2012).
At trial, the State presented evidence that on January 3, 2011, S.A. went to the apartment of his friend, B.G., in St. Cloud. When S.A. arrived, B.G. was having a party and there were approximately twenty [273]*273people at the apartment. At some point, S.A. told B.G. he was interested in getting some marijuana and was directed toward Dylan Kelley. S.A. approached Kelley, and Kelley told him he could buy some marijuana from a friend. S.A. drove Kelley to the friend’s house, and Kelley purchased the marijuana and gave it to S.A. S.A. complained that it was an inadequate amount of marijuana, and Kelley responded that his friend would bring the rest to B.G.’s apartment.
Kelley’s friend arrived at B.G.’s apartment later that evening, and told S.A. that Kelley wanted to see him outside. S.A. agreed and approached Kelley, who was leaning into the window of a parked car. Kelley quickly turned around and hit S.A. in the face. Kelley’s friend held S.A. from behind and Kelley repeatedly hit S.A. in the face. When S.A. fell to the ground, they both kicked him. After they kicked S.A. in the face and fractured five of his teeth, S.A. lost consciousness. Kelley and his friend stole S.A.’s cigarettes, lighter, cell phone, and car keys. Subsequently, S.A. regained consciousness and returned to the apartment and told those present what happened. S.A. later discovered that his wallet, which contained about $240, was missing from his car. S.A. was taken to the hospital that morning for treatment.
At the close of the evidence the State requested that the district court instruct the jury on accomplice liability. Kelley argued that the accomplice liability instruction should not be given because he was charged as a principal and was not charged with aiding and abetting. Kelley further argued that the State had not offered any evidence regarding who committed the crime or who Kelley aided and abetted. The district court overruled Kelley’s objection and the standard accomplice liability jury instruction was given to the jury.
The jury found Kelley guilty of both offenses. Subsequently, the district court entered judgment of conviction for first-degree aggravated robbery and third-degree assault and sentenced him to the presumptive sentence of 58 months.
The court of appeals affirmed, even though it concluded that the accomplice liability instruction given to the jury was legally erroneous because it failed to explain the “intentionally aiding” element of accomplice liability as required by State v. Milton, 821 N.W.2d 789, 806 (Minn.2012), which was decided after Kelley’s conviction but before he filed his appellate brief. State v. Kelley, 832 N.W.2d 447, 451-52 (Minn.App.2013). But the court further concluded the error was not plain because at the time of the trial, the obligation of the district court to explain the “intentionally aiding” element of accomplice liability was unsettled and did not become settled in favor of Kelley until the time of appeal. Id. at 456-57.
I.
Kelley argues that the accomplice liability instruction given to the jury for the offense of first-degree aggravated robbery failed to accurately state the law. Kelley acknowledges that he did not object to the instruction on this specific basis, and therefore we review the instruction for plain error.
The three requirements that an appellant must satisfy under the plain-error doctrine were first articulated in United States v. Olano, 507 U.S. 725, 732, 113 S.Ct. 1770, 123 L.Ed.2d 508 (1993), and later clarified in Johnson v. United States, 520 U.S. 461, 466-67, 117 S.Ct. 1544, 137 L.Ed.2d 718 (1997). We adopted those requirements in State v. Griller, 583 N.W.2d 736, 740 (Minn.1998). Under the plain-error doctrine, the appellant must [274]*274show that there was (1) an error; (2) that is plain; and (3) the error must affect substantial rights. Griller, 583 N.W.2d at 740 (citing Johnson, 520 U.S. at 467, 117 S.Ct. 1544). If the appellant satisfies the first three prongs of the plain-error doctrine, “we may correct the error only if it ‘seriously affeet[s] the fairness, integrity, or public reputation of judicial proceedings.’” State v. Crowsbreast, 629 N.W.2d 433, 437 (Minn.2001) (quoting Johnson, 520 U.S. at 467, 117 S.Ct. 1544); accord Griller, 583 N.W.2d at 740. We consider each prong of the plain-error doctrine in turn.
