Saucedo pleaded guilty to possession of methamphetamine, but post-plea laboratory testing has shown that the substance he possessed was actually methylethcathinone. Saucedo argues that, in light of this new evidence, his guilty plea was involuntary. Under the Court's holdings in Ex parte Palmberg and Ex parte Broussard , I cannot agree that Saucedo's plea was involuntary. However, I do think that his conviction violates due process on other grounds. I write separately to explain my rationale.
I. FACTS
Saucedo was initially charged with two felony offenses: first-degree felony possession of a controlled substance, over 400 grams of diluted codeine; and second-degree felony possession of a controlled substance, between 4 and 200 grams of methamphetamine. At his first court appearance on January 27, 2014, Saucedo pleaded guilty to these charges and, pursuant to the terms of a plea-bargain agreement, was placed on a six-year term of deferred-adjudication probation. At the time he entered his plea, no confirmatory laboratory testing had been done on any of the substances Saucedo possessed.
On March 21, 2014, what was then called the Houston Police Department (HPD) Crime Laboratory prepared a report identifying some, but not all, of the substances from Saucedo's cases. The HPD Crime Laboratory reported that the substance supporting Saucedo's second-degree felony charge, initially thought to be methamphetamine, was in fact "methylethcathinone," a "derivative of 2-aminopropanal." There were 9.18 grams of this substance in the sample analyzed. Methamphetamine is a penalty group one substance, and possession of 4-200 grams of a penalty group one substance is a second-degree felony.1 Methylethcathinone is a penalty group two substance, and possession of 4-400 grams of a penalty group two substance is also a second-degree felony.2
On January 26, 2017, Saucedo's probations were revoked, and he was sentenced to ten years' imprisonment on each charge. These sentences were ordered to run concurrently. The habeas judge made a finding that, both at the time of Saucedo's plea and at the time his probation was revoked, Saucedo was unaware of the lab report indicating that the substance he possessed was actually methylethcathinone.
On June 14, 2017, the Houston Forensic Science Center (HFSC) prepared another report pertaining to Saucedo's cases. In this report, the HFSC reported that the substance supporting Saucedo's first-degree felony charge, initially thought to be codeine, was in fact promethazine. Promethazine is not within any of the penalty groups in the Controlled Substances Act; it is, at most, a dangerous drug.3 Possession *713of any amount of a dangerous drug is a Class A misdemeanor offense.4
Shortly thereafter, Saucedo contemporaneously filed two applications for writs of habeas corpus in the convicting court. In his first, -01 writ, Saucedo argued that (1) his guilty plea as to possession of codeine was involuntary under Ex parte Mable ,5 and (2) his conviction violated due process because, "[i]f Applicant committed an offense, it is not the specific offense Applicant was accused of and convicted of violating." We granted relief in the -01 writ, expressly citing Mable .6 Whether to grant relief from Saucedo's conviction for possession of methamphetamine is the issue before us in this -02 writ.
II. LAW
A. Due Process, Involuntary Pleas, and Ex parte Mable .
If a guilty plea is not "voluntary and knowing, it has been obtained in violation of due process and is therefore void."7 For the plea to be voluntary and knowing, the accused must have a "sufficient awareness of the relevant circumstances and likely consequences" of his plea.8 To meet this standard, "the defendant must have an actual awareness of the nature and gravity of the charges against him and of the constitutional rights and privileges that he necessarily relinquishes-in short, 'a full understanding of what the plea connotes and of its consequences.' "9
Applying these principles in controlled-substance cases in which the defendant pleads guilty before the identity of the substance has been confirmed by laboratory testing has been a unique, recurring challenge for this Court-one that we first confronted head-on in Ex parte Mable .10 Mable pleaded guilty to possession of a controlled substance, but a post-plea laboratory analysis indicated that the substance he possessed "did not actually contain any illicit materials."11 In light of this newly acquired evidence, Mable claimed that he was "actually innocent,"12 but we disagreed. We noted that, under State v. Wilson , the phrase "actual innocence" applies "only in circumstances where the accused did not actually commit the charged offense or any possible lesser included offenses."13 We concluded that it was possible, even in light of the post-plea laboratory report, that Mable had "attempted to possess a controlled substance," and had thereby committed "a lesser included offense[ ] of possession."14
Nevertheless convinced that Mable was entitled to relief, we opted instead to invalidate his plea as insufficiently knowing and voluntary.15 We noted that "all parties involved, *714including the applicant, incorrectly believed the applicant had been in possession of drugs"-a fact that we described as "crucial."16 "[W]hile operating under such a misunderstanding," we said, "the applicant cannot be said to have entered his plea knowingly and intelligently."17
B. Mable 's progeny.
Mable quickly invited a spate of litigation. What other sorts of adjudicative facts, known-to-be-unknown at the time of the plea, might later be deemed "crucial" to the guilty-plea process, such that miscalculating them would render the plea invalid?
