Johnnie Dunning, Appellant, pled guilty to aggravated sexual assault of a child under fourteen years of age pursuant to a plea bargain and was sentenced as a habitual offender to 25 years' confinement. He appealed, but his conviction was affirmed. Dunning v. State , No. 2-99-311-CR (Tex. App.-Fort Worth Feb. 22, 2001, pet. ref'd) (not designated for publication). Later, he filed a Chapter 64 motion for post-conviction DNA testing. The trial judge held a hearing in February 2017 and entered a non-favorable finding. Appellant appealed that finding, and the court of appeals reversed and ordered the trial court to enter a favorable finding. Dunning v. State , 544 S.W.3d 912, 922 (Tex. App.-Fort Worth 2018). The State filed a petition for discretionary review, which we granted in part. The three issues that we agreed to review are,
(1) Whether the court of appeals properly determined that the post-conviction DNA testing results established a reasonable probability that [Appellant] would not have been convicted had they been available at the time of trial?
(2) Whether the court of appeals gave proper deference to the trial court's determination of historical facts and application-of-law-to-fact issues that turn on credibility or demeanor?
(3) Whether the court of appeals considered all of the evidence before the trial court in making its article 64.04 finding before determining that post-conviction DNA testing results established a reasonable probability that the appellant would not have been convicted had they been available at the time of trial?
Because we conclude that the court of appeals erred, we will reverse its judgment vacating the trial court's non-favorable finding.
BACKGROUND1
a. Trial Proceedings
In 1999, Appellant was charged with aggravated sexual assault of a twelve-year-old, allegedly intellectually disabled child by inserting his penis into the victim's anus. After the jury was sworn in, and Appellant pled not guilty, the trial court granted the State's motion in limine to exclude evidence of prior convictions supporting the defense's theory that Lorne Clark, a registered sex offender and the stepfather of the victim, was the actual perpetrator. Clark lived with his stepson and was one of the first people his stepson told about the assault. Clark had also been convicted of first-degree sexual abuse of his stepdaughter in Arkansas, and a few *689weeks before Appellant's trial was set to begin, he pled guilty to sexually assaulting two female children who lived in the same apartment complex. Appellant wanted to argue that Clark, a white man, sexually assaulted his stepson and manipulated him into accusing a "black man" of raping him to deflect attention from himself. Appellant is black. After the court granted the State's motion in limine, Appellant made an offer of proof and pled guilty pursuant to a plea bargain. He entered a written judicial confession and confessed to committing the crime on the stand.2 The parties also agreed to the following stipulation:
[STATE]: Your Honor, at this time, the State and the defense have stipulated to the following facts as being true and correct: We believe the testimony and the evidence would show at trial that the victim of this offense ... at no time, has ever made any allegation regarding Lorne Clark sexually abusing him. There has been no allegation, formally or informally to anyone at all, no police agency, not CPS, no family member. At no time [has] [the victim] ever accused Lorne Clark of sexually abusing him in any fashion.
And that would conclude our stipulation.
[DEFENSE]: We agree to stipulate to that, Your Honor.
Appellant was sentenced to 25 years' confinement as a habitual offender but was allowed to appeal the trial court's ruling.
The State maintained custody of the sexual assault kit and the white shorts that the victim wore during and after the assault. The victim was still wearing the shorts when he went to the hospital two days after the assault. The police report showed that the victim had not bathed or washed his genitals after the assault. No serology or DNA testing was performed before Appellant pled guilty.
b. Chapter 64 Proceedings
In 2010, Appellant moved for post-conviction DNA testing. The Department of Public Safety (DPS) Crime Laboratory technician, Nicole Mullins, tested (1) an anal swab, (2) a perianal swab, (3) a swab from the back-waistband of the shorts, and (4) a swab from the inside front-crotch area of the shorts. She did not find any interpretable DNA profiles. The trial court entered a non-favorable finding because the results were inconclusive. Appellant appealed, but during the pendency of his appeal, the trial court appointed counsel to represent him. A motion to dismiss was filed, which the court of appeals granted. Dunning v. State , No. 02-15-00222-CR, 2015 WL 5722605 (Tex. App.-Fort Worth Aug. 26, 2015) (mem. op., not designated for publication).