Under the plain-error doctrine, an “error” is a “[deviation from a legal rule [ ] unless the rule has been waived.”4 Olano, 507 U.S. at 732-33, 113 S.Ct. 1770. In the context of jury instructions, a district court has broad discretion. State v. Anderson, 789 N.W.2d 227, 239 (Minn.2010). But a district court abuses that discretion if its jury instructions confuse, mislead, or materially misstate the law. State v. Vang, 774 N.W.2d 566, 581 (Minn.2009); State v. Moore, 699 N.W.2d 733, 736 (Minn.2005). We review the jury instructions as a whole to determine whether the instructions accurately state the law in a manner that can be understood by the jury. State v. Scruggs, 822 N.W.2d 631, 642 (Minn.2012).
Accomplice liability is defined by statute in Minn.Stat. § 609.05 (2012). The statute provides:
Subdivision 1. Aiding, abetting; liability. A person is criminally hable for a crime committed by another if the person intentionally aids, advises, hires, counsels, or conspires with or otherwise procures the other to commit the crime.
Subd. 2. Expansive liability. A person liable under subdivision 1 is also liable for any other crime committed in pursuance of the intended crime if reasonably foreseeable by the person as a probable consequence of committing or attempting to commit the crime intended.
Id. The statute does not define the phrase “intentionally aids.” The district court gave the standard accomplice liability jury instruction.5
In State v. Milton, the defendant argued for'the first time on appeal that an instruction on accomplice liability was legally er[275]*275roneous because the district court did not require the jury to find that he knowingly and intentionally aided another to commit the aggravated robbery. 821 N.W.2d 789, 805 (Minn.2012). We explained that the law regarding such a requirement was unsettled because our previous discussions of “intentionally aiding” had been in the context of sufficiency-of-the-evidence claims, not the adequacy of jury instructions. Id. at 807. We concluded that an accomplice liability instruction must explain that the intentionally aiding element requires that the jury find beyond a reasonable doubt that the defendant knew his alleged accomplice was going to commit a crime and the defendant intended his presence or actions to further the commission of that crime. Id. at 808. We nevertheless affirmed Milton’s conviction, explaining that he failed to establish an error that was plain. Id. at 807.
The instruction on accomplice liability in this case failed to explain the intentionally aiding element as required by Milton, and therefore was error. Indeed, the State concedes that Kelley has satisfied the first prong of the plain-error doctrine.
II.
The crux of the dispute is whether the plainness of the error is examined at the time of the district court’s error or at the time of appellate review. Kelley argues that plain error is determined at the time of appellate review and the State counters that plain error is determined at the time of the district court’s error. For the reasons that follow, we conclude that plain error is determined at the time of appellate review.
A.
In State v. Baird, we considered whether a defendant could satisfy the second prong of the plain-error doctrine by establishing the error was plain at the time of the appeal where the law was unsettled at trial but settled in favor of a defendant during the pendency of the appeal. 654 N.W.2d 105, 113 (Minn.2002). Specifically, we considered whether the district court’s unobjeeted-to jury instruction that Baird had a duty to retreat from his home before using self-defense against a co-resident constituted an error that was plain for the purposes of the plain-error doctrine. Id. We acknowledged that at the time of Baird’s trial, the law regarding the duty to retreat was unsettled. Id. at 109 n. 2. While Baird’s appeal was pending before the court of appeals, the law became settled in Baird’s favor due to our decision in State v. Glowacki, 680 N.W.2d 392, 402 (Minn.2001), which held that there is no duty to retreat from one’s own home before using self-defense against a co-resident. Baird, 654 N.W.2d at 112.
We concluded in Baird that plain error is determined at the time of appellate review when the law was unsettled at the time of the error but settled in favor of the defendant during the pendency of the appeal. Id. at 113. We quoted directly from Griller stating that “ ‘[t]o satisfy the second prong [of the plain-error doctrine] it is sufficient that the error is plain at the time of the appeal.’ ” Id. (quoting Griller, 583 N.W.2d at 741). We applied the plain-at-the-time-of-appeal rule announced in Gril-ler, and held that “[b]ecause Glowacki’s holding ... rendere[d] the duty-to-retreat instruction given in error at Baird’s trial clear and obvious, the error [was] plain.” Id.