In Ex parte Palmberg , just as in Mable , the defendant pleaded guilty to possessing a controlled substance before the laboratory had completed an analysis of the substance he possessed.18 The lab eventually informed the parties, post-plea, that "there was no unprocessed sample left over for the laboratory to analyze," the arresting officer having used up all of the substance found on Palmberg for field testing.19 While Palmberg sought relief under Mable , we considered his situation to be "factually distinguishable" from Mable's.20 While Mable pleaded guilty under a misapprehension that was ultimately "crucial" to his case, Palmberg had merely "overestimated the State's ability to ... prove he was guilty in the absence of his judicial confession."21 We rejected the idea that Mable requires the defendant to "kn[o]w every fact relevant to the prosecution of his case" before he may enter a voluntary plea.22
In Ex parte Broussard , the defendant pleaded guilty to delivery of cocaine, but laboratory tests later determined that the substance he delivered was actually methamphetamine.23 Broussard claimed that his plea was involuntary under Mable , but we again disagreed. We reiterated that "[e]very defendant that pleads guilty does so with the implicit understanding that conviction at trial is never certain."24 So, we reasoned, a defendant should not ordinarily be allowed to withdraw his plea "simply because his good-faith evaluations of the facts turned out to be incorrect."25 Because Broussard failed to articulate any "reason why he would not have accepted the plea bargain other than the bare fact that the illicit substances were different," we denied relief.26
III. ANALYSIS
A. Mable should be overruled.
We ordinarily follow the doctrine of stare decisis in order to "promote judicial efficiency and consistency, encourage reliance upon judicial decisions, and contribute to the integrity of the judicial process."27 But these laudable interests are not advanced when we continue to follow a precedent that was either poorly reasoned from the outset or that has, in the ensuing years, proven itself "unworkable."28 In those situations, it is appropriate for the *715Court to overrule its prior precedent and find a better way going forward.
I. Mable was poorly reasoned.
In Mable , the Court's desire to resolve the case on involuntary-plea grounds was clearly based on the belief that it simply could not grant relief on Mable's claim of actual innocence.29 This belief was, in turn, based upon language from our opinion in State v. Wilson , wherein we said that "the term 'actual innocence' shall apply ... only in circumstances in which an accused did not, in fact, commit the charged offense or any lesser-included offenses."30
There were two problems with this approach. In the first place, Mable never claimed that his plea was involuntary. In his writ application, Mable grounded his claim for post-conviction habeas relief on a single legal theory: that he was "actually innocent" of possessing a controlled substance. The closest Mable came to arguing involuntariness was an observation that a claim of actual innocence "is much like a traditional involuntary plea" claim.31 But that is a far cry from Mable claiming that his own plea was coerced or that he did not understand, when he pleaded guilty, the consequences of his plea or the rights that he was necessarily relinquishing. Although the Court's instinct to invalidate Mable's conviction as inconsistent with due process was correct, for reasons that I explain below,32 involuntariness was clearly not the fundamental due-process defect that Mable asked us to cure. The Court should not have reached that issue sua sponte .
Secondly, and more importantly, Mable ' s description of what Wilson held was woefully incomplete. In Wilson , the appellee pleaded guilty to felony DWI.33 He later came to realize that, at the time he pleaded guilty, one of his prior DWI convictions was not eligible for use as a jurisdictional enhancement. In a post-conviction habeas application under Code of Criminal Procedure Article 11.072, Wilson complained that he was "actually innocent" of felony DWI, and both the trial court and court of appeals agreed.34
We disagreed that Wilson's situation could be characterized as "actual innocence," holding that that nomenclature should apply "only in circumstances in which an accused did not, in fact, commit the charged offense or any of the lesser-included offenses."35 That was the only passage that Mable quoted from Wilson , but Wilson said much more:
In cases such as this one, in which the issue is the offense of which [the applicant] is, in fact, guilty, thus implicating the legality of his sentence, the appropriate terms are "guilty only of" a lesser-included offense and "ineligible for" the sentence assessed. For the purpose of this case, and other pending cases in which the issue is the offense of which the accused is, in fact, guilty, the sentence assessed, or both, we will interpret a claim of "actual innocence" to mean "guilty only of" a lesser-included offense or "ineligible for" the sentence assessed, or both.36
*716Notably, we went on to affirm the lower courts' decisions granting Wilson habeas relief, despite the fact that Wilson was claiming actual innocence in a situation where, by virtue of his guilty plea, he was still guilty of a lesser-included offense-misdemeanor DWI. And the remedy we afforded was not simply reforming the judgment to reflect a misdemeanor conviction and ordering Wilson re-sentenced. We affirmed the lower courts' decisions vacating Wilson's felony conviction altogether, at which point he was presumably subject to re-prosecution for the misdemeanor offense.