Appellant then sought to conduct his own DNA testing at the Serological Research Institute (SERI). The trial court granted Appellant's motion and rescinded its non-favorable finding. Serologist Amy Lee performed the testing. She reached the following conclusions,
(1) The anal-swab extract[3 ] contained a single source male DNA profile *690matching the victim at all tested loci.
(2) The perianal-swab extract contained a single weak male DNA profile from which the victim is included as a possible source. The defendant is excluded as a possible contributor to that profile.
(3) A single weak male DNA profile was obtained from the white-shorts swab that includes the victim as a possible source with the chance that another random person unrelated to him could be similarly included is approximately one in 330,000. The defendant is excluded as a possible contributor to that profile.
(4) A mixture of at least two individuals was obtained from the front-crotch swab and extract. The victim is included as the major contributor to both mixtures and the chance that another random person unrelated to him could be similarly included is approximately one in one billion. The defendant is excluded as a possible contributor to both mixtures.
(5) A mixture of at least two individuals was obtained from the back-waistband swab with both the victim and the defendant excluded as possible contributors to its major portion. There is insufficient information in its minor component to draw any conclusions.
(6) The extract from the back-waistband swab contained a weak mixture of at least two individuals, including at least one male, but there is insufficient information for any further conclusions to be made.
(7) No DNA was obtained from either of the reagent blank extracts.
The State asked Doctor Bruce Budowle to review Lee's conclusions, and he agreed with all of them except Conclusion # 5. He disagreed that the victim could be excluded as the contributor to the major portion of the mixture found on the back-waistband swab. The importance of the disagreement is that, if Appellant and the victim are excluded, then that DNA from the major portion of the mixture came from a third party, but if the victim cannot be excluded, as Budowle asserted, then the major contributor could be the victim or another person. (i.e., the result is inconclusive). He also cautioned against overstating the probative value of finding trace third-party DNA on the crotch swab and in the crotch extract. The trial court then entered a new non-favorable finding after "reviewing the evidence presented before it" because the DNA results did "not cast affirmative doubt on the defendant's guilt."
c. The Court of Appeals
On appeal, the State argued that Appellant's exclusion as a major contributor to all of the tested items was not exculpatory because it did not prove that he was not the assailant. Dunning , 544 S.W.3d at 919. It further asserted that, even if the results were exculpatory to some degree, Appellant still could not prevail because the victim identified him in a photographic array and because Appellant entered a judicial confession and confessed on the stand to committing the assault. Id. It also argued that the trial court's finding should be upheld because, in entering a non-favorable finding, the trial court must have credited Budowle's testimony, not Lee's, that the test results from the back-waistband *691swab were inconclusive and that the presence of minor DNA profiles on the front-crotch area swab did not show that there was an alternative suspect. Id. at 919-20.
The court of appeals found that the exclusion of Appellant as a major contributor to all of the tested items, and the results showing the existence of DNA from a third person in the crotch of the shorts, were strongly exculpatory. Id. at 920. It reasoned that this is not a case in which the defendant's DNA is simply absent or where it is inconclusive as to whether the defendant was a contributor.4 It also distinguished this case from LaRue v. State , 518 S.W.3d 439, 446 (Tex. Crim. App. 2017) because, here, the effect of the exculpatory DNA evidence does not merely "muddy the waters."
