OPINION

delivered the opinion of the Court

We granted the State’s petition for discretionary review in this cause to address the question whether the court of appeals can reverse a conviction on the basis of factual insufficiency of the evidence even when it believes that the evidence preponderates in favor of guilt, consistent with our opinion in Zuniga v. State.1 Moreover, we granted discretionary review on our own motion to address the further question whether we should undertake further clarification of the Clewis2 standard for factual-sufficiency review as articulated in Zuniga. We now recognize that the Zuniga re-articulation of the Clewis standard represents an inadvertent distortion of factual-sufficiency review as historically conducted by this Court and its predecessors, and to that extent we overrule it.

I. FACTS AND PROCEDURAL POSTURE

The appellant was prosecuted for the burglary of a habitation during which he allegedly attempted to commit sexual assault. The jury convicted him of this offense and assessed his punishment at thirteen years’ confinement in the penitentiary. On appeal he challenged both the legal and factual sufficiency of the evidence. The Tenth Court of Appeals found the evidence legally sufficient.3 However, finding the evidence factually insufficient, the court of appeals reversed the conviction and remanded the cause for a new trial.4 On the authority of our opinion in Zuniga, the court of appeals held that, even though the evidence preponderated in favor of finding an intent to commit sexual assault, it was nevertheless factually insufficient to support a finding of such an intent to a level of confidence beyond a reasonable doubt.5 We granted the State’s petition for discretionary review to determine whether Zuniga does *406indeed authorize reversal when the legally sufficient evidence supporting the verdict “equals or outweighs the contrary evidence[.]” On our own motion, we expanded review to include the question whether we should re-examine Zuniga. In order to gain a proper perspective on these issues, we think it useful to review the origins and history of factual-sufficiency review in Texas.

II. FACTUAL-SUFFICIENCY REVIEW BEFORE CLEWIS

The authority of criminal appellate courts in Texas to reverse criminal convictions on the basis of factually insufficient evidence did not begin with our opinion in Clewis. Criminal appellate courts in Texas have always had, within their constitutional “appellate jurisdiction,” the authority to review criminal convictions for what we now call “factual sufficiency” of the evidence. This includes the authority to reverse cases we would now regard as supported by “legally sufficient” evidence when the verdict nevertheless seems to be against the great weight and preponderance of the evidence. That authority has been codified in all former incarnations of what is now Article 44.25 of the Code of Criminal Procedure, authorizing criminal appellate courts to “reverse” (nee “revise”) criminal judgments “as well upon the law as upon the facts[.]”6 Clewis did not invent factual sufficiency. Instead, the Court in Clewis, with an assist from our earlier opinions in Meraz v. State,7 and Bigby v. State,8 rediscovered factual sufficiency, and resurrected it from a brief, court-imposed dormancy.

A. Bailey, Smith and Bishop

The authority of criminal appellate courts in Texas to review factual sufficiency derives from the Texas Constitution. Both Bigby, and later Clewis, cite a trio of very old Texas Supreme Court cases for the proposition that criminal appellate courts in Texas have the authority, within their constitutionally endowed “appellate jurisdiction,” to conduct a factual review. None of these cases, Bailey v. Haddy,9 Republic of Texas v. Smith,10 and Bishop v. State,11 purports to hold that such a factual review embraces what we now call a “factual sufficiency” analysis. But each confirms that, first under the 1836 Constitution of the Republic, and then under succeeding state constitutions, the Supreme Court’s appellate jurisdiction over criminal cases included the power to revise the case “upon the law as upon the facts.”12 This is precisely how a majority of this Court would come to construe these cases in Bigby13

*407None of these earliest cases expressly declared that factual-sufficiency review, per se, was authorized in criminal cases. But later cases construed the criminal appellate “fact” jurisdiction in Texas, first recognized in these earliest cases, to include what we would now call factual-sufficiency review. Moreover, Old Code Article 744, and all later iterations of that provision up until the 1981 amendment to present-day Article 44.25,14 expressly contemplated that criminal appellate courts had the power to “reverse[ ] for the reason that the verdict is contrary to the weight of the evidence[.]”

B. Tollett and Loza

Among the very earliest cases to address sufficiency of the evidence were Tol-lett v. State15 and Loza v. State.16 In each of these cases, the respective courts conducted an analysis that, to the contemporary ear, sounds much like what we would now call a “legal sufficiency” analysis. Neither Tollett nor Loza explicitly holds that the factual jurisdiction of criminal appellate courts in Texas includes the power to conduct a factual-sufficiency review, per se. But they do not rule it out either, and, like the earliest Texas Supreme Court cases, they do confirm the general factual jurisdiction of criminal appellate courts. Nevertheless, other cases from this period from the Texas Court of Appeals, the direct predecessor of this Court, demonstrate that the Court well “understood the nature and extent of its power to review jury verdicts for evidentiary sufficiency, albeit not then labelled ‘legal sufficiency’ and ‘factual sufficiency.’ ”17 Some of those cases utilize language that sounds distinctly as though it belongs in the latter category.18

C. Walker

In Walker v. State, the Texas Court of Appeals (the direct predecessor of this Court) undertook a detailed analysis of the prior case law touching on the question of evidentiary sufficiency.19 From that analysis the Court distilled the following “rules of practice”:

From a careful consideration of the cases in which this subject has been discussed, we deduce the following rules of practice governing this court, viz:

*408First. Where the evidence is conflicting, and there is sufficient, if believed, to prove the case of the State, the jury-being the exclusive judges of the credibility of the testimony, their verdict will not be set aside unless it dearly appears to be wrong.

Second. Where there is no testimony to support it, the verdict will be set aside.

Third. Where the evidence is insufficient to rebut the presumption of innocence, the verdict will be set aside.

Fourth. Where the verdict is contrary to the weight of the evidence, it will be set aside.20

The discussion that the Texas Court of Appeals undertook in Walker just before it set out these “rules of practice” seems manifestly to contemplate multiple and variable standards of sufficiency review, not variations on a single standard. The Walker rules even include a standard whereby the Court was authorized to reverse a conviction “though there was evidence sufficient to support it.”21 In its entirety, omitting string citations, the Walker discussion reads as follows:

With reference to the authority of this court to set aside a verdict when that verdict is, in our judgment, against the weight of the evidence, or not supported by it, we think the statute confers it. Article 870 of the [1879] Code of Criminal Procedure provides: “The Court of Appeals may revise the judgment in a criminal action, as well upon the law as upon the facts; but when a cause is reversed for the reason that the verdict is contrary to the weight of evidence, the same shall, in all cases, be remanded for a new trial.”

With reference to trials by jury it is provided: “The jury are the exclusive judges of the facts in every criminal cause.” ([1879] Code Crim. Proc., Art. 676) And again it is provided: “The jury in all cases are the exclusive judges of the facts proved, and of the weight to be given to the testimony, except,” etc. ( [1879] Code Crim. Proc., Art. 728.)

While Article 870, above quoted, expressly confers the authority to revise the facts, and to reverse the judgment for the reason that the verdict is contrary to the weight of evidence, it has been the general practice of this court to refuse to set aside a verdict where the evidence was conflicting, but where there was sufficient, if believed, to support the finding.

But even in such case, where it was manifest that the verdict was wrong, and it was clear that injustice had been done the defendant, it has been set aside, though there was evidence sufficient to support it.

And it has never been doubted, but has always been considered by this court, not only that it had the authority, but that it was its duty to set aside a verdict where that verdict was contrary to the evidence, or unsupported by it, though it is with reluctance that the court will disturb a verdict where there is any evidence to sustain it.22

It could not be more clear from this discussion that the Texas Court of Appeals deemed itself authorized to reverse convictions under a number of distinct scenarios, *409and that the “rules of practice” that immediately follow are not meant to be understood as variations on a single insufficiency theme, but as, indeed, separate “rules.”

It is also imperative to note the Texas Court of Appeals’s reference to the predecessors of present-day Articles 36.18 and 38.04 of the Code of Criminal Procedure.23 These are the provisions that reserve the fact-finding function to the jury. That the Texas Court of Appeals would juxtapose a reference to these statutory provisions with its discussion of its own authority to reverse a conviction that is against the weight of the evidence, or manifestly wrong or unjust, notwithstanding “there was evidence sufficient to support it,” demonstrates that the statutory provisions have peacefully co-existed with that appellate authority for at least a hundred and twenty-three years. From their placement in Chapters 36 and 38 of the Code, which govern trial practice and procedure, it is apparent that Articles 36.13 and 38.04 were meant merely to allocate the fact-finding function at the trial level and do not purport to affect appellate review.24 It is true that they have often been cited for the proposition that jury verdicts should not ordinarily be disturbed when supported by competent evidence. But criminal appellate courts in Texas have never interpreted them to wholly prohibit the limited factual review that is otherwise inherent in their appellate jurisdiction, and authorized by Article 44.25 and its predecessors.

After Walker, in cases of conflicting testimony or evidence, the Texas Court of Appeals continued to measure sufficiency of the evidence by inquiring, inter alia, whether the verdict was against the great weight of the evidence.25 In several of those cases the Court reversed the convictions, expressly finding the evidence to be, indeed, against the weight of the evidence.26

D. Factual-Sufficiency in the Texas Court of Criminal Appeals

When the Texas Court of Appeals became the Texas Court of Criminal Appeals in 1891, we continued to exercise our constitutionally endowed factual review authority to review cases for factual suffi*410ciency. During its very first term, this Court reversed a conviction for theft of a horse on the basis of factual insufficiency of the evidence, in a cased styled Fore-sythe v. State.27 There, a certain witness, one Ellard, testified that on the day of the theft he had observed Foresythe in the vicinity of the pasture from which the horse was stolen. The horse was later found in Foresythe’s possession. Other witnesses placed Foresythe in a different county on the day of the theft, and it was shown that he had bought the horse from a man named Wats, with a bill of sale and corroborating testimony introduced to that effect. This Court reversed, observing:

We believe that the verdict of the jury is against the great weight of the testimony; that every apparent inculpatory fact has been met by the testimony for the defendant and made to consist with his innocence, except Ellard’s testimony that he saw defendant in the pasture. This we believe to be a mistake, and a very reasonable and common one.28

Nothing about this reversal is consistent with what we today regard as a “legal sufficiency” analysis. It was a reversal on factual insufficiency, pure and simple.

Again, in 1912, we reversed a conviction for the crime of seduction on account of insufficiency, in Murphy v. State.29 Although the complainant, a teenage girl, testified that Murphy promised to marry her before she submitted to have intercourse with him, the corroborating testimony placed any promise of marriage at a time after the intercourse occurred. For his part, Murphy denied the sexual intercourse even occurred, and the circumstances tended to bear out his denial. We observed that:

While the jury is the judge of the credibility of the witnesses and the weight to be given the testimony, yet we are authorized to look to see if there is any testimony, or that it is so slight as not to authorize a finding therein against the great preponderance of the testimony.30

Although it is arguable that the case could have been reversed on legal insufficiency, ie., the failure of proof of corroboration statutorily required to convict, the Court nevertheless clearly resorted to the language of factual sufficiency.

