In re Richards

LIU, J., Dissenting.

Before 1975, a petitioner seeking habeas corpus relief could generally prevail only where the prosecution knowingly offered perjured testimony or the petitioner was able to produce new evidence pointing *971unerringly toward innocence. (See maj. opn., ante, at p. 960.) In 1975, the Legislature amended the Penal Code to provide new grounds for relief when a conviction or guilty plea rested on false evidence. (See Stats. 1975, ch. 1047, § 2, p. 2466.) The statute now in effect makes no distinction with respect to how false evidence was introduced, nor does it condition relief upon the petitioner’s ability to produce new evidence that was not available at trial. Instead, the petitioner is entitled to relief whenever “[f]alse evidence that is substantially material or probative on the issue of guilt or punishment was introduced.” (Pen. Code, § 1473, subd. (b), hereafter section 1473(b); all further statutory references are to the Penal Code.) As the court observes, “under the present version of section 1473 it does not matter why evidence is false or whether any party to the proceeding knew it was false. So long as some piece of evidence at trial was actually false, and so long as it is reasonably probable that without that evidence the verdict would have been different, habeas corpus relief is appropriate.” (Maj. opn., ante, at p. 961.)

Although the false evidence statute makes no distinction between lay and expert testimony, today’s decision imposes novel burdens on a petitioner who seeks relief under section 1473(b) where false evidence was introduced through expert testimony. As explained below, there is no reason to treat expert testimony differently from lay testimony under section 1473(b). I would hold, consistent with our precedents, that “false evidence” within the meaning of section 1473(b) is established when a petitioner shows by a preponderance of the evidence either the falsity of an expert’s testimony or the falsity of an underlying fact essential to an expert’s testimony. Because petitioner has met that burden here, and because the false evidence was sufficiently prejudicial to cast doubt on the verdict of guilt, I respectfully dissent.

I.

Section 1473(b) provides in relevant part: “A writ of habeas corpus may be prosecuted for, but not limited to, the following reasons: [f] (1) False evidence that is substantially material or probative on the issue of guilt or punishment was introduced against a person at any hearing or trial relating to his incarceration . . . .” This court has repeatedly considered the applicability of section 1473(b) to cases of lay testimony, particularly eyewitness identification. (See, e.g., In re Bell (2007) 42 Cal.4th 630 [67 Cal.Rptr.3d 781, 170 P.3d 153] [eyewitness identification]; In re Roberts (2003) 29 Cal.4th 726 [128 Cal.Rptr.2d 762, 60 P.3d 165] (Roberts) [eyewitness identification]; In re Malone (1996) 12 Cal.4th 935 [50 Cal.Rptr.2d 281, 911 P.2d 468] (Malone) [prison informant]; In re Sassounian (1995) 9 Cal.4th 535 [37 Cal.Rptr.2d 446, 887 P.2d 527] (Sassounian) [prison informant]; In re Hall (1981) 30 Cal.3d 408 [179 Cal.Rptr. 223, 637 P.2d 690] (Hall) [eyewitness identification].) The *972question now before us is how to apply the same statutory provision to expert testimony that is alleged to be false.

Section 1473(b) makes no distinction between lay and expert testimony. Yet today’s opinion effectively narrows the availability of habeas corpus relief when the allegedly false testimony was offered by an expert rather than a lay witness. The court does so by announcing the following rule: “When . . . there has been a generally accepted and relevant advance in the witness’s field of expertise, or when a widely accepted new technology has allowed experts to reach an objectively more accurate conclusion, a strong reason may exist for valuing a later opinion over an earlier opinion. If, and only if, a preponderance of the evidence shows that an expert opinion stated at trial was objectively untrue, the false evidence standard applies. In that narrow circumstance, if it is reasonably probable that the invalid opinion given at trial affected the verdict, then habeas corpus relief is appropriate.” (Maj. opn., ante, at p. 963, fn. omitted.) I see two problems with this approach.

A.