Our reliance in Baird on the plain-at-the-time-of-appeal rule announced in Gril-ler was well-founded and reasonable.6 Be[276]*276cause we reversed the conviction based on plain error, our discussion of each prong of the plain-error doctrine was essential to our decision, and therefore not dicta. See Carlton v. State, 816 N.W.2d 590, 614 (Minn.2012) (explaining the statement in question was “dicta” because the resolution of the question was not necessary to our ultimate holding).
Our analysis in Griller relied upon Johnson, 520 U.S. 461, 117 S.Ct. 1544, to explain plain error. We explained that
[i]n Johnson, the [United States Supreme] Court considered whether the error was plain when at the time of trial the district court correctly stated the law, but later that same law became incorrect based on a case decided during the appeal. The Court concluded that to satisfy the second prong it is sufficient that the error is plain at the time of the appeal.
Griller, 583 N.W.2d at 741. We applied the Johnson analysis in Griller, concluding that “[u]nder Pendleton, which was released after Griller’s conviction and while his case was on appeal, the defense-of-dwelling instruction given is now in error, and thus the error is plain.” Id. In other words, the court in Griller concluded that the plain-at-the-time-of-appeal rule applies when a district court correctly states the law at the time of trial, but later that same law becomes incorrect based on a case decided during appeal.
The United States Supreme Court in Henderson v. United States, — U.S. -, 133 S.Ct. 1121, 185 L.Ed.2d 85 (2013) recently reaffirmed the plain-at-the-time-of-appeal rule. In Henderson, the defendant appealed his sentence on the ground that the district court plainly erred in sentencing him to a prison term of 60 months, which was an upward durational departure, so that Henderson could participate in a prison drug rehabilitation program. Id. at -, 133 S.Ct. at 1125. At the time of the sentence, the circuits were split on whether the sentence was proper, and the circuit in which defendant was sentenced (the Fifth Circuit), had not ruled on the issue. Id. at -, 133 S.Ct. at 1125. While the Henderson case was on appeal, the Supreme Court decided Tapia v. United States, 564 U.S. -, 131 S.Ct. 2382, 180 L.Ed.2d 357 (2011), concluding that a court errs when it imposes or lengthens a prison sentence solely for rehabilitative purposes. Henderson, — U.S. at -, 133 S.Ct. at 1125.
In a 6-3 decision, the Henderson Court extended Johnson, concluding that regardless of whether the legal question was settled or unsettled at the time of trial, the second prong of the plain-error doctrine is satisfied if the error is plain at the time of appellate review. Id. at -, 133 S.Ct. at 1130-31. The Court reasoned that assessing the error at the time of appellate review advances the general rule “that an appellate court must apply the law in effect at the time it renders its decision.”7 [277]*277Id. at -, 133 S.Ct. at 1126 (citing Thorpe v. Hous. Auth. of Durham, 393 U.S. 268, 281, 89 S.Ct. 518, 21 L.Ed.2d 474 (1969)). Moreover, the plain-at-the-time-of-appeal approach is consistent with the purpose of the plain-error doctrine, to provide a fairness-based exception to the forfeiture doctrine. Id. at -, 133 S.Ct. at 1127-28. Additionally, to distinguish and treat more harshly cases where the law was unsettled at trial but then settled in favor of a defendant during the pendency of the appeal (the Henderson-scenaño), versus cases where the law was settled against the defendant at trial but then changed in the defendant’s favor during the pendency of the appeal (the Johnson-scenario), “would simply promote arguments about whether the law of the circuit was unclear.” Id. at -, 133 S.Ct. at 1128. Thus, the Court in Henderson determined that regardless of whether an error is made under settled or unsettled law, as long as the error is “plain” at the time of appellate review, it is “plain” for purposes of the plain-error doctrine. Id. at -, 133 S.Ct. at 1130-31.
In sum, we have previously determined that plain error is determined as of the time of appellate review in three circumstances. The first circumstance is when the settled law is the same at the time of trial and appellate review. State v. Dobbins, 725 N.W.2d 492, 513 (Minn.2006); see Olano, 507 U.S. at 730-34, 113 S.Ct. 1770. The second is when the law is settled at the time of trial and the settled law has been reversed as of the time of appellate review. Griller, 583 N.W.2d at 741; see Johnson, 520 U.S. at 464-67, 117 S.Ct. 1544. The third is when the law is unsettled at the time of the district court’s error and the law has become settled in the defendant’s favor at the time of appellate review. Baird, 654 N.W.2d at 113; see Henderson, — U.S. at -, 133 S.Ct. at 1128-31.