In both Rochetszky v. State,31 and Green v. State,32 the prosecution presented eyewitness testimony of dubious reliability. In the former, a prosecution for sale of intoxicating liquor, the defendant “presented unusually strong testimony of alibi.”33 In the latter, a rape case, the complainant had made no outcry until it was discovered that she was pregnant, and her account struck the Court as, “to say the least, slightly out of line.”34 While reversing both cases, we did not resort specifically to language with respect to the weight of the evidence. But both cases invoked then-Article 939 of the Code of Criminal Procedure (now Article 44.25) as authority for reversing a conviction “as well upon the law as upon the facts.” Indeed, in Green we expressly acknowledged that, under the statute, “this court may reverse a judgment upon the facts; we hesitate to interfere with the verdict of the jury in any case, but it is our duty to do so when there *411is not sufficient testimony to support it.”35 By the semicolon in this quoted passage, it is obvious that the type of sufficiency review we were referring to was a factual-sufficiency review, not a legal sufficiency review. And in Rochetszky, we expressly held that a reversal was necessary to avoid an injustice, “where the verdict was manifestly wrong.”36 Of course, this is classic language of factual-sufficiency review, of the type that Walker would classify as the third of its “rules of practice.” Both opinions survived stout assertions on rehearing that the Court had erred to invade the jury’s province by deciding the cases “upon questions of fact[.]”37

In Villareal v. State,38 the defendant was convicted of murder and sentenced to death. The evidence was circumstantial as to exactly how the homicide occurred, the witnesses having only heard the offense, but not seen it. The defendant testified that he killed the deceased in self defense, while the deceased was on top of him and choking him. A physician testified for the State that the fatal wound had come from behind the deceased, thus contradicting the defendant’s account, and supporting the State’s case of murder with malice, a capital offense. Nevertheless, this Court reversed the conviction. Invoking Article 848 of the 1925 Code of Criminal Procedure,39 now Article 44.25, we observed:

[W]e are not to be understood as saying that the evidence may not be strong enough to support a conviction for murder, but that the mind of this court upon an examination of this record is left in such a condition that we are desirous that the evidence should be passed upon again by another jury.

[[Image here]]

In reaching our conclusion in this case we have not overlooked the general rule, frequently announced, that the credibility of witnesses and the weight to be given their testimony exclusively rests with the jury and the lower court. Further, that this court will not pass on the sufficiency of the evidence unless there is an entire failure of proof; and that the amount of punishment is the jury’s special province. However, we do find that this court has upon occasion applied such rules with a degree of laxity where, under the facts of such cases, it appeared that the conviction was unjust or that the punishment was harsh or unreasonable.40

It is hard to imagine a more starkly candid exercise of criminal appellate factual juris*412diction than this. While it may be the case, as has been suggested, that factual sufficiency went into hiding in the late Forties and early Fifties,41 that does not mean that it never existed.

E. The Ascendancy of Legal Sufficiency

In 1979, the United States Supreme Court recognized for the first time, in Jackson v. Virginia,42 a minimum federal due process standard for sufficiency of the evidence below which a state court conviction cannot be sustained.43 The Supreme Court also held that this legal standard for evidentiary sufficiency could be vindicated by the federal courts in federal habeas corpus proceedings.44 But while Jackson v. Virginia does impose upon the states a constitutionally minimum legal sufficiency standard,45 it does not (and could not, consistent with principles of federalism) prevent the states from applying sufficiency standards that are more solicitous of defendants’ rights.46 Thus, nothing about Jackson impacts the legitimacy or fate of factual-sufficiency review in criminal cases in Texas.

Nonetheless, within six months after Jackson was decided, this Court announced in White v. State,47 apparently for the first time, that we lack factual-review authority, and therefore “we do not have jurisdiction to pass upon the great weight and preponderance of the evidence^]”48 The Court predicated this holding on the so-called “factual eonclusivity clause” in Article V, Section 6 of the Texas Constitution.49 That provision makes the resolution of factual issues by the intermediate courts of appeals conclusive upon discretionary review courts in Texas, such as this Court and the Texas Supreme Court.50 What we failed to take into account in White was the fact that, in 1979, this Court was still exclusively a direct appellate court, not a discretionary review court.51 For that reason we were mistaken to think that our own factual appellate jurisdiction could be circumscribed by the factual con-clusivity clause.

Three years later we mistakenly relied upon White to support the observation, in Combs v. State,52 that we lack the authority to make “determinations of sufficiency of the evidence ... passing upon the weight and preponderance of the evidence.” 53 Following that observation, we immediately dropped a footnote in which we globally (and equally mistakenly) de-*413dared: “We perceive no other standard may be utilized by the Court of Appeals in reviewing criminal convictions other than sufficiency of the evidence [by which is evidently meant legal sufficiency] to support the conviction.”54 We then identified the applicable legal sufficiency standard to be that which the Supreme Court had just articulated in Jackson v. Virginia.55 Thus, in the course of three short years we stripped both ourselves (in our capacity as a direct appellate court in capital cases) and the intermediate courts of appeals of the fact jurisdiction we had always previously considered to be inherent in the respective courts’ “appellate jurisdiction,” as conferred by Article V, Sections 5 and 6 of the Texas Constitution.56

F. Factual Sufficiency Makes a Comeback

In 1994, in Bigby, we expressly overruled White, and reinvigorated our own appellate jurisdiction to review facts and make factual-sufficiency determinations.57 The question in Bigby was whether this Court had authority to conduct a factual review of a jury’s determination of an affirmative defense. In addressing this question, we corrected the mistakes of White and Combs to hold that criminal appellate courts in Texas have always had factual review jurisdiction, conferred by both the Texas Constitution and by Article 44.25 and its predecessors.58 The reality is that our factual review jurisdiction had never disappeared. It was merely temporarily obscured by White and Combs. Neither of those opinions ever disputed (or even referenced) our inherent constitutional fact jurisdiction, or purported to overrule the plethora of previous cases recognizing our constitutional and statutory authority to conduct fact reviews.59

Clewis simply continued the restoration of appellate factual jurisdiction in *414criminal cases by expressly holding that the courts of appeals may review the evidence for factual sufficiency as well as legal sufficiency — effectively overruling the footnote in Combs, albeit sub silentio.60 And we finished the job in Jones v. State,61 when we extended our own fact jurisdiction to permit a review of the factual sufficiency of the evidence to establish the guilt of the accused in capital murder cases, and not just to review whether the jury’s rejection of an affirmative defense was against the great weight and preponderance of the evidence, as in Bigby. Thus, the law as it stands today in Texas is that, in the direct appeal of a criminal case, the courts of appeals (and this Court, in the direct appeal of capital murder cases) have the statutory and constitutional authority to entertain a claim of factual insufficiency and to reverse the conviction and remand the cause for a new trial in the event they find the evidence to be, indeed, factually insufficient.

III. FACTUAL-SUFFICIENCY REVIEW SINCE CLEWIS

The basic ground rules for post-Clewis factual-sufficiency review were well articulated in Cain v. State.62 First, the appellate court should be mindful that a jury has already passed on the facts, and convicted, and that the court should never order a new trial simply because it disagrees with the verdict, but only where it seems to the court to represent a manifest injustice, though supported by legally sufficient evidence.63 Second, the appellate court should support its judgment that a manifest injustice has occurred by explaining in exactly what way the State’s evidence, while legally sufficient, is nevertheless too weak to withstand scrutiny, or in exactly what way it perceives the conflicting evidence greatly to preponderate against conviction.64 Third, although viewing all the evidence, as it would in a legal sufficiency analysis, the appellate court should review that evidence, not in the light most favorable to the verdict, but in a neutral light.65

The only substantive adjustments to these basic ground rules have come in two cases, Johnson v. State,66 and Zuniga v. State.67 In Johnson, the Court broke down factual-sufficiency analysis into two prongs. The first prong asks whether the evidence introduced to support the verdict, though legally sufficient, is nevertheless “so weak” that the jury’s verdict seems *415“clearly wrong and manifestly unjust[.]”68 The second prong asks whether, considering conflicting evidence, the jury’s verdict, though legally sufficient, is nevertheless against the great weight and preponderance of the evidence.69 This adjustment simply brings the Clewis standard in line with civil practice, and with factual-sufficiency review in criminal cases from the earliest days in Texas. After all, the Johnson adjustment, in its essentials, merely reflects the first and fourth “rules of practice” set out more than a hundred and twenty years ago in Walker.

IV. ZUNIGA

The far more radical adjustment was made in Zuniga. Zuniga represents an attempt, not to alter the factual-sufficiency standard of previous cases, but to re-articulate it in terms specific to the particular level of confidence that due process requires in criminal cases, viz: proof beyond a reasonable doubt. Our express intention was to eschew any formulation of the standard that would incorporate language of “preponderance of the evidence,” since that is a lesser burden of proof, not applicable in criminal cases. Thus we said:

There is only one question to be answered in a factual sufficiency review: Considering all of the evidence in a neutral light, was a jury rationally justified in finding guilt beyond a reasonable doubt?70

Thus articulated, the standard for factual sufficiency seems barely distinguishable from a Jackson v. Virginia legal sufficiency analysis. The only apparent difference is that the appellate court, using the Zuni-ga standard, is to measure the rationality of the jury’s verdict by viewing all of the evidence from a “neutral” standpoint, rather than “in the light most favorable to the verdict.” That is to say, it need not indulge the presumption that is built-in to a legal sufficiency analysis that the jury, “faced with a record of historical facts that supports conflicting inferences ... resolved such conflicts in favor of proseeution[.]”71

Nevertheless, the distinction is a real one, and might have served as an acceptable reformulation of the factual-sufficiency standard, standing alone. But we apparently created a problem when we elaborated on our reformulation, first in the following vein:

However, there are two ways in which the evidence may be insufficient. First, when considered by itself, evidence supporting the verdict may be too weak to support the finding of guilt beyond a reasonable doubt. Second, there may be both evidence supporting the verdict and evidence contrary to the verdict. Weighing all the evidence under this balancing scale, the contrary evidence may be strong enough that the beyond-a-reasonable-doubt standard could not have been met, so the guilty verdict should not stand.72

This reformulation of the standard begins to make sense only when one remembers that it is not a legal sufficiency standard. In other words, we are not now asking whether the jury could rationally have found the defendant guilty beyond a reasonable doubt, having presumably resolved all conflicts or ambiguities in the evidence in favor of the prosecution. (Presumably the jury has, or in most cases the appellate *416court would not have reached the issue of factual sufficiency, having acquitted the defendant.) The fact is that rational people can disagree whether even legally sufficient evidence is persuasive to a level of confidence beyond a reasonable doubt. “That rational men disagree is not in itself equivalent to a failure of proof by the State, nor does it indicate infidelity to the reasonable doubt standard. Jury verdicts finding guilt beyond a reasonable doubt are regularly sustained even though the evidence was such that the jury would have been justified in having a reasonable doubt[.]”73 This means that an appellate court can both affirm the legal sufficiency of the evidence, and at the same time harbor its own, admittedly subjective, reasonable doubt about the accuracy of the jury’s verdict.