The court says that a petitioner must “show[] that an expert opinion stated at trial was objectively untrue” in order to meet the false evidence standard. (Maj. opn., ante, at p. 963.) By this, the court means that the ultimate opinion rendered by the expert—here, Dr. Sperber’s trial testimony that the lesion on the victim’s hand was “consistent with” petitioner’s teeth—must be proven objectively untrue. Petitioner fell short of this proof, the court says, because his experts in the habeas corpus proceeding “could not definitively rule out petitioner’s teeth as a possible source of the lesion.” (Maj. opn., ante, at p. 964, original italics.)

By focusing exclusively on whether the expert’s ultimate opinion has been proven untrue, the court overlooks an important and conventional manner in which trial testimony may later be proven false. Consider the facts of Hall, supra, 30 Cal.3d 408. Three brothers—Victor Lara, Daniel Lara, and Jesse Ortiz—were assaulted by several youths. One of the youths had a gun and fired several shots, “slightly wounding Victor and Daniel, and killing Jesse with a single bullet in the head.” (Id. at p. 414.) As eyewitnesses to the crime, Victor and Daniel testified at trial that the petitioner was the gunman. (Ibid.) After trial, however, the Lara brothers recanted their testimony, believing that someone else was the gunman. Victor signed “a declaration stating in part, T never got a good look at the person who shot my brother. I was mistaken when I later identified [petitioner] Hall as the killer . . . .’ Daniel Lara testified in the same fashion, explaining his erroneous identification by the fact that he was distraught about his brother’s death and careless in deciding *973which of the . . . youths were responsible.” (Id. at p. 417.) While acknowledging that “we routinely view recantations with suspicion . . . ,” the court found the recantations credible and sustained the habeas corpus referee’s finding to that effect. (Id. at p. 418.) The court noted that this was not a case of perjured trial testimony: “there is no evidence whatever of ulterior motive or dishonesty on the part of either brother.” (Id. at pp. 417-418.) On these facts, the court held under section 1473(b) that “the trial testimony of the Laras identifying petitioner as the killer apparently was false, albeit unintentionally so.” (30 Cal.3d at p. 424.)

The Lara brothers’ trial testimony was false evidence because the perceptual basis for their eyewitness testimony turned out to be false. Victor said “ T never got a good look at the person who shot my brother,’ ” and Daniel said he had been “distraught” and “careless” in identifying the gunman. (Hall, supra, 30 Cal.3d at p. 417.) In prevailing on his false evidence claim, the petitioner in Hall did not need to show that the ultimate fact to which the Lara brothers testified at trial—that the petitioner was the gunman—was untrue. It was enough to show that the perceptual basis of the trial testimony—that the Lara brothers, as eyewitnesses to the crime, were well situated to identify the gunman—was untrue. When a lay witness in good faith gives testimony that the witness later concedes he or she had no perceptual basis to give, that witness has given false evidence within the meaning of section 1473(b). And that is so regardless of whether other evidence demonstrates the truth or falsity of the ultimate fact to which the witness testified.

There is no reason to treat expert testimony differently. Just as the truth or falsity of eyewitness testimony under section 1473(b) depends on the truth or falsity of underlying facts concerning the witness’s perceptual abilities, the truth or falsity of expert testimony depends on the truth or falsity of underlying facts essential to the expert’s inferential method and ultimate opinion.

The present case illustrates the point. Dr. Sperber’s trial testimony culminated in his opinion that the lesion on the victim’s hand was consistent with petitioner’s abnormal dentition. To reach that conclusion, Dr. Sperber relied on a series of intermediate steps. He explained how forensic odontologists use crime scene or autopsy photographs to make a dental comparison. He explained how he determines whether an injury is in fact a bite mark, and he concluded from a single photograph of the victim’s hand that the lesion was a human bite mark. He explained why he believed the mark came from a lower not upper jaw. He noted a gap in the lesion where it appears an incisor tooth failed to leave a mark. He then showed the jurors a mold of petitioner’s teeth, explaining that petitioner’s right incisor was abnormal and would not *974have left a mark. He further explained how he compared the mold from petitioner’s mouth with the photograph of the lesion. Based on these steps, Dr. Sperber concluded that the mark was consistent with petitioner’s teeth.