Mable did not explain any of this. Consistent with Wilson , we should not have simply rejected Mable's claim of actual innocence and then proceeded to address a due-process defect that Mable never complained about. Instead, we should have construed Mable's claim of actual innocence as a claim that he was "guilty only of" a lesser-included offense or that he was factually "ineligible for" the punishment he received. Under either of these theories, Mable's claim was self-evidently meritorious. Furthermore, the remedy we would have afforded Mable under this approach would have been exactly the same as the remedy we afforded him anyway: His plea would have been withdrawn, and he would have been returned to the district court to answer the charges as set out in the indictment.37
Mable is also out-of-step with the United States Supreme Court's description of voluntariness in the plea-bargain context. In United States v. Ruiz , the Supreme Court was asked to decide "whether the Constitution requires ... disclosure of impeachment information" before a plea agreement may be entered into.38 In concluding that it does not, the Court noted that "the law ordinarily considers a waiver knowing, intelligent, and sufficiently aware if the defendant fully understands the nature of the right and how it would likely apply in general in the circumstances-even though the defendant may not know the specific detailed consequences of invoking it."39 Thus, a defendant may validly "waive his right to remain silent, his right to a jury trial, or his right to counsel even if the defendant does not know the specific questions the authorities intend to ask, who will likely serve on the jury, or the particular lawyer the State might otherwise provide."40
In like fashion, so long as a criminal defendant "fully understands the nature" of his right to plead not guilty, the fact that he was unaware that the State's evidence would ultimately fail to prove his guilt at trial means only that the "specific detailed consequences" of waiving that right eluded him.41 As Ruiz illustrates, the Constitution does not consider this ignorance of particulars to be so odious as to invalidate the "grave and solemn act" of a guilty plea.42 On the contrary, "the Constitution, in respect to a defendant's awareness of relevant circumstances, does not require complete knowledge ... but permits a court to accept a guilty plea ...
*717despite various forms of misapprehension under which a defendant might labor."43
Even to the extent that Ruiz draws a distinction between mere "impeachment information" and exculpatory evidence,44 we are not dealing with a situation where state actors knowingly suppressed exculpatory evidence. In each of these cases- Mable , Palmberg , Broussard , and this one-the State, defense counsel, and trial court all believed that the evidence would ultimately show that the defendant was guilty of the charged offense. There was, then, no fault to be found in the actions of any of those institutional entities. If there had been, I would be much more open to the idea that the defendant's plea was involuntary.45 Instead, we are dealing with a plea proceeding that could rightly be described as "error-free"46 -yet another reason why claims of this nature are better analyzed under an actual-innocence framework rather than a voluntary-plea framework. After all, an "error-free" proceeding is the precise situation that a bare claim of innocence was designed to account for.47
By way of explaining the claim that there are "various forms of misapprehension" under which a defendant might permissibly labor, Ruiz noted the case of McMann v. Richardson .48 In McMann , the defendant had pleaded guilty under the law then existing and, some time after the plea, the law changed so as to make his plea bargain less advantageous than he initially planned.49 McMann claimed that this change in the law rendered his plea involuntary, but the Supreme Court concluded that these kinds of subsequent legal developments do not vitiate the voluntary nature of an otherwise lawful guilty plea.50 Why? Because "when the defendant waives his state court remedies and admits his guilt, he does so under the law then existing; further, he assumes the risk ... in either his or his attorney's assessment of the law and facts."51 We ourselves recently, unanimously, and rightly reaffirmed this constitutional analysis.52
Mable , then, would seem to allow for an illogical double-standard within our plea jurisprudence: Subsequent legal developments cannot retroactively render a defendant's guilty plea involuntary, but subsequent factual developments can. In other words, whether a guilty plea was voluntarily entered depends on (1) the law in existence at the time of the plea, and (2) the facts in existence at the time the defendant *718seeks to rescind his plea. I can think of no good reason why this should be.