Thus, when the Court in Zuniga referred to evidence that is factually insufficient because it is “too weak to support a finding of guilt beyond a reasonable doubt[,]” we must have meant that the evidence would not satisfy the appellate court’s own threshold of proof beyond a reasonable doubt. And when the Court said that “the contrary evidence may be strong enough that the beyond a reasonable doubt standard could not have been met,” we must have meant that the weight and credibility of the exculpating evidence is so great that the appellate court itself feels compelled to harbor a reasonable doubt as to the defendant’s guilt. This understanding is problematic, however, because it smacks of an appellate court simply opting to “disagree” with the jury’s verdict — something we have never before tolerated even in the “factual sufficiency” context. That an appellate court would have acquitted a defendant on the same facts that convinced a rational jury to convict has not ever, by itself, met our criteria of a “manifest injustice.”

The problem is exacerbated by what our Zuniga opinion says next:

This standard acknowledges that evidence of guilt can ‘preponderate’ in favor of conviction but still be insufficient to prove the elements of the crime beyond a reasonable doubt. Stated another way, evidence supporting guilt can ‘outweigh’ the contrary proof and still be factually insufficient under a beyond-a-reasonable-doubt standard.74

This re-re-articulation of the standard is particularly troublesome. The first sentence (upon which the court of appeals relied in the instant case) suggests that evidence legally sufficient to convict may still be factually insufficient if it merely “preponderates” in favor of guilt, rather than establishing guilt beyond a reasonable doubt. Thus, even if, viewing all of the evidence in the light most favorable to the verdict, the appellate court concludes that a rational jury could find guilt to the requisite level of confidence beyond-a-reasonable-doubt, that same appellate court should order a new trial if, viewing that same evidence neutrally, it is subjectively convinced by no more than a preponderance of the evidence. The second sentence suggests that even when the State’s evidence of guilt is more weighty and more credible than the defendant’s contrary evidence, and therefore “preponderates” in favor of guilt, an appellate court could find that it does not meet its own threshold level of confidence beyond a reasonable doubt, and reverse and remand on that basis.

It is in the very nature of a factual-sufficiency review that it authorizes an ap*417pellate court, albeit to a very limited degree, to act in the capacity of a so-called “thirteenth juror.” Indeed, it is this characteristic of a factual-sufficiency review that justifies the conclusion that a reversal on the basis of factually insufficient evidence has no jeopardy consequences.75 But prior to Zuniga, as our review of the origins and history of factual-sufficiency review has demonstrated, we always required a high level of skepticism about the jury’s verdict in order to justify granting a new trial. Any holding that a criminal appellate court can reverse and remand for a new trial even when the evidence “preponderates” in favor of a conviction is inconsistent with that historically required high level of skepticism.

We therefore disavow such language in Zuniga and reiterate that it is not enough that the appellate court harbor a subjective level of reasonable doubt to overturn a conviction that is founded on legally sufficient evidence. An appellate court judge cannot conclude that a conviction is “clearly wrong” or “manifestly unjust” simply because, on the quantum of evidence admitted, he would have voted to acquit had he been on the jury. Nor can an appellate court judge declare that a conflict in the evidence justifies a new trial simply because he disagrees with the jury’s resolution of that conflict. We have always held that an appellate court must first be able to say, with some objective basis in the record, that the great weight and preponderance of the (albeit legally sufficient) evidence contradicts the jury’s verdict before it is justified in exercising its appellate fact jurisdiction to order a new trial. We have never, at least until Zuniga, interpreted the factual review jurisdiction of criminal appellate courts to include the ability to overturn a jury verdict and remand for a new trial when the greater weight and preponderance of the evidence actually favors conviction! To the extent that Zuniga may be to read to allow such a reversal, it was flawed at the outset and has a clear potential to cause far more reversals for factual insufficiency than was ever contemplated by either Clewis or the many cases within the last 125 years that have inquired into factual sufficiency.76 It is therefore, to that extent, overruled.

CONCLUSION

Accordingly, we reverse the judgment of the court of appeals and remand the cause to that court for a reconsideration of the appellant’s factual-sufficiency claim in light of this opinion.

COCHRAN, J., filed a dissenting opinion in which KELLER, P.J., and KEASLER and HERVEY, JJ., joined.

dissenting in which KELLER, PJ., KEASLER and COCHRAN, JJ., joined.

In another decision meant to “clarify” Clewis,1 the Court overrules a less than three year-old, 7-judge majority opinion in Zuniga,2 which was also meant to “clarify” Clewis. The reader should notice that, in addition to overruling Zuniga, the Court’s opinion also overrules a significant portion *418of Clewis and a significant portion of Cain,3 which was another decision meant to “clarify” Clewis.

The Court’s opinion disavows that part of Clewis stating that Texas courts apply the Jackson v. Virginia4 legal sufficiency standard “in such a way that the only evidence a reviewing court considers is the evidence that supports the verdict.”5 See Watson v. State, 204 S.W.3d at 414 (Jackson v. Virginia legal sufficiency standard requires a review of “all the evidence”) (emphasis in original). This misstatement of the Jackson v. Virginia standard by Clewis formed a very significant portion of Clewis’ rationale that a factual sufficiency standard was needed to overturn convictions that are “unjust” under Jackson v. Virginia with Clewis’ 40-nun hypothetical6 used to illustrate such an “unjust” conviction. This, of course, requires the reader to accept that a verdict that is rational under the beyond a reasonable doubt standard under Jackson v. Virginia can at the same time be “clearly wrong” or “manifestly unjust.” See Watson, Op. at 417.7

The Court’s opinion also disavows a significant portion of Cain. With the Court recognizing that the Jackson v. Virginia standard requires an appellate court to consider all the evidence, this means that the only difference between a Jackson v. Virginia legal sufficiency review and a Clewis factual sufficiency review is that the former requires the appellate court to defer to the jury’s credibility and weight determinations8 while the latter permits the appellate court to disagree with the jury’s credibility and weight determina*419tions.9

But, authorizing appellate courts to disagree with a jury’s credibility and weight determinations via Clewis’ factual sufficiency review is inconsistent with Articles 38.04 and 36.13, Tex.Code CRIM. PROC., which unambiguously, without limitation, make the jury the exclusive judges of the credibility and weight of the evidence. The Court’s opinion, however, states that these Code provisions do not apply to appellate review of evidentiary sufficiency questions because “their placement in Chapters 36 and 38 of the Code, which govern [sic] trial practice and procedure” make it apparent that these provisions are “meant merely to allocate the fact-finding function at the trial level and do not purport to affect appellate review.” See Watson, Op. at 409; accord Clewis, 922 S.W.2d at 134-35 (import of these Code provisions is “to distinguish the role of the jury from the role of the judge at trial”).10 But, this Court’s decision in Cain made no such distinction, and, Cain expressly relied on these Code provisions to state, contrary to what Clewis originally stated, that “the appellate court must defer to jury findings, and may find the evidence factually insufficient only where necessary to prevent manifest injustice.” See Cain, 958 S.W.2d at 407 (emphasis supplied).11 In stating that these Code provisions do not apply to the appellate process, the Court’s opinion overrules a significant portion of Cain.12

Notwithstanding the foregoing, the issue of whether Texas appellate courts have factual sufficiency jurisdiction in criminal cases is a matter of legislative intent subject to a Boykin analysis. This requires an interpretation of Article 44.25, Tex. Code Crim. Proc., against the backdrop of *420the legal landscape as it existed in 1981 when the Legislature amended Article 44.25 to its current form and granted the intermediate appellate courts criminal jurisdiction. When this occurred in 1981, the Supreme Court had recently decided Burks/Greene,13 which provided the remedy of an appellate acquittal on evidentiary insufficiency grounds, and Jackson v. Virginia, which sets out the now familiar due process standard for reviewing evidentiary sufficiency.

Prior to 1981, Article 44.25 and its statutory predecessors provided that appellate criminal courts may reverse a conviction “upon the law as upon the facts.”14 These pre-1981 statutory predecessors to current Article 44.25 also authorized the appellate remedy of a remand for a new trial when a conviction was reversed because it was contrary to the weight of the evidence.15 But, it is very significant that the 1981 legislative amendments to Article 44.25 deleted this language authorizing the appellate remedy of a new trial when a conviction was reversed because it was contrary to the weight of the evidence. See Footnotes 13, 14. The effect of these 1981 legislative amendments to Article 44.25 deleting this language was to withdraw an appellate criminal court’s jurisdiction to provide the remedy of a remand for a new trial when a conviction was reversed because it was contrary to the weight of the evidence.16

It, therefore, seems clear that the Legislature very carefully insured that there would be no factual sufficiency jurisdiction in criminal cases when it amended Article 44.25 to its current form in 1981.17 It seems equally clear that the Legislature intended that appellate criminal courts apply only the evidentiary sufficiency standard set out in Jackson v. Virginia and provide only the appellate remedy of an acquittal set out in Burks/Greene upon a determination that the evidence is insufficient to support the conviction. And, as the Court’s opinion seems to recognize, Texas appellate criminal courts applied only the Jackson v. Virginia standard after 1981 until this Court some years later “rediscovered factual sufficiency, and resurrected it from a brief, court-imposed dormancy.” See Watson, Op. at 405.

Finally, while purporting to retain a [factual] sufficiency standard that is “more solicitous of defendants’ rights,”18 the Court’s retention of this standard is actual*421ly less solicitous of defendants’ rights. Clewis’ 40-nun hypothetical illustrates this as well. According to Clewis, this 40-nun hypothetical illustrates a legally sufficient but factually insufficient conviction. See Clewis, 922 S.W.2d at 133 n. 12. This means that the State can retry this hypothetical defendant, and the appellate court would probably have to affirm this hypothetical defendant’s conviction the second time around on the same evidence as that presented at his first trial. See Tibbs v. Florida, 457 U.S. 31, 43 n. 18, 102 S.Ct. 2211, 72 L.Ed.2d 652 (1982). But, this hypothetical defendant would have been entitled to an appellate acquittal the first time around under a proper application of Jackson v. Virginia.

With these comments, I join Judge Cochran’s dissent.

dissenting in which KELLER, P.J., and KEASLER and HERVEY, JJ„ joined.

