filed this dissenting opinion in which JOHNSON, J., joined.
OPINION
I dissent to the majority’s conclusion that Bailey’s prosecution is not barred by double jeopardy. (For simplicity’s sake, I will refer to the Appellants as “Bailey,” even though Brenda Sue Bailey was only one of the defendants.)
I.
The Double Jeopardy clause of the Fifth Amendment “protects against a second prosecution for the same offense after acquittal.” 1 As the majority notes, the issue in this case is whether this prosecution of Bailey, naming the City of Houston as the victim, constitutes prosecuting her for the “same offense” as the previous prosecution, in which C & C Services was the named victim. The majority begins its analysis with a peculiar line of Texas authority which holds that an acquittal based on a variance does not bar re-prosecution. I will address this authority in turn, but I think the starting place in this case is federal double jeopardy jurisprudence.
The “cornerstone of double jeopardy jurisprudence” 2 on the issue of whether two prosecutions are for the “same offense” is Ball v. United States.3 In Ball, three defendants were prosecuted for murder. One was acquitted, the other two were convicted. The judgments of the latter two were reversed because the indictments were defective for not alleging “when and where” the victim died.4 The State then prosecuted all three again for the same murder. The prosecution which concerns us is the one against the defendant who had been previously acquitted.
The Supreme Court recognized the old English law that “[i]f any person, who is indicted for an offence, shall on his trial be acquitted upon the ground of a variance between the indictment and the proof, or upon any exception to the form or to the substance of the indictment, he may be arraigned again on a new indictment, and may be tried and convicted for the same offence, notwithstanding such former acquittal.” 5 But the Court explicitly rejected this doctrine as being “unsatisfactory in the grounds on which it proceeds, as well as unjust in its operation upon those accused of crime.”6 The Court emphasized that the Fifth Amendment’s “prohibition is not against being twice punished, but against being twice put in jeopardy; and the accused, whether convicted or acquitted, is equally put in jeopardy at the first trial.”7 The Court quoted at length from *130the dissent in People v. Barretft 8 which explained that the State should not get a second chance at a prosecution simply because it erred in its first indictment. This was characterized as “presenting the] novel and unheard-of spectacle of a public officer, whose business it was to frame a correct bill, openly alleging his own inaccuracy or neglect as a reason for a second trial, when it is not pretended that the merits were not fairly in issue on the first.”9 The Court concluded that, “[h]ow-ever it may be in England, in this country a verdict of acquittal, although not followed by any judgment, is a bar to a subsequent prosecution for the same offense.” 10
Ball governs today’s case. Bailey was acquitted, and the State now seeks to prosecute her for the same offense, this time alleging a different victim. Under Ball, this prosecution is barred. The majority concludes that this is not the “same offense” because this indictment alleges a different victim. But Ball rejects the notion that making a minor change in an indictment creates a different offense. In Ball, the first indictment did not allege the “where and when” of the murder, and the second indictment did. This did not create a new offense. There was only one offense in Ball — the murder — regardless of the particulars of the second indictment as compared with the first.
There is not much authority to assist courts in determining whether a defendant is being prosecuted twice for the “same offense” under the same statute. Generally, courts analyze whether two offenses are the “same offense” under the Blockburger11 test by determining whether each offense requires proof of an element which the other does not. But that test is only applicable when the defendant is prosecuted under two different statutes for the same conduct. It does not apply when, as here, the defendant is prosecuted twice under the same statute.
And the issue is even more confusing when considering multiple victims. The fact is, in some cases, two prosecutions based on two different victims will be permissible, while in other cases it will not. It depends on the facts of each case and whether those facts constitute one offense or two offenses.
We have held that, for double jeopardy purposes, the allowable unit of prosecution for robbery is each victim.12 The same could be said for engaging in organized criminal activity by committing theft — the allowable unit of prosecution would be each victim. We have also held that “same offense” means the “identical criminal act.”13 Essentially, if there is one victim, there is one criminal act, and one offense; but if there are two victims, there are two offenses. For example, if Bailey had committed two thefts by stealing money from two entities, both the City of Houston and C & C Services, then there would be two victims in this case and therefore two offenses. The theft from one is not the “same offense” as the theft from the other. They are different offenses because there are two different victims.