We conclude that for purposes of applying the plain-error doctrine the court examines the law in existence at the time of appellate review, not the law in existence at the time of. the district court’s error, to determine whether an error is plain. Our conclusion is supported by our decision in Baird and the U.S. Supreme Court’s decision in Henderson. Additionally, our conclusion simplifies the law by adopting a unified standard for the scenarios discussed in Olano, Johnson, and Henderson.
We next review whether the jury instruction given in this case was plainly erroneous. Kelley was convicted 8 months before we decided Milton, and therefore the district court’s failure to comply with the Milton rule was not plain at the time of Kelley’s conviction. Nevertheless, Kelley filed his brief in the court of appeals after Milton was decided, and therefore the district court’s failure to comply with the Milton rule was plain at the time of appellate review. Because the failure to comply with the Milton rule was plain at the time of appellate review, Kelley has [278]*278satisfied the second prong of the plain-error doctrine.
B.
The concurrence contends we should adopt a plain-at-the-time-of-trial rule on the grounds that it is most consistent with the purpose of the plain-error doctrine, that Rairdon v. State, 557 N.W.2d 318 (Minn.1996), supports a conclusion that “we have never actually decided which rule — plain-at-the-time-of-trial or plain-at-the-time-of-appeal — controls under Rule 31.02,” and that Milton and Kelley are similarly situated, and therefore giving Kelley the benefit of the Milton rule would be unfair. For the reasons that follow, we conclude that the arguments asserted by the concurrence lack merit. We will discuss each argument in turn.
The concurrence first alleges that the plain-at-the-time-of-trial rule is most consistent with the purpose of the plain-error doctrine, which the concurrence claims “provides an incentive for criminal defendants to object at trial by limiting the review of unpreserved errors on appeal and making relief discretionary,” “raises the bar for relief based on unpreserved errors,” and “discourages the strategic withholding of objections in order to gain the proverbial ‘second bite at the apple’ on appeal.” Infra at C-3, C-4. The concurrence’s argument rests upon the mistaken premise that the plain-error doctrine serves the same purpose as the common-law forfeiture doctrine. The argument lacks merit.
The doctrines of forfeiture and plain error have different purposes and are guided by different principles. Under the forfeiture doctrine, “ ‘a constitutional right,’ or a right of any other sort, ‘may be forfeited in criminal as well as civil cases by the failure to make timely assertion of the right before a tribunal having jurisdiction to determine it.’ ” Olano, 507 U.S. at 731, 113 S.Ct. 1770 (quoting Yakus v. United States, 321 U.S. 414, 444, 64 S.Ct. 660, 88 L.Ed. 834 (1944)); see also State v. Williams, 794 N.W.2d 867, 874 (Minn.2011) (explaining that “[w]e ordinarily do not consider issues raised for the first time on appeal, even when those issues are constitutional questions of criminal procedure or are challenges to the constitutionality of a statute”); State v. Goodloe, 718 N.W.2d 413, 422 n. 6 (Minn.2006) (explaining that the term “forfeiture” most accurately described the effect of failing to bring an alleged error to the attention of the district court). The forfeiture doctrine reflects the “need to encourage all trial participants to seek a fair and accurate trial the first time around.” United States v. Young, 470 U.S. 1, 15, 105 S.Ct. 1038, 84 L.Ed.2d 1 (1985) (citation omitted) (internal quotation marks omitted); accord State v. Ramey, 721 N.W.2d 294, 299 (Minn.2006). Put differently, the forfeiture doctrine encourages defendants to object while in the district court so that any errors can be corrected before their full impact is realized.