I respectfully dissent. For the reasons set out below, I believe that further efforts to clarify, refine, or revise the Clewis standard are as unlikely to succeed as our previous attempts. The Clewis factual-sufficiency review was a well-intentioned but ultimately unworkable effort to incorporate civil standards of review on elements of a crime that must be proven beyond a reasonable doubt. I would overrule Clewis and return to the single standard of review for sufficiency of the evidence in a criminal case as set out by the United States Supreme Court in Jackson v. Virginia.1 Therefore, I would reverse the court of appeals and uphold the trial court’s judgment because the evidence supporting appellant’s conviction was legally sufficient under the Jackson standard.

I.

Appellant was charged with burglary by entering Jennifer Romo’s home without consent and attempting to commit or committing sexual assault. The evidence at trial showed that, on a warm June afternoon, Jennifer Romo was at her kitchen screen door, watching her two pre-school children play in the yard outside her Waco apartment, when she noticed appellant jogging across the parking lot toward her. Jennifer turned around and went back into her kitchen. When she turned around again, appellant was standing at her screen door. He stuck his head and body halfway through the door and put his foot in the doorway. Appellant tried to have a conversation with Jennifer and kept saying that he knew her. Appellant was wearing gym shorts, and Jennifer could see that he had an erection. Finally she told him to leave, and he did.

Jennifer went into her living room, but a few seconds later she heard the screen door open again. She turned around and saw appellant standing inside her kitchen. She was “extremely” angry and a little bit scared. She started forward, screaming at him to get out of her house, but he didn’t move. “He just stood there, and he looked at me, and he just dropped his shorts, and they just fell.” His penis was sticking out. Jennifer turned around and grabbed a butcher knife. She testified, “I pointed it at him, and I told him that he better leave my house, that he better get out.... He just looked at me, and he grabbed the end of my knife.” Jennifer twisted the knife to get it out of his hands. Finally, he let go of it. Jennifer said, “I guess he thought I meant business because he picked his shorts up and he walked out of the house.” Jennifer thought that the reason appellant *422left was because he thought she was going to stab him.

Jennifer watched as he walked back across the parking lot and sat down on a porch. Jennifer called her father, who told her to call the police. When Officer Arnold arrived a few minutes later, appellant was still sitting on the porch. After talking with Jennifer and taking her statement, Officer Arnold arrested appellant, who gave him a false name. Officer Arnold formed the opinion that appellant was trying to cover up his culpability.

The defense presented no evidence, but appellant made a motion for instructed verdict. He argued that the State failed to offer sufficient evidence to prove appellant “attempted to commit or actually committed a sexual assault[.]” The State conceded that there was no evidence that appellant actually committed sexual assault, but argued that

there are plenty of things that he did that go to his attempt to commit a sexual assault. He pulled down his pants. He exposed his erect penis, and, critically, he grabbed for a knife when she had grabbed a knife to protect herself. He didn’t turn and run away. He put his hand out, and he grabbed that knife, and that tells me that his intent was something more than just to expose himself to her.... There was a reason why he came into her house. He pulled his pants down. He was ready to do something to her. You heard her testify that she had one second — in one second, she had that knife in her hand. Just because he never got the opportunity to grab her or pull her clothes off, attempt to sexually assault her any further, doesn’t mean that the pulling the pants down and the grabbing of the knife were not an attempt to sexually assault her. And that is some evidence. It’s enough evidence that the jury could find beyond a reasonable doubt that he did attempt that offense.

The trial judge denied the motion for instructed verdict, and the jury convicted appellant of burglary and sentenced him to thirteen years’ imprisonment.

On appeal, appellant argued that the evidence was both legally and factually insufficient. In a published opinion, a majority of the court of appeals held that, although the evidence was legally sufficient, it was factually insufficient because “the evidence which tends to support the verdict ‘preponderates’ in favor of conviction but still [is] insufficient to prove the elements of the crime beyond a reasonable doubt.”2 Chief Justice Gray dissented, setting out the inculpatory evidence and concluding, “I guess my response to the majority opinion — ‘You have got to be kidding’ — makes me irrational, along with 12 jurors and the trial court.”3

II.

Should a Texas jury’s “guilty” verdict be subjected to two distinct “sufficiency of the evidence” hurdles on appellate review? The first hurdle, Jackson, is the constitutionally required review for legal sufficiency of the evidence.4 The second hurdle was created by this Court in Clems and draws from Texas civil standards of review. Since Clewis was decided ten years ago, we have made frequent, but largely futile, attempts to clarify the differences between these two theoretically distinct sufficiency hurdles in criminal cases. It is *423crucial to distinguish clearly the constitutionally mandated hurdle of legal sufficiency from the judicially created hurdle of factual sufficiency. The result of finding one rather than the other insufficient is dramatically different. If the evidence is legally insufficient, the defendant is entitled to an acquittal by the appellate court, and he cannot be retried. If the evidence is factually insufficient, he is entitled only to a second trial in which the State may produce additional evidence to satisfy whatever gaps or conflicts a court of appeals found in the first trial.5 It is essential that evidentiary hurdles like Jackson and Clewis be easily distinguishable. If they are not easily and logically distinguishable, allowing both to apply muddles the law and violates the cardinal principle of appellate justice — the “uniform and coherent enunciation and application of the law.”6

In Jackson v. Virginia, the Supreme Court of the United States held that constitutional due process requires reviewing courts to “determine whether the record evidence could reasonably support a finding of guilt beyond a reasonable doubt.”7 In assessing the sufficiency of evidence, “the relevant question is whether, after viewing the evidence in the light most favorable to the prosecution, any rational trier of fact could have found the essential elements of the crime beyond a reasonable doubt.”8 The Supreme Court explained:

This familiar standard gives full play to the responsibility of the trier of fact fairly to resolve conflicts in the testimony, to weigh the evidence, and to draw reasonable inferences from basic facts to ultimate facts. Once a defendant has been found guilty of the crime charged, the factfinder’s role as weigher of the evidence is preserved through a legal conclusion that upon judicial review all of the evidence is to be considered in the light most favorable to the prosecution.9

Under Jackson, appellate courts review a claim attacking the sufficiency of evidence in a criminal case to determine whether the State produced enough evidence to support a “rational” factfinder’s guilty verdict beyond a reasonable doubt. A “rational” verdict is one that is “based on reason”; it is “not foolish or silly”; it is “sensible.”10 It is a verdict “based on, or guided by reason, principle, fairness, [and] logic[.]”11 To ensure that the original factfinder’s weight and credibility decisions are given sufficient deference, the Supreme Court requires reviewing courts to view all of the evidence in the light most favorable to the prosecution.12

*424In Clems, this Court authorized Texas courts of appeals to conduct a second review of the evidence: a /acfetai-suffieiency review under the following test:

In conducting a factual sufficiency review, the court of appeals “views all the evidence without the prism of ‘in the light most favorable to the prosecution’ ... [and] sets aside the verdict only if it is so contrary to the overwhelming weight of the evidence as to be clearly wrong and unjust.” 13

On its face, this test differs from the Jackson standard in two ways: (1) the evidence is not viewed in the light most favorable to the State (which is simply a proxy for deference to the factfinder, either judge or jury); and (2) the verdict, even if rational and reasonable under Jackson, is examined to see if it is so contrary to the overwhelming weight of the evidence, that it is manifestly wrong and unjust. The issues that have bedeviled this Court and the courts of appeals for the past ten years include: (1) what, precisely, do those two prongs mean; (2) how does the Clewis standard differ from the Jackson standard; and (3) to what extent is the court of appeals statutorily or constitutionally empowered to act as a thirteenth (and in absentia) juror? More broadly, do these standards differ sufficiently to require such a dramatic difference in result — acquittal versus retrial? And finally, as a matter of public policy, what is the injustice or legal problem that a factual-sufficiency review is intended to solve? Does it do so satisfactorily? I do not think so.

To adequately address these issues, however, one must examine the historical development of both Texas and federal appellate review standards for the sufficiency of evidence in criminal cases.

A. Origins of Texas Appellate Review of the Sufficiency of the Evidence.

Even in the days of the Texas Republic, the Texas Supreme Court could review the sufficiency of evidence to support a criminal conviction. The earliest known case is Republic of Texas v. Smith,14 an 1841 decision, in which the Supreme Court of Texas addressed the threshold question of whether a criminal defendant had a right to appeal “the facts” as well as “the law.”15 Finding that it had jurisdiction “to have the facts as well as the law, at [the defendant’s] election, opened for re-examination,” 16 the supreme court reviewed the defendant’s claim that, although he was charged with dealing faro17 in Austin, venue was improper in Travis County because, at the time of the offense, Austin was in the territorial jurisdiction of Bastrop County.18 The supreme court said that it would “consider the decision of the district judge, with an eye both to the law and the facts as shown by the record.”19 It rejected the defendant’s venue claim.20

*425The next important decision dealing with “the facts” in a criminal case, was Bishop v. State,21 decided in 1875. A jury convicted Bishop of stealing a cow from Mrs. Lizzie Hill. There was conflicting evidence on whether he “appropriated” the cow in good faith by purchasing it from a man named Keel, or he fraudulently took it from Mrs. Hill.22 The jury charge, however, omitted the statutory word “appropriation”23 and its definition. Given the conflicting evidence on the sole disputed fact issue, this was a material error.24 The trial judge committed a second material error when he told the jury that, in civil cases, “the jury must believe all evidence given before them until the same is impeached, and in a criminal case they may discard from their consideration any part of the whole of the testimony that they may regard as improbable.”25 The supreme court condemned that statement as infringing upon the jury’s right to make its own credibility decisions, and it noted that “many improbable things actually do occur, and when detailed in evidence as part of any matter under investigation they are not to be discarded from the consideration of the jury because they are improbable.” 26 It is the jury, not the trial judge (and not the appellate court), that makes credibility decisions.27

These two cases have been cited as supporting the view that Texas appellate courts are authorized to conduct a review of the factual sufficiency of evidence in criminal cases.28 They provide no such support. Instead, Smith explained that the Texas Supreme Court had jurisdiction over an “appeal” from a criminal case, including both the facts and the law, not merely over a “writ of error” which, under the early English common law, was a for*426mal point of law that accepted the “facts” uncritically.29 And Bishop emphasized the importance of the jury’s factfinding role once it has been properly instructed on the applicable law. These cases hold that appellate courts may re-examine the facts— the evidence in the trial record — in addressing the legal issues.

