MEYER, Justice
(concurring).
Although I concur with the outcome reached by the plurality, I disagree with the plurality’s formulation of the appellate review standard and analysis of the underlying claim. The plurality’s formulation of the standard for review of convictions based on circumstantial evidence is a far-reaching departure from our traditional rule, which calls for careful review of the rationality of the jury’s inferences from the evidence and their sufficiency to overcome the presumption of innocence and establish guilt beyond a reasonable doubt. Furthermore, I agree with the opinion of Justice Paul H. Anderson, who concludes that the plurality’s rule denies meaningful review.
I.
In reviewing claims of insufficient evidence, the trial court on a motion for judgment of acquittal, and the appellate court in reviewing the claim on appeal, apply the same basic standard of review. “Even the trial court, which has heard the testimony of witnesses firsthand, is not to weigh the evidence or assess the credibility of witnesses when it judges the merits of a motion for acquittal.” Burks v. United States, 437 U.S. 1, 16, 98 S.Ct. 2141, 57 L.Ed.2d 1 (1978). The trial court “is to submit a case to the jury if the evidence and inferences therefrom most favorable to the prosecution would warrant the jury’s finding the defendant guilty beyond a reasonable doubt.” Id. “Obviously a federal appellate court applies no higher a standard; rather, it must sustain the verdict if there is substantial evidence, viewed in the light most favorable to the Government, to uphold the jury’s decision.” Id. at 17, 98 S.Ct. 2141. As Tibbs v. Florida explained, there is a difference between a conviction set aside based on the weight of the evidence and a conviction set aside based on insufficient evidence. 457 U.S. 31, 42-43, 102 S.Ct. 2211, 72 L.Ed.2d 652 (1982). The reviewing court’s finding that the evidence was “legally insufficient” to support conviction “ ‘means that the government’s case was so lacking that it should not have even been submitted to the jury.’ ” Id. at 41, 102 S.Ct. 2211 (quoting Burks, 437 U.S. at 16, 98 S.Ct. 2141) (emphasis in original).
According to the plurality’s analysis, however, a somewhat different standard applies on appeal: first, the appellate court identifies the facts implicit in the verdict, disregards conflicting evidence, and then determines the inferences from the facts implicit in the verdict. The plurality maintains this rule is based on an analytical framework established in our cases. But I read our cases as demonstrating that although we accept the State’s evidence, we make an independent evaluation of the inferences derived from the circumstantial evidence, assessing rationality and whether they established guilt beyond a reasonable doubt. E.g., State v. McArthur, 730 N.W.2d 44, 50 (Minn.2007) (affirming murder conviction following independent evaluation of circumstantial evidence of premeditation); State v. Jones, 516 N.W.2d 545, 549 (Minn.1994) (reversing assault conviction following independent evaluation of circumstantial evidence); State v. Loss, 295 Minn. 271, 281, 204 N.W.2d 404, 409-10 (1973) (affirming manslaughter conviction following independent evaluation of circumstantial evidence).
Cases cited by the plurality in support of its formulation of the rule turned on credibility. E.g., State v. Colbert, 716 N.W.2d 647, 654 (Minn.2006) (concluding that “the jury was free to, and evidently did, reject” defendant’s claim that someone else had the murder weapon); State v. Gorman, 219 Minn. 162, 168-69, 17 N.W.2d 42, 45-46 (1944) (concluding that credibility of expert witnesses and weight of their testimony were for the jury); State v. Johnson, 173 Minn. 543, 544-45, 217 N.W. 683, 683-84 (1928) (concluding that credibility of alibi witnesses and weight of their testimony were for the jury).
Other cited cases might have had a different result under the plurality’s analysis. For example, in State v. Webb, 440 N.W.2d 426, 427 (Minn.1989), defendant Webb was found guilty of first-degree murder. Webb, who suffered from mental illness, was seen talking to the victim at a drop-in center on Easter Sunday. Id. The victim’s body was found the next day about a block away from Webb’s apartment. Id. at 427-28. Nearby was a plastic bag containing a bedspread. Id. During an interrogation, Webb was shown a photograph of the bedspread and repeatedly asked if it belonged to him. Id. at 428. Webb said he did not know. Id. When one of the officers feigned anger, Webb admitted ownership. Id. However, immediately after that officer left the room, Webb told the other officer that he admitted ownership “because that is what the other officer wanted him to say.” Id. at 428-29. The bedspread was identical to some 70-100 other bedspreads used by a hotel in Duluth. Id. at 429. One witness testified he saw a similar bedspread in Webb’s apartment. Id. Other witnesses said that Webb did not have a bedspread. Id.