The plain-error doctrine serves a very different purpose: providing a means for appellate courts to remedy forfeited errors. The plain-error doctrine was first articulated by the United States Supreme Court in Wiborg v. United States, 163 U.S. 632, 16 S.Ct. 1127, 41 L.Ed. 289 (1896). In Wiborg, the defendants failed to request “that the jury be instructed to find for [the] defendants.” Id. at 658, 16 S.Ct. 1127. The Court in Wiborg explained that although the jury instruction issue “was not properly raised [in the trial court], yet if a plain error was committed in a matter so absolutely vital to defendants, we feel ourselves at liberty to correct it.” Id. The Court’s holding in Wiborg was later memorialized in Fed.R.Crim.P. 52(b). See Fed.[279]*279R.Crim.P. 52(b) advisory committee notes — 1944. Federal Rule of Criminal Procedure 52(b) reads: “A plain error that affects substantial rights may be considered even though it was not brought to the court’s attention.” Fed.R.Crim.P. 52(b). The purpose of Rule 52(b) is “to enable the courts of appeals to review prejudicial errors so that any miscarriage of justice may be thwarted.” Young, 470 U.S. at 15 n. 12, 105 S.Ct. 1038 (citation omitted) (internal quotation marks omitted). A serious miscarriage of justice occurs when the error “seriously affects the fairness, integrity or public reputation of judicial proceedings.” Olano, 507 U.S. at 736, 113 S.Ct. 1770 (citation omitted) (internal quotation marks omitted). Based on Fed.R.Crim.P. 52(b),8 we promulgated Minn. R.Crim. P. 31.02, which provides that a “[pjlain error affecting a substantial right can be considered by the court on motion for new trial, posttrial motion, or on appeal even if it was not brought to the trial court’s attention.” Minn. R.Crim. P. 31.02.
The United States Supreme Court has observed, “The plain-error doctrine of Federal Rule of Criminal Procedure 52(b) tempers the blow of a rigid application of the contemporaneous-objection requirement.” Young, 470 U.S. at 15, 105 S.Ct. 1038 (quoted in State v. Ramey, 721 N.W.2d 294, 299 (Minn.2006)). The Court in Young further explained that Rule 52(b) carefully balances the “need to encourage all trial participants to seek a fair and accurate trial the first time around against [the] insistence that obvious injustice be promptly redressed.” Young, 470 U.S. at 15, 105 S.Ct. 1038 (citation omitted) (internal quotation marks omitted). Later, in Olano, the Court emphasized the competing purposes of the forfeiture and plain-error doctrines and explained that because “[a] rigid and undeviating judicial! ]” application of the forfeiture doctrine “would be out of harmony with ... the rules of fundamental justice,” Fed.R.Crim.P. 52(b) reflects an appellate court’s “limited power to correct errors that were forfeited because not timely raised in district court.” 507 U.S. at 731-32, 113 S.Ct. 1770 (citation omitted) (internal quotation marks omitted).
In sum, it is well established that the forfeiture and plain-error doctrines are based on the competing purposes of encouraging timely objections at trial and providing appellate courts a means to remedy unobjected-to errors. The concurrence correctly points out that the plain-at-the-time-of-trial rule encourages timely objections at trial. But that argument is neither relevant nor material because the purpose of the plain-error doctrine is to provide a means for an appellate court to remedy unobjected-to errors.
Not only is the plain-at-the-time-of-appellate-review rule a better rule for allowing an appellate court to correct unobject-ed-to errors, but it also is more consistent with the United States Supreme Court’s analysis of the second prong of the plain-error doctrine in Johnson, 520 U.S. at 467, 117 S.Ct. 1544. Our discussion of the Johnson Court’s analysis of the second prong is informed by a brief review of the Court’s decision in Olano, 507 U.S. 725, 113 S.Ct. 1770. The alleged error in Ola-no was plain both at the time of trial and at the time of appellate review, and therefore the Court could have announced the rule proposed by the concurrence — the plainness prong is only satisfied when the error is plain at the time of trial and the time of appellate review. But the Court [280]*280instead observed that it “need not consider the special case where the error was unclear at the time of trial but becomes clear on appeal because the applicable law has been clarified.” Id. at 734, 113 S.Ct. 1770. The Court ultimately held that an appellate court “cannot correct an error ... unless the error is clear under current law.” Id. (emphasis added).9 Keeping in mind the analysis in Olano, we consider an important clarification to the plainness prong made by the United States Supreme Court in Johnson.