In 1875, the supreme court did review the sufficiency of evidence in Tollett v. State,30 and, in 1877, the newly created court of appeals did so in Loza v. State.31 In both cases, the reviewing court held that there was not “legal and competent evidence” establishing an element of the offense: in Tollett, the prosecution failed to prove that the defendant participated in a theft from a storehouse;32 in Loza, the prosecution failed to prove that the defendant had the intent to permanently deprive the owner of his gelding.33 The courts used a single sufficiency-of-the-evidence standard:

whether or not there has been adduced before the jury a sufficient amount of legal and competent evidence as would render it safe to allow the verdict to stand and become a precedent in the adjudication of offenses under the law. The performance of this duty on the part of the court is the exercise of a legal discretion and judgment as to what facts should be sufficient to rebut the legal presumption of innocence to which every one is entitled who is put upon his trial for an offense.34

This sounds very much like the Jackson standard appellate courts now use for assessing legal sufficiency of the evidence. And it was the only standard used. There were not two distinct “sufficiency of the evidence” hurdles in the early Texas years. But, like Smith and Bishop, both Tollett and Loza have been cited as supporting the view that early Texas courts were authorized to consider “factual” as well as “legal” sufficiency of the evidence in a criminal case.35 In fact, Smith, Bishop, Tollett, and Loza were all properly decided under a Jackson-like standard alone; a separate factual-sufficiency analysis was neither made nor necessary to their resolution.

The most important case discussing sufficiency of the evidence was the court of appeals’s 1883 decision in Walker v. State,36 in which the court set out four categories or zones to use in analyzing the *427question, “Does the evidence support the verdict of the jury?”

First. Where the evidence is conflicting, and there is sufficient [evidence], if believed, to prove the case of the State, the jury being the exclusive judges of the credibility of the testimony, their verdict will not be set aside unless it clearly appears to be wrong.

Second. Where there is no testimony to support it, the verdict will be set aside. Third. Where the evidence is insufficient to rebut the presumption of innocence, the verdict will be set aside. Fourth. Where the verdict is contrary to the weight of the evidence, it will be set aside.37

Walker was apparently the first Texas case to specifically set out the “zone” theory of sufficiency analysis, one that has historically been used to determine whether a party has shouldered his burden of proof by a preponderance of the evidence.38 In zone one, under Walker, the appellate court upholds a criminal conviction when there is sufficient evidence to prove the State’s case (even though the evidence may be conflicting), unless “it clearly appears to be wrong.”39 Walker’s zone two is a “no evidence” standard, and that means no evidence or a mere modicum or scintilla. Zone three is “insufficient evidence,” more than a mere scintilla but not enough to “rebut the presumption of innocence,” ie., not enough to meet Jackson’s “beyond a reasonable doubt” standard. Zone four is a verdict contrary to the “weight” of the evidence; this zone translates to the modem Texas civil standard of “against the great weight and preponderance” of the evidence.40 Missing is zone 5 — the modern civil standard of “con-*428elusive evidence” — because the State is never entitled to a directed verdict of guilty. The Walker court explained that its duty was to determine which zone the evidence fell into.

In Walker, the court held that the evidence was “insufficient” (zone 3) to prove that the murder victim had actually died because the State failed to show that the body retrieved 600-900 miles down the Brazos River was actually that of the missing man — W.B. Mathis.41 Thus, the court reversed the defendant’s murder conviction and remanded the case for another trial, giving the State hefty hints on how to produce sufficient evidence in the retrial.42 Walker has been cited as supporting both a legal and factual sufficiency of the evidence review,43 even though the court’s decision was fully consistent with the Jackson standard alone.44

*429In sum, case law before the creation of the Texas Court of Criminal Appeals held that the evidence in a criminal case was required to be “sufficient” to support a conviction and that evidence was measured against one single standard, but there were various ways of expressing that standard. When the evidence was insufficient, the case was remanded for a second trial.

B. Texas Appellate Review of Sufficiency of the Evidence from 1891-1976.

The 1876 Texas Constitution stripped the Texas Supreme Court of jurisdiction over criminal cases and created the Texas Court of Appeals that had jurisdiction over all criminal, county court, and probate appeals.45 A majority of the delegates to the Constitutional Convention of 1875 were progressives and largely anti-government; they were not particularly interested in the judiciary or the structure of state courts, but they had great faith in juries— their fellow citizens.46 Indeed, every Texas constitution had enshrined the right to a trial by jury in both civil and criminal cases, and, under the 1876 Constitution, that right to a jury trial “shall remain inviolate.”47

*430In 1891, the Texas Constitution was amended in two important respects. First, because the overburdened supreme court was so far behind on its docket, the legislature created three courts of civil appeals which had intermediate appellate jurisdiction.48 “The theory in creating the courts of civil appeals apparently was that their decisions would be final in most civil cases.”49 To help ensure that goal, the constitutional provision creating the courts of civil appeals also stated that “the decision of said courts shall be conclusive on all questions of fact brought before them on appeal or error.”50 This “factual con-clusivity” provision did not grant any new, implicit factfinding authority to the courts of civil appeals; instead, it limited the jurisdiction of the supreme court.51 The supreme court was to confine itself to weighty legal principles and conflicts in the law.52 The courts of civil appeals could then consistently apply those enduring legal principles to varying facts, but whenever the law otherwise permits or requires a determination of “the facts” for some purpose in an appeal, the decision by a court of civil appeals on those questions of fact was conclusive.53 The “factual conclusivity” provision applies only to those factual questions which a statute or other law requires or permits.54

A second 1891 constitutional amendment created the Texas Court of Criminal Appeals that had exclusive appellate jurisdiction over criminal cases.55 The court of criminal appeals reviewed issues concerning the sufficiency of the evidence exactly as the Texas Supreme Court and the Texas Court of Appeals had done before it: one single standard of review, but many different ways of articulating that standard. Here are some examples:

* “If the guilt of appellant is not made to appear to a reasonable certainty, it is the duty of this court to reverse the judgment”; and [every element] “must be proven beyond a reasonable doubt; proved by facts and circumstances which lead with reason*431able certainty to the conclusion sought, and not left to speculation and surmises[.]”56

* “This court will not reverse unless there is such a manifest lack of evidence as to make it apparent that the verdict was the result of prejudice or that such verdict is against the great weight of the evidence.”57

* “The sanctity of the verdict of the jury is such that it must not lightly be annulled in any case, but the law does not contemplate that one shall suffer because of the verdict when the evidence, viewed in its strongest light from the standpoint of the state, fails to make his guilt reasonably certain.”58

* “[Tjhis court will reverse a case on the insufficiency of the evidence only when the testimony, viewed in its strongest light from the standpoint of the State, fails to make guilt of the accused reasonably certain.”59

* “Ordinarily if facts be proved from any source which, if believed, would warrant the conclusion of guilt, the evidence is sufficient to support the conviction.”60

* “The jury is not bound to accept all or any of the testimony of a witness, even though uncontradicted. It is not the prerogative of the appellate court to interfere with the jury’s finding on questions of fact where there is evidence authorizing their verdict.”61

* “[T]he jury has been made the exclusive judges of the facts proven and the weight to be given to the testimony ... this court does not, and should not, assume to exercise the right to reverse on the facts, unless the evidence, when viewed in its strongest light from the standpoint of the State, fails to make guilt reasonably certain.”62

* “In reviewing the sufficiency of the evidence to support the conviction, *432we must view the evidence in the light most favorable to the verdict. In doing so, the verdict will be sustained if there is any evidence which, if believed, shows the guilt of the accused.”63

Until Clewis in 1996, this Court had consistently used a single standard (although the precise phrasing varied) and reviewed the evidence in the light most favorable to the factfinder, giving great deference to the jury’s credibility and weight determinations. But it did not hesitate to reverse a conviction if the evidence failed to prove a defendant’s guilt “with reasonable certainty,” or “beyond a reasonable doubt.” Until Jackson in 1979, the result of a finding of insufficient evidence in Texas was always the same: remand and retrial. There was no acquittal on appeal when the evidence was insufficient. So matters stood until the United States Supreme Court mandated a constitutional, due process review of sufficiency of the evidence for all courts.

C. The United States Supreme Court Announces Constitutional Standards of Review and Double Jeopardy Consequences of Insufficient Evidence.

In 1979, the Supreme Court, in Jackson v. Virginia, held that the Due Process Clause of the federal constitution requires the government to produce “sufficient evidence to justify a rational trier of the facts to find guilt beyond a reasonable doubt.”64 The Court explicitly rejected a “no evidence” standard of review that it had applied nineteen years earlier in Thompson v. Louisville,65 In Thompson, the Supreme Court had concluded that “we find no evidence whatever in the record to support these convictions. Just as ‘Conviction upon a charge not made would be sheer denial of due process,’ so is it a violation of due process to convict and punish a man without evidence of his guilt.”66 In Jackson, the Court explained that the “no evidence” doctrine “secures to an accused the most elemental of due process rights: freedom from a wholly arbitrary deprivation of liberty!,]”67 but that standard is inadequate for “a question of evidentiary ‘sufficiency.’ ”68 Instead, the correct standard must incorporate the prosecution’s burden of proof — beyond a reasonable doubt — in a due-process review. The Court noted that a “ ‘reasonable doubt’ has often been described as one ‘based on reason which arises from the evidence or lack of evidence.’ ”69 For example, a reasonable doubt might arise because the verdict is manifestly against the great weight and *433preponderance of the evidence or because there is nothing more than a mere scintilla of evidence to support some element of the offense. But, of course, the reviewing court does not “ ‘ask itself whether it believes that the evidence at the trial established guilt beyond a reasonable doubt.’ ”70 Rather, it “gives full play to the responsibility of the trier of fact fairly to resolve conflicts in the testimony, to weigh the evidence, and to draw reasonable inferences from basic facts to ultimate facts.”71 Thus, “all of the evidence is to be considered in the light most favorable to the prosecution” because the reviewing court may impinge upon “ ‘jury’ discretion only to the extent necessary to guarantee the fundamental protection of due process of law.”72

In sum, due process does not permit legal sufficiency of the evidence to be measured by a “no evidence” (zone 1) standard; it must be measured in terms of the prosecution’s constitutionally-mandated “beyond a reasonable doubt” burden of proof (which does not correlate to any of the “zones” in civil review standards fashioned on the “by a preponderance of the evidence” burden of proof), but giving great deference to the jury’s role in assessing credibility, weight, and reasonable inferences.73

The Jackson “legal sufficiency” standard is the one required by the United States Constitution, and every state and federal court must comply with that standard.74 It is the only standard required for reviewing sufficiency of the evidence. While noting that “states are free to set higher standards of review[,]” this Court quickly adopted the Jackson standard for the review of sufficiency of the evidence.75 We adhered to that single, constitutionally mandated standard for seventeen years.76 *434Similarly, the Texas Legislature declined hints in Griffin that it could set a higher standard of review; it did not enact any statutes to heighten the Jackson sufficiency review.