But that is not this case. This case involves a single theft from a single victim, *131but the State alleged the incorrect victim in its first indictment. There is a single offense in this case. The theft from C & C Services is the “same offense” — the identical criminal act — as the theft from the City of Houston because there was no separate theft from the City of Houston. It was all one offense (though it occurred over a period of time and through multiple time sheets). This indictment, then, charges Bailey for the “same offense” for which she was already acquitted.
Another way of looking at it is to ask: Would we allow this subsequent prosecution if Bailey had been convicted, rather than acquitted, under the first indictment? Surely not. If Bailey had been convicted under the first indictment naming the City of Houston as the victim, a subsequent prosecution alleging C & C Services as the victim would no doubt be barred. This is because, again, there were not two victims in this case. There was one victim. The result should not be any different just because Bailey was acquitted. There was just one offense, and this indictment charges Bailey with that “same offense,” regardless of who is named as the victim.
II.
So federal double jeopardy principles bars this prosecution. But the majority allows it based on a peculiar fine of authority from this Court. The majority finds this line of authority relevant because it characterizes Bailey’s first trial as results ing in an acquittal due to a “variance.”14
First, I disagree with the majority’s characterization of Bailey’s acquittal in the first trial as being based on a variance. I have previously noted the havoc that the word “variance” has caused in our case-law.15 But even given the reality that variance law is here to stay, the fact is that not every acquittal or reversal due to insufficient evidence can rationally be characterized as a variance case. We have stated that “a ‘variance’ occurs when there is a discrepancy between the allegations in the charging instrument and the proof at trial.”16 While this broad definition covers the facts of this case, it includes too much. The word “variance” ought to be used to describe instances in which there is a minor discrepancy between the facts alleged and those proved, such as a difference in spelling, in numerical digits, or in some other minor way.17 When the variance pertains to a person, the vast majority of the cases do not involve any dispute as to the identity of the person.18 The person the State alleges is the same person as that which it proves, but the indictment contains a clerical error regarding that person’s name.19 Essentially, a variance case ought to refer to a case in which the State makes a minor clerical error in the indictment. Indeed, the Supreme Court’s Ball opinion says as much when it basically equates acquittals based on a variance with those based on an exception to the form or substance of the indictment.20
This case does not consist of a clerical or typographical error in the indictment — it involves a substantive failure in proof. In Bailey’s first trial the State alleged the City of Houston as the owner but, according to the trial court, it proved that C & C *132Services was the owner. The “City of Houston” and “C & C Services” are not even remotely similar. They are not similarly spelled. They are not the same institution. They are not even the same type of institution, one being a government entity and the other being a private business. The majority states that the fact that the victims in this case were “two different people” does not distinguish it from other variance cases,21 but I disagree. Rather than a clerical error in the indictment, this case involves a flaw in the State’s investigation before trial and a substantive failure in proof at trial. I would not use the word “variance” to describe this type of insufficient-evidence case.
III.
The majority’s use of the word “variance” leads it to follow a line of authority in which we have indicated that an acquittal based on a variance does not bar re-prosecution on a new indictment alleging that version of the offense that the State’s evidence proved at the first trial.22 I will refer to this as the “Swindel Rule” since Swindel v. State 23 seems to be the first case in which we articulated it. Because I don’t believe this case involves a variance, I don’t believe the Swindel Rule is even applicable. But assuming that Bailey’s acquittal was due to a variance and the Swindel Rule applies, I believe the Swindel Rule is wrong. Besides being almost blatantly contradictory to Ball v. United States, the Swindel Rule also conflicts with more reasoned authority from this Court, has been criticized by commentators, is poorly reasoned, and is unworkable.
A.
First, it conflicts with Ball. As noted above, the Supreme Court in Ball specifically rejected the very notion on which Swindel is based.24 The majority attempts to find the Swindel Rule constitutional by characterizing this prosecution as being for a different offense, but the Supreme Court recognized in Ball that a subsequent prosecution after an acquittal based on a variance was inherently for the “same offense” — merely changing the wording of the indictment does not create a new offense. The Swindel Rule and the majority’s opinion do not survive under Ball. And even if this case is properly considered a variance case because this case involved poor drafting of the indictment, Ball plainly rejects the notion that a clerical error in the indictment reheves the State of the responsibility of complying with double jeopardy principles.25
B.
Swindel also conflicts with more reasoned authority from our Court. Although our variance law is not a model of clarity, we recently re-affirmed in Gollihar v. State 26 that a variance will not result in a conviction’s reversal unless it is material. I expressed my disagreement with that opinion at the time,27 but the Court has spoken on the issue and Gollihar is law.