Perhaps the Court's decision to retain this dual standard would be understandable if the Supreme Court had yet to say whether subsequent factual developments can retroactively render an otherwise-voluntary guilty plea involuntary. The problem is that the Supreme Court has weighed in on this issue-and found, contrary to Mable , that they cannot. In Puckett v. United States , the Supreme Court was asked to decide whether, "[w]hen the government breaks a promise that was made to a defendant in the course of securing a guilty plea, the knowing and voluntary character of that plea retroactively vanishes, because (as it turns out) the defendant was not aware of its true consequences."53 Its response was telling:
Although the analogy may not hold in all respects, plea bargains are essentially contracts. When the consideration for a contract fails-that is, when one of the exchanged promises is not kept-we do not say that the voluntary bilateral consent to the contract never existed, such that it is automatically and utterly void; we say that the contract was broken. The party injured by the breach will generally be entitled to some remedy, which might include the right to rescind the contract entirely, but that is not the same thing as saying the contract was never validly concluded.54
All of the reasons Puckett gives for concluding that plea-bargain breaches do not retroactively render a defendant's plea involuntary apply with equal force to other kinds of subsequent factual developments. Under general contract principles, "[e]rror in predicting a future fact known to be uncertain is not the kind of mistake which will relieve a party from a contract."55 To keep our jurisprudence in harmony with these common-law contract principles,56 we must reject Mable 's assumption that certain facts are so "crucial" to the guilty-plea process that a defendant can only validly speculate about them if his speculations ultimately prove correct.
Consistent with Puckett , as we shall soon see, there may yet be a due-process remedy available to a person who can demonstrate with new evidence that his conviction is unsupportable, and that remedy may sometimes include rescinding the plea agreement. "[B]ut that is not the same thing as saying the contract was never validly concluded."57 To the extent that the Court continues to insist otherwise-that, not only are pleas such as these unsupportable in their current state, they were never lawfully entered in the first place-it renders our plea jurisprudence both internally inconsistent and at odds with binding Supreme Court precedent.
ii. Mable has become unworkable.
For a number of reasons, Mable has also become unworkable. First, the Court's voluntary-plea *719jurisprudence following Mable has begun to undermine Mable 's central thesis. Palmberg held that, even if a defendant is ignorant of certain facts, he may still enter a knowing and voluntary plea, despite that ignorance, so long as he is aware that those facts represent "an unknown variable in his prosecution."58 In other words, his ignorance of facts he "knew ... he did not know" should not invalidate his otherwise voluntary decision to plead guilty.59 Likewise, Broussard held that a defendant cannot "withdraw his plea simply because his good-faith evaluations of the facts turned out to be incorrect."60
These holdings, applied to the facts of Mable , would have counseled against the conclusion that Mable's plea was somehow unknowing or involuntary. Like Palmberg and Broussard , Mable was aware, at the time he pleaded guilty, that the identity of the substance he possessed had not yet been confirmed. He "knew ... he did not know" whether the State would be able to prove its case at trial, and he made a conscious, uncoerced decision to plead guilty despite his ignorance in that regard.61 Under the logic of Palmberg and Broussard (to say nothing of Brady , McMann , Ruiz , Puckett , etc. ), Mable's plea was both voluntary and adequately informed.
I suppose the argument could be made that, to the extent that Palmberg and Broussard conflict with Mable , Mable ought to control, and those opinions ought to be abandoned-if for no other reason than that Mable came first. The problem with this argument is that, as I have already suggested, Palmberg and Broussard are more consistent with Supreme Court precedent than Mable is. This should come as no surprise. Palmberg went to great lengths to explain its holding in reference to the Supreme Court's plea jurisprudence,62 as did Broussard .63 Mable , in contrast, contained very little in the way of careful legal analysis.
Second, because of its inherent malleability, Mable has produced, and will likely continue to produce, blatantly contradictory outcomes. For example, in Ex parte Brooks , we allowed an applicant to withdraw his guilty plea because "he was not made aware of credibility issues pertaining to certain State's witnesses."64 We expressly cited Mable in support of that decision. This despite the fact that, in Ruiz , the Supreme Court had already rejected the idea that "the Constitution requires ... disclosure of impeachment information" before a plea may be voluntarily entered into.65 Perhaps in recognition of this obvious deviation from binding Supreme Court precedent, six months later, in Ex parte Sneed , we denied relief on identical facts-and cited Ruiz .66 We did so over a dissenting *720opinion urging us to follow Mable .67 The fact that Mable can reasonably be cited in support of a holding that directly contradicts binding Supreme Court precedent ought to amply demonstrate its unworkability.