In weighing the test results against the inculpatory evidence, the court of appeals accorded Appellant's guilty plea and testimonial confession little value because it concluded that Appellant's plea was involuntary. According to the court, Appellant pled guilty only because the trial court granted the State's motion in limine excluding evidence of Clark as a possible alternative suspect-the "platform" of Appellant's defense-because he was indicted as a habitual offender, and because the State agreed to recommend the minimum term of confinement of 25 years. The court also cited the fact that Appellant was prepared to go to trial before the motion in limine ruling.
The court of appeals acknowledged that the victim identified Appellant as his rapist in a photographic array, but it said that it also had to consider that "[the victim] was twelve years old, was mentally impaired and hearing impaired, lived with Clark, and according to [Appellant]'s trial counsel, could have been easily manipulated by Clark to deflect suspicion away from himself. Clark had also spoken to police and reported that his stepson had said that 'a black man raped me.' "5
The court went on to explain that,6
[E]xamining the entire record, giving almost total deference to the trial court's resolution of disputed historical fact issues supported by the record and applications-of-law-to-fact issues turning on witness credibility and demeanor, the record before us reflects:
• that [Appellant] had pleaded not guilty;
• that [Appellant] was prepared to begin trial before a jury;
• that [Appellant] signed a judicial confession the morning of trial only after the trial court ruled he could not present Clark's Arkansas conviction to the jury or mention or present arguments concerning Clark;
• that [Appellant] faced up to a life sentence and that, in exchange for his guilty plea and judicial confession, the State agreed to the minimum 25-year sentence;
• that within three weeks of his guilty plea [Appellant] filed a pro se motion for new trial explaining that his decision to plead guilty was an error and was made based on the exclusion on the morning of trial of any evidence or arguments concerning Clark-the *692"platform" of his case-which had left him "frantically scrambling";
• that identity was an issue-in fact, the only issue;
• that the DNA test results established the absence of [Appellant]'s DNA on all tested items-including the crotch of complainant's shorts worn during the sexual assault and not removed until the complainant reached the hospital;
• that the DNA test results established that not only was [Appellant]'s DNA not present in the "shorts crotch swab," but that another person's DNA was present there along with the complainant's DNA; and
• that [Appellant]'s and the State's experts both agreed that another person-who was not [Appellant] and not the complainant-had contributed DNA to the "shorts crotch swab" tested and to the "shorts crotch extract" tested.
In light of all these facts-including [Appellant]'s judicial confession and the complainant's identification of [Appellant] from a photographic lineup-applying a de novo standard of review to the application-of-the-law-to-the-fact-issue of whether [Appellant] has proved that had the post-conviction DNA test results we now have been available during the trial of the offense it is reasonably probable that he would not have been convicted, we hold that he has so proven by a preponderance of the evidence ....
Id. at 921-22.
ANALYSIS
The State asked us to review the decision of the court of appeals, and after doing so, we conclude that it erred when it decided that the trial court should have entered a favorable finding. TEX. CODE CRIM. PROC . art. 64.04. It reached the wrong conclusion because it accorded too much weight to the test results excluding Appellant as a major contributor to the tested items and showing the presence of third-party trace DNA in the crotch area of the shorts and because it gave too little weight to the inculpatory evidence.
a. Standard of Review and Implicit Findings
The trial court entered a non-favorable finding. When reviewing an Article 64.04 favorability finding, we apply a bifurcated standard of review, affording almost total deference to a trial court's resolution of historical facts and mixed questions that turn on credibility and demeanor, but we review de novo mixed questions that do not turn on credibility and demeanor and questions of law. Rivera v. State , 89 S.W.3d 55, 59 (Tex. Crim. App. 2002). When the trial court does not enter separate findings, we imply findings necessary to support the ruling so long as they are reasonably supported by the record. Guzman v. State , 955 S.W.2d 85, 87 (Tex. Crim. App. 1997).
b. Test Results
At the Chapter 64 hearing, Budowle agreed with Lee that there was third-party DNA on the crotch swab and in the crotch-swab extract. The two primary issues litigated at the hearing were whether third-party DNA was present on the back-waistband swab and the probative value of third-party trace DNA on the tested items.