We concluded the State’s case was “thin in several respects,” including ownership of the bedspread, and reversed the conviction. Id. at 429-31. In so doing, we noted:
No physical evidence was discovered linking the victim with the appellant or his apartment despite a painstaking investigation for such evidence. No hair, fibers, blood or body fluids were found on the victim, or on the bedspread or sheet to indicate that they had ever been in appellant’s apartment. Similarly, no paint flakes were found on the bedspread which would indicate that it was used to drag the victim out of the apartment, across the painted wood porch floor.
Id. at 431. The plurality’s reformulated rule, however, would have resolved ownership of the bedspread for the State as implicit in the verdict and disregarded the other evidence in conflict with the verdict.
II.
Typically, in assessing the sufficiency of the evidence, we review the evidence to determine “whether the facts in the record and the legitimate inferences drawn from them would permit the jury to reasonably conclude that the defendant was guilty beyond a reasonable doubt of the offense of which he was convicted.” State v. Moore, 481 N.W.2d 355, 360 (Minn.1992). But when a conviction is based on circumstantial evidence, “that evidence must be consistent with the hypothesis that the accused is guilty and inconsistent with any other rational hypothesis except that of guilt.” McArthur, 730 N.W.2d at 49 (internal quotation marks omitted).
“ ‘Direct evidence’ is that which proves a fact without an inference or presumption and which in itself, if true, establishes that fact.” 1 Barbara E. Bergman & Nancy Hollander, Wharton’s Criminal Evidence § 1:8 (15th ed. 1997) (internal quotation marks omitted). “Circumstantial evidence is evidence from which the factfinder can infer whether facts in dispute existed or did not exist. The factfinder is permitted to draw this inference if a reasonable relationship exists between the known facts and circumstances and the facts sought to be proved.” Id.
As explained in my concurrence in Tscheu, the rational hypothesis standard for evaluating circumstantial evidence was derived from Commonwealth v. Webster, 59 Mass. (5 Cush.) 295 (1850), a murder prosecution based entirely on circumstantial evidence. State v. Tscheu, 758 N.W.2d 849, 869-70 (Minn.2008) (Meyer, J., concurring). In Webster, Chief Justice Shaw described the advantages and disadvantages of each mode of proof, cautioned against the use of circumstantial evidence in criminal cases and set out the following rule:
[T]he circumstances taken together should be of a conclusive nature and tendency, leading on the whole to a satisfactory conclusion, and producing in effect a reasonable and moral certainty, that the accused, and no one else, committed the offense charged. It is not sufficient that they create a probability, though a strong one; and if, therefore, assuming all the facts to be true which the evidence tends to establish, they may yet be accounted for upon any hypothesis which does not include the guilt of the accused, the proof fails. It is essential, therefore, that the circumstances taken as a whole, and giving them their reasonable and just weight, and no more, should to a moral certainty exclude every other hypothesis.
59 Mass. (5 Cush.) at 319. Courts throughout the country, including Minnesota, adopted the Webster rule as a jury instruction for assessing circumstantial evidence. State v. Staley, 14 Minn. 105, 121-22, 14 Gil. 75, 90 (1869); Irene Merker Rosenberg & Yale L. Rosenberg, “Perhaps What Ye Say is Based Only on Conjecture” — Circumstantial Evidence, Then and Now, 31 Hous. L. Rev. 1371, 1392 (1995). Some 100 years later, the Supreme Court held that a reasonable doubt instruction obviated the need for a special circumstantial evidence instruction. Holland v. United States, 348 U.S. 121, 139-40, 75 S.Ct. 127, 99 L.Ed. 150 (1954). Although we eventually abandoned the special jury instruction, State v. Turnipseed, 297 N.W.2d 308, 313 (Minn.1980), we retained the traditional rational hypothesis review standard. E.g., McArthur, 730 N.W.2d at 49.
Holland involved a prosecution for willful tax evasion based on a “net worth” theory, a method of proof involving “something more than the ordinary use of circumstantial evidence in the usual criminal case.” 348 U.S. at 124, 75 S.Ct. 127. Although the Court permitted net worth prosecutions, it emphasized the “dangers that must be consciously kept in mind in order to assure adequate appraisal of the specific facts in individual cases.” Id. at 127, 75 S.Ct. 127. The Court cautioned appellate courts to “review the cases, bearing constantly in mind the difficulties that arise when circumstantial evidence as to guilt is the chief weapon of a method that is itself only an approximation.” Id. at 129, 75 S.Ct. 127. The Court then made an exhaustive examination of the government’s evidence, including the government’s negation of the defendants’ claims inconsistent with guilt, detailed investigations of leads that if true would have established innocence and evidence of a consistent pattern of underreporting to establish willfulness. Id. at 133-37, 75 S.Ct. 127. Notably, the Court did not determine, at the outset, facts implicit in the verdict, disregard conflicting evidence and then determine the inferences from the facts implicit in the verdict. Although the Court rejected a special circumstantial evidence instruction, it basically employed a rational hypothesis analysis. See id. at 135, 75 S.Ct. 127 (“When the Government rests its case solely on the approximations and circumstantial inferences of a net worth computation, the cogency of its proof depends on its effective negation of reasonable explanations by the taxpayer inconsistent with guilt.”).