In Gollihar, the Court stated that a variance is material if it is “prejudicial to a *133defendant’s substantial rights.’”28 The Court explained that, to determine whether a defendant’s substantial rights are prejudiced, we consider “[1] whether the indictment ... informed the defendant of the charge against him sufficiently to allow him to prepare an adequate defense at trial, and [2] whether prosecution under the deficiently drafted indictment would subject the defendant to the risk of being prosecuted later for the same crime.”29 The important part of that statement for our purposes is part (2): A variance is material if a prosecution under the defi-ciently drafted indictment would subject the defendant to the risk of being prosecuted later for the same crime. Of course, Gollihar did not make this up. We said essentially the same thing many times before Gollihar, at least as far back as 1910 in Rowan v. State.30 There we stated, “A variance is not now regarded as material unless it is such as might mislead the defense, or might expose the accused to the danger of being put twice in jeopardy for the same offense.”31 I’ll call this the Rowan Rule.
So the Rowan Rule states that a material variance occurs when the defendant is at risk of being put twice in jeopardy for the same offense. But that runs directly contrary to the Swindel Rule. Under the Swindel Rule, if there is a material variance, the defendant will be acquitted, but he can be re-prosecuted. Under the Rowan Rule, if the variance is material and the defendant is re-prosecuted, then the defendant is put in jeopardy twice for the same offense, which violates double jeopardy.
Essentially, we have been articulating two different double jeopardy rules. The Rowan Rule is usually articulated in cases in which the conviction is affirmed because the variance is found to be immaterial.32 In those cases, we generally reason that the variance was immaterial because there was no risk that the defendant would be re-prosecuted on a subsequent indictment for the same offense. In contrast, the Swindel Rule is usually mentioned in cases in which the conviction is reversed because the variance is material.33 In that instance, we generally conclude that re-prosecution on a subsequent indictment for the same offense is permissible.
The two rules are not distinguishable under any theory that one of them involves cases involving a re-prosecution for a “different” offense while the other involves cases involving a re-prosecution for the “same” offense. In both Rowan cases and Swindel cases, the Court has faced the same issue regarding whether the subse*134quent prosecution is for the “same offense.” The two Rules have been applied in variance cases that are virtually indistinguishable. Swindel applied in Nance v. State,34 where the variance was between Pauline Leitz and Pauline Seitz. But Rowan applied in Reyes v. State,35 where the variance was between Modesta Espino-sa and Modesta Espinosa. In either case, the question of whether re-prosecution under an indictment which correctly spelled the person’s name constitutes the “same offense” for double jeopardy purposes ought to be answered the same; it either is the same offense in both cases or it is a different offense in both cases. And yet, the cases are treated differently.
C.
Commentators agree with me that Swindel is flawed. The majority cites Professors Dix and Dawson’s treatise as support for the Swindel Rule. But those professors do not endorse it. On the contrary, they explain that whether this rule “would survive attack based on federal double jeopardy law is problematic at best.”36 They indicate that the split Supreme Court in United States v. Dixon 37 would very likely unite in their opposition to “cases such as Fulmer where the State relies upon the same behavior of the accused as constituting the same statutory crime but seeks a second opportunity to convict because in the first proceeding its evidence failed to prove the version of the facts it chose to plead.”38 A contrary result allows the State “to prosecute the defendant for what is a different offense only in the narrowest technical sense.”39
D.
Moreover, Swindel itself is poorly reasoned. In Swindel, the defendant was charged with stealing a horse, but was discharged because the evidence showed that he stole a gelding instead.40 The Court concluded that a discharge was essentially the same as an acquittal by a jury, and that double jeopardy did not bar re-prosecution for theft of the gelding because “[t]he theft of any one of the animals constitute a separate offense.”41 That is no doubt true, but the facts of Swindel reflected that the defendant stole only one animal, not two. The animal was a gelding, which, of course, is a type of horse. The Swindel Court reasoned that there were two separate offenses here because the statute distinguished between a gelding and a horse.42 But the fact that the statute made that distinction does not mean that the defendant committed two offenses. The Swindel Court cited no authority for the proposition that the defendant could be prosecuted twice for stealing the same animal when the State “erred” in its initial indictment by alleging a horse instead of a gelding.
E.