Because Mable was poorly reasoned and because it established what I consider to be an unworkable standard, I would take this opportunity to overrule it.68 I would hold that if a person pleads guilty knowing that a certain fact is as-yet unknown, then in the absence of any fraud, coercion, misconduct, or ineffective assistance of counsel, he may still render a knowing and voluntary guilty plea-even if subsequent factual developments ultimately show that the plea was inaccurate or untrue. On this analysis, Palmberg and Broussard would both remain good law.
B. Wilson remains good law.
I hasten to add that I also think Wilson is still good law. So if an applicant can demonstrate that he is "guilty only of" a lesser-included offense or that he is factually "ineligible for" a particular sentence, he might still be allowed to withdraw his plea of guilty.69 But if an applicant obtained relief in this way, it would not be because, in light of the new evidence, his plea was somehow retroactively rendered unknowing or involuntary.70 Neither would it be because the applicant demonstrated that he was "actually innocent," with all the additional trappings and baggage that accompany that nomenclature.71 Rather, it would be because due process simply will not tolerate punishing as a felon a person who, by every objective measure of the law and facts, committed only a misdemeanor.72 Neither will it tolerate punishing as a first-degree felon one who committed at most a second-degree felony; as a second-degree felon one who committed at most a third-degree felony; and so forth. These are the basic premises behind a post-conviction claim of "illegal sentence."73
To be sure, even Wilson noted the interrelatedness of "guilty only of" claims and claims of actual innocence.74 Naturally, then, a claim for post-conviction relief under Wilson should function much like a claim of actual innocence: The applicant must show, by clear and convincing evidence, that no reasonable juror would have convicted him of the greater offense, or made findings consistent with the heightened sentence, in light of the new evidence.75 That being said, although a Wilson claim is functionally similar to a claim of actual innocence, I would not deem an applicant who met the Wilson criteria to be "actually innocent." I would prefer to call a person in that situation "actually not guilty" of the charged offense, rather than "actually innocent" of any offense.76
*721C. Although Saucedo's plea was not involuntary, he is nevertheless entitled to relief under Wilson .
In this case, Saucedo claims that, under Mable , his plea of guilty to the offense of possession of methamphetamine was involuntary because a subsequent lab report showed that the substance he possessed was actually methylethcathinone. Because I would overrule Mable , I need not decide whether the difference between possessing methamphetamine and possessing methylethcathinone is a "crucial" fact, such that misapprehending it would retroactively render Saucedo's guilty plea involuntary.77 I need only observe that Saucedo pleaded guilty with the knowledge that the identity of the substance he possessed was as-yet unconfirmed. So, following Palmberg and Broussard , Saucedo "cannot now invalidate his guilty plea because his prediction of the State's ability to prove he possessed [methamphetamine] was incorrect."78 There is no suggestion that Saucedo was fraudulently misled or coerced into pleading guilty or that his plea counsel was ineffective. The record shows that Saucedo was duly advised of the consequences of a guilty plea and of the rights he was necessarily relinquishing by pleading guilty. Under these circumstances, Saucedo's plea was knowing and voluntary.79 That claim should be denied.
In a separate ground for relief, Saucedo also claims that his "right to due process" was violated because, "[i]f Applicant committed an offense, it is not the specific offense Applicant was accused of and convicted of violating." I would construe this claim in accordance with Wilson : Saucedo claims that, in light of new evidence, he was "guilty only of" possessing methylethcathinone, and that a judgment reflecting a conviction for possession of methamphetamine violates due process.
I agree with Saucedo that this conviction violates due process. No rational juror, weighing the lab report indicating "methylethcathinone" against the charge of possession of methamphetamine, would have found Saucedo guilty of the latter. In relation to the charged offense, the most the jury could have found is that, under Saucedo's mistaken understanding of the facts, he attempted to possess methamphetamine. That offense is one degree lower than the offense of which he was convicted.80 In this regard, as presently charged, Saucedo is factually ineligible for second-degree felony punishment.