The legal consequence of a finding of insufficient evidence under the Jackson standard is an acquittal on appeal; any retrial is barred by double jeopardy principles.77 However, in its 1982 decision in Tibbs v. Florida,78 the Supreme Court held that, if a state chooses to conduct a factual, as well as legal, sufficiency review, double jeopardy does not bar a retrial.79 The Supreme Court explained,

A reversal on this ground, unlike a reversal based on insufficient evidence, does not mean that acquittal was the only proper verdict. Instead, the appellate court sits as a “thirteenth juror” and disagrees with the jury’s resolution of the conflicting testimony. This difference of opinion no more signifies acquittal than does a disagreement among the jurors themselves.... Similarly, an appellate court’s disagreement with the jurors’ weighing of the evidence does not require the special deference accorded verdicts of acquittal.80

Thus, an appellate reversal based on the weight of the testimony “simply affords the defendant a second opportunity to seek a favorable judgment” because such a reversal can occur only after the State has produced sufficient evidence to convict and has persuaded the jury to convict.81

Ironically, the Florida Supreme Court had, in the Tibbs case itself, jettisoned its judicially created “factual sufficiency” review because it was conceptually flawed and unworkable in practice.82 Although it noted that a state rule of criminal procedure allowed a trial judge to grant a motion for new trial if the jury verdict is “contrary to the law or weight of the evidence,” the Florida Supreme Court concluded that “an appellate court should not retry a case or reweigh conflicting evidence submitted to a jury or other trier of fact.”83 “Legal sufficiency alone, as opposed to evidentiary weight, is the appropriate concern of an appellate tribunal.”84 The court recognized that the language in some earlier opinions was ambiguous and could be interpreted as calling for a review of factual sufficiency, but the notion of factual-sufficiency review had “a questionable historical foundation.”85 Furthermore, the court concluded that “[cjonsiderations of policy *435support, if not dictate” elimination of factual-sufficiency review on appeal.86 First, the court noted that the proper role of appellate courts in Florida criminal cases was to leave questions of the weight of the evidence to the trier of fact.87 Second, eliminating factual-sufficiency review “will avoid disparate appellate results” or further review to determine whether a reversal is based on legally or factually insufficient evidence.88 Third, abolishing a factual-sufficiency review will “eliminate any temptation appellate tribunals might have” to reverse for factual insufficiency to avoid the double-jeopardy bar on retrials for legal insufficiency.89 The Florida Supreme Court concluded that twelve ordinary citizens sitting as jurors are better able to determine the credibility of the witnesses and the weight of the testimony than are appellate judges reading the written record. A factual-sufficiency review on appeal contradicts that fundamental fact.

D. Texas Standards for Reviewing the Sufficiency of Evidence to Reject an Affirmative Defense.

The first time that this Court directly addressed the concept of factual-sufficiency review was in Schuessler v. State,90 a 1986 case in which the court of appeals had held that the jury’s rejection of his affirmative defense of insanity was against the great weight and preponderance of the evidence.91 This Court rejected that standard, even for affirmative defenses in which the defendant bears the burden of proof by a preponderance of the evidence — a civil law standard — because there can be “no reweighing or reclassifying of the evidence by the appellate court.”92 This Court then applied — to the rejection of an affirmative defense — the same Jackson standard of legal sufficiency that it routinely applied to the elements of the criminal offense.93

Judge Clinton dissented because the issue of insanity is not an element of the criminal offense and thus is not governed by the due process standard set out in Jackson.94, He explained,

When an affirmative defense is not an element of an offense charged and Win-ship is not implicated by allocation of burden of proving an affirmative defense, it follows that “the critical inquiry on review of the sufficiency of the evidence to support a criminal conviction” prescribed by Jackson v. Virginia is irrelevant to a review of sufficiency of evidence relating to an affirmative defense. The trier of fact was not called on to view the evidence in the light most favorable to the prosecution in order to find “the essential elements of the crime beyond a reasonable doubt.” Rather, its *436function was to determine whether an accused has proved an affirmative defense “by a preponderance of the evidence.” 95

Judge Clinton’s position was adopted four years later in Meraz v. State,96 a case in which the defendant appealed the jury’s rejection of his plea of incompetence. In Meraz, this Court applied the civil law factual-sufficiency standard when the party who bears the burden of persuasion by a preponderance of the evidence complains that the rejection of his issue is “against the great weight and preponderance of the evidence.”97 This Court reasoned that

the utilization of the standard of review constitutionally reserved for the courts of appeals and required by the Texas Supreme Court to be applied in situations where the burden of proof was on the individual claiming factual insufficiency does not impede a defendant from seeking a factual review relevant to his affirmative defense nor does it preclude him from a sufficiency review as to whether there was sufficient evidence to warrant a conviction. The two reviews are mutually exclusive.98

Indeed, this Court was exactly correct: the two different types of review are mutually exclusive. Review for legal sufficiency under the Jackson “beyond a reasonable doubt” standard is required for every element that supports a conviction, but a sufficiency-of-the-evidence review under civil standards is appropriate when reviewing affirmative defenses in which the defendant bears a burden of “preponderance of the evidence.”99 In Meraz, this Court held that the “factual conclusivity clause” of the Texas Constitution applies when the defendant bears the burden of proof by a preponderance of the evidence.100

And this Court followed Meraz in Bigby v. State,101 by reviewing the sufficiency of the evidence that supported the jury’s rejection of the defendant’s affirmative defense of insanity in a capital murder trial. In Bigby, as in Meraz, the burden of proof was on the defendant to establish his defense by a preponderance of the evidence.102 In that context, use of the civil *437standards adopted by the Texas Supreme Court is entirely appropriate. In these cases, Texas courts have the power and authority to conduct a sufficiency review of the evidence under controlling supreme court precedent because: (1) both the burden of production and persuasion is on the defendant; and (2) that burden is the civil standard of proof “by a preponderance of the evidence.”103 Both Meraz and Bigby are in line with the historical and jurisprudential underpinning of Texas civil law and the Texas Constitution. We followed our brethren on the Texas Supreme Court in applying the factual conclusivity clause to those “questions of fact” for which the proponent bears the burden of proof by a preponderance of the evidence.104 And we should continue to follow that court’s development of sufficiency-of-the-evidence review when the proponent bears the burden of proof by a preponderance of the evidence.105

The real issue is not whether “factual sufficiency” review exists in Texas law. The issue is whether “factual sufficiency” review applies to the elements of a criminal offense that the State must prove beyond a reasonable doubt. And, of course, both this Court and the courts of appeals have the authority to review “the facts” as well as the law.106 If “the facts” do not establish every element beyond a reasonable doubt, those “facts” require an appellate court to acquit the defendant under Jackson.

But two years after this Court delivered Meraz, one court of appeals, in Stone v. State,107 applied the Meraz reasoning to a factual-sufficiency review of the elements of the offense, even though those elements must be proven beyond a reasonable doubt and are subject to a constitutional sufficiency review under Jackson.108 The court of appeals stated that, under the “factual conclusivity clause,” it had the power to review the factual sufficiency of the evidence to support a conviction as well as the power to review the factual sufficiency to support affirmative defenses:

[T]he Jackson test does not logically preclude factual-sufficiency review. Jackson requires the appellate court to view the evidence in the light most favorable to the prosecution. Jackson permits the states to devise tests for factual-sufficiency review that do not indulge in presumptions in favor of either side. Because Jackson does not preclude factual-sufficiency review and because we see nothing in the Meraz opinion that limits the pre-existing factual-sufficiency review power of the courts of appeals to matters the defendant must prove, we conclude that a criminal defendant is entitled to a factual-sufficiency review as a matter of state constitu*438tional law.109

Jackson does not, of course, preclude a state from enacting a factual-sufficiency review, but neither does it suggest that such review is necessary, appropriate, or useful. Until the court of appeals’s decision in Stone, no Texas court had found a constitutional right to a factual-sufficiency review of the elements of a criminal offense.110 There is nothing in the “factual conclusivity” clause of the Texas Constitution that grants authority to the courts of appeals to create a factual-sufficiency review of criminal convictions absent some statutory enactment or other law. Because that provision does not grant any new authority to the courts of appeals, a defendant does not have a constitutional “right” to a factual-sufficiency review of his conviction based solely upon the “factual conclusivity” clause. Any right to a factual-sufficiency review of the elements of a criminal offense must exist elsewhere in the law, if at all. Although the Texar-kana Court of Appeals adopted the Stone reasoning and result,111 other courts of appeals rejected it.112

*439The reasoning and result in Stone were premised upon the proposition that the intermediate courts of appeals are vested with the authority to review factual sufficiency in criminal cases under Article Y, section 6 of the Texas Constitution (the factual conclusivity clause).113 However, the constitutional provision says nothing about any Texas court’s authority to conduct a factual-sufficiency review of the evidence in a criminal case. As discussed above, it is only when the law otherwise allows or requires the review of a factual question that the conclusivity clause applies.114

In 1979, this Court had held, in White v. State, that we have “no fact jurisdiction as do the Courts of Civil Appeals, and cannot ‘unfind’ a vital fact finding by a jury. Since we do not have the jurisdiction to pass upon the great weight and preponder-anee of the evidence,”115 the defendant’s contention that his court could review factual sufficiency was rejected.116 If this Court, in 1979, did not have jurisdiction to conduct a factual-sufficiency review, the courts of appeals could not gain any “factual conclusivity” power from us when, two years later, they were authorized to review criminal cases.

This conclusion was reiterated by the Texas Legislature in 1981 when it amended what is now section 22.225 of the Texas Government Code. That provision states, “A judgment of a court of appeals is conclusive on the facts of the case in all civil cases.”117 This section says nothing about the conclusivity of factual findings in criminal cases.118 Furthermore, this Court had, until Clewis, always held that neither we nor the courts of appeals had any jurisdie*440tion to assess the factual sufficiency of evidence in a criminal case:

It is well settled that our Court does not have jurisdiction to pass upon the weight and preponderance of the evidence or “unfind” a vital fact. More specifically, our determinations of sufficiency of the evidence have never involved passing upon the weight and preponderance of the evidence.... In deciding sufficiency of the evidence questions this Court views the evidence in the light most favorable to the verdict.... Sufficiency of the evidence as determined by this Court is a question of law. It is irrelevant whether we as a court believe the evidence, or believe that defense evidence “outweighs” the State’s Evidence. If there is any evidence that establishes guilt beyond a reasonable doubt, and if the trier of fact believes that evidence, we are not in a position to reverse the judgment on sufficiency of the evidence grounds.119

Furthermore, the Code of Criminal Procedure contains two provisions that explicitly state that the jurors are the exclusive judges of facts in jury trials.120 Articles 36.13 and 38.04 are clear statutory mandates which cannot be reconciled with a judicially created doctrine that authorizes the courts of appeals to redetermine the facts and the weight of the testimony or “unfind” facts that the jury has found. There are no analogous mandatory provisions like these articles for civil cases.121 Thus, trial courts and courts of appeals may have implicit authority to redetermine or “unfind” facts and to reweigh evidence in civil cases, but they do not have similar authority in criminal eases because articles 36.13 and 38.04 clearly forbid any such reweighing or “unfinding.”