In addition, Swindel has recently demonstrated its unworkability in the face of double jeopardy law. In Ex parte Coleman,43 we were forced to carve out a sig*135nificant exception to the Swindel Rule. In that case, the State first alleged that the defendant stole property from a person “unknown to the grand jury.” The defendant was acquitted. The State then re-prosecuted him for the same offense, this time alleging Home Depot as the property owner. The defendant alleged that the subsequent prosecution violated double jeopardy, but the State argued it was permissible based on Fulmer v. State 44 and Smotherman v. Stated 45 As the majority notes, Fulmer and Smotherman are both Swindel cases.
We agreed with the defendant.46 We noted that in Fulmer and Smotherman, the State had “made a mistake in pleading the complainants.”47 We explained that “the State could retry the defendants in those cases for offenses committed against patently different people.”48 We distinguished Coleman because in that case we could “be certain that ... [the defendant] was acquitted of the same offense in the first trial for which the State wants to try her in the second, though the first indictment stated that the complainant was unknown.” 49
The distinction between Coleman on the one hand and Fulmer and Smotherman on the other is not persuasive. Whether the State alleges the victim as a particular person whose name is misspelled, or as a person unknown to the grand jury, in either case, if the defendant is acquitted the first time and then re-prosecuted under a corrected indictment, the defendant is prosecuted twice for the same offense. Moreover, the Coleman Court’s characterization of Fulmer and Smotherman as permitting subsequent prosecutions “against patently different people” is misleading. Though Smotherman did involve a prosecution against a totally different person, Fulmer did not. And more importantly, the Swindel Rule has almost uniformly been applied in cases in which the second prosecution was plainly against the same person, not a completely different person.50 Smotherman is in the minority in that regard.
Professors Dix and Dawson agree that “Coleman ⅛ effort to distinguish situations involving allegations of matters unknown is not convincing.”51 They also note that Coleman’s approach is “the only reasonable one if the prohibition against double jeopardy is to be taken seriously.”52 The distinction between Coleman and Swindel is unpersuasive, and our need in Coleman to limit the Swindel Rule demonstrates the unworkability of that rule in the face of federal double jeopardy principles.
F.
We should not lightly overrule precedent. But when one of our cases conflicts with Supreme Court authority, we have no choice but to overrule it.53 Swindel con*136flicts with Ball v. United, States. Additionally, when precedent conflicts with a decision that is found to be more soundly reasoned, we may resolve the inconsistency in favor of the more soundly reasoned decision.54 Swindel conflicts with Rowan, which is more soundly reasoned. Finally, if one of our decisions is poorly reasoned or unworkable, we should not follow it.55 Swindel is poorly reasoned, and the fact that we have recently found it necessarily to limit it indicates its unworkability. We should not follow it today. If Swindel is applicable, then we should overrule it and, to the extent that they follow it, Smother-man and Fulmer.
IV.
Again, I have my doubts as to whether this ease even involves a variance. But if it does, then we should take this opportunity today to continue in our recent efforts to clarify variance law.56 We now recognize that a variance is not material and should not result in a reversal unless the defendant was surprised.57 We should today continue with the logical extension of that rule: if the conviction is reversed because of a material variance, or if a defendant is acquitted because of such a variance, then a subsequent prosecution for that same offense is barred by double jeopardy. And simply correcting the indictment error which resulted in the variance in the first place does not create a “new offense.” To hold otherwise is to grant the State unlimited opportunities to prosecute an individual, and this cannot be what the Fifth Amendment means.
V.
I would reverse the judgment of the Court of Appeals and hold this prosecution barred by double jeopardy. Because the majority concludes otherwise, I dissent.
*137APPENDIX A
[[Image here]]
*138[[Image here]]
*139[[Image here]]
. North Carolina v. Pearce, 395 U.S. 711, 717, 89 S.Ct. 2072, 23 L.Ed.2d 656 (1969).
. 5 LaFave et al, Criminal Procedure § 25.3(b) (2d ed.1999).
. 163 U.S. 662, 16 S.Ct. 1192, 41 L.Ed. 300 (1896).
. Id. at 664, 16 S.Ct. 1192.
. Id. at 669, 16 S.Ct. 1192.
. Id.
. Id.
. 1 Johns. 66.
. Ball, 163 U.S. at 667, 16 S.Ct. 1192.
. Id. at 671, 16 S.Ct. 1192.
. Blockburger v. United States, 284 U.S. 299, 52 S.Ct. 180, 76 L.Ed. 306 (1932).