It is also in some sense true that, in light of this new evidence, Saucedo is still guilty of an offense that is co-equal to the one he was charged with and convicted of. Possession of 4-400 grams of a penalty group two controlled substance is, like possession of 4-200 grams of a penalty group one substance, a second-degree felony. Saucedo judicially confessed to possessing just over 9 grams of a controlled substance. In this sense, Saucedo is still theoretically eligible for second-degree felony punishment, because he is still in theory guilty of possession of a penalty group two controlled substance.
But no institutional trial-level fact finder, be it judge or jury, has yet passed upon *722Saucedo's guilt or innocence of the separate and distinct offense of possession of methylethcathinone. For us to take judicial notice, or else simply declare, that the facts demonstrate that Saucedo is guilty of that offense as a basis for denying him relief from his conviction for possession of methamphetamine would be to bypass, among many other things, his Sixth-Amendment right to a jury trial. I would not have the Court go down that road.
So, while I acknowledge that Saucedo is theoretically eligible for second-degree felony punishment for the offense of possession of methylethcathinone, to turn that theoretical possibility into a reality, the State must charge Saucedo with that offense and prosecute him for it. I can think of no impediment to the State doing so in this case. But I would not allow Saucedo to remain convicted of an offense that he plainly did not commit simply because there is evidence to suggest that, independent of that offense, he also committed another.
IV. CONCLUSION
I realize that the framework I have proposed today will, in the vast majority of cases going forward, lead to the same bottom-line results as Mable . But every judge should care, in every case, not only that the right result be reached, but that it be reached for the right reasons. Confronted with a knotty legal issue, a careful, reasoned untangling is what's expected of us-not simply that, for any reason or no reason at all, the knot be cut.
I would deny Saucedo's claim that his guilty plea was involuntary. Instead, I would hold that Saucedo is entitled to relief because he has established, by clear and convincing evidence, that no rational juror would have found him guilty of possession of methamphetamine in light of this newly available lab report. Finally, although I would allow Saucedo to withdraw his plea as inconsistent with due process, I would not declare him actually innocent.
With these thoughts, I concur in the Court's judgment.
Hervey, J., filed a concurring opinion in which Keasler, J., joined.
I wholeheartedly agree with Judge Keasler that Mable has proven itself poorly reasoned over time and that it should be overruled. I also agree with the new analytical framework that he proposes. I write separately, however, to point out that the moral of this long-running story is that defense attorneys and prosecutors should not allow a defendant to plead guilty until the laboratory report comes back. Defense attorneys might be subject to ineffective-assistance-of-counsel claims, and the State might have to re-prosecute defendants.
With these comments, I join Judge Keasler's concurring opinion and concur in the judgment of the Court.
Newell, J. filed a concurring opinion in which Richardson, Walker and Slaughter, JJ., joined.
This is another case where an applicant pleaded guilty to possession of a controlled substance before laboratory testing on the substance was completed. I agree with the Court that Applicant is entitled to relief because the testing now shows he did not possess the controlled substance the State alleged he had possessed. In these types of situations, it should be enough that the State, the applicant, and the habeas court all agree that the applicant is entitled to relief. I write separately to address the suggestion of overruling Ex parte Mable.1
*723In Mable , we held that a guilty plea to possession of a controlled substance was involuntary when the applicant did not know that the seized substances contained no illicit materials.2 We reasoned that this "fact [wa]s crucial to th[e] case, and while operating under such a misunderstanding, the applicant cannot be said to have entered his plea knowingly and intelligently."3
We have previously said that the Court should not frivolously overrule established precedent.4 "We follow the doctrine of stare decisis to promote judicial efficiency and consistency, encourage reliance on judicial decisions, and contribute to the integrity of the judicial process."5 Overruling precedent, however, is acceptable under certain circumstances when the goals of stare decisis would not be achieved.6 But those special circumstances are not present in the aftermath of Mable , and there is no reason to believe that adhering to Mable fails to achieve the goals of stare decisis.