In Johnson, the Court considered for the first time a situation in which the law changed between the time of trial and appellate review. 520 U.S. at 464, 117 S.Ct. 1544. The issue in Johnson was whether the district court committed plain error at the defendant’s perjury trial when the court, rather than the jury, determined that the statement at issue was “material.” Id. At the time of trial, near unanimous precedent in the circuits held that the question of materiality was for the judge to decide, but before the case was appealed, the Supreme Court decided United States v. Gaudin, 515 U.S. 506, 115 S.Ct. 2310, 132. L.Ed.2d 444 (1995), concluding that the question of materiality must be submitted to the jury. Johnson, 520 U.S. at 464, 468 n. 1. The Court noted that “[i]n the case with which we are faced today, the error is certainly clear under ‘current law, ’ but it was by no means clear at the time of trial.” Id. at 467, 117 S.Ct. 1544 (emphasis added). Having clarified that an appellate court cannot correct an error unless the error is clear under the law at the time of appellate review, the Court considered the Government’s argument that to satisfy the plainness prong, an error must be “plain” both at the time of trial and at the time of appellate review. Id. The Court acknowledged that such a rule would certainly encourage timely objections in the trial court but it questioned the usefulness of objections to rulings that were plainly supported by existing precedent. Id. at 467-68, 117 S.Ct. 1544. The Court ultimately rejected the Government’s argument, which demonstrates that an appellate court’s power to address a manifest injustice is not limited to cases where a trial court should have recognized and sua sponte corrected an unobjected-to error. Indeed, the Court held “that in a case such as this — where the law at the time of trial was settled and clearly contrary to the law at the time of appeal — it is enough that an error be ‘plain’ at the time of appellate consideration.” Id. at 468, 117 S.Ct. 1544.
Because the Court’s analysis in Johnson demonstrates that an appellate court’s power to address a manifest injustice is not limited to cases where a trial court should have recognized and corrected the error without the parties’ help, it substantially undercuts the concurrence’s arguments. The plain-at-the-time-of-trial rule is incompatible with Johnson because it limits an appellate court’s power to address a manifest injustice to cases where the trial court should have recognized and sua sponte corrected an unobjected-to error.[281]*28110 In sum, the concurrence’s argument that the plain-at-the-time-of-trial rule is more consistent with the purpose of the plain-error doctrine lacks merit because it mischaracterizes the purpose of the plain-error doctrine and ignores the United States Supreme Court’s decision in Johnson, 520 U.S. 461, 117 S.Ct. 1544.
Additionally, the concurrence relies upon language in Rairdon, 557 N.W.2d at 323, to allege that “we have never actually decided which rule — plain-at-the-time-of-trial or plain-at-the-time-of-appeal — controls under Rule 31.02.” The argument is without merit because our statement in Rair-don was mere dicta.
In Rairdon, the defendant filed a petition for postconviction relief nine years after his conviction. 557 N.W.2d at 322. As part of his petition, he claimed the prosecutor committed plain error in closing argument. Id. at 322-23. Acknowledging that after Rairdon’s trial our response to prosecutorial misconduct had grown “more stringent,” we opined that Rairdon “may not reap any benefit from such decisions merely because he waited nine years to seek review.” Id. at 323. It was in the context of discussing Rairdon’s delay in seeking review that we said: “[T]he proper standard for overturning Rairdon’s murder convictions is found in precedent existing at the time of his conviction.” Id. Importantly, we did not decide in Rairdon whether the error was plain. Instead, we affirmed the district court’s denial of postconviction relief based on the third prong of the plain-error doctrine, which requires a showing that the alleged error affected the defendant’s substantial rights. Id. at 324-25. More specifically, we determined that although the prosecutor’s statements “may have constituted misconduct at the time of Rairdon’s conviction,” “[a]ny improper conduct by the prosecutor ... was not so prejudicial that the defendant was denied a fair trial.” Id. at 324 (emphasis added). In summarizing our analysis, we emphasized that we had assumed that the error was plain and then addressed the third prong of the plain-error doctrine. More specifically we said that “even if Rairdon has identified misconduct plain enough to overcome his failure to object, such misconduct is insufficient to vacate Rairdon’s murder eonvic-[282]*282tions when viewed against all the evidence and in the context of the prosecutor’s entire closing statement.” Id. at 325. Consequently, our statement that “the proper standard for overturning Rairdon’s murder convictions is found in precedent existing at the time of his conviction” was dicta because it was not essential to our decision.11 See Carlton, 816 N.W.2d at 614 (explaining the statement in question was “dicta” because the resolution of the question was not necessary to our ultimate holding); Black’s Law Dictionary 485 (8th ed.2004) (defining judicial dicta as “[a]n opinion by a court on a question that is directly involved, briefed, and argued by counsel, and even passed on by the court, but that is not essential to the decision”).