. Ex parte Hawkins, 6 S.W.3d 554, 560 (Tex.Crim.App.1999).
. Ex parte Goodbread, 967 S.W.2d 859, 860 (Tex.Crim.App.1998); Luna v. State, 493 S.W.2d 854, 855 (Tex.Crim.App.1973).
. Ante, op. at 126-127.
. Gollihar v. State, 46 S.W.3d 243, 260-62 (Tex.Crim.App.2001) (Keasler, J., concurring & dissenting).
. Gollihar, 46 S.W.3d at 246.
. See, e.g., Appendix A.
. Id.
. Id.
. See Ball, 163 U.S. at 669, 16 S.Ct. 1192.
. Ante, op. at 127.
. Ante, op. at 127, citing 43 Dix et al, Criminal Practice and Procedure § 31.233 (2d ed.2001).
. 32 Tex. 102, 103-04 (1869).
. See Ball, 163 U.S. at 669, 16 S.Ct. 1192.
. See id. at 667, 16 S.Ct. 1192.
. 46 S.W.3d at 248-49.
. Id. at 260-62 (Keasler, J., concurring & dissenting).
. Id. at 248.
. Id.
. 57 Tex.Crim. 625, 124 S.W. 668, 673 (1910).
. Id. See also Fuller v. State, 73 S.W.3d 250 (Tex.Crim.App.2002); Santana v. State, 59 S.W.3d 187, 195 (Tex.Crim.App.2001); Middleton v. State, 476 S.W.2d 14, 16 (Tex.Crim.App.1972); Martin v. State, 152 Tex.Crim. 261, 213 S.W.2d 548 (1948); Raven v. State, 149 Tex.Crim. 294, 193 S.W.2d 527 (1946); Jones v. State, 115 Tex.Crim. 418, 27 S.W.2d 653, 656 (1930); Davis v. State, 88 Tex.Crim. 7, 224 S.W. 510 (1920); Reyes v. State, 81 Tex.Crim. 588, 196 S.W. 532, 535 (1917); Zweig v. State, 74 Tex.Crim. 306, 171 S.W. 747, 750 (1913); Feeny v. State, 62 Tex.Crim. 585, 138 S.W. 135, 138 (1911) (op. on reh'g).
. See, e.g., Fuller, 73 S.W.3d at 254; Santana, 59 S.W.3d at 195; Middleton, 476 S.W.2d at 16; Martin, 213 S.W.2d at 548; Raven, 193 S.W.2d at 527.
. See, e.g., Fulmer v. State, 731 S.W.2d 943 (Tex.Crim.App.1987); Kinney v. State, 67 Tex.Crim. 175, 148 S.W. 783, 783-84 (1912); Branch v. State, 20 Tex.App. 599 (Tex.Ct.App.1886); Nance v. State, 17 Tex.App. 385, 388-89 (Tex.Ct.App.1885); Parchman v. State, 2 Tex.App. 228, 241-42 (Tex.Ct.App.1877).
. 17 Tex.App. 385.
. 196 S.W. at 535.
. Dix et al, supra at § 31.233.
. Id.
. Id.
. Id.
. 32 Tex. at 102-03.
. Id. at 103.
. Id. at 103-04.
. 940 S.W.2d 96 (Tex.Crim.App.1996).
. 731 S.W.2d 943.
. 415 S.W.2d 430 (Tex.Crim.App.1967).
. Coleman, 940 S.W.2d at 98.
. Id. at 99.
. Id.
. Id.
. See, e.g., Fulmer, 731 S.W.2d at 946; Reynolds v. State, 58 Tex.Crim. 273, 274, 124 S.W. 931 (1910); Branch, 20 Tex.App. 599; Nance, 17 Tex.App. at 388-89; Parchman, 2 Tex.App. at 241-42.
. Dix et al, supra at § 31.234.
. Id.
. See Hernandez v. State, 988 S.W.2d 770, 771 (Tex.Crim.App.1999); State v. Guzman, 959 S.W.2d 631, 633 (Tex.Crim.App.1998). See also Malik v. State, 953 S.W.2d 234, 236 (Tex.Crim.App.1997).
. Awadelkariem v. State, 974 S.W.2d 721, 725 (Tex.Crim.App.1998).
. Paulson v. State, 28 S.W.3d 570, 571-72 (Tex.Crim.App.2000).
. See Gollihar, 46 S.W.3d 243.
. Id.