The suggestion for overruling Mable seems to flow from Ex parte Broussard , where we held that, in a possession-of-a-controlled-substance case, the identity of the controlled substance isn't a crucial fact.7 But we ignored our prior cases Watson v. State and Mable , both of which held that the identity of the controlled substance is an essential element of the offense of possession of a controlled substance.8 We are better served by following that precedent than by jerry-rigging a different test that will ultimately result in the same relief as a straight-up application of Mable.9
To that end, I see no reason to overrule Mable. It is not enough to simply say that Applicant "knew that he did not know" the identity of the substance he possessed because that is not the whole picture. In these types of cases, both the applicant and the State are operating on the assumption that the identity of the substance possessed is exactly what the State has alleged. This situation is distinguishable from that in Ex parte Palmberg.10 In Palmberg , the fact that there was no substance left to be tested did not undermine both parties' assumption that Palmberg had possessed cocaine. In that sense, the lack of remaining evidence was not a crucial fact. But in Mable , the affirmative evidence showing that the substance possessed was not a controlled substance at all made it clear that both Mable and the State had wrongly assumed that Mable *724had actually possessed a certain controlled substance.11
The same holds true in this case. Applicant and the State both operated under the assumption at the time of the plea that Applicant had possessed methamphetamine, but later testing made it clear that both parties were mistaken. Applicant had actually possessed methylethcathinone.
The outlier, therefore, is not Mable -it's Broussard.12 There is no meaningful distinction between the facts of this case and those in Broussard.13 In Broussard , both the State and Broussard operated under the assumption at the time of the plea that Broussard had delivered cocaine, but subsequent testing revealed that he had actually delivered methamphetamine.14 And yet, we distinguished Broussard from Mable by holding that the parties' mistake about the identity of the controlled substance did not render the plea involuntary.15 If we are going to get rid of one of these cases, it should be Broussard rather than Mable.
Discovering new information crucial to the plea bargain can retroactively invalidate the underlying agreement. Under the well-established doctrine of mutual mistake, a contract can be rescinded "[w]here a mistake of both parties at the time the contract was made as to a basic assumption on which the contract was made has a material effect on the agreed exchange of performances."16 When both parties enter into a plea bargain while awaiting lab results, both parties operate on the assumption that the testing will confirm what the State has alleged. Even if the parties are "rolling the dice," they are both betting on the same outcome based upon that assumption. If affirmative evidence reveals that they were both mistaken about the underlying assumption of the identity of the controlled substance, the plea bargain may be involuntary because of that mutual mistake. Given the inability to rely on sight alone to accurately identify most controlled substances17 and the State's control of the investigation, analysis, and charging processes, it is appropriate to treat the plea as involuntary based on the mutual mistake about the identity of the controlled substance at issue.
I also disagree that Mable is unworkable. It provides a clear ruling that is easy to follow: when laboratory testing subsequent to a plea establishes that a defendant did not possess the drug he was alleged to have possessed, he gets his plea back. We are the ones who have continued to file and set these cases even though the *725State and the applicants repeatedly agree (and make their agreements clear to the habeas courts and this Court) that relief is warranted under Mable. We can quite easily apply Mable here to determine the outcome.18 Just because some don't want to doesn't make Mable unworkable.
What will be unworkable is holding that the identity of the controlled substance is not an essential element of a possession-of-a-controlled-substance offense. If that were so, a defendant who possesses multiple different controlled substances that fall within the same penalty group would be guilty of only one offense. Would the State have to prove that the defendant knew which penalty group the possessed substance belongs to? And would the State ever even need to allege the specific controlled substance? I see nothing wrong with holding the State to the burden of proving the specific identity of the controlled substance it alleges in the indictment.
In this case, the State and Applicant agree that relief is warranted under Mable. The trial court agrees and recommends that we grant relief. Rather than overrule Mable and apply a due-process framework that will nevertheless invariably lead to the same results as Mable , I would hold that the identity of the controlled substance at issue is a material or "crucial" fact underlying the plea bargain and give the parties what they ask for. In light of this understanding, I believe Applicant's plea was involuntary because both parties in this case were mistaken at the time of the plea about a material fact-the identity of the controlled substance possessed. With these thoughts, I concur.
Keller, P.J., filed a dissenting opinion in which Keel, J., joined
Applicant knowingly possessed a controlled substance. He knew at the time of his plea that the substance had not yet been tested. But the State was giving him a great deal-six years' deferred adjudication-so he pled guilty. Testing on the substance was complete less than two months after the plea,1 but for almost three years, Applicant did nothing to challenge his conviction. Now that his guilt has been adjudicated and he has been sentenced to ten years, Applicant wants a do-over, and the Court gives him one. There are three reasons this is a mistake.
First, under Ex parte Broussard , Applicant has not shown that his guilty plea was involuntary.2 The testing did not disprove Applicant's guilt of possessing an illegal drug; it just proved that he possessed a different illegal drug than the one charged.3
Second, Applicant has not shown harm. Under the doctrine of transferred intent, he is still guilty of culpably possessing the drug that the testing revealed, and the punishment range for that drug is the same as the one with which he was charged. Consequently, Applicant is not actually innocent nor has he been subjected to a harsher punishment range than the offense that he actually committed would support.