A crucial fact underlying all of Dr. Sperber’s testimony was that the single photograph of the victim’s hand was alone a sufficient basis for reaching his ultimate conclusion. He acknowledged problems with the photograph, including distortion, and he said additional photographs would have helped him render a more definite conclusion. But he never wavered from the essential premise that the single photograph was itself a sufficient basis to render an opinion. Based on that premise, Dr. Sperber offered his expert opinion that the lesion was a bite mark consistent with petitioner’s teeth. That underlying premise—akin to the perceptual premise of the eyewitness testimony shown to be false in Hall—has now been shown to be false by petitioner’s habeas corpus evidence.

The testimony at the habeas corpus proceeding shows by a preponderance of the evidence that the photograph alone did not provide a sufficient basis for comparison with petitioner’s teeth. Petitioner first presented testimony by Dr. Sperber himself. After reviewing additional autopsy photographs not provided to him by the prosecution at the time of trial, Dr. Sperber was asked, “Doctor, had you had those photographs prior to your testimony in 1997 at the trial of Mr. Richards, would they have affected your opinion?” He answered, “Very definitely.” In light of those additional photographs and after considering other problems with his analysis, Dr. Sperber testified, “I don’t know for sure that that lesion or that photograph [presented to the jury] depicts a bite mark.” And based on those same factors, Dr. Sperber concluded that “[m]y opinion today is that [petitioner’s] teeth, as we have seen, are not consistent with the lesion on the hand” and that “I would rule him out basically on the evidence as I’ve seen now in hindsight.”

Petitioner also called Dr. Golden, who had testified at trial that petitioner’s teeth were consistent with the lesion but that the quality of the photograph was too poor to make a definitive determination as to whether the lesion was or was not consistent with petitioner. At the habeas corpus hearing, Dr. Golden testified about advances in technology that allowed experts to correct distortion in photographs like the one relied on by both him and Dr. Sperber at trial. He further testified that since that technology came into use, he has personally used that technology to correct photo distortion in cases in which he testified for the county. Based on this technology, Dr. Golden testified that he would still view the lesion as consistent with a human bite mark, though it could also have come from a variety of sources. In addition, Dr. Golden said he “would tend to rule out Mr. Richards ... as the suspected biter” and that “[a]fter all the subsequent analyses with the photographic correction and the digital analysis, my opinion is that I would exclude him.”

*975Petitioner called two additional dental experts: Dr. Johansen and Dr. Bowers. Dr. Johansen testified at length about the procedure he followed to correct the distortion in the photograph of the victim’s hand using Adobe Photoshop. He also looked at other crime scene photographs not made available to Dr. Sperber at the time of trial. Even after correcting the photo distortion, Dr. Johansen concluded that “the very, very poor quality of the bite mark” precluded him from either including or excluding petitioner as the source of the mark. In his opinion, it was just as likely that the lesion was caused by fencing material located under the victim’s body as by petitioner’s teeth.

Petitioner also called Dr. Bowers to testify about the process of correcting photo distortion. After correcting the photograph, Dr. Bowers compared the size of the lesion to the span of petitioner’s lower teeth from one side to the other, and he concluded that the lesion was too small to have been made by petitioner’s teeth. Comparing petitioner’s individual teeth to the lesion, Dr. Bowers found three teeth that matched the lesion and three that did not. Dr. Bowers testified that such a match “doesn’t have any forensic significance in identifying [the] biter.” Finally, Dr. Bowers testified that the other crime scene and autopsy photographs of the victim “raised significant doubt. . . that the hand injury in this original Richards case was caused by teeth.”

Significantly, the Attorney General offered no competing evidence. This is not a case in which a habeas corpus evidentiary hearing has devolved into a fresh battle of the experts. (Cf. maj. opn., ante, at p. 966, fn. 5.) Instead, all of the experts provided testimony refuting critical facts underlying Dr. Sperber’s trial testimony. In essence, all four experts testified that the photograph did not provide a sufficient basis for Dr. Sperber to conclude at trial that the lesion was a bite mark consistent with petitioner’s teeth.