III.
“Circumstantial evidence must form a complete chain that, in view of the evidence as a whole, leads so directly to the guilt of the defendant as to exclude beyond a reasonable doubt any reasonable inference other than guilt.” State v. Taylor, 650 N.W.2d 190, 206 (Minn.2002). Even with this strict standard, we have recognized that the jury is in the best position to determine credibility and weigh the evidence. Moore, 481 N.W.2d at 360. In applying these rules, I would analyze this case as follows.
The evidence was that on June 2, 2005, at 4:01 a.m., Mound police responded to a burglary at a home north of Lynwood Boulevard and established a perimeter around the area. At 4:17 a.m., police were dispatched to a second burglary in the vicinity of the first burglary. The police adjusted the perimeter, extending it west to Westedge Boulevard and east to Langdon Lane, and called for the emergency response unit, a small SWAT team. Around 4:30 a.m., the New Hope K-9 patrol started tracking from the site of the second burglary south through wooded areas down to the railroad bed and west along the north shore of Langdon Lake. The dog at some point went off track, but continued leading the officers south on Westedge where he seemed to be picking up the trail again.
At 5:15 a.m., a third burglary call came in from a home less than two blocks from the K-9 patrol on Westedge. The victim had ripped the shirt off the intruder and chased after him as he was fleeing her home. She could see his bare back, shorts and boxers visible above the shorts. She secured the doors to her home, looked outside to see if he was gone, and saw him coming back. As he was trying to reenter her home, the victim was screaming at him, telling him the police would be there soon. The intruder left, running east toward the neighbor’s house.
The New Hope K-9 patrol was near the area of the third burglary when that call came in, but because the officers had no idea where the home was located, they let the dog lead. They “trailed” east on Beachwood Road to where they picked up a trail in a yard covered with cottonwood seeds. It was obvious by sight that someone had run through the yard. The dog tracked north through the yard and down to the shoreline. The dog wanted to go out into the water, and the handler’s partner thought he saw a splash and someone swimming in the lake. By this time Trooper 7, a helicopter patrol, had arrived and was asked to do a perimeter of the lake, but nothing “positive” came of that search. The New Hope K-9 patrol stopped tracking because the dog was tired.
At 6:10 a.m., a Mound resident saw a shirtless young man in the area of the first burglary, north of the lake. He was wearing jeans “hanging kind of low,” with plaid boxers visible above the shorts, and a baseball cap on backwards. He was wet “only” or “maybe” up to his thighs. He walked north to a neighbor’s house and then turned around, heading back toward the trees. The resident noted there was a helicopter patrol “flying around and he wasn’t leaving.”
At around 6:30 a.m., a shirtless young man was seen running between houses on Sunnybrook Circle, about three-quarters of a mile north of the first suspicious-person sighting. He was wearing loose pants with darker underwear visible, and a baseball cap on backwards. He crossed over a small bridge and continued north in the direction of the Westonka public schools. Sometime after 7:20 a.m., the Minnetonka K-9 patrol tracked from the small bridge to the elementary school and continued north toward the high school. The officers stopped tracking for safety reasons, given the time of day and number of students around.
Meanwhile, at about 7:00 a.m., a Minne-trista resident saw a shirtless young man running across the yard, east-to-west toward Sunnyfield Road. A bank of trees blocked the resident’s view to the north, but because the young man did not turn south, the resident presumed he turned north.
After the tracking concluded, the police conducted a grid search, during which they encountered J.B. J.B. and his brother M.B. lived on Lynwood, in the area of the first two burglaries. J.B. had heard helicopters “buzzing” over his house, turned on the news, and walked outside. He told the police he had some friends over the night before and provided their names, including appellant’s. Because the K-9 trackings and suspicious-person sightings were heading in the direction of appellant’s home, his name was “significant.” The police located appellant at his workplace. He had scratches on his arms and neck and was taken into custody. Photos taken at the police department that day showed appellant also had scratches on his chest, back, and legs.
There was evidence that appellant was living with his brothers in a house on Sunnyfield Road. He slept on the living room couch. On June 1, appellant got a ride to a gathering at J.B. and M.B.’s home. En route, at 10:08 p.m., appellant completed an ATM transaction. Around 3:30 a.m., appellant left the gathering in J.K’s truck. J.K. dropped appellant off near Lynwood and Langdon Lane, drove south to his home, and went to bed.