The general rule in post-conviction habeas proceedings is that the convicted person *726has the burden to show harm.4 To understand how harm could be shown in a case such as this, we begin with the fact that Applicant's conviction was the result of a guilty plea. A guilty plea constitutes an admission of guilt to the charged crime.5 In pleading guilty to possession of methamphetamine, Applicant admitted to all of the elements of that offense. His claim now is that he is not guilty of possession of methamphetamine because the testing conclusively establishes that he possessed a different drug-methylethcathinone. This claim does not satisfy the requirements for showing actual innocence because Applicant could still be guilty of the lesser-included offense of attempted possession of methamphetamine.6
Applicant's guilty plea to the methamphetamine offense and his habeas pleading and proceeding necessarily establish all the elements of the methylethcathinone offense. Under those circumstances, Applicant has effectively admitted his guilt of the methylethcathinone offense.
Base drug-possession offenses7 ordinarily contain four elements: (1) the possession of a substance, (2) the amount of that substance, (3) the identity of that substance as a particular drug, and (4) the culpable mental state for possession of the drug.8 The possession of a substance and the amount of that substance were admitted by Applicant at his guilty plea, and he does not now challenge those admissions. Instead, the claim is that the substance possessed was a different drug than the one alleged, namely methylethcathinone, instead of methamphetamine. So Applicant's guilty plea to the methamphetamine offense necessarily satisfied elements (1) and (2) of the methylethcathinone offense.
As for element (3)-the identity of the substance as methylethcathinone-that element was necessarily satisfied by Applicant's habeas pleading and this habeas proceeding. To even make his claim on habeas corpus, Applicant has to concede that the substance he possessed was methylethcathinone. He cannot claim, for example, that the chemist tested the wrong substance or that the testing procedure was flawed, because such claims would undermine his contention that the testing shows he did not possess methampethamine. Applicant cannot use the identity of that substance when it operates in his favor but call its identity in question when it operates against him. And consistent with Applicant's habeas pleading, the evidence in this habeas proceeding shows the substance to be methylethcathinone.
That leaves element (4), the culpable mental state. We have said in the past that, for a possession-of-controlled-substance offense, the State is required to prove that the accused "knew the matter [possessed] was contraband."9 If that is *727literally all the State is required to prove, then Applicant's guilty plea to intentionally and knowingly possessing methamphetamine satisfies the culpable-mental-state element, and all four elements have been established.
But methamphetamine and methylethcathinone fall within different penalty groups,10 and it could be argued that the culpable mental state applies at least to the penalty group that the substance is in. If that were so, then the intent or knowledge with respect to a drug in penalty group 2 might differ from the intent or knowledge with respect to a drug in penalty group 1.11 But this difference ultimately does not matter because of the law of transferred intent.
The Penal Code contains two "transferred intent" doctrines, but the one applicable here involves the transfer of intent from one offense to another:
A person is nevertheless criminally responsible for causing a result if the only difference between what actually occurred and what he desired, contemplated, or risked is that ... a different offense was committed.12
We have applied this version of the "transferred intent" doctrine to drug offenses.13 Consequently, Applicant's culpable mental state with respect to methamphetamine, admitted to by his guilty plea, transfers to the methylethcathinone offense.
Because Applicant has effectively admitted, through his plea and his current pleadings, to all the elements of possession of methylethcathinone, and because that offense has the same punishment range as the methamphetamine offense to which he pled guilty, Applicant has suffered no harm.
Finally, Applicant should be estopped from challenging his plea because he waited to do so until his guilt was adjudicated. A person who receives probation as a part of a plea agreement should be estopped from challenging the validity of his plea when he enjoys the benefits of being on probation and raises his challenge only after probation is revoked.14 Despite his allegation that he would not have pled guilty if he had known the relevant circumstances, Applicant enjoyed the benefit of his plea bargain for almost three years after the substance was tested.15 The habeas *728court found that Applicant was never made aware of the lab report after his initial plea or during the motion to adjudicate. Even so, Applicant could have inquired as to the result of testing, and had he done so earlier, could have filed a habeas application under Article 11.072, while he was still on probation.16 Granting relief under such circumstances creates a windfall for defendants who have nothing to lose by laying behind the log. I would hold that Applicant's claim is barred by estoppel.17
For these reasons, I would deny relief.