There will no doubt be cases where it is more difficult to determine whether the falsity of one or more facts underlying an expert’s trial testimony renders the entire opinion false. In many cases, an expert opinion may be based on a variety of facts, and the falsity of one fact might not undermine the expert’s ultimate conclusion. But this is not such a case. Here, the critical underlying fact—that the single uncorrected photograph provided Dr. Sperber with a sufficient basis for matching petitioner’s teeth to a lesion on the victim’s hand—was proven false. Without that premise, Dr. Sperber’s trial testimony that the lesion was consistent with petitioner’s teeth was false evidence. Just as the eyewitness testimony in Hall was false because it depended crucially on the witnesses having seen something that it turns out they did not actually see, the expert testimony here was false because it depended crucially on Dr. Sperber having seen something—a true photographic representation of the lesion on the victim’s hand—that it turns out he *976did not actually see. In sum, because petitioner has shown by a preponderance of the evidence that the essential premise of Dr. Sperber’s trial testimony was false, it follows that the testimony was false evidence under section 1473(b).

The court suggests that this approach might warrant habeas corpus relief even if the petitioner’s experts at the habeas corpus proceeding fully agreed with the trial expert’s conclusion but showed that the expert “relied on a false assumption to reach that conclusion.” (Maj. opn., ante, at p. 963, fn. 2.) This concern is unwarranted. First, the petitioner would have to show not only that the trial expert relied on an underlying fact or methodology that turned out to be false, but also that the falsity was sufficient to undermine the trial expert’s conclusion. Second, the petitioner must show prejudice to obtain relief, and it is difficult to see how the petitioner could make this showing if the petitioner’s habeas experts agree with the trial expert’s conclusion.

B.

Instead of focusing on the factual underpinnings of the expert’s trial testimony, the court looks only at the expert’s ultimate conclusion. As explained above, section 1473(b) does not require petitioner to show that Dr. Sperber’s ultimate conclusion that the bite mark was consistent with petitioner’s teeth was objectively untrue; it is enough that petitioner has shown the falsity of an underlying premise essential to Dr. Sperber’s conclusion. Nevertheless, showing Dr. Sperber’s ultimate conclusion to be objectively untrue is an alternative means of establishing false evidence. On the facts here, I believe petitioner has met that burden as well. The court reaches a contrary conclusion by unjustifiably heightening the standard of proof required to show the falsity of expert testimony.

Under settled law, a petitioner seeking habeas corpus relief “bears the ‘burden ... of alleging ... the facts on which he relies in support of his claim [or claims] for relief . . . .’ [Citation.] He also ‘bears the burden of proving [those] facts ... by a preponderance of the evidence.’ [Citation.]” (Sassounian, supra, 9 Cal.4th at pp. 546-547.) Thus a petitioner seeking relief based on false evidence must prove—not “definitively” or “absolutely,” but by a preponderance of the evidence—that false evidence was offered against him at trial.

We made clear the proper application of this standard to a false evidence claim in Malone, supra, 12 Cal.4th 935. There, a prison informant testified at the petitioner’s trial that the petitioner had confessed to him several crimes. The petitioner sought relief after the informant recanted his original trial testimony, then recanted his recantation. We said: “From this record, it is *977impossible to know with certainty whether [the informant] lied or told the truth at petitioner’s trial. Petitioner’s burden, however, is only to prove his claims by a preponderance of the evidence. [Citation.] On the basis of all the evidence, we agree with the referee [that the informant] probably lied when he testified petitioner confessed to him . . . .” (Id. at p. 962.) We did not require the petitioner in Malone to definitively show that he never confessed to the informant. It was enough that the informant “probably lied” when he testified to the contrary.

In the present case, Dr. Sperber testified at trial that the lesion on the victim’s hand was “consistent with” petitioner’s teeth. This testimony is false if the lesion was in fact inconsistent with petitioner’s teeth. Under Malone, it is enough for petitioner to show the inconsistency by a preponderance of the evidence. Petitioner’s habeas corpus evidence meets this standard. As previously noted, Dr. Sperber examined additional evidence at the habeas proceeding and disavowed his trial testimony, saying; “I would rule [petitioner] out basically on the evidence as I’ve seen now . ...” A second expert, Dr. Golden, used new technology to remove angular distortion from the photo of the lesion and said, “After all the subsequent analyses with the photographic correction and the digital analysis, my opinion is that I would exclude him.” A third expert, Dr. Bowers, observed three areas of inconsistency between petitioner’s teeth and the lesion, and he also testified that the lesion “was too small ... to have been created by Mr. Richards’ lower teeth in terms of dimension of that area.” Dr. Bowers also said he had “significant doubt . . . that the hand injury in this original Richards case was caused by teeth.” A fourth expert, Dr. Johansen, said that “due to the very, very poor quality of the bite mark and the very little information contained within the bite mark, ... I cannot exclude [petitioner] or disexclude him from the population of possible individuals that could have made that mark.” He also said it was just as likely the lesion was caused by nearby fencing material as by petitioner’s teeth.

Thus, two experts expressly said they would “rule out” or “exclude” petitioner. A third expert likewise “found no match” (maj. opn., ante, at pp. 957-958) between the lesion and petitioner’s teeth based on size and other physical features, and “doubted whether the hand lesion was a human bite mark” (id. at p. 958). A fourth expert found the quality of the photograph too poor to rule in or rule out petitioner’s teeth as a possible source of the lesion. Based on the totality of the experts’ testimony, petitioner has shown by a preponderance of the evidence that the lesion was inconsistent with his teeth.

Though purporting to apply the preponderance of the evidence standard, the court concludes that petitioner did not meet his burden because the habeas corpus experts failed to conclude definitively that he was not the source of *978the bite. The court says the experts at the habeas corpus hearing “could not definitively rule out petitioner’s teeth as a possible source of the lesion” (maj. opn., ante, at p. 964), “still could not definitively rule out petitioner’s teeth as a possible source of the mark” after enhancing the photograph (id. at p. 965), and in the end “did not absolutely rule out petitioner’s teeth as a possible source of the mark” (id. at pp. 965-966). The assertion that the experts could not “definitively” or “absolutely” rule out petitioner’s teeth is what leads the court to conclude that petitioner has not proven Dr. Sperber’s trial testimony to be objectively untrue. (Ibid.)

As noted, three of the four habeas corpus experts did rule out petitioner’s teeth, and the fourth expert declined to draw any conclusion from the photo of the lesion. In concluding this is not enough, the court effectively requires petitioner to show that all of the habeas corpus evidence, without exception, would “definitively” or “absolutely” rule out his teeth. But if that is petitioner’s burden, then the court has “confuse[d] the standard of proof with the question of what must be proved” (maj. opn., ante, at p. 966, fn. 5) by requiring petitioner to prove the inconsistency to a virtual certainty rather than by a preponderance of the evidence. In Malone, we held that proof of falsity by a preponderance of the evidence was satisfied by the referee’s finding that the informant “probably lied,” even as we acknowledged that “it is impossible to know with certainty whether [the informant] lied or told the truth at petitioner’s trial.” (Malone, supra, 12 Cal.4th at p. 962.) Because proof of falsity by a mere preponderance of the evidence always admits the possibility that the contested trial evidence “might be true” (maj. opn., ante, at p. 966, fn. 5), the court’s objection to the straightforward application of the preponderance of the evidence standard to this case suggests, in essence, the court’s disagreement with the standard itself.

At the core of the court’s reasoning that petitioner’s evidence must “definitively” or “absolutely” rule out his teeth is the premise that experts “testify ‘in the form of an opinion’ ([Evid. Code], § 801, italics added)” and that “[t]he word ‘opinion’ implies a subjective component to expert testimony.” (Maj. opn., ante, at p. 962.) But the court’s emphasis on the word “opinion” is a miscue. The opinions typically given by experts are not subjective in the “chocolate tastes better than vanilla” sense. Instead, opinion testimony may include a subjective component in the sense that experts use professional judgment in the course of applying a methodology or drawing inferences to reach an ultimate conclusion. But the ultimate conclusion itself—the evidence that is either true or false within the meaning of section 1473(b)—is often, as here, a representation of objective fact.

*979In the present case, Dr. Sperber gave an expert opinion at trial, but he was opining on a question of fact. We call his testimony an “opinion” because of the inferences and judgments he used to derive an ultimate conclusion. But the ultimate conclusion itself—the lesion on the victim’s hand was “consistent with” petitioner’s teeth—is a claim of objective fact. Either the lesion was consistent with petitioner’s teeth, or it wasn’t. Either way, this is a matter of objective fact that is subject to falsification like other matters of objective fact. Malone holds that the standard for showing false evidence is proof of falsity by a preponderance of the evidence. The petitioner in Malone did not definitively prove the informant lied about the confession; the petitioner only showed it was more likely than not. Similarly here, petitioner need not definitively prove the lesion was inconsistent with his teeth; to prove the expert’s conclusion false, he only needs to show that the lesion was inconsistent with his teeth by a preponderance of the evidence. Petitioner has met that burden.

II.

I now consider whether the false evidence was substantially material and probative on the issue of guilt. “ ‘False evidence is “substantially material or probative” if it is “of such significance that it may have affected the outcome,” in the sense that “with reasonable probability it could have affected the outcome . . . .” [Citation.] In other words, false evidence passes the indicated threshold if there is a “reasonable probability” that, had it not been introduced, the result would have been different. [Citation.] The requisite “reasonable probability,” we believe, is such as undermines the reviewing court’s confidence in the outcome.’ ” (Malone, supra, 12 Cal.4th at p. 965, italics added by Malone; see Roberts, supra, 29 Cal.4th at pp. 741-742.) We make such a determination based on the totality of the relevant circumstances. (Malone, at p. 965.)

One notable aspect of this case is the history of two hung juries before the third was able to reach a verdict. The importance of a trial history that includes hung juries is underscored by this court’s discussion in People v. Gonzalez (2006) 38 Cal.4th 932 [44 Cal.Rptr.3d 237, 135 P.3d 649] (Gonzalez). In Gonzalez, we upheld the judgment of guilt and the special circumstance finding of multiple murder and personal use of a firearm. But the court found error in the penalty phase where the judge denied reciprocal discovery of the prosecution’s evidence in rebuttal after the defendant disclosed his evidence in mitigation. In light of that erroneous ruling, the defendant opted not to present mitigating evidence as he did in the first penalty phase. The first penalty phase trial had resulted in a hung jury.

*980Because the state law error in Gonzalez occurred at the penalty phase of a capital trial, the question was whether there was a reasonable possibility (not probability) that the error affected the verdict. (See People v. Brown (1988) 46 Cal.3d 432, 446-448 [250 Cal.Rptr. 604, 758 P.2d 1135].) The court found that there was “a strong suggestion that the defense would have been different” in the absence of the trial court’s erroneous ruling. (Gonzalez, supra, 38 Cal.4th at p. 962.) We then said: “We also find a reasonable possibility the verdict would have been different had defendant presented the proffered mitigating evidence. Although the crime here was egregious, a death verdict was not a foregone conclusion. Indeed, the first penalty trial ended with a hung jury. The aggravating evidence of defendant’s other crimes (possession of an assault weapon, two assaults on inmates, and possession of a shank in jail), although serious, was not overwhelming. Father Horan’s proffered evidence regarding the ability of persons to change was not very compelling, but defendant presented similar evidence in mitigation at the first penalty trial and obtained a hung jury. The main difference between the two trials was that defendant presented mitigating evidence at the first trial that he did not present at the second trial. Under the circumstances, we find it reasonably possible the verdict at the second trial would have been different had defendant presented similar mitigating evidence at the second trial.” (Ibid.)

Unlike Gonzalez, which considered an evidentiary issue at the penalty phase, here we must determine whether, without the prosecution’s expert testimony, there is a reasonable probability the verdict would have been different. The standard is more exacting, but the present record supports such a finding for much the same reason as in Gonzalez.

Today’s opinion accurately presents the evidence supporting petitioner’s guilt. (Maj. opn., ante, at pp. 952-956, 967.) I agree this evidence was substantial, and the conviction would survive a sufficiency-of-the-evidence review. At the same time, all of the prosecution’s evidence was circumstantial, and a guilty verdict was not a foregone conclusion. Petitioner’s defense highlighted the limited amount of time petitioner would have had to commit the crime even if he had sped home from his workplace that night, driving much faster than the posted speed limit. Other evidence showed that the victim had answered her phone throughout the day but then stopped answering the phone hours before petitioner came home, with no indication she had left the premises. And although a prosecution witness testified that bloodstains on petitioner’s pants and shoes were consistent with spatter as a result of standing near the victim when the injuries were inflicted, a defense expert testified that the spatter was consistent with contact or transfer when petitioner cradled the body postmortem.

The first two trials to reach jury deliberations both ended with hung juries. The final jury also declared it was deadlocked and sought further instruction *981on the meaning of “reasonable doubt.” The trial court denied petitioner’s motion for a mistrial and refused further instruction on the meaning of “reasonable doubt,” sending the jury back to deliberate. Three and a half hours later, it returned a verdict of guilt.

The main difference between the two trials ending in hung juries and the final trial ending in a guilty verdict was the bite mark evidence, which was offered only at the final trial. The purported bite mark was the evidence that most directly linked petitioner to the crime. Moreover, the bite mark evidence was not limited to Dr. Sperber’s verbal testimony that a lesion on the victim’s hand was a bite mark matching petitioner’s unusual dentition. Dr. Sperber also prepared a mounted photograph of the lesion along with a plastic overlay created from dental molds of petitioner’s lower teeth, which could be flipped up and down to demonstrate the “match” between the two. The photograph, overlay, and dental molds were all admitted into evidence and available to the jurors during deliberations. Further, Dr. Sperber estimated that only one or two out of a hundred people share petitioner’s dental abnormality. Even taking into account Dr. Sperber’s admission that he did not know of any scientific studies to back up that estimate, his expert testimony on the uniqueness of that feature, which was undisputed by petitioner’s trial expert, increased the probative value of Dr. Sperber’s testimony at the final trial.

When viewed in the context of the remainder of the evidence, Dr. Sperber’s testimony was substantially material and probative of petitioner’s guilt. Indistinct as it was, the lesion on the victim’s hand provided a direct and visceral link between petitioner and the injuries inflicted upon the victim. And while Dr. Sperber concluded only that the lesion was “consistent” with petitioner’s teeth, he coupled that finding with additional testimony that significantly narrowed the range of possible suspects by focusing on petitioner’s unusual dentition. In particular, Dr. Sperber’s estimate that just one or two out of 100 individuals share the same unusual dentition amplified the probative value of the evidence. Near the end of his closing argument, the prosecution relied on this piece of evidence: “And, oh, this person [who committed the crime] also just happened to share the same dental abnormality as William Richards, [which] is only shared by two percent of the population, [f] . . . [][] That’s what you are looking at. And that, folks is unreasonable. That doesn’t wash.” Without the bite mark evidence, two juries hung. Even with that evidence, a third jury deadlocked before returning a guilty verdict. The totality of the circumstances leads me to conclude that Dr. Sperber’s testimony was sufficiently probative and material to cast doubt on the outcome.

*982In sum, the underlying basis of Dr. Sperber’s trial testimony as well as the testimony itself have been proven false by a preponderance of the evidence, and it is reasonably probable that the verdict at the final trial would have been different without Dr. Sperber’s testimony. Accordingly, I would reverse the Court of Appeal’s decision and grant petitioner’s request for habeas corpus relief.

Werdegar, J., and Chin, J. concurred.

Petitioner’s petition for a rehearing was denied February 13, 2013. Werdegar, J., Chin, J., and Liu, 1, were of the opinion that the petition should be granted.