1. Back-Waistband Swab
Budowle disagreed with Lee's conclusion that the victim could be excluded as a possible contributor to the mixture on the back-waistband swab because she did not consider all of the testing data. Under SERI's protocols, if an allelic peak fell below a certain threshold-called the stochastic *693threshold7 -the technician had to ignore the data even if the peak was barely below it.8 Budowle acknowledged that the threshold serves an important purpose, but he thought that SERI's protocols should call for a more holistic review. When examining all of the allelic peaks from the back-waistband swab, including peaks that were just below the stochastic threshold, Budowle concluded that the major DNA profile from the swab is consistent with the victim's profile. Lee agreed with Budowle that, using his analysis, the victim could not be excluded, but she explained that she was constrained to follow SERI protocols.
2. Third Party Trace DNA on the Crotch Swab and in the Extract
With respect to the third-party DNA found in the crotch area, Budowle cautioned against jumping to the conclusion that the third-party DNA was from the "real" perpetrator. He said that the DNA samples from this case were all low-level trace DNA, also referred to as touch-transfer DNA or "touch" DNA, and he explained that there are many innocent ways by which low-level trace DNA can be transferred to an item of clothing. For example, he testified that it can be transferred between clothes that are washed together. Because of that, Budowle thought that the probative value of the sample was low because it merely showed that someone else had touched the shorts at some point, and that, even then, the DNA was not necessarily from the person who touched the shorts. We have expressed similar concerns: "[T]ouch DNA poses special problems because 'epithelial cells are ubiquitous on handled materials,' because 'there is an uncertain connection between the DNA profile identified from the epithelial cells and the person who deposited them,' and because 'touch DNA analysis cannot determine when an epithelial cell was deposited.' " Hall v. State , No. AP-77, 569 S.W.3d 646, 658, 2019 WL 361900, at *8 (Tex. Crim. App. Jan. 30, 2019). Consequently, we said, "the significant possibility of [touch] DNA being deposited by an innocent person reduces the probative value of any [touch] DNA test result." Id.
3. Discussion
While Appellant is excluded as a major contributor to the tested items, that in and of itself, is insufficient to meet his burden under Article 64.04.9 The court of appeals, however, emphasized that the test results were strongly exonerating, not only because Appellant was excluded as a contributor to all the tested items, but also because the technicians agreed that third-party DNA was on the crotch swab and in the crotch-swab extract. The court of appeals accorded too much value to the presence *694of the third-party DNA found in the crotch area,10 but to reach that conclusion, it had to disbelieve Budowle's testimony about the low probative value of the touch DNA. This might be a proper view of the evidence had the trial court entered a favorable finding, but it did not. It entered a non-favorable finding.11
Appellant complains that Budowle improperly suggested that the shorts were contaminated with trace DNA after the police seized them12 (i.e., that the chain of custody had been compromised), but that was not Budowle's testimony. He testified that, generally, touch DNA is typically low-level, and as a result, is difficult to analyze, and that it is easily transferred to clothes. According to Budowle, he finds mixtures of touch DNA from innocent people on clothing "all the time."
We agree with the State that the court of appeals erred because it failed to defer to the trial court's implicit determination that Budowle's testimony was credible. When that determination is accorded its proper deference, the results revealing third-party touch DNA from the crotch area of the shorts were significantly less probative than the court of appeals concluded.13
c. Inculpatory Evidence
1. Guilty Pleas and "During Trial"
During the trial of Appellant's offense, he pled guilty.14 He also entered a judicial confession and expressly admitted on the stand that he inserted his penis into the victim's anus. Under Article 64.04, the question is whether the inculpatory evidence-including Appellant's guilty plea, judicial confession, his testimonial confession, and other inculpatory evidence adduced at the Chapter 64 hearing-is so undermined by the test results that it is reasonably probable that a fact finder would not have convicted Appellant had the tests been available at trial despite the inculpatory evidence. TEX. CODE CRIM. PROC. art. 64.04 (the convicting court shall "make a finding as to whether, had the results been available during the trial of the offense, it is reasonably probable that the person would not have been convicted").