It is true that we have cited Rairdon for the proposition that an alleged error must be plain at the time of conviction. See, e.g., State v. Tscheu, 758 N.W.2d 849, 863 (Minn.2008) (“[F]or plain error to exist, the trial error must have been so clear under applicable law at the time of the conviction....”) (citation omitted) (internal quotation marks omitted). But in each of those cases the law had either not changed during the pendency of the appeal or the defendant was denied relief based on another of the plain-error prongs.12 Consequently, the issue of when an error must be plain was not material to our analysis. Moreover, our reference to the Rairdon dicta in a few cases does not call into question our decision in Baird, 654 N.W.2d at 113, which squarely addressed the issue before us in this case.
Further, the rule proposed by the concurrence does not promote fairness. The concurrence argues that Kelley and the defendant in Milton are similarly situated, and therefore it is unfair for the court to conclude that the failure to explain the “intentionally aiding” element of accomplice liability to the jury was not “plain” error in Milton, but is “plain” error here. We disagree.
The concurrence contends Milton and Kelley are similarly situated because then-appeals “arrived at this court in an identical procedural posture.” Infra at C-1. If we were announcing a “plain-at-the-time-of-filing-the-appeal” rule, the concurrence’s argument might have merit. But we are not announcing such a rule. Instead, we are adopting a “plain-at-the-time-of-appellate-review” rule. The phrase “appellate review” in this context means appellate review by a court on motion for new trial, posttrial motion, or on appeal. See Minn. R.Crim. P. 31.02. At the time we reviewed Kelley’s conviction, he and Milton were no longer similarly situated because the law had changed during the pendency of Kelley’s appeal. Because we conclude Milton and Kelley were not similarly situated at the time of appellate review, the concurrence’s fairness argument is not persuasive.
[283]*283Moreover, any perceived unfairness in this case is the result of the law regarding retroactivity, not plain error. Under the Teague rule, “new rules” apply to (1) cases pending on direct appeal at the time of the new rule’s announcement, and (2) cases arising after the rule is announced. Chambers v. State, 831 N.W.2d 311, 323 (Minn.2013). Absent certain exceptions, “new rules” do not apply to defendants whose convictions were final at the time the new rule was announced. Id. Whenever a retroactivity line is drawn, an unfairness argument can be made. For example, suppose Milton and Kelley’s appeals arrived at our court on the same day and in the same procedural posture. The opinions in Milton and Kelley are issued a week apart, neither defendant petitions for certiorari, and consequently Milton’s conviction becomes “final” a week before Kelley’s conviction. If the United States Supreme Court announces a new rule during the intervening week, Kelley will receive the benefit of the new rule because, unlike Milton, Kelley’s conviction has not yet become final. In such an example, the only difference between Milton and Kelley is that we decided Milton’s case first. Under the concurrence’s argument, fairness would require us to deny Kelley the benefit of the new rule, even though his conviction had not yet become final. Followed to its logical conclusion, the concurrence’s fairness argument effectively modifies our retroactivity jurisprudence by limiting the application of a new rule to cases that arise after the new rule is announced. Neither our case law nor the principles underlying retroactivity support such a limitation.
The concurrence’s fairness argument restricts the relief available under Minn. R.Crim. P. 31.02 to errors that were plain at the time of trial. If we were to adopt the narrow interpretation of Rule 31.02 proposed by the concurrence, state defendants would receive less relief than their federal counterparts. See Henderson, — U.S. -, 133 S.Ct. 1121 (extending the relief available under the federal plain-error rule, Fed. R.Crim. P. 52(b), to errors that were not plain at the time of trial but had become plain by the time of review). Although we have on occasion provided greater protection or relief to criminal defendants, we have never provided less relief than that available in federal court. See State v. Ramey, 721 N.W.2d 294, 301-02 (Minn.2006) (shifting the burden of proving the impact of prosecutorial misconduct from the defendant to the State); State v. Borst, 278 Minn. 388, 397, 154 N.W.2d 888, 894-95 (1967) (extending right to court-appointed counsel to persons charged with misdemeanor offenses).
III.
Kelley next argues that the plain error affected his substantial rights. Specifically, Kelley argues that the evidence indicated S.A. did not know who actually took his property, and that a properly instructed jury could have concluded that Kelley’s unidentified friend, and not Kelley, actually robbed S.A.
To convict Kelley as an accomplice of first-degree aggravated robbery, the State had to prove beyond a reasonable doubt that Kelley (1) knew his friend was going to commit the robbery, and (2) intended his presence to further the commission of the crime. Milton, 821 N.W.2d at 806; State v. Mahkuk, 736 N.W.2d 675, 682 (Minn.2007). To establish that the erroneous accomplice liability jury instruction affected his substantial rights, Kelley has the heavy burden of proving that “there is a reasonable likelihood that giving the instruction in question had a significant effect on the jury verdict.” State v. Gomez, 721 N.W.2d 871, 880 (Minn.2006); accord Griller, 583 N.W.2d at 741. An [284]*284erroneous jury instruction will not ordinarily have a significant effect on the jury’s verdict if there is considerable evidence of the defendant’s guilt. See, e.g., Montanaro v. State, 802 N.W.2d 726, 733 (Minn.2011) (concluding that the self-defense jury instruction could not have had a significant effect on the jury’s verdict because “no reasonable jury could find [defendant’s] actions to be a reasonable use of force”); State v. Larson, 787 N.W.2d 592, 601 (Minn.2010) (holding that defendant’s substantial rights were not affected by allegedly erroneous accomplice liability jury instruction because there was “considerable evidence” of the defendant’s intent that the victim be murdered); Gomez, 721 N.W.2d at 881 (“Given the totality of the evidence, it seems unlikely that the jury would have reached a different verdict.”).
We conclude there is no reasonable likelihood that the erroneous jury instruction had a significant effect on the jury verdict because there is considerable evidence of Kelley’s guilt, and his defense did not focus on accomplice liability. Specifically, there is considerable evidence in the record that Kelley knew his friend was going to commit the robbery and intended his presence to further the commission of the crime. Kelley met with his friend for a half hour shortly before the incident, and then they went outside together and Kelley’s friend asked S.A. to come outside to meet with Kelley. The meeting beforehand and the friend’s request that S.A. meet with Kelley support the conclusion that Kelley and his friend planned the ensuing attack and robbery.
Moreover, S.A. testified that Kelley and his friend “were searching through my pockets ... then they rolled me over because they were going through my pockets ... and they were asking me ‘where’s your wallet.’ ” After Kelley and his friend had taken all S.A.’s belongings from his pockets, they continued kicking and hitting S.A. Kelley’s presence and active participation in punching and kicking S.A., helping roll S.A. over to gain access to his back pockets, and rummaging through his pockets is strong proof that Kelley intended his presence to further the commission of the robbery. See State v. Pierson, 530 N.W.2d 784, 788 (Minn.1995) (stating that “presence, companionship, and conduct before and after the offense are circumstances from which a person’s participation in the criminal intent may be inferred”).
Significantly, Kelley’s defense at trial did not focus on accomplice liability. Indeed, in his closing argument, Kelley’s defense counsel only briefly mentioned accomplice liability. Instead, Kelley focused his defense on the theory that S.A. had mistakenly identified him as one of the assailants, and he was not guilty as a principal. Kelley did not argue that he did not know the other person was going to commit the crime, or that he did not intend his presence to further the commission of the crime. Given Kelley’s strategy to focus on liability as a principal and not as an accomplice, the error in the accomplice liability jury instruction was not prejudicial. See State v. Davis, 820 N.W.2d 525, 538 (Minn.2012) (concluding that the defendant’s trial strategy impacted whether trial error was prejudicial).
In sum, Kelley has failed to satisfy the third prong of the plain-error doctrine, and therefore he is not entitled to a new trial.13 [285]*285We therefore affirm Kelley’s conviction, although on different grounds than the court of appeals.
Affirmed as modified.