E. Clewis.

In 1996, this Court resolved the split in the courts of appeals on the issue of factual-sufficiency review. In a fractured 5-4 opinion in Clewis v. State,122 this Court *441rejected Justice LaGarde’s thorough explanation in the court of appeals of why factual-sufficiency review in criminal cases was neither authorized nor appropriate.123 Instead, the Court adopted the reasoning and result in Stone.124 This Court cited an example to explain why factual-sufficiency review was an important safeguard against runaway juries:

The prosecution’s sole witness, a paid informant, testifies that he saw the defendant commit a crime. Twenty nuns testify that the defendant was with them at the time, far from the scene of the crime. Twenty more nuns testify that they saw the informant commit the crime. If the defendant is convicted, he has no remedy under Jackson because the informant’s testimony, however incredible, is legally sufficient evidence.125

Of course, this was a law-school hypothetical; no one had heard of such a bizarre jury which would reject the eyewitness testimony of forty nuns and believe a single paid informant instead. Would such a result be the product of a “rational” or “reasonable” jury under Jackson? Would the United States Supreme Court (or this Court) uphold such a verdict under its carefully crafted language that, given this evidence, any “rational trier of fact could have found the essential elements of the crime beyond a reasonable doubt?” Under the hypothetical as presented, it would appear to be an irrational verdict. But suppose that the defendant were a priest? Or that the defendant had just donated a million dollars to the local convent? Or that the nuns were inconsistent and contradictory on all of the underlying details of their testimony? We defer to the jury’s credibility and weight determinations because they have seen and assessed the witnesses in person, and they have heard all of the evidence, not just the bare-bone facts of a hypothetical.

Nonetheless, in Clems, this Court judicially created a new appellate standard for reviewing the factual sufficiency of evidence on top of the legal sufficiency review which had, under various phrasing, existed in Texas for 150 years. Under that newly minted standard,

an appellate court reviews the factfin-der’s weighing of the evidence and is authorized to disagree with the factfin-der’s determination. This review, however, must be appropriately deferential so as to avoid an appellate court’s substituting its judgment for that of the jury.126

Thus, Clems empowered the courts of appeals to act as a “thirteenth juror” (one who was not even present to see and hear the witnesses) and disagree with the fact-finder’s determination, but to be deferential to the factfinder’s judgment as it did so. This standard was ambiguous and contradictory from the beginning.

*442In the past ten years Texas courts have seen how it works in practice. Our decision in Clewis has been cited, distinguished, explained, criticized, limited, clarified, or applied in more than 5,000 cases, eleven law review articles, and nine treatise sections.127 It has kept lawyers, judges, and paper mills busy. We have “clarified''' or “refined” Clewis in at least six cases ourselves because we did not initially explain precisely how a factual-sufficiency review differs from a legal sufficiency review. It is to those cases I now turn.

F. Attempts to Clarify Clewis.

The first clarification came in the same year that Clewis was decided. In Jones v. State,128 a capital murder case, this Court stated that a factual-sufficiency review requires the appellate court to review all of the evidence, not just the evidence which supports the verdict.129 “The appellate court reviews the evidence weighed by the jury which tends to prove the existence of the elemental fact in dispute, and compares it to the evidence which tends to disprove that fact.”130 But the reviewing court must be appropriately deferential to the factfinder,

finding fault only when “the verdict is against the great weight of the evidence presented at trial so as to be clearly wrong and unjust.” Examples of such a wrong and unjust verdict include instances in which the jury’s finding is “manifestly unjust,” “shocks the conscience,” or “clearly demonstrates bias.”131

This Court’s conscience was not shocked by the jury’s verdict in Jones. We rejected the defendant’s factual-sufficiency claim because there was conflicting evidence of whether he intended to shoot the murder victim.132

In 1997, the year after Clewis was decided, this Court held, in Cain v. State,133 that the courts of appeals must continue to give considerable deference to the original fact-finder under a factual-sufficiency review.134 In that case we reversed the court of appeals because “it was not deferential to the jury’s determination of witness eredi-*443bility.”135 Thus, the weight to be given to “contradictory testimonial evidence is within the sole province of the jury, because it turns on an evaluation of credibility and demeanor.”136 Under this deferential standard, however, the testimony of the forty nuns in the Clewis hypothetical could well be rejected because acceptance of their testimony turns on an evaluation of credibility.137

Our next major foray into an explication of Clewis was in Johnson v. State138 in 2000. There we found that the court of appeals had correctly found the evidence factually insufficient to support the aggravated sexual assault victim’s identification of the defendant because she was not “a hundred percent positive” that he was her assailant and DNA tests showed that the defendant and assailant were both members of a group consisting of 8.5% (according to the State’s expert) to 26% (according to the defense expert) of the black population.139 We stated:

The degree of deference a reviewing court provides must be proportionate with the facts it can accurately glean from the trial record. A factual sufficiency analysis can consider only those few matters bearing on credibility that can be fully determined from a cold appellate record. Such an approach occasionally permits some credibility assessment but usually requires deference to the jury’s conclusion based on matters beyond the scope of the appellate court’s legitimate concern. Unless the available record clearly reveals a different result is appropriate, an appellate court must defer to the jury’s determination concerning what weight to give contradictory testimonial evidence because resolution often turns on an evaluation of credibility and demeanor, and those jurors were in attendance when the testimony was delivered.140

After discussing both Cain and Johnson, Professor Dix, for one, concluded that “it is arguable that little meaningful difference exists between Clewis factual-sufficiency review and traditional review of the evidence for legal sufficiency.”141

In 2001, we took another stab at elarify-*444ing Clewis in Goodman v. State.142 In that case, we acknowledged that “[pjerhaps the problem is that this Court has not always been crystal clear in explaining how the courts of appeals should analyze factual-sufficiency questions in a criminal case.”143 We relied heavily upon Chief Justice Calvert’s 1960 law review article explaining “no evidence” and “insufficient evidence” standards in civil cases144 and distinguished the two types of civil factual-insufficiency zones. In zone two, “the evidence in support of the existence of a vital fact, considered standing alone, is factually too weak to support it.”145 In zone four, there is conflicting evidence, “[s]ome evidence supports a positive inference, and some evidence supports a negative inference.”146 When the conflicting evidence is in “fair equipóse,” a reviewing court may not conclude that it is against the great weight and preponderance of the evidence or that it is manifestly unjust. Only when the evidentiary scales tip “radically” toward a negative finding on an essential element may the reviewing court exercise any “thirteenth juror” role and conclude that the jury’s verdict is “clearly wrong.”147 The problem with Goodman, like that of all of our Texas factual-insufficiency cases, is that it attempted to force a square peg into a round hole. That is, the Jackson legal-sufficiency test already incorporates the State’s burden of proof and persuasion — beyond a reasonable doubt — into its analysis while the civil standard uses a zone theory of analysis to inferentially incorporate a “preponderance of the evidence” burden of proof. The zone theory may be useful for analyzing sufficiency issues for a defendant’s affirmative defense because those defenses must be proven by a preponderance of the evidence — the same standard as in Texas civil law. But it does not translate into a second-tier sufficiency review when the State must prove every element beyond a reasonable doubt and the reviewing court must ensure that there is sufficient evidence to support a reasonable, rational verdict with proof beyond a reasonable doubt.

Under Clewis, the reviewing court views all of the evidence in a neutral or impartial light, not in the light most favorable to the prosecution (or the jury’s verdict).148 This “neutral light” review seems to be the driving force behind Clewis: it circumvents Jackson’s “in the light most favorable to the prosecution” requirement. While this phrase might sound like pro-prosecution rhetoric, in fact it defers only to the factfinder’s verdict, because the only criminal trials reviewed for sufficiency of the evidence on appeal are those in which the factfinder (whether judge or jury) agreed with the prosecution. The prosecution can never appeal the sufficiency of evidence when a judge or jury acquits a defendant, regardless of how “manifestly wrong,” “unjust,” or irrational the factfin-der’s verdict may appear to be.

In our 2003 Swearingen v. State149 decision, we reiterated the importance of giving deference to the factfinder’s credibility and weight determinations, i.e., the review-*445mg court must look at the evidence in the light most favorable to the jury’s assessment of credibility and weight.

Although authorized to disagree with the jury’s determination even if probative evidence exists which supports the verdict, a reviewing court must give due deference to the fact finder’s determinations concerning the weight and credibility of the evidence and will reverse the fact finder’s determination only to arrest the occurrence of a manifest injustice. In assessing the factual sufficiency of the evidence to support a conviction we consider all the evidence in a neutral light.150

These two sentences are internally inconsistent. One cannot view the evidence in a neutral light while at the same time giving deference to the factfinder’s determinations of weight and credibility.

In our 2004 decision in Zuniga v. State,151 we again marched into the field of factual insufficiency, once more acknowledging that the evolution of factual-sufficiency review “has been somewhat confusing.” 152 In Zuniga, we further refined the *446zone four “against the great weight of the evidence” prong of factual insufficiency, and stated:

We will attempt to resolve some of the confusion created by the standard that has developed since Clewis by: 1) linking the burden of proof at trial to the standard of review and 2) avoiding language suggestive of a preponderance-of-the-evidence burden of proof.... There is only one question to be answered in a factual-sufficiency review: Considering all of the evidence in a neutral light, was a jury rationally justified in finding guilt beyond a reasonable doubt? However, there are two ways in which the evidence may be insufficient. First, when considered by itself, evidence supporting the verdict may be too weak to support the finding of guilt beyond a reasonable doubt. Second, there may be both evidence supporting the verdict and evidence contrary to the verdict. Weighing all the evidence under this balancing scale, the contrary evidence may be strong enough that the beyond-a-reasonable-doubt standard could not have been met, so the guilty verdict should not stand. This standard acknowledges that evidence of guilt can “preponderate” in favor of conviction but still be insufficient to prove the elements of the crime beyond a reasonable doubt. Stated another way, evidence supporting guilt can “outweigh” the contrary proof and still be factually insufficient under a beyond-a-reasonable-doubt standard.153

But incorporating the “beyond a reasonable doubt” standard into factual-sufficiency review returned us to the Jackson legal sufficiency standard. Compare the wording of the two standards:

* “the relevant question is whether, after viewing the evidence in the light most favorable to the prosecution, any rational trier of fact could have found the essential elements of the crime beyond a reasonable doubt”154 (legal sufficiency);

* “Considering all of the evidence in a neutral light, was a jury rationally *447justified in finding guilt beyond a reasonable doubt?”155 (factual sufficiency).

One is hard-pressed to adequately explain the distinction between these two standards of review, yet a finding of legal insufficiency leads to an acquittal, while factual insufficiency leads to a retrial. The only way that these two standards can be distinguished is by viewing — in a factual-sufficiency review — all of the evidence with divine neutrality and refusing to give any deference to the jury’s assessment of weight and credibility. This type of review explicitly makes the reviewing court a “thirteenth juror” who makes an independent, de novo determination of credibility and the weight to be given the testimony and the inferences to be drawn from the base facts. Such a review violates articles 36.13 and 38.04 of the Code of Criminal Procedure because the legislature has explicitly decreed that the jury is “the exclusive judge of the facts proved, and of the weight to be given to the testimony.”156 Therefore, I certainly agree with the Court’s decision today to overrule, or at least modify, Zuniga.