A guilty plea is typically strong evidence supporting a non-favorable finding because "[c]onvicting courts should ... give great respect to knowing, voluntary, and intelligent pleas of guilty."15 Ex parte Tuley , 109 S.W.3d 388, 391 (Tex. Crim. App. 2002). While a guilty plea does not *695preclude a favorable finding under Article 64.04,16 as the court of appeals correctly pointed out, if there is no reason to believe that a defendant's plea was inaccurate or unreliable, that plea should be given great weight. See id. If the test results are exceedingly exculpatory, however, "it is difficult to conclude that a prior guilty plea was accurate or reliable."17 Id. In such a case, the defendant's guilty plea might be outweighed.
Here, the court of appeals gave little weight to Appellant's guilty plea, judicial confession, and testimonial confession because it concluded that his guilty plea was involuntary. But Appellant asserts that he has never argued that his guilty plea was involuntary, only that he pled guilty because of the trial court's ruling on the morning before the trial. That shows only that Appellant was prepared to go to trial so long as he was allowed to present the defense of his choice,18 not that his guilty plea, judicial confession, and testimony are inaccurate or unreliable. The plea-hearing transcript also shows that Appellant knew the character of the State's evidence when he pled guilty. He knew that there was no DNA or serology testing and that the State's case was based on the identifications by the allegedly intellectually disabled and hearing-impaired victim who was under fourteen years of age. Yet, he nonetheless chose to plead guilty.
The post-conviction DNA test results are no more helpful in showing that Appellant's plea was inaccurate or unreliable. Even though Appellant was excluded as a major contributor to all the items tested, and testing revealed the presence of third-party touch DNA in the crotch area of the victim's shorts, the results are not the sort that would call into question the accuracy or reliability of his guilty plea because they only "muddy the waters."19 In short, the court of appeals erred when it gave little weight to Appellant's guilty plea, judicial confession, and testimonial confession because Appellant has not shown that his guilty plea was inaccurate or unreliable.20
*6962. Inculpatory Evidence from the Chapter 64 Hearing
The vast majority of the evidence adduced at the Chapter 64 hearing was inculpatory. Specifically, the State introduced into evidence a copy of its investigative file, and it argues that the court of appeals erred because it did not consider the inculpatory information even though it was admitted at the Article 64.04 hearing. From our review of the lower court's opinion, it is not clear whether it relied on the investigative file, but to the extent that it might not have considered it or other inculpatory evidence before the trial court at the Article 64.04 hearing, it erred. Asberry v. State , 507 S.W.3d 227, 228-29 (Tex. Crim. App. 2016).
The investigative file shows that the victim spoke with James Oliver by the apartment-complex pool after he was raped and that the victim pointed to a black man and said that he was a "raper." When interviewed by police, Oliver described the man's appearance, including that the man had a "goatee type beard and mustache ...." and that he was wearing beige slacks, a blue and white striped shirt, and a baseball hat. The victim's mother and stepfather told the police that Oliver might be talking about a man named Allen Beavers, who frequented Apartment # 145. The victim also told police that his assailant was a "big black man with a beard and a mustache."