Other evidence included testimony that appellant was not sleeping on the couch when his brothers left for work, around 6:15-6:30 a.m. on June 2; that his cell phone records showed consistent usage throughout the evening of June 1 and morning of June 2, except for a break between a call placed at 3:17 a.m. and 6:20 a.m., when no calls were placed; and that the 6:20 a.m. call was to appellant’s own cell phone. There was an ATM surveillance video showing appellant wearing a baseball cap, blue jeans, and a black T-shirt with a white tag at the bottom and a pocket; testimony that the T-shirt seen in the video matched the Dickies T-shirt recovered from the third victim; and testimony that the DNA profile generated from the T-shirt was consistent with a mixture and that the predominant contributor matched the profile from the biological sample provided by appellant.
The reasonable inferences that I believe may be drawn from this evidence are that the intruder was on foot; that he fled shirtless from the third home; that he ran to the southeast corner of the lake; that he traveled north around the lake and through the perimeter set up by the police; that he continued on a course toward Sun-nyfield Road and then turned north. Appellant was on foot in the area at the time of the first two burglaries; he was wearing a black Dickies T-shirt that matched the shirt obtained in the third burglary; and his DNA was on that shirt. In addition, appellant lived a few houses north of the last suspicious-person sighting, and he had scratches on his arms, chest, back, and legs, consistent with moving shirtless through brush and wooded areas in an effort to elude the manhunt.
Appellant contends that where the “burglar escaped by swimming in Langdon Lake,” his cell phone would have been “ruined”; yet appellant’s cell phone was still working even when he was in jail, where he made cell phone calls until the police “took it away from him.” Although the New Hope K-9 partner thought he saw someone swimming in the lake, the helicopter patrol’s contemporaneous surveillance of the lake indicated otherwise; and the testimony was that the suspicious person moving through the police perimeter was wet “only” or “maybe” up to his thighs.
As to appellant’s alternative perpetrator evidence — that J.K., J.B., or M.B. committed the burglaries — I would conclude that has more to do with the credibility of witnesses than the circumstantial evidence rule. J.K. testified that he drove home and went to bed. J.B. was excluded by DNA evidence and testified that he was also home in bed. M.B., unaware that he had been called by the defense to implicate himself as a suspect, testified that he did not commit the burglaries. The credibility of these witnesses and the weight of their testimony were for the jury.
Nevertheless, giving deference to the jury’s evaluation of witness credibility and the weight of their testimony does not, in my view, absolve us from taking seriously our obligation to scrutinize the inculpatory circumstantial evidence. The ultimate determination of guilt is based also on inferences from the evidence, but the jury is given no guidance by way of instruction in the analysis of this evidence. As I said in Tscheu, our rational hypothesis appellate review standard allows us “ ‘to oversee whether the jury’s inferences from circumstantial evidence were in fact rational and sufficient to establish proof beyond a reasonable doubt.’ ” 758 N.W.2d at 871 (Minn.2008) (Meyer, J., concurring) (quoting Rosenberg & Rosenberg, supra, at 1415). I would conclude that the inferences drawn from the evidence in appellant’s case were rational and formed a complete chain, in view of the evidence as a whole, leading so directly to the guilt of appellant as to exclude beyond a reasonable doubt any reasonable inference other than guilt.
IV.
The circumstantial evidence rule “is founded upon the presumption of innocence and its demand that guilt be proved beyond a reasonable doubt.” State v. Kaster, 211 Minn. 119, 121, 300 N.W. 897, 899 (1941). The reasonable-doubt standard “is a prime instrument for reducing the risk of convictions resting on factual error.” In re Winship, 397 U.S. 358, 363, 90 S.Ct. 1068, 25 L.Ed.2d 368 (1970). But neither the reasonable doubt jury instruction nor the basic appellate review standard “comes to grips with the separate issue addressed by use of the reasonable hypothesis of innocence test at the trial and appellate levels, namely, the proper evaluation of circumstantial evidence.” Rosenberg & Rosenberg, supra, at 1420. Our traditional rational hypothesis review standard provides added assurance against wrongful convictions. I join in the opinion of Justice Paul H. Anderson because I agree the plurality’s standard denies meaningful appellate review.
. By plurality, I am referring to the opinion authored by Chief Justice Magnuson, joined in by Justice G. Barry Anderson and Justice Gildea.
. See 2k Charles Alan Wright and Peter J. Henning, Federal Practice and Procedure § 467 (4th ed. 2009); 6 Wayne R. LaFave et al., Criminal Procedure § 24.6(c) (3d ed. 2007); see also 23 C.J.S. Criminal Law § 1504 (2006 & Supp. 2009) (noting jurisdictions that apply a special rule for cases based only on circumstantial evidence).
. In the event the plurality’s view proves correct, meaning that we have abandoned our traditional appellate review standard, then I would reassert my plea that "we should revisit Turnipseed to consider cautionary instructions so that the jury will be advised that circumstantial evidence must exclude every rational hypothesis except that of guilt.” Tscheu, 758 N.W.2d at 871 (Meyer, J., concurring).