But I would go further. I, like the Florida Supreme Court did in Tibbs, would conclude that there is little historical basis in Texas for creating a factual-sufficiency review as an addition to the constitutionally mandated legal-sufficiency standard. I would further conclude that neither Clems nor its progeny adequately explain the dynamics of a factual-sufficiency review or how appellate courts are to view the evidence both “neutrally” and with deference to the factfinder.

Ironically, at the very time we were attempting to devise a workable structure by which to review the elements of a criminal offense under a civil-law standard that depended upon a preponderance-of-the-evidence burden of proof, our brethren on the Supreme Court of Texas were attempting to devise a workable standard for reviewing the sufficiency of evidence when a party has a heightened burden of proof— ie., when the burden of proof was “clear and convincing.” They, like us, had a difficult time trying to force a square peg into a round hole of factual sufficiency review. In Southwestern Bell Telephone Co. v. Garza,157 that court seemingly came to the same conclusion that I would reach: the factual conclusivity clause applies only to “questions of fact” for determining the sufficiency of evidence when the proponent bears a burden of proof by a preponderance of the evidence.158 And, after City of *448Keller, even that proposition may be doubtful.

I turn now to the question of whether the creation and application of a factual-sufficiency review by appellate judges is nonetheless good public policy that should be maintained despite its manifold problems.

III.

The creation of a factual-sufficiency review in Texas criminal cases has come at an enormous jurisprudential price. As noted above, Clewis has been cited, discussed, or applied in almost 6,000 Texas criminal cases since 1996. This Court has spent considerable time and energy in its attempts to clarify Clewis. Litigants and the courts of appeals have expended enormous resources in attempting to understand Clewis and apply the factual-sufficiency construct to the cases before them. However, as Professor Dix has noted, the courts of appeals have found the evidence factually insufficient in a minuscule number of cases.159

Nonetheless, such an enormous expenditure of judicial time and resources might be a worthwhile investment if a factual-sufficiency review were directed toward a widely acknowledged “evil” in the Texas criminal justice system, and if it solved that serious problem effectively and efficiently. Unfortunately, there is nothing in Clewis (or its progeny) that points to any systemic criminal justice problem that a factual-sufficiency review might have been designed to resolve. Although this Court set out the “forty nuns” hypothetical in Clewis to show how a Jackson legal-sufficiency review might uphold such a conviction whereas a factual-sufficiency review would presumably reverse such a conviction, I am unable to find any “forty nun” type cases. Professor Dix has cited only twelve published cases in which the courts of appeals have reversed on factual insufficiency grounds,160 and most of these were split decisions, while others were vacated by this Court because the lower court used the “wrong” Clewis analysis.

Furthermore, the “forty nuns” scenario does not pose a serious threat to the integrity of the criminal-justice system in the rest of the United States. I am aware of only two other jurisdictions, New York and the United States military courts, that currently employ a factual-sufficiency review. In each of those jurisdictions, a statute explicitly authorizes such a review.161 There is no such statute in Texas.

*449The jurisprudential evil to which a factual-sufficiency review might legitimately be a solution is the conviction of an innocent person by a biased, irrational jury. But that is precisely what the Jackson legal sufficiency review is designed to prevent. I am unaware of any evidence that this standard is insufficient protection of a defendant’s right to legally sufficient proof of every element beyond a reasonable doubt.

One of the practical problems that the Florida Supreme Court noted when it rejected its factual-sufficiency doctrine was the “temptation appellate tribunals might have to direct a retrial merely by styling reversals as based on ‘weight’ when in fact there is a lack of competent substantial evidence to support the verdict or judgment and the double jeopardy clause should operate to bar retrial.”162 Although I am confident that an appellate court would not knowingly reject a valid complaint of legal insufficiency and instead grant a factual-insufficiency claim to avoid a double jeopardy bar, there may be a natural disinclination to award an appellate acquittal to a defendant who has been convicted by a jury of twelve citizens. Given the barely visible distinction between these two standards of review, a defendant who should be entitled to a complete acquittal may be forced, because of a “compromise” appellate factual-sufficiency review, to spend the time, money, and mental anguish to undergo another trial and perhaps another jury conviction.

To the extent that Clems addresses a problem in the Texas criminal-justice system — manifestly unjust convictions that shock the conscience of the appellate courts — that problem can be resolved more satisfactorily in other ways.

First, reviewing courts must apply the Jackson legal sufficiency standard robustly, taking into account all of the evidence, although viewed in the light most favorable to the jury’s verdict.163 If that evidence supports a rational and reasonable finding of guilt beyond a reasonable doubt, it cannot be said that the jury’s verdict is manifestly unjust or shocks the conscience of the reviewing court. The verdict is either rational and reasonable or it is not; it cannot be “semi-rational” and still meet the Jackson standard. There is no jurisprudential value in reversing a rational, reasonable verdict and forcing the parties to go back and do it again.

Second, if the legislature is concerned by potential “runaway” juries that render guilty verdicts which are simultaneously rational and reasonable, yet biased and manifestly unjust, it may enact a statute mandating “factual sufficiency” review. The most appropriate “thirteenth juror”— the person authorized to second guess the jury’s credibility and weight decisions— would be the trial judge who has at least seen all of the witnesses first-hand, and judged their demeanor and credibility on a personal basis. He, much more than the distant appellate courts relying solely upon the cold written record, can take the pulse *450of the jury, the advocates, the witnesses, and the defendant. Further, he would be privy to in-chamber or side-bar discussions not available to the jurors. If there is to be a factual-sufficiency review, it ought to be pursued in a motion for new trial under legislatively enacted statutes and standards.

Third, in the very same year that this Court created “factual sufficiency” review in Clewis, it also recognized “actual innocence” claims in post-conviction writs of habeas corpus.164 Although such claims depend upon newly discovered evidence that was unavailable at the time of trial, this is certainly one mechanism to exonerate those whom a jury should not have convicted.165

Last (and least), as Judge White pointed out in his Clewis dissent, “[i]f a defendant’s verdict of guilt confining him to prison were manifestly unjust due to a failure of the sufficiency of the evidence, would not this qualify as a ground of innocence to warrant a grant of executive clemency?” 166

In sum, there are alternate methods by which the criminal justice system can address the “40 nuns” hypothetical — the problem of a rational jury which reaches a unanimous guilty verdict, based on evidence that proves every element of the offense beyond a reasonable doubt, which is, nevertheless, “clearly wrong and unjust.”

I think that the Clewis “factual sufficiency” review is internally inconsistent and denigrates the role of juries in the state of Texas by requiring a “neutral” review of the evidence, while simultaneously espousing deference to the jury’s exclusive role in assessing credibility and the weight to be given to the evidence. It has not contributed to the integrity of the appellate review process; it has led to inconsistent results; and it has required numerous, but futile, attempts to clarify its content and application. Furthermore, the dual legal- and factual-sufficiency review mandated by Clewis upended over 150 years of Texas case law that had employed a single standard for the review of sufficiency of the evidence in a criminal case.

IV.

The present case exemplifies the problems created by the Clewis factual-sufficiency review because it deals with conflicting inferences that might reasonably be drawn from the same basic facts. Twelve jurors, one trial judge, and the Chief Justice of the Waco Court of Appeals all concluded that it was a reasonable inference that appellant attempted to commit sexual assault upon Jennifer Romo after he entered her home without consent, dropped his shorts to expose his erect penis, came toward her as she grabbed a butcher knife to defend herself, and attempted to pull it out of her hands.

Two justices of the Waco Court of Appeals concluded that an even more reasonable inference was that appellant did not attempt to commit sexual assault because (1) he “did not say anything threatening or anything to indicate he intended to sexually assault [Jennifer] Romo”; (2) “his conduct in grabbing the knife could also be viewed as an effort to defend himself’; (3) “he did not flee the apartment complex”; and (4) Jennifer’s father “testified that her emotional state at the time was one primarily of anger and confusion rather than *451fear.”167 All of these basic facts are true. A rational trier of fact might well conclude from these facts that appellant was not guilty of attempting to sexually assault Jennifer Romo. This jury did not do so. It rejected this inference and instead concluded, from the totality of the basic facts proven, that appellant did attempt to commit sexual assault. Drawing inferences from basic facts to the elements of the offense is the jury’s job as long as those inferences are rational and reasonable. The mistake that Clewis invites is the appellate court’s substitution of what it considers “a more reasonable” inference from the basic facts than the reasonable inference that the jury did, in fact, draw.

The issue of whether appellant attempted to sexually assault Jennifer Romo was well presented by both the State and defense in this case. Both advocates cogently argued their respective positions that the evidence did or did not support a finding, beyond a reasonable doubt, that appellant intended to sexually assault Jennifer Romo and had attempted to do so. Appellant emphasized that Jennifer Romo was “mad,” not “scared” by appellant’s actions, and she “had no idea what [appellant’s] intent was.” He noted that appellant never touched Jennifer or her clothes. The jury rejected the inferences that appellant asked it to draw. It chose to agree with the inferences drawn from this evidence that support the conclusion that appellant did, in fact, attempt to sexually assault Jennifer Romo. Two members of the court of appeals agreed with appellant’s argument. But that agreement does not render the evidence either factually or legally insufficient, nor does it mean that the jury’s verdict was the product of bias or the result was manifestly unfair. Factual disagreements and disputes lie at the heart of our jury system.

We are most likely to invoke the collective wisdom of a jury when the basic facts are in dispute or when the logical inferences to be drawn from base facts conflict. Trials are generally unnecessary in the “easy” cases, the “whale in a barrel” cases; they generally result in a plea bargain. It is when the direct testimony conflicts or the logical inferences that may be drawn from the circumstantial evidence differ, that a jury is needed to resolve those conflicts. However, it is precisely in those cases that an appellate court is most likely to second guess a jury, but it does so without considering whether the jury’s credibility and weight decisions were made on the basis of irrational bias and prejudice as opposed to rational deduction or induction. The fact that a jury chose one rational alternative as opposed to another rational alternative is not a basis for reversal, but that is precisely what the Clewis standard permits. It permits a fortuitous “lightning strikes” reversal based on whether an appellate court (without personally seeing or hearing the witnesses) differs with a jury on its choice of what inference should be drawn or what weight should be given to certain direct testimony. Although reversals under Clewis are rare, they are also random, inconsistent, and based primarily on “the luck of the draw.”

I would therefore overrule Clewis and all of its progeny, including Zuniga. Because the majority lops off only the most recent branch of Clewis and does not address the root of the problem, I most respectfully dissent.