The next day, the victim told his mother that his assailant was sitting on the stairs in front of Apartment # 145. When the two walked over there, the victim pointed to Appellant. The apartment-complex manager contacted police, and when they arrived, several people also identified Appellant. According to witnesses, Appellant was a daily visitor to Apartment # 145, and when the officer spoke to him, Appellant "seemed to know quite a bit of info[rmation] about the offense." The victim's mother and sister also gave statements to the police implicating Appellant. His sister wrote that she was coming home from a friend's house the night of the offense when she saw "Joe," and he was wearing a white and striped shirt with black slacks. She described him as a black man with slicked back hair with curls in the back, and she said that he carries a baseball cap. She also said that, "when my brother describe[d] him to me I knew it was the guy who assaulted my brother because he was wearing the same thing my brother told me he was wearing and what he looked like." The mother wrote that her son told her that "a black man had sex with him in the [laundromat] ...." when she came home from work. She also said that Jim (Oliver) was at the pool when her son approached him and pointed at his assailant *697and said "he is a rapist." The man was wearing a white shirt with stripes, black pants, and headphones. She realized the next day that the man Oliver described regularly went to the Stop-N-Go, where she worked, and he carried a blue backpack and headphones. She also realized that he came into her store the night of the offense and was wearing a white shirt with stripes and black pants and was carrying a baseball cap.
Doctor Leah Lamb of the Cook Children's Hospital examined the victim. Her findings supported that he had been sexually assaulted. Glenda Wood from the Tarrant County District Attorney's Office interviewed the victim, and he told her the same story that he had originally told police. At the police station, the victim picked out Appellant as his assailant in a photographic array. When he did so, he said "that's the man that had sex with me." The officer separately showed the photographic array to the victim's mother and sister; both identified Appellant as the person that the victim described to them and as someone who loiters at the apartment complex at Apartment # 145.
According to the court of appeals, the victim's identification of Appellant was not conclusive of the favorability issue because we have stated that "any overwhelming eye-witness identification and strong circumstantial evidence ... supporting guilt is inconsequential when assessing whether a convicted person has sufficiently alleged that exculpatory DNA evidence would prove his innocence under Article 64.03(a)(2)(A)." Dunning , 544 S.W.3d at 921 (citing Esparza v. State , 282 S.W.3d 913, 922 (Tex. Crim. App. 2009) ). That is true, but in Esparza we were dealing with Article 64.03(a)(2)(A) -entitlement to DNA testing-not Article 64.04-whether DNA test results are favorable. When a defendant files a Chapter 64 motion to obtain DNA testing, he must include an affidavit in which he alleges facts showing that the test results will be so exonerating that he would not have been convicted at trial had the results been available.21 Under Article 64.04, the judge does not consider the defendant's claims about how exculpatory he thinks the results will be if the testing is performed , as in Esparza ; he must consider the results after testing is performed to determine the exculpatory value, if any, of the test results and whether the results meet the standard set out in Article 64.04.22 In relying on Esparza in its Article 64.04 analysis, the court of appeals erred.
It also erred when it dismissed the victim's multiple identifications of Appellant as his rapist because "according to [Appellant]'s trial counsel, [the victim] could have been easily manipulated by Clark to deflect suspicion away from himself," and "Clark had spoken to police and reported that the complainant had said that 'a black man raped me.' "
*698Dunning , 544 S.W.3d at 921. There was no evidence that the victim had been coached to falsely accuse Appellant or that he could be easily manipulated. That was just trial counsel's theory of the case.23 Moreover, in reaching its conclusion, the court of appeals's wrongly credited the testimony of Appellant's trial counsel even though the trial court entered a non-favorable ruling. In entering a non-favorable finding, the trial court must not have believed trial counsel's claim that the victim could be easily manipulated into falsely accusing an innocent person of sexual assault.
CONCLUSION
In summary, the court of appeals did not defer to implicit credibility and demeanor determinations made by the trial court that weighed against Appellant. After properly deferring to those determinations, the test results are significantly less exculpatory than the lower court found. Moreover, the court of appeals gave too little weight to Appellant's guilty plea and testimonial confession. When the true exculpatory value of the test results are weighed against all of the inculpatory evidence, we conclude that Appellant has not shown that, had the results been available during the trial of the offense, it is reasonably probable that he would not have been convicted. Because the court of appeals concluded otherwise, we reverse its judgment vacating the trial court's non-